TESIS DOCTORAL - ccpo.odu.eduklinck/Reprints/PDF/delbarrioPHD2015.pdf · Resumen en español 17 Capítulo I: Introducción y antecedentes de la investigación 19 1.1 Motivación de

UNIVERSIDAD DE CANTABRIA

DEPARTAMENTO DE CIENCIAS Y TÉCNICAS

DEL AGUA Y DEL MEDIO AMBIENTE

TESIS DOCTORAL

DESARROLLO E IMPLEMENTACIÓN DE

HERRAMIENTAS MATEMÁTICAS PARA EL MODELADO

DE LA EUTROFIZACIÓN EN SISTEMAS COSTEROS

SEMI-ENCERRADOS EUTRÓFICOS E

HIPEREUTRÓFICOS.

Ph.D. THESIS

DEVELOPMENT AND IMPLEMENTATION OF

EUTROPHICATION MODELING TOOLS FOR

EUTROPHIC AND HYPERTROHIC SEMI-ENCLOSED

COASTAL SYSTEMS.

AUTORA

Pilar del Barrio Fernández

DIRECTORES

Andrés García Gómez

José Antonio Revilla Cortezón

Santander, 2015

UNIVERSIDAD DE CANTABRIA

E.T.S INGENIEROS DE CAMINOS, CANALES Y PUERTOS

DPTO. DE CIENCIAS Y TÉCNICAS DEL AGUA Y DEL MEDIO AMBIENTE

TESIS DOCTORAL

DESARROLLO E IMPLEMENTACIÓN DE HERRAMIENTAS

MATEMÁTICAS PARA EL MODELADO DE LA EUTROFIZACIÓN EN

SISTEMAS COSTEROS SEMI-ENCERRADOS EUTRÓFICOS E

HIPEREUTRÓFICOS.

Ph.D. THESIS

DEVELOPMENT AND IMPLEMENTATION OF EUTROPHICATION

MODELING TOOLS FOR EUTROPHIC AND HYPERTROPHIC SEMI-

ENCLOSED COASTAL SYSTEMS.

Autora: Pilar del Bario Fernández

Dirigida por: Andrés García Gómez

José Antonio Revilla Cortezón

Santander, 2015

A David

A mis padres

A toda mi familia

Si buscas resultados distintos, no hagas siempre lo mismo.

Albert Einstein.

AGRADECIMIENTOS

Quiero agradecer en primer lugar a la Universidad de Cantabria y al Gobierno de

Cantabria por haberme dado la oportunidad de realizar la Tesis doctoral disfrutando de

una beca predoctoral.

En segundo lugar a mis directores de Tesis, Andrés García Gómez y José Antonio Revilla

Cortezón por la confianza depositada en mí y por permitirme realizar este trabajo. Sin su

ayuda y apoyo este estudio no podría haberse realizado.

Me gustaría agradecer en concreto a Andrés García Gómez por su dedicación, ayuda,

paciencia y esfuerzo. Por sus revisiones y comentarios, con los que siempre he aprendido,

y por inculcar en mí el sentido del rigor académico, sin el cual no podría tener una

formación completa como investigadora.

Quiero agradecer especialmente a Neil K. Ganju por su ayuda y por darme la oportunidad

de realizar una estancia de investigación en el United States Geological Survey (USGS),

Woods Hole (Massachusetts, USA). También me gustaría darle las gracias por todo lo

que me ha enseñado y por las contribuciones realizadas a este trabajo.

Agradezco también a Sonia Castanedo, Raúl Medina, Giovanni Coco y a José Antonio

Juanes por su apoyo, respaldo y motivación para ayudarme a terminar este proyecto.

Quiero agradecer también a todos los compañeros de trabajo que me han ayudado, en

especial a Javier García Alba, a Alfredo Aretxbaleta y a Javier Barcena por su ayuda, en

especial en todas las dudas que he tenido sobre la implementación de los códigos en los

distintos lenguajes de programación, pero más especialmente por sus ánimos y su sentido

del humor. Me habéis enseñado los entresijos de la programación, y de la creación de

muchísimos tipos de figuras con distintos programas y la verdad es que sin vosotros esta

Tesis no hubiera sido posible.

Gracias también a todos los coautores de los artículos que forman parte de esta Tesis,

además de los ya mencionados, y a todos los investigadores que han aportado

conocimientos y datos imprescindibles a este estudio, especialmente a César Álvarez,

Melanie Hayn, Robert W. Howarth, y Aina García.

Agradecer también a Elvira, Tamara, Lara, Mar, Miriam, Alba, Andrea, Fernando y Diego

por su amistad. Gracias por apoyarme, ayudarme, escucharme y por estar a mi lado

siempre que lo he necesitado. ¡Muchas gracias a todos!

También agradecer al resto de mis compañeros Bea, Zhen, Rafa, Javi, Iñaki, Maitane,

Pablo A., Pablo H., María, Marisa, Sheila, Bárbara, Carolina, Imen, Nabil y muchos más.

Gracias por vuestra ayuda cuando lo he necesitado.

También agradecer a mis amigas y amigos por estar siempre ahí, entendiendo que me

encerrara a escribir la Tesis, y siempre haciendo planes para pasar un buen rato todos

juntos. Por esos momentos inolvidables que hemos compartido durante todo este tiempo,

gracias.

Finalmente, quiero dar las gracias a toda mi familia, por su apoyo, paciencia y ayuda.

Ellos son las principales personas en mi vida y sin las cuales esta Tesis nunca hubiera

sido posible. Agradecer en especial a mis padres por todas las oportunidades que me han

dado en la vida, por invertir en mi educación, por su apoyo incondicional y por haberme

enseñado a ser la persona que soy. A mi hermana y a Steven, por su ayuda y apoyo siendo

al tiempo familia y amigos. Y a Ingrid, por toda la alegría y felicidad que me ha dado

durante estos últimos años. También quiero agradecer a Carmen y a Daniel por todos los

ánimos y el apoyo en todo el proceso, por su comprensión y cariño en todo momento.

Por último, quiero agradecer especialmente a David, mi marido, compañero y mejor

amigo. Gracias por todo, por tu apoyo, ayuda, paciencia, por animarme siempre, y por

creer en mí. Gracias por tu ayuda incondicional y por estar siempre a mi lado, en los

buenos y malos momentos. Por eso y por mucho más quiero dedicarte a tí especialmente

este trabajo.

1

Contents

CONTENTS

List of figures 7

List of tables 13

Resumen en español 17

Capítulo I: Introducción y antecedentes de la investigación 19

1.1 Motivación de la investigación 19

1.2 Objetivos 24

Capítulo II: Descripción y análisis de las zonas de estudio. 26

Capítulo III. Desarrollo e implementación de un modelo simplificado para

sistemas costeros semi-encerrados hipereutróficos. Aplicación a una laguna

costera fuertemente regulada. 30

3.1 Descripción abreviada del modelo 31

3.2 Datos de campo y establecimiento del modelo 34

3.3 Calibración y validación. 35

3.3.1 Análisis de sensibilidad 35

3.3.2 Calibración del modelo 37

3.3.3 Validación del modelo 41

3.4 Resultados y discusión. 42

3.5 Conclusiones 46

Capítulo IV. Desarrollo e implementación de un sistema de modelado

ecológico para sistemas costeros eutróficos semi-encerrados. Aplicación a un

estuario alimentado por aguas subterráneas con vegetación acuática

sumergida. 48

4.1 Métodos y resultados del muestreo 50

4.2 Descripción resumida del modelo 52

4.3 Calibración del modelo 55

4.3.1 Calibración de los modelos biogeoquímicos y de irradiancia. 55

4.3.2 Calibración del modelo bio-óptico de zostera marina. 58

4.4 Escenarios de carga de nitratos y subida del nivel del mar 61

4.5 Discusión 66

2

Contents

4.6 Conclusiones 68

Capítulo V. Conclusiones y futuras líneas de investigación 70

5.1 Conclusiones generales 70

5.2 Conclusiones del análisis de sensibilidad 71

5.3 Conclusiones del modelado de luz 72

5.4 Conclusiones de calibración y resultados 73

5.5 Futuras líneas de investigación 75

Chapter I. Introduction and research background 79

1.1 Motivation of the research 81

1.2 History of coupled-linked models in estuaries and SECS 88

1.3 Review of hydrodynamic, water quality, ecosystem and bio-optical

irradiance models with coupled-linking capabilities: description and

applications. 91

1.3.1 Hydrodynamic models 92

1.3.1.1 Regional Ocean Modelling System (ROMS) 93

1.3.1.2 Estuarine and Coastal Ocean Model (ECOM) 94

1.3.1.3 Finite Volume Coastal Ocean Model (FVCOM) 94

1.3.1.4 MIKE 3 95

1.3.1.5 Mohid Water Modelling System (MWMS) 96

1.3.1.6 HAMSOM 97

1.3.1.7 Delft3D-Flow 98

1.3.1.8 Environmental Fluid Dynamics Code (EFDC) 99

1.3.1.9 Estuary and Lake Computer Model (ELCOM) 100

1.3.1.10 TELEMAC 100

1.3.1.11 H2D/H3D 101

1.3.1.12 Discussion 102

1.3.2 Water quality and ecosystem models 103

1.3.2.1 Water Quality Analysis Simulation Program (WASP) 106

1.3.2.2 CE-QUAL-ICM 107

1.3.2.3 DELWAQ 107

1.3.2.4 WQ 108

3

Contents

1.3.2.5 MIKE3-WQ (ECO LAB) 109

1.3.2.6 European Regional Seas Ecosystem Model (ERSEM) 110

1.3.2.7 Computational Aquatic Ecosystem Dynamics Model (CAEDYM) 110

1.3.2.8 Pelagic Interaction Scheme for Carbon and Ecosystem Studies

(PISCES) 111

1.3.2.9 ECOPATH with ECOSIM (EwE) 112

1.3.2.10 Integrated Generic Bay Ecosystem Model (IGBEM) 113

1.3.2.11 Ecological North Sea Model, Hamburg (ECOHAM) 114

1.3.2.12 Flexible Biological Module (FBM) 115

1.3.2.13 Mohid Water Quality Module 116

1.3.2.14 CE-QUAL-W2 117

1.3.2.15 NEUTRO 117

1.3.2.16 EnvHydrEM 118

1.3.2.17 Intermittently Closed and Open Lakes or Lagoons (ICOLLS) model

119

1.3.2.18 SEACOM 120

1.3.2.19 Phytoplankton-Zooplankton (P-Z) Models 120

1.3.2.20 Nutrient-Phytoplankton-Zooplankton (NPZ) Model 121

1.3.2.21 North Pacific Ecosystem Model for Understanding Regional

Oceanography (NEMURO) 123

1.3.2.22 Port Phillip Bay Model (PPBM) 124

1.3.2.23 Nutrient-Phytoplankton-Zooplankton-Detritus (NPZD) Model 124

1.3.2.24 Fasham 125

1.3.2.25 Fennel 125

1.3.2.26 System Wide Eutrophication Model (SWEM) 126


1.3.3 Bio-optical irradiance models with coupling or linking capabilities 129

1.3.3.1 Hydrolight-Ecolight v5 (HE5) 130

1.3.3.2 Fuji et al.’s model 131

1.3.3.3 Gallegos et al.’s model 132

1.3.3.4 Zimmerman model 133

4

Contents


1.4 Balancing spatial and temporal resolution 134

1.4.1 Low spatial, low temporal 135

1.4.2 High spatial, low temporal 136

1.4.3 Low spatial, high temporal 137

1.4.4 High spatial, high temporal 137

1.5 Modelling Tradeoffs 138

1.6 Discussion 140

1.7 Objectives 142

1.8 Layout of Thesis 143

Chapter II. Study sites 145

2.1 Introduction 147

2.2 Albufera of Valencia 148

2.3 West Falmouth Harbor 154

2.4 Study sites comparison and modeling strategies 160

Chapter III. Development and implementation of a simplified model for

semi-enclosed hypertrophic coastal systems. Application to a heavily

regulated coastal lagoon. 163


3.2 Materials and methods 169

3.2.1 The hydrodynamic model 169

3.2.1.1 The long wave model 170

3.2.1.2 The wind model 172

3.2.2 Eutrophication model 173

3.2.2.1 Transport equation 176

3.2.2.2 Chemical and biological interactions 176

3.2.2.3 Phytoplankton growth 177

3.2.2.4 Phytoplankton death 179

3.2.2.5 Chlorophyll-a concentration 180

3.3 Numerical techniques 181

5

Contents

3.3.1 The numerical grid 181

3.3.2 Field data and model set up 182

3.4 Calibration and validation 185

3.4.1 Sensitivity analysis 186

3.4.2 Model calibration 188

3.4.3 Model validation 194

3.5 Results and discussion 197

3.6 Conclusions 202

Chapter IV. Development and implementation of a coupled ecological

modelling system for semi-enclosed eutrophic coastal systems. Application

to a groundwater-fed estuary with submerged aquatic vegetation. 205


4.2 Observational methods 212

4.3 Observational results 214

4.4 Model description 217

4.4.1 Physical model 219

4.4.2 Biogeochemical model 221

4.4.3 Irradiance model 223

4.4.4 Bio-optical seagrass model 224

4.5 Model skill assessment 227

4.5.1 Biogeochemical and irradiance model assessment 227

4.5.2 Seagrass bio-optical model assessment 232

4.6 Nitrate loading and sea-level rise scenarios 235

4.7 Discussion 243

4.8 Conclusions 246

Chapter V. Conclusions and future research 247


5.2 Conclusions 250

5.1.1 General conclusions 250

5.1.2 Sensitivity analysis conclusions 251

6

Contents

5.1.3 Light modelling conclusions 251

5.1.4 Conclusions of calibration and results 252

5.2 Future research 254

5.3 Thesis impact and dissemination 257

5.3.1 Research articles 257

5.3.2 Communications in conferences and workshops 258

References 259

7

Lists of figures and tables

LIST OF FIGURES

Figura 1. Albufera de Valencia, canales de regadío y localización de las estaciones

de muestreo. 26

Figura 2. West Falmouth Harbor, estaciones de muestreo y cargas contaminantes. 27

Figura 3. Resumen gráfico del capítulo III 31

Figura 4. Diagrama de flujo del modelo 33

Figura 5. Concentración de clorofila-a media calculada en el análisis de sensibilidad

en la Albufera para los valores mínimos y máximos de los principales parámetros del

modelo. 37

Figura 6. Distribución del flujo de fósforo soluble reactivo (SRP) desde el sedimento

a la columna de agua en la Albufera de Valencia. 38

Figura 7. Comparación entre los resultados obtenidos por el modelo (líneas) y los

datos de campo (puntos) en las estaciones de muestreo y en toda la laguna para cada

periodo de calibración. 40

Figura 8. Evolución de la concentración de clorofila-a simulada (línea continua) y

los datos observados (puntos negros) en cada estación de muestreo, para los periodos

de calibración del año hidrológico 2005/2006. 41

Figura 9. Evolución de la concentración de clorofila-a promedio para la laguna

calculada por el modelo (línea) para todo año hidrológico 2005/2006 y los datos

observados (puntos) para los periodos de validación. 42

Figura 10. Distribución espacial de clorofila-a en la Albufera de Valencia en el año

hidrológico 2005/2006. 44

Figura 11. Balance de masas del fósforo soluble reactivo (SRP) que entra y sale de

la Albufera. 46

Figura 12. Resumen gráfico del capítulo IV 49

Figura 13. Histogramas de los datos de campo de clorofila y Kd en Outer y Snug

Harbors. 52

8


Figura 14. Diagrama de flujo del sistema de modelado y sus interacciones. En el

panel inferior se muestra que el ratio P/R>1 indica hábitat potencial de zostera

marina, mientras que el P/R<1 indica perdida potencial de hábitat de zostera marina.

U y V son las velocidades, h la profundidad de la columna de agua, T la temperatura

del agua y η la variación de la superficie libre. 54

Figura 15. Análisis de sensibilidad del modelo biogeoquímico de Fennel et al. (2006)

con un nuevo módulo integrado de irradiancia espectral. 55

Figura 16. a) Variación de la concentración vertical de clorofila-a en tres capas del

modelo; b) Promedio temporal y vertical de clorofila-a en el estuario completo. 58

Figura 17. a) Distribución del ratio Producción/Respiración; b) Distribución del ratio

Producción/Respiración aplicando el criterio de P/R>1, y comparación con los datos

del campo (línea negra continua); c) Detalle de la distribución de P/R en Snug y Outer

para P/R>1 y comparación con datos de campo (línea negra continua); d)

Distribución de zostera marina obtenida con la ecuación de limitación de

profundidad (Duarte et al., 2007). El área blanca representa las zonas donde se

descarta la presencia de zostera marina (P/R<1), el área verde es el hábitat potencial

de zostera marina (P/R>1) y la línea negra continua delimita el área de presencia de

zostera marina medida en la campaña de campo. 60

Figura 18. Variación espacial de P/R debido a la reducción de nutrientes (NR), al

aumento del nivel del mar (SLR) y a los escenarios combinados (CS). 62

Figura 19. Variación de P/R debida a la reducción de nutrientes y aumento del nivel

del mar. 63

Figura 20. Variación de clorofila-a debido a la reducción de nitratos y al SLR. 64

Figura 21. Variación espacial de clorofila-a debido a la reducción de nutrientes

(NR), aumento del nivel del mar (SLR) y escenarios combinados (CS). 65

Figura 22. Clorofila-a, Kd, P/R y variación del área de z. marina para los escenarios

combinados. 66

9


Figure 1.1. Examples of different kind of SECS: a) Open (Bay of Wismar,

Germany); b) Leaky (Venice Lagoon, Italy); c) Restricted (Quanzhou Bay, China);

d) Choked (Étang de Thau, France). 82

Figure 1.2. Trophic status. 85

Figure 1.3. Seagrass loss and pressures 86

Figure 1.4. Seagrass changes due to sea-level rise (SLR) 88

Figure 1.5. Water quality and ecosystem models coupling and/or linkage

possibilities. 104

Figure 1.6. Model tradeoff between generality, precision, and realism, adapted from

Levins (1966). 140

Figure 2.1. a) Aerial view of the Albufera of Valencia Natural Park (abc.es) and b)

West Falmouth Harbor (fineartamerica.com). 147

Figure 2.2. Albufera of Valencia, irrigation channels and sampling stations location.

148

Figure 2.3. Some of the pollutant pressures that surround the Albufera of Valencia. 149

Figure 2.4. Rice fields and irrigation channels surrounding the Albufera of Valencia.

150

Figure 2.5. Albufera of Valencia connection with the sea (“golas”), a) Gola Pujol, b)

Gola Perellonet, c) Gola Perelló. 151

Figure 2.6. a) Submerged aquatic vegetation that used to be at the Albufera of

Valencia (potamogetun pectinatus); b,c,d) water column and bottom of the Albufera

nowadays at different areas of the lagoon. 153

Figure 2.7. West Falmouth Harbor, site locations and input loads. 154

Figure 2.8. West Falmouth Harbor connection with the sea (a) and view from the

marked point (b) 155

Figure 2.9. West Falmouth Harbor closed shellfish activity due to pollution. 156

Figure 2.10. West Falmouth Harbor sailing activity. 156

10


Figure 2.11. West Falmouth Harbor seagrass (area delimited by red line)

disappearance. Adapted from http://buzzardsbay.org/historical-eelgrass-west-

falmouth.htm. 157

Figure 2.12. Outer and Snug Harbors seagrass presence in 2010 (green area) and

2012 (blue area). Adapted from Hayn (2012). 158

Figure 2.13. West Falmouth Harbor seagrass meadows at outer harbor (a and b),

seabed covered by macroalgae at south cove (c) and with some Ulva lactuca at Snug

Harbor (d) in 2012. 159

Figure 3.1. Graphical Abstract 166

Figure 3.2. Model flow chart 175

Figure 3.3. Variation of the light extinction coefficient (Ke) with the chlorophyll-a

concentration. 183

Figure 3.4. Average cloudiness variation during the hydrological year 2005/2006. 184

Figure 3.5. Comparison between calculated and observed lagoon water surface

during the period October 2005-September 2006 in a point of the lagoon located in

front of gola Pujol. 185

Figure 3.6. Mean chlorophyll-a (Chl-a) concentration calculated for the period of the

sensitivity analysis in the Albufera of Valencia for the minimum and maximum

calibration parameter values. 187

Figure 3.7. Distribution of soluble reactive phosphorus (SRP) flux from the sediment

to the water column into the lagoon. 191

Figure 3.8. Comparison between results obtained by the model (solid lines) and the

observed data (black dots) in the sampling stations and the whole lagoon for each

calibration period. 194

Figure 3.9. Evolution of the simulated chlorophyll-a concentration (“full line”,

calculated) and the observed data (“black dots”, observed) in each sampling station,

for the validation periods of the hydrological year 2005/2006. 196

11


Figure 3.10. Evolution of the lagoon-averaged chlorophyll-a concentration

calculated by the model (“full line”, calculated) for the hydrological year 2005/2006

and the observed values (“black dot”, observed) for the validation periods. 197

Figure 3.11. Chlorophyll-a spatial distribution in the Albufera of Valencia in the

hydrological year 2005/2006. 199

Figure 3.12. Mass balance of the soluble reactive phosphorus loads that comes in

and out of the Albufera of Valencia. 201

Figure 3.13. Comparison between calculated and observed chlorophyll-a calibration

and validation data 202

Figure 4.1. Graphical Abstract 209

Figure 4.2. Field survey stations and equipment 212

Figure 4.3. Groundwater fluxes and nitrate concentrations at West Falmouth Harbor.

Arrows indicate main contributions from Falmouth Wastewater Treatment Plant

(FWTP). 214

Figure 4.4. Chlorophyll-a and Kd field data histograms in Outer and Snug Harbors.

Data collected from sensors deployed during summer 2012 with a 5 minutes

sampling interval. 216

Figure 4.5. Modeling system flowchart and interactions. In the bottom panel, P/R

ratio > 1 indicates potential seagrass habitat; P/R < 1 indicates potential loss of

seagrass habitat. U and V are the velocities, h the water depth, T the water

temperature and η the water surface variation. 218

Figure 4.6. Comparison of hourly model and field data values of chlorophyll-a at the

sampling stations. 227

Figure 4.7. Comparison of hourly model and field data values of chlorophyll-a at the

sampling stations. 229

Figure 4.8. Comparison of mean model and field data values of Chlorophyll-a and

Kd at the sampling stations. 230

12


Figure 4.9. Spectral analysis of chlorophyll-a based on model results at the sampling

stations 231

Figure 4.10. a) Vertical chlorophyll-a variation in three layers of the model; b) Time-

averaged and mean vertical chlorophyll-a concentration for the whole estuary. 232

Figure 4.11. a) Photosynthesis/Respiration ratio distribution; b)

Photosynthesis/Respiration ratio distribution applying the P/R>1 criterion and

comparison with field data (black solid line); c) Detail of Snug and Outer P/R

distribution with P/R> 1, and comparison with field data (black solid line) ; d)

Seagrass distribution obtained with the depth-limited equation (Duarte et al., 2007).

The white area represents where seagrass presence is discouraged (P/R<1), the light

green area the potential seagrass habitat (P/R >1), and the black solid line delimits

the seagrass presence area measured in the field survey. 234

Figure 4.12. P/R spatial variation under nitrate reduction (NR), sea-level rise (SLR)

and combined (CS) scenarios. See Table 4.9 for an explanation of the scenarios

nomenclature. 237

Figure 4.13. P/R variation due to nitrate reduction and sea-level rise 238

Figure 4.14. Chlorophyll-a variation due to nitrate reduction and sea-level rise 240

Figure 4.15. Time-averaged and mean vertical chlorophyll-a spatial variation under

nitrate reduction (NR), sea-level rise (SLR) and combined (CS) scenarios. See Table

4.9 for an explanation of the scenarios nomenclature. 241

Figure 4.16. Chlorophyll-a, Kd, P/R and seagrass area variation for the combined

scenarios (CS). See Table 4.9 for an explanation of the scenarios nomenclature. 243

13


LIST OF TABLES

Tabla 1. Características principales de la Albufera de Valencia y West Falmouth

Harbor. 28

Tabla 2. Rango de variación de los principales parámetros y valores asignados en la

calibración. 36

Tabla 3. Errores obtenidos para todas las estaciones de muestreo en cada periodo de

calibración. 39

Tabla 4. Errores globales obtenidos con la concentración de clorofila-a media para

cada periodo de calibración en toda la laguna. 40

Tabla 5. Parámetros principales del modelo biogeoquímico y de irradiancia y valor

asignado. 56

Tabla 6. Valores medios de clorofila-a y Kd, desviación estándar (Std), y BIAS para

Outer, Snug y South. Los datos de campo de clorofila-a y Kd fueron obtenidos

procesando los datos de los sensores desplegados durante el verano de 2012. Los

resultados del modelo fueron obtenidos para el mismo periodo de tiempo. 57

Table 1.1. Characteristics of physical models with linked or coupled ecological

models. 103

Table 1.2. Some relevant characteristics of the main water quality and ecosystem

models. 129

Table 1.3. Some relevant characteristics of the main bio-optical models. 134

Table 2.1. Albufera of Valencia and West Falmouth Harbor main characteristics

comparison. 161

Table 3.1. Range of variation and assigned calibration value of the main

eutrophication parameters of the model. 186

Table 3.2. Error formulations applied in the calibration process. Фi is the calculated

concentration in cell i, Фiobs is the observed concentration in cell i and N is the number

of cells analyzed. 189

14


Table 3.3. Errors obtained with the chlorophyll-a values of all the sampling stations

in each calibration period. 192

Table 3.4. Global errors obtained with the mean chlorophyll-a concentration of each

calibration period in the whole lagoon. 193

Table 3.5. Errors of the different sampling stations obtained with the validation

periods. 195

Table 4.1. Mean values and standard deviation (Std) of measurements. 215

Table 4.2. Mean values, standard deviation (Std) and percentile 84 of measured

optical data during daylight hours. 215

Table 4.3. Physical model main equations and parameters 220

Table 4.4. Irradiance model main equations and parameters 222

Table 4.5. Irradiance model main equations and parameters 224

Table 4.6. Bio-optical seagrass model main equations and parameters 226

Table 4.7. Main biogeochemical and irradiance model parameters and chosen value.

228

Table 4.8. Mean values of chlorophyll-a and Kd, standard deviation (Std), and BIAS

for Outer, Snug and South Harbors. Field values of chlorophyll-a and Kd were

obtained processing data from sensors deployed during summer 2012. Model results

were obtained for the same time-period. 230

Table 4.9. Nitrate reduction and sea level rise scenarios, being CS_0/ NR_0/

SLR_2012 the initial scenario. 235

15


16


17

Resumen en español

Resumen en español

18

Resumen en español

19

Resumen en español

Resumen en español

De acuerdo a la normativa de estudios de doctorado de la Universidad de Cantabria en

relación a los requerimientos exigidos para aquellas Tesis redactadas en un idioma

diferente al español, aprobada por la Junta de Gobierno de 12 de marzo de 1999 y

actualizada a 17 de diciembre de 2012, a continuación se presenta un resumen de la Tesis

redactada en inglés.

Capítulo I: Introducción y antecedentes de la investigación

1.1 Motivación de la investigación

Los sistemas costeros semi-encerrados (SECS, semi-enclosed coastal systems) engloban

lagunas costeras y aguas de transición (Newton et al., 2013), siendo importantes sistemas

ecológicos con un considerable valor socio-económico (Lassere, 1979). Sin embargo, la

geomorfología de los SECS los hace especialmente vulnerables a cambios globales, tales

como el aumento del nivel del mar, variaciones de temperatura, tormentas, sequías,

inundaciones y cambios en la dinámica del sedimento (Newton et al., 2013). Además, las

actividades humanas, el desarrollo agrícola e industrial y la navegación provocan cambios

que afectan a la estructura y función de estos ecosistemas costeros. Por otra parte, se

caracterizan por sus propiedades de intercambio hidrodinámico con el sistema acuático

adyacente y pueden clasificarse como abiertos, permeables, restringidos o estrangulados

(Newton et al., 2013). Además, los SECS son ecosistemas complejos con una alta

productividad. Soportan una flora y fauna autóctona rica y variada por lo que

normalmente son áreas protegidas, siendo habitualmente lugares de especial importancia

para la alimentación y anidación de una multitud de especies de aves. La gama de

servicios ecosistémicos proporcionados por los SECS es extensa e incluye la provisión

de alimento y protección a larvas, moluscos y peces, así como la producción de oxígeno

20

Resumen en español

que realiza la vegetación acuática sumergida. Además proporcionan servicios culturales,

como la recreación y el ecoturismo entre otros (Millenium Ecosystem Assessment, 2005),

y ayudan a la regulación de los flujos de nutrientes, partículas y organismos entre los

sistemas de agua dulce y el océano. Sin embargo, estos valiosos ecosistemas están siendo

sometidos a fuertes presiones antropogénicas como el marisqueo, la piscicultura,

descargas de aguas residuales, el turismo, o la urbanización. Una de las principales

consecuencias de las presiones antropogénicas sobre estos sistemas es la pérdida de

hábitats como las praderas marinas, que pueden actuar como viveros y atrapar partículas

en suspensión. Además, estas presiones también han provocado el problema de la

eutrofización cultural, suponiendo una grave amenaza para la conservación de dichos

hábitats.

La eutrofización es el proceso mediante el cual un cuerpo de agua adquiere una alta

concentración de nutrientes, especialmente fosfatos y nitratos, generando un crecimiento

de algas excesivo. Este enriquecimiento de nutrientes puede ocurrir de forma natural o

puede ser el resultado de la actividad humana, produciéndose en este caso "eutrofización

cultural", que generalmente es provocada por la descarga de fertilizantes y aguas

residuales al sistema (Lawrence et al., 1998). Debido a los efectos indeseables que la

eutrofización tiene sobre el agua, se considera una forma de contaminación que perjudica

gravemente la calidad del agua, ya que incrementa el crecimiento de floraciones algales,

el agotamiento de oxígeno, y provoca la pérdida de vida en el fondo del sistema. En

consecuencia, el estudio y la determinación del estado trófico es un tema de suma

importancia debido a los efectos que este proceso tiene en ecosistemas costeros semi-

encerrados. Los términos que se utilizan para su determinación son (OCDE, 1982):

oligotrófico, mesotrófico, eutrófico, e hipereutrófico.

Cabe destacar que los estados que causan mayor preocupación son los eutróficos e

hipereutróficos, por lo que en los sistemas que presentan estos niveles de eutrofización

conviene realizar un análisis profundo de las causas que han provocado dicha situación,

http://www.microsofttranslator.com/bv.aspx?from=en&to=es&a=http%3A%2F%2Ftoxics.usgs.gov%2Fdefinitions%2Fnutrients.html

21

Resumen en español

y de los procesos que tienen lugar en el mismo. De hecho, los procesos de los sistemas

que presentan estos dos tipos de estados tienen similitudes y diferencias, siendo una de

las principales diferencias entre ellos la penetración de la luz en el agua. En sistemas

eutróficos puede llegar a existir vegetación acuática sumergida, ya que la luz podría

penetrar a través de la columna de agua tanto como permita la concentración de

fitoplancton, expresada en función de clorofila-a, entre otros factores; por lo que podrían

vivir distintas especies de vegetación, aunque su crecimiento estaría limitado por la luz

de manera diferente en función de los requisitos lumínicos de cada especie. Sin embargo,

en un sistema hipereutrófico la presencia de fitoplancton, y por tanto, la concentración de

clorofila-a en la superficie del agua es tan alta, que la penetración de la luz a través de la

columna de agua está muy limitada. En consecuencia, en sistemas hipereutróficos la

supervivencia de vegetación sumergida es poco probable, ya que en general, tiene altos

requisitos de luz, siendo habitualmente reemplazada por macroalgas oportunistas con

menores necesidades lumínicas. Además, el exceso de nutrientes en sistemas

hipereutróficos también produce un mayor crecimiento de epífitos que contribuyen a una

mayor atenuación de la luz.

Por otra parte, para determinar el crecimiento y la productividad de la vegetación acuática

sumergida, es importante tener en cuenta la atenuación espectral de la luz en el agua para

poder reproducir con exactitud el proceso de fotosíntesis llevado a cabo por la

misma. Esto es debido a que de la radiación que llega a la superficie, sólo la de longitudes

de onda entre 400 y 700 nm puede ser utilizada para la fotosíntesis, denominándose

radiación fotosintéticamente activa (PAR, Photosynthetically Active Radiation). La luz

llega a la superficie del agua previamente atenuada por los componentes atmosféricos, el

ozono, el vapor de agua, y los aerosoles marinos entre otros. Una vez en la columna de

agua, se atenúa por diversos factores, como por ejemplo la profundidad, la clorofila-a, los

pigmentos del fitoplancton, la turbidez y la materia orgánica disuelta coloreada (CDOM,

Colored Dissolved Organic Matter). Estas sustancias producen absorción, dispersión y

22

Resumen en español

retrodispersión de la luz en el agua, variando la cantidad de luz que llega a la planta; por

lo que, es importante determinar con precisión el ambiente lumínico en sistemas con

vegetación acuática sumergida, donde el hábitat puede existir, pero es extremadamente

vulnerable, como es el caso de los SECS eutróficos. Mientras que en el caso de los SECS

hipereutróficos, estos procesos no son tan relevantes ya que hay una menor disponibilidad

de luz por los altos niveles de concentración de clorofila-a que atenúan la misma, lo que

origina una escasez de vegetación acuática sumergida. Por consiguiente, el carácter

tridimensional que presentan los sistemas eutróficos no es significativo en sistemas

hipereutróficos, ya que en estos últimos, la penetración de la luz en el agua es muy baja.

Sin embargo, ambos tipos de sistemas presentan una alta variabilidad horizontal ya que

los gradientes horizontales de la distribución del fitoplancton son generalmente muy altos.

Otro aspecto que influye en la atenuación de la luz en sistemas eutróficos es la subida del

nivel del mar (SLR, Sea Level Rise) debido al cambio climático, que podría modificar el

hábitat de las praderas marinas a largo plazo. Cuando el nivel del mar aumenta, se produce

una variación de las condiciones lumínicas a medida que la profundidad se incrementa,

pudiendo cambiar la distribución de las praderas marinas. De hecho, las partes más

profundas de los sistemas presentan menor disponibilidad de luz, por lo que las praderas

marinas podrían desaparecer de estas zonas, mientras que en las zonas someras podría

crearse nuevo hábitat y algunas plantas migrarían a esas áreas. Sin embargo, la creación

de nuevo hábitat vendrá limitado en cada sistema por factores tales como barreras físicas,

la inadecuación del sustrato, construcciones litorales, aumento de la competencia entre

especies, la salinidad, y la temperatura entre otros. De manera que, aunque las condiciones

de luz fueran adecuadas para el crecimiento de praderas marinas, existirían otros

parámetros que podrían influir en la distribución de los hábitats.

La complejidad y variabilidad espacial de estos sistemas, hacen que el estudio de las

relaciones causa-efecto entre las distintas acciones humanas, hidrográficas,

hidrodinámicas y los procesos ecológicos sea una tarea difícil. De hecho, para estudiar

23

Resumen en español

estas relaciones se suelen utilizar modelos que ayudan al análisis del comportamiento de

los SECS, y que son útiles para la determinación y predicción del comportamiento de los

mismos ante diversas perturbaciones. Los modelos complejos, con un gran número de

formulaciones y de parámetros son ampliamente utilizados para describir SECS,

independientemente del nivel de eutrofización de los mismos. No obstante, muy pocos

presentan una formulación espectral de la atenuación de la luz en la columna del agua, y

casi ninguno tiene formulaciones para calcular el hábitat de vegetación acuática

sumergida. Además, a la hora de utilizar estos modelos suele hacer falta una gran cantidad

de datos de entrada que dificultan el proceso, así como amplios conocimientos específicos

del modelo. Por todo ello, existe la necesidad de desarrollar herramientas

computacionales simplificadas para la gestión de SECS eutróficos e hipereutróficos,

teniendo en cuenta la compleja hidrodinámica de estos sistemas, y los procesos

específicos que rigen cada tipo. En los últimos años, modelos acoplados y/o conectados

han demostrado ser una buena solución para integrar de manera flexible y simplificada

los procesos más importantes para cada problemática. Por otra parte, el compromiso entre

simplificación, realismo y precisión, ha dado buenos resultados en la descripción y

predicción del comportamiento de sistemas mediante modelos, por lo que serán factores

a tener en cuenta en el presente estudio.

A su vez, cabe resaltar que en el Chapter I de la Tesis se ha realizado un análisis sobre la

historia de los modelos acoplados y conectados, una profunda revisión bibliográfica de

modelos hidrodinámicos, ecológicos, de irradiancia y bio-ópticos; y un análisis sobre el

equilibrio entre resolución espacial y temporal, y el compromiso que debe cumplirse entre

generalidad, realismo y precisión en la definición de la complejidad de un modelo.

El análisis del estado del arte revela que la complejidad de la mayoría de los modelos

ecológicos existentes aplicados a SECS hace difícil su utilización y comprensión, siendo

además necesario un equilibrio entre generalidad, realismo y precisión. Por otra parte, se

ha puesto de manifiesto la necesidad de ahondar en el modelado de la interacción entre la

24

Resumen en español

eutrofización y la vegetación acuática sumergida. De hecho, el fitoplancton producido

por un incremento en la disponibilidad de nutrientes, la turbidez, las concentraciones de

CDOM y el aumento de la profundidad del agua debido a la subida del nivel del mar son

cuestiones fundamentales que afectan a la atenuación de la luz en la columna de agua y

que limitan el proceso de fotosíntesis y crecimiento de la vegetación acuática

sumergida. Con todo, el estudio de este proceso y el modelado espectral de la luz de forma

tridimensional con alta resolución espacial apenas se ha llevado a cabo para estrategias

de gestión, ya que los modelos existentes que podrían simular estas características son

muy complejos y difíciles de utilizar.

El grado de simplificación y la selección de los procesos que deben tenerse en cuenta en

un modelo es una tarea compleja que depende del sistema y que requiere una comprensión

completa de los procesos que lo controlan. De hecho, los modelos complejos de

eutrofización son generalmente utilizados tanto para sistemas hipereutróficos como

eutróficos, aunque la problemática y los procesos que rigen cada uno de ellos son

diferentes. Asimismo, el gran número de ecuaciones y parámetros de estos modelos

conducen a que normalmente la obtención de datos para su calibración y validación sea

un proceso exigente y caro, y su ajuste difícil y tedioso, por lo que no serían apropiados

para analizar estrategias de gestión que deben ser evaluadas en un corto periodo de

tiempo.

De lo antes dicho, se desprende la necesidad de desarrollar nuevas herramientas de

modelado ecológico, surgiendo los objetivos de la presente Tesis.

1.2 Objetivos

El objetivo general de esta Tesis es: desarrollar nuevas herramientas de modelado

ecológico para evaluar y describir el comportamiento de sistemas costeros semi-

encerrados eutróficos e hipereutróficos. Además, esta Tesis permite ahondar en el

conocimiento del funcionamiento de dos sistemas costeros semi-encerrados, la Albufera

25

Resumen en español

de Valencia, que es un sistema hipereutrófico, y West Falmouth Harbour, que es un

sistema eutrófico.

Los objetivos específicos de esta Tesis son:

Analizar diferencias entre sistemas eutróficos e hipereutróficos y las limitaciones

de los modelos existentes.

Diseñar, conectar e implementar un modelo de calidad del agua simplificado para

un sistema hipereutrófico costero semi-encerrado fuertemente regulado.

Diseñar, acoplar e implementar un sistema de modelado para describir el

comportamiento de un sistema eutrófico costero semi-encerrado y las

implicaciones de la eutrofización en la atenuación de la luz y en la vegetación

acuática sumergida.

Evaluar la sensibilidad de los modelos, calibrarlos a partir de datos de campo, y

aplicarlos a ecosistemas costeros semi-encerrados con problemas de eutrofización

cultural y con importancia socio-económica y ambiental.

Analizar diferentes factores, tales como los efectos de la reducción de nutrientes

y el aumento del nivel del mar en un sistema eutrófico semi-encerrado, y las cargas

de entrada y salida en un sistema hipereutrófico mediante un balance de masas.

Analizar las limitaciones de los modelos desarrollados y proponer futuras líneas

de investigación.

26

Resumen en español

Capítulo II: Descripción y análisis de las zonas de estudio.

La Albufera de Valencia (ver Figura 1) y West Falmouth Harbor (ver Figura 2) son dos

SECS con similitudes y diferencias. Por un lado, ambos son SECS con un problema de

eutrofización cultural, están cerca de una zona poblada, y tienen una gran importancia

económica y ambiental. Por otro lado, sus características específicas son diferentes, como

puede observarse en la Tabla 1.

Figura 1. Albufera de Valencia, canales de regadío y localización de las estaciones de muestreo.

La Albufera de Valencia está deteriorada en un mayor grado que West Falmouth Harbor

debido en parte a que está muy cerca de una gran ciudad (Valencia, España). Es un SECS

muy cerrado y constreñido, cuya conexión con el mar solamente se efectúa unas pocas

veces al año mediante tres canales artificiales (golas) regulados por compuertas. A su vez,

está rodeado de campos de arroz lo que hace que la carga de nutrientes de entrada sea

excesiva. Estos factores han provocado el estado hipereutrófico del sistema, que ha

27

Resumen en español

llevado a la desaparición de muchas especies de fauna y flora. Una de las principales

razones de esta desaparición es la limitada penetración de la luz a través de la columna

de agua.

Por el contrario, West Falmouth Harbor es un SECS eutrófico cerca de la ciudad de

Falmouth (Massachusetts, EEUU), con mucha menor población alrededor que la

Albufera. Una de las mayores fuentes contaminantes de este SECS son las cargas de

nutrientes procedentes de una Estación Depuradora de Aguas Residuales (EDAR) que

llegan al SECS a través de aguas subterráneas con largos tiempos de viaje (hasta 10

años). Esta bahía es un SECS restringido, pero esta restricción está provocada por dos

diques situados en la bocana (ver Figura 2), por lo que la conexión con el mar es

constante.

Figura 2. West Falmouth Harbor, estaciones de muestreo y cargas contaminantes.

28

Resumen en español

En este estuario, la vegetación acuática sumergida ha ido desapareciendo durante años

debido principalmente a la limitación de la penetración de la luz a través de la columna

de agua y a la eutrofización. Todos estos factores explican la desaparición y disminución

del área cubierta por zostera marina, existiendo también una preocupación por las

consecuencias de la subida del nivel del mar en esta zona debido al cambio climático y a

la constante aunque restringida conexión con el mar.

Tabla 1. Características principales de la Albufera de Valencia y West Falmouth Harbor.

CARACTERÍSTICAS ALBUFERA DE VALENCIA WEST FALMOUTH HARBOR

SUPERFICIE (Km2) 23.2 0.7

PROFUNDIDAD (m) 0.9 1

FUENTES DE CONTAMINACIÓN

- Vertidos de aguas residuales - Fertilizantes y pesticidas

- Vertidos de aguas residuales

ESTADO TRÓFICO Hipereutrófico Eutrófico

NUTRIENTE LIMITANTE P N

TIPO DE SECS Estrangulado Restringido

CONEXIÓN CON EL MAR Regulado por 3 canales con compuertas Limitado por 2 diques

PRADERAS MARINAS No Sí

Las diferencias entre los sistemas estudiados hacen que la estrategia de modelado pueda

ser diferente, aunque ambos son SECS con serios problemas de eutrofización. Para un

sistema hipereutrófico podría utilizarse un modelo simple tipo NPZ, ya que el crecimiento

del fitoplancton debido a la sobreabundancia de nutrientes, y su mortalidad por el

herbivorismo del zooplancton son los principales procesos que rigen el sistema. Sin

embargo, en un sistema eutrófico la penetración de la luz a través de la columna de agua

podría permitir la supervivencia de praderas marinas, por lo que no sería suficiente

solamente con un modelo biogeoquímico simple, sino que además sería necesario el uso

de un modelo bio-óptico. Sumado a esto, las condiciones hidrodinámicas son bastante

diferentes en los dos SECS, siendo necesario tener en cuenta en la Albufera la regulación

antropogénica con el mar, mientras que en West Falmouth Harbor el efecto de la marea

y el aumento del nivel del mar podrían tener consecuencias en los hábitats de praderas

marinas. Finalmente, West Falmouth Harbor presenta un carácter tridimensional debido

29

Resumen en español

a su hidrodinámica, a las diferencias de concentración de clorofila-a en diferentes capas

de la columna de agua, y a la distribución de la luz que penetra en el agua. Mientras que,

la Albufera presenta un carácter bidimensional, ya que la comunicación con el mar es

muy limitada y la luz apenas penetra en la columna de agua debido a la alta concentración

superficial de fitoplancton, y por tanto de clorofila-a.

30

Resumen en español

Capítulo III. Desarrollo e implementación de un modelo

simplificado para sistemas costeros semi-encerrados

hipereutróficos. Aplicación a una laguna costera fuertemente

regulada.

Se desarrolló un modelo simplificado de eutrofización bidimensional para simular las

variaciones temporales y espaciales de clorofila-a en sistemas costeros hipereutróficos

semi-encerrados. Este modelo considera una conexión con el mar muy limitada y regulada

antropogénicamente. También tiene en cuenta la entrada y salida de cargas variables de

nutrientes, el flujo de los sedimentos a la columna de agua, y la cinética de crecimiento y

mortalidad del fitoplancton. El modelo fue calibrado y validado aplicándolo a la Albufera

de Valencia, un SECS hipereutrófico cuya conexión con el mar está fuertemente regulada

por un sistema de compuertas. Los resultados de calibración y validación presentan un

acuerdo significativo entre el modelo y los datos obtenidos. La exactitud se evaluó

mediante un análisis cuantitativo, en el cual la incertidumbre promedio de la predicción

del modelo fue menos del 6%. Los resultados confirmaron un bloom de fitoplancton en

abril y octubre, alcanzando unos valores máximos alrededor de 250 µg L-1 de clorofila-

a. Un balance de masas reveló que el proceso de eutrofización está magnificado por la

limitada conexión de la laguna con el mar, y por el flujo de sedimentos existente a la

columna de agua. Este estudio ha demostrado que el modelo desarrollado es una

herramienta eficaz para describir el problema de la eutrofización en sistemas costeros

hipereutróficos semi-encerrados.

En este capítulo, se presenta de manera abreviada la descripción del modelo desarrollado

(ver Figura 3), el análisis de sensibilidad de los principales parámetros del mismo, y la

calibración, validación y aplicación del modelo a un SECS hipereutrófico, la Albufera de

Valencia. La información completa sobre el desarrollo, evaluación y aplicación de este

31

Resumen en español

modelo puede encontrarse en el Chapter III de la presente Tesis. También puede

encontrarse una descripción más detallada de la zona de estudio en el Chapter II.

Figura 3. Resumen gráfico del capítulo III

3.1 Descripción abreviada del modelo

El comportamiento hidrodinámico de la laguna está controlado por varios factores, como

las entradas de agua dulce de los canales de riego, el equilibrio entre la precipitación y

evaporación, el viento y las salidas a través de las golas. Este último factor es fuertemente

dependiente del régimen de apertura de las compuertas y de la diferencia del nivel de agua

entre la laguna y el mar. Con el fin de evaluar estos procesos se utilizaron dos modelos

32

Resumen en español

hidrodinámicos, un modelo de onda larga y un modelo de viento. El primero de ellos, un

modelo bidimensional promediado en profundidad, fue utilizado para caracterizar la

circulación del agua en la Albufera, los flujos de agua entrando en la laguna a través de

los canales de regadío, y los de salida y entrada que se producen a través de las

tres golas al mar. El segundo, un modelo cuasi-tridimensional, fue aplicado para calcular

las corrientes de viento en el sistema. Ambos modelos consideran todo el dominio

(canales de riego, laguna, golas, mar), estando el efecto de la regulación de las salidas de

agua a través de las golas específicamente incluido en el modelo de onda larga. Este

modelo, resuelve las ecuaciones de Navier-Stokes con la aproximación promediada de

Reynolds (RANS, Reynolds Averaged Navier Stokes). Esta aproximación, propuesta por

Reynolds en 1895, está basada en la descomposición de las variables de flujo en un valor

medio y otro fluctuante. Aunque en ingeniería los flujos son turbulentos en su mayoría,

muchos de estos flujos pueden ser considerados como muy poco dependientes del tiempo,

con unas fluctuaciones superpuestas a la corriente principal estacionaria. En dicho caso,

solo interesarán las magnitudes promediadas en el tiempo, en lugar de los detalles de

variación con el tiempo. Por ello, las ecuaciones de Navier Stokes con la aproximación

de Reynolds se resuelven para los valores medios, que son los más interesantes en muchas

aplicaciones, como es el caso de la Albufera de Valencia.

En cuanto al modelo de eutrofización desarrollado, es un modelo numérico simplificado

bidimensional que resuelve la ecuación de advección-dispersión para cada variable de

calidad de agua seleccionada. La poca profundidad que generalmente caracteriza los

SECS hipereutróficos junto con su baja variación vertical justifica la simplificación de

promediar en vertical. El modelo de eutrofización desarrollado simula la calidad del agua

con respecto a la concentración de fitoplancton y fósforo soluble reactivo (SRP, Soluble

Reactive Phosphorus) en la columna de agua. En este sentido, el fitoplancton es un

indicador de la clorofila-a presente en el lago, mientras que el fósforo soluble reactivo es

el nutriente limitante del sistema. Por ello, el crecimiento de fitoplancton se calculó como

33

Resumen en español

una función del fósforo soluble reactivo, de la temperatura, y de la intensidad de luz en la

columna de agua; y su consumo se centró principalmente en la respiración endógena y en

el herbivorismo del zooplancton. La Figura 4 describe el diagrama de flujo del modelo

desarrollado y los principales procesos considerados.

Figura 4. Diagrama de flujo del modelo

34

Resumen en español

3.2 Datos de campo y establecimiento del modelo

Este estudio se llevó a cabo para el año hidrológico 2005/2006, para el cual se tenían

datos tanto de aportes de flujos y contaminantes al lago, como de las principales variables

de calidad de agua dentro del mismo, de la apertura y cierre de las golas y de las

condiciones ambientales y mareales. Los principales datos de aportes y de calidad del

agua del modelo fueron medidos por la Entidad Pública de Saneamiento de Aguas de

Valencia (EPSAR) entre Octubre de 2005 y Septiembre de 2006, en siete estaciones

distribuidas a lo largo del lago. Los datos recolectados en dichas estaciones fueron

clorofila-a, temperatura, SRP y la profundidad del disco de Secchi. Estos datos se

recogieron todos los meses de Octubre 2005 a Septiembre de 2006, obteniéndose 12

muestras para cada variable y cada estación a lo largo del año hidrológico, sumando un

total de 336 muestras. La localización de las estaciones de muestreo, de los canales de

regadío, y de las golas puede observarse en la Figura 1. Además, el SRP de los canales

de regadío principales fue medido durante cada mes del periodo de estudio para describir

las concentraciones de nutrientes que se descargan al lago, obteniéndose un total de 156

muestras de SRP. Los valores máximos de SRP se encontraron en las acequias del norte,

debido a que el origen de este tipo de nutriente es urbano e industrial.

La concentración media de clorofila-a obtenida fue de 115.7 µg L-1, lo que significa que

la Albufera de Valencia es un sistema hipereutrófico. El promedio de profundidad del

disco de Secchi varía entre 0.12 y 0.36 metros, lo que significa que la penetración de la

luz está altamente atenuada en la columna de agua. Con los datos observados de clorofila-

a y de profundidad del disco de Secchi en las siete estaciones de muestreo distribuidas

alrededor del lago, se obtuvo una expresión para describir la variación del coeficiente de

extinción luz en el agua (Ke) con la concentración de clorofila-a (ver Eq. 30 Chapter III)

que fue implementada en el modelo.

35

Resumen en español

Los datos climatológicos para el año hidrológico 2005/2006, tales como nubosidad y

viento, fueron obtenidos de la estación meteorológica Valencia Viveros (a 10 Km de la

Albufera). La nubosidad fue medida por la Agencia Estatal de Meteorología en una escala

de 0 a 8 oktas, siendo 0 oktas la mínima cobertura de nubosidad y 8 oktas la máxima. El

viento más frecuente fue del sureste, siendo en general de baja intensidad. Sin embargo,

en algunos casos, el viento alcanzó intensidades por encima de 3 m s-1.

Otro parámetro importante de entrada que ha sido tenido en cuenta en este trabajo es el

flujo de fósforo soluble reactivo (SRP) del sedimento a la columna de agua. Este

parámetro fue obtenido a partir de un estudio específico con el fin de caracterizar la

influencia del flujo de SRP desde el sedimento a la columna de agua en la eutrofización

del sistema, para lo cual se muestrearon datos en 17 estaciones.

3.3 Calibración y validación.

La calibración del modelo de eutrofización implica ajustar los parámetros de éste, de

manera que sus resultados se ajusten a los datos medidos en los periodos de calibración.

Previamente, se llevó a cabo un análisis de sensibilidad con el objeto de cuantificar el

efecto de la variación de los parámetros del modelo en los resultados del mismo. Una vez

calibrado, el modelo fue validado para un año hidrológico (año 2005/2006), de cara a

confirmar el valor asignado a los parámetros del modelo durante la calibración.

3.3.1 Análisis de sensibilidad

Un análisis de sensibilidad fue utilizado para identificar los parámetros más influyentes

en la variación de los resultados del modelo. Los rangos de variación de cada parámetro

se muestran en la Tabla 2.

36

Resumen en español

Tabla 2. Rango de variación de los principales parámetros y valores asignados en la calibración.

PARAMETRO DESCRIPCION UNIDADES RANGO DE

VARIACIÓN

VALOR

ASIGNADO

apc Ratio fosforo-carbono gP gC-1 0.011-0.025a,d 0.011***, d

Kr Tasa de respiración endógena del fitoplancton

día -1 0.05-0.5b,a,d 0.12***, g, h, d

Kmp Constante de semi-saturación del fósforo

mgP L-1 0.001-0.005c 0.0027***, d

aC/CHL-a Ratio carbono/clorofila-a mgC mgChl-a-1 50 – 133d 88***, d

Gmax Tasa de crecimiento máximo del fitoplancton

día-1 1.5-2.5c 1.5***, i

Is Intensidad de saturación de la luz en el agua del fitoplancton

Ly día-1 100-400c *(ver Eq. 33 Chapter III)

Fs Flujo de fósforo soluble reactivo del sedimento a la columna de

agua.

mgP m-2 día-1 5-50e 20.72 **

Cg Tasa de herbivorismo (y filtracion) del zooplancton

LmgC-1día-1 0.05-0.3f 0.3 ***, f

Source: a(Ambrose, 1988), b(Di Toro and Matystik, 1980), c(Thomann and Mueller, 1987), d(Martín, 1998), e field data, f (Chau and Haisheng, 1998); g (Lindenschmidt, 2006); h (Ambrose et al., 1993); i (Parslow et al., 1999).

*The values were assigned by empiricism;**The values were field data; ***The values were verified by calibration and literature.

Para realizar el análisis de sensibilidad, se fijaron ocho parámetros en sus valores mínimos

y máximos, definidos en los rangos marcados por la bibliografía (ver Tabla 2). Cada

simulación fue llevada a cabo fijando uno de los parámetros en su valor mínimo o

máximo, y dejando el resto de parámetros en el valor medio de su rango de variación

correspondiente. Este proceso se repitió para cada parámetro, por lo que la variación

media de la concentración de la clorofila-a sobre toda la laguna fue utilizada para evaluar

la sensibilidad del modelo a la variación de cada parámetro.

El histograma de la Figura 5 revela que los parámetros para los cuales los resultados de

clorofila-a del modelo presentan mayor sensibilidad son Kr, Cg, Gmax, Fs, y apc. Es

importante mencionar que Kr, Cg y Gmax alteran directamente las tasas de crecimiento del

37

Resumen en español

fitoplancton, mientras que Fs y apc afectan a la posible utilización o captamiento de fósforo

por parte del fitoplancton. Estos resultados están de acuerdo con los obtenidos en otros

estudios (Schladow and Hamilton, 1997; Wu et al., 2009), en los que Kr era el parámetro

de sensibilidad que más afectaba al crecimiento del fitoplancton, y por tanto a la

concentración de clorofila-a.

Figura 5. Concentración de clorofila-a media calculada en el análisis de sensibilidad en la Albufera para

los valores mínimos y máximos de los principales parámetros del modelo.

3.3.2 Calibración del modelo

Los periodos seleccionados para la calibración fueron Octubre, Febrero, Mayo y Julio.

Dado que en Octubre y Mayo suele ocurrir un bloom de fitoplancton, y durante Febrero

y Julio la concentración de clorofila-a suele ser más baja en comparación con el resto del

año. La comparación entre las medidas de campo y las predicciones del modelo se llevó

a cabo mediante el cálculo de diferentes tipos de errores. Los errores calculados fueron el

error absoluto (AE), el error relativo (RE), el error relativo medio (MRE), el error

cuadrático medio (RMSE), el error normalizado cuadrático medio (PRMSE), el error

absoluto medio (MAE), el error absoluto medio normalizado (NMAE), y el BIAS.

La calibración del modelo fue llevada a cabo teniendo en cuenta los parámetros más

influyentes resultantes del análisis de sensibilidad y las variables estacionales. A pesar de

38

Resumen en español

la compleja naturaleza y la alta variabilidad de la Albufera, el valor asignado a los

parámetros utilizados en el modelo se mantuvo constante en todos los periodos. Sin

embargo, la intensidad de saturación de luz del fitoplancton (Is), varía dependiendo de la

estación y la temperatura, siendo mínima en invierno y máxima en verano (Macedo et al.,

2001) (ver Eq.33 Chapter III).

Otro parámetro importante que ha sido calibrado en este trabajo es el flujo de SRP desde

el sedimento a la columna de agua. Este parámetro fue obtenido mediante una campaña

de campo específica en la que, como ya se ha indicado, se muestrearon datos en 17

estaciones para representar la distribución espacial del flujo de SRP en el lago. Los datos

medidos fueron interpolados utilizando el método geoestadístico Kriging (Kitanidis,

1997). Los resultados obtenidos, los datos medidos, y la localización de las estaciones de

muestreo del flujo de SRP están representadas en la Figura 6. Como puede verse en dicha

figura, la zona norte del lago presenta el flujo máximo difusivo, debido a las altas

descargas de SRP que durante años llevan depositándose en esa zona, siendo el valor

medio de flujo de SRP para todo el lago de 20.72 mgPm-2día-1.

Figura 6. Distribución del flujo de fósforo soluble reactivo (SRP) desde el sedimento a la columna de

agua en la Albufera de Valencia.

39

Resumen en español

Además, se realizaron diversas simulaciones para asignar un valor a los parámetros más

significativos del modelo, los cuales fueron ajustados para obtener el menor error entre

los datos observados y los resultados del modelo. Los valores asignados se pueden

observar en la Tabla 2, encontrándose en los rangos recogidos en la literatura. Los errores

entre los resultados del modelo y los datos observados en las campañas de campo se

resumen en la Tabla 3. Estos errores se calcularon en las siete estaciones de muestreo para

cada mes, siendo el mes con el menor error relativo medio Febrero, mientras que el de

mayor error relativo medio fue Mayo. Se ha calculado también el error cuadrático medio

porcentual (PRMSE) para cada periodo de calibración, y este se ha comparado con el

error absoluto medio normalizado (NMAE). Como resultado, se obtiene que los valores

más bajos de estos errores se han encontrado en el mes de Mayo, siendo el PRMSE 7.30%

y el NMAE de 8.50%.

Tabla 3. Errores obtenidos para todas las estaciones de muestreo en cada periodo de calibración.

Periodo MRE (%) RMSE(μg L-1) PRMSE (%) MAE(μg L-1) NMAE (%) BIAS(μg L-1)

Octubre 2.52 26.94 11.95 23.10 10.70 2.41

Febrero 1.75 10.46 19.26 8.52 16.13 -1.22

Mayo -6.56 16.12 7.30 12.45 8.50 8.36

Julio -3.62 6.23 14.76 4.84 14.01 -0.28

El BIAS también se ha calculado, siendo Julio el mes en el que se ha obtenido un valor

más cercano a cero. De hecho, los valores calculados por el modelo fueron en torno a 0.28

μg L-1 más bajos que los valores observados. Después se calcularon los errores globales

utilizando los valores medios observados y calculados. Las simulaciones mostraron que

para Octubre, Febrero, Mayo y Julio, los errores relativos obtenidos comparando la media

de los datos observados con la de los resultados del modelo están entre un 3 y un 5 %.

Los errores absolutos se encuentran entre 1.86 y 7.15 µg L-1, como puede verse en la

Tabla 4.

40

Resumen en español

Tabla 4. Errores globales obtenidos con la concentración de clorofila-a media para cada periodo de

calibración en toda la laguna.

Periodo EA (μg L-1) ER (%)

Octubre 7.15 3.17

Febrero 1.86 -3.42

Mayo 7.13 4.27

Julio 2.23 -5.28

Como puede verse en la Figura 7, la comparación entre los resultados del modelo

numérico y los datos medidos dan un ajuste adecuado. Además, la Figura 7 muestra la

evolución de la concentración de la clorofila-a en las siete estaciones de muestreo y la

concentración media para toda la laguna en los periodos de calibración. Respecto a la

concentración de clorofila-a, el mejor ajuste se ha obtenido en la estación A2.

Figura 7. Comparación entre los resultados obtenidos por el modelo (líneas) y los datos de campo

(puntos) en las estaciones de muestreo y en toda la laguna para cada periodo de calibración.

41

Resumen en español

3.3.3 Validación del modelo

Una vez calibrado, el modelo numérico fue validado en los meses que no se habían tenido

en cuenta en la calibración (Noviembre, Diciembre, Enero, Marzo, Abril, Junio, Agosto

y Septiembre), calculándose una serie de errores en cada estación. Como puede

observarse en la Figura 8, la estación C2 es la que presenta un mejor ajuste, con un error

relativo medio de 1.02%, mientras que el peor ajuste ocurre en A1, con un error relativo

medio de 47.74%. Además, todas las estaciones excepto A1, A2 y C1 tienen errores

relativos medios por debajo de 16 %, que es un valor aceptable para la validación espacial.

Figura 8. Evolución de la concentración de clorofila-a simulada (línea continua) y los datos observados

(puntos negros) en cada estación de muestreo, para los periodos de calibración del año hidrológico

2005/2006.

42

Resumen en español

En lo que concierne a la validación, la Figura 9 muestra la comparación entre los datos

promediados espacialmente, calculados y observados, para el año hidrológico 2005/2006,

en los meses donde se ha llevado a cabo la validación. El error relativo global obtenido

tiene un valor de 5.81%, lo cual indica que el modelo es capaz de reproducir

adecuadamente el comportamiento del sistema.

Figura 9. Evolución de la concentración de clorofila-a promedio para la laguna calculada por el modelo

(línea) para todo año hidrológico 2005/2006 y los datos observados (puntos) para los periodos de

validación.

3.4 Resultados y discusión.

Los resultados evidencian que el modelo propuesto es una herramienta efectiva para

describir la distribución de la clorofila-a en un SECS hipereutrófico con alta resolución

espacial (ver Figura 10). Además, la distribución espacial está de acuerdo con la

evolución temporal como se muestra en las Figuras 9 y 10, donde Octubre y Abril son los

meses que presentan unas concentraciones de clorofila-a mayores, mientras que Febrero

y Marzo son los que presentan las menores concentraciones. Esto es coherente con Romo

et al. (2008), que encontraron los menores niveles de clorofila-a en la Albufera de

Valencia en Febrero y Marzo y los mayores en Octubre y Abril. Además, los valores de

43

Resumen en español

concentración de clorofila-a calculados en este estudio son consistentes con los resultados

obtenidos por Romo et al. (2005) con valores medios máximos comprendidos entre 200

y 250 µg L-1.

El aumento de la concentración de clorofila-a en Octubre es producido por la cálida

temperatura del agua (23 ºC), el hidrodinamismo, y las altas cargas de nutrientes durante

ese periodo. La cosecha del arroz se produce entre Septiembre y Octubre, por lo tanto una

gran cantidad de nutrientes entran al lago durante esa época (Onandia et al., 2014),

favoreciendo el proceso de eutrofización. El mismo comportamiento fue observado por

Menéndez et al. (2002) en la laguna Buda (España). En abril y Mayo un gran número de

nutrientes llegan a la Albufera procedentes de fertilizantes y pesticidas utilizados en la

preparación de los campos de arroz que rodean al sistema. Febrero y Marzo por otra parte

presentan una menor concentración de clorofila-a. En Febrero, la mayoría de las

compuertas están abiertas, por lo que la hidrodinámica del SECS aumenta

considerablemente. En Marzo también se renueva el agua, y la entrada de cargas de

nutrientes es menor que en Febrero.

Además, esta reducción de nutrientes produce un aumento de la especie zooplanctonica

Daphnia magna, que es un crustáceo Cladocero responsable de una gran parte de la

consumición del fitoplancton durante ese periodo (Romo et al., 2005; Onandia et al.

2015). Como consecuencia de esto, el efecto del herbivorismo del zooplancton aumenta,

y se produce una “fase clara” en ese periodo, dando lugar a la menor concentración de

clorofila-a de todo el año.

En vista a esos resultados, la conexión de la laguna con el mar, el herbivorismo del

zooplancton y las cargas de entrada de nutrientes afectan directamente a la eutrofización

de esta laguna costera altamente regulada, dando lugar a un SECS hipereutrófico. Para

evaluar la influencia del SRP en la evolución de la clorofila-a, se ha llevado a cabo un

44

Resumen en español

balance de masas, que determina las cargas de SRP que entran y salen, y la cantidad neta

que se queda en el sistema cada mes.

Figura 10. Distribución espacial de clorofila-a en la Albufera de Valencia en el año hidrológico

2005/2006.

45

Resumen en español

El análisis del balance neto de masas de los contaminantes que entran en el sistema menos

los que salen del mismo dan una medida de como el SECS está acoplado con los sistemas

adyacentes, como en este caso con el Mar Mediterráneo. El balance de masas revela que

la mayor parte del SRP que entra en el sistema permanece en él. La carga total de SRP

que entra en la Albufera es de 26.4 t año-1, y la de salida es de 5.8 t año-1. Estos resultados

son similares a los de Burger et al (2008) en el Lago Rotorua, que es un lago altamente

eutrófico que tiene una carga de SRP de entrada de 27.5 t año-1.

Utilizando el método descrito anteriormente, se obtuvo que Abril es uno de los meses con

mayor carga de entrada de SRP (ver Figura 11), aumentando por tanto su concentración

de clorofila-a debido tanto a la alta carga de entrada de SRP, como a la cálida temperatura

del agua. Durante el otoño, la carga de SRP que entra en el lago es también elevada (ver

Figura 11). En Octubre tanto la concentración de SRP como el caudal de entrada de las

acequias de regadío aumenta, haciendo que sea uno de los meses con mayores cargas de

entrada de SRP. Como puede verse en la Figura 11, en todos los meses del año

hidrológico, la carga de entrada de SRP ha sido considerablemente mayor que la de salida,

especialmente en Octubre y Abril, donde no había flujo de salida al estar las golas casi

completamente cerradas. Además, en estos meses, la entrada de carga de SRP fue

considerablemente mayor que en otros meses, y es igual a la carga neta acumulada en la

Albufera en dichos periodos, lo cual produce serios problemas de eutrofización.

Por otra parte, es importante resaltar que en la Albufera de Valencia la carga de entrada

de SRP que viene de las acequias de regadío es aproximadamente el 35 % de la carga

total de SRP que entra en la columna de agua, mientras que el flujo de fósforo desde el

sedimento a la columna de agua constituye el 65 % (IHCantabria, 2009). Por tanto, se

puede concluir que los blooms de fitoplancton están directamente afectados no solo por

la temperatura, sino también por la carga de nutrientes de entrada, el flujo de fósforo

soluble reactivo desde el sedimento y la conexión del sistema con el mar.

46

Resumen en español

Figura 11. Balance de masas del fósforo soluble reactivo (SRP) que entra y sale de la Albufera.

Finalmente, los valores simulados por el modelo están correlacionados positivamente con

los valores medidos, con un coeficiente de correlación de Pearson de 0.93 para los

periodos de calibración y de 0.92 para los periodos de validación. Además, también se

calculó el coeficiente de eficiencia de Nash-Sutcliffe (Moriasi et al., 2007), que arrojó un

valor de 0.96, que es considerado excelente de acuerdo a Usaquen et al. (2012).

3.5 Conclusiones

En este capítulo se ha desarrollado un modelo simplificado bidimensional para SECS

hipereutróficos. Este modelo se ha aplicado satisfactoriamente a la Albufera de Valencia,

un sistema hipereutrófico cuya conexión con el mar está altamente regulada. La

concentración de clorofila-a en la Albufera de Valencia fue utilizada para calibrar y

validar el modelo propuesto en distintos periodos, así como para evaluar la sensibilidad

del mismo. Tras el análisis de sensibilidad, se puede concluir que los parámetros para los

cuales la concentración de clorofila-a en el modelo es más sensible son: Kr, Cg, Gmax, Fs

y acp. Siendo la tasa de respiración endógena del fitoplancton, Kr, el parámetro dominante

que más afecta a la concentración de clorofila-a, ya que afecta directamente al crecimiento

del fitoplancton.

0

1

2

3

4

O N D J F M A M J J A S

SRP

(t/m

on

th)

IN

OUT

NET

47

Resumen en español

El modelo nos proporciona una mayor comprensión del comportamiento del sistema. De

hecho, los resultados del modelo concluyen que los blooms de fitoplancton en Abril y

Octubre, no se deben solamente a la temperatura, sino también a las cargas de nutrientes

y a la conexión entre la laguna y el mar en dichos periodos. Además, el modelo es capaz

de reproducir la existencia de una “fase clara” en torno al mes de Marzo, que se debe

principalmente a la reducción de nutrientes, a los cambios hidrodinámicos y al efecto del

herbivorismo del zooplancton.

Como se demostró en el balance de masas, las cargas de entrada en el sistema son mayores

que las de salida, por lo que la limitada conexión con el mar magnifica la eutrofización

del sistema. Además, el flujo de SRP desde el sedimento a la columna de agua contribuye

a mantener la alta concentración de clorofila-a.

Por otra parte, un análisis estadístico cuantitativo fue aplicado para calcular la

incertidumbre y eficiencia del modelo, obteniéndose valores excelentes que demuestran

que un modelo simplificado puede caracterizar la eutrofización en un SECS

hipereutrófico.

Los resultados confirman que el modelo es una herramienta válida para la gestión de la

eutrofización en SECS hipereutróficos altamente regulados como la Albufera de

Valencia, siendo capaz de describir, con alta resolución temporal y espacial, y bajos

requerimientos computacionales, la evolución de la concentración de clorofila-a durante

un año hidrológico completo.

48

Resumen en español

Capítulo IV. Desarrollo e implementación de un sistema de

modelado ecológico para sistemas costeros eutróficos semi-

encerrados. Aplicación a un estuario alimentado por aguas

subterráneas con vegetación acuática sumergida.

La eutrofización en SECS ha provocado numerosos cambios ecológicos, entre los que se

incluye la desaparición de praderas marinas. Una causa potencial de esta pérdida es la

reducción de la disponibilidad de luz debido a la creciente atenuación de la luz en el agua

por el fitoplancton. El cambio climático y el consecuente aumento del nivel del mar

también tenderán a reducir la penetración de la luz en el agua y a modificar el hábitat de

zostera marina.

Por todo ello, se ha integrado un modelo de irradiancia espectral dentro de un modelo

biogeoquímico acoplado al Sistema de Modelado Regional Oceanográfico (ROMS), que

se conectaron a su vez a un modelo bio-óptico capaz de calcular la producción y

respiración de la zostera marina en el agua, para poder evaluar y predecir el hábitat

potencial de zostera marina en SECS hipereutróficos.

El sistema de modelado se aplicó a West Falmouth Harbor, que es un estuario poco

profundo semi-encerrado, localizado en Cape Cod (Massachusetts) donde los nitratos que

llegan al estuario a través de aguas subterráneas y su limitada pero permanente conexión

con el mar ha causado la eutrofización del sistema y la consecuente pérdida de praderas

de zostera marina. Para la calibración y aplicación del sistema de modelado se realizaron

medidas de campo de clorofila-a, turbidez, atenuación de la luz, y cobertura de zostera

marina durante un verano completo. La concentración media de clorofila-a medida varió

desde 28 µg L-1 en las zonas más interiores del estuario a 6.5 µg L-1 en las zonas más

exteriores, mientras que la atenuación de la luz se movió en un rango de 0.86 a 0.45 m-1.

El modelo reproduce la variabilidad espacial de la clorofila-a y de la atenuación de la luz

en el agua con un error cuadrático medio de 3.72 µg L-1 y 0.07 m-1 respectivamente. Se

49

Resumen en español

simularon también distintos escenarios para estudiar el efecto de una futura reducción de

nutrientes y del aumento del nivel del mar en el estuario, y los resultados revelaron que

con una reducción de un 75% de la carga de nitratos se conseguiría una mejoría

considerable de las condiciones lumínicas en el agua. Este sistema de modelado puede

ser útil para evaluar la variación de clorofila-a y del hábitat potencial de zostera marina

desde la perspectiva de las condiciones lumínicas y de la atenuación de la luz en el agua,

así como para describir la variabilidad temporal y espacial del sistema. En la Figura 12

se puede ver un resumen gráfico de los temas abordados en este apartado. Una mayor

descripción del sistema de modelado, así como de la problemática bajo estudio, de la

evaluación del modelo y de sus resultados puede ser encontrada en el Chapter IV de la

presente Tesis.

Figura 12. Resumen gráfico del capítulo IV

50

Resumen en español

4.1 Métodos y resultados del muestreo

Se tomaron datos de campo en West Falmouth Harbor para medir datos meteorológicos,

hidrodinámicos, de calidad del agua y de condiciones lumínicas durante el verano de 2012

en tres puntos de control (Outer, Snug y South) (ver Figura 2). Los datos meteorológicos

fueron medidos a intervalos de 1 minuto por una estación meteorológica Onset desde el

28 de Junio de 2012 hasta el 11 de Septiembre de 2012. Los parámetros incluidos fueron

la dirección del viento, la velocidad del viento, la presión atmosférica, la humedad

relativa, la radiación de onda corta, el PAR y la temperatura del aire. En cuanto a los datos

hidrodinámicos, de luz y de calidad del agua dentro de la bahía, se instaló una plataforma

subacuática que consistía en un Nortek Aquadopp ADCP (para medir velocidad del agua),

un SeaBird SeaCat (para medir presión), una YSI 6600 multisonda (para la salinidad,

temperatura, clorofila-a, turbidez, oxígeno disuelto), y dos sensores WetLabs ECO-

PARSB (para medir el PAR). Todos los sensores se colocaron a 0.3 m de profundidad

medidos desde el fondo (mab, meters above bed) excepto el de el PAR superior que está

colocado a 0.8 mab. Los valores de clorofila-a fueron obtenidos por un sensor YSI 6025

localizado en la multisonda YSI 6600. Las medidas se tomaron en intervalos de 5 minutos

del 3 al 19 de julio 2012 en Outer Harbor, del 19 de julio 2012 al 9 de Agosto de 2012 en

South Cove, y del 9 al 27 de agosto de 2012 en Snug Harbor. Debido al hecho de que no

se observaron cambios intra-estacionales entre julio y agosto durante una campaña de

campo previa en 2010 (Ganju et al., 2011), los datos tomados en cada localización fueron

considerados representativos de la estación de verano en cada punto de muestreo. Por otra

parte, se calculó el coeficiente de atenuación difusa de la luz en el agua, Kd, siguiendo las

formulaciones de Gallegos et al. (2011) (ver Eq. 34, Chapter IV).

Para determinar la extensión superficial de las praderas de zostera marina en West

Falmouth Harbor se realizaron campañas de campo durante Junio de 2012 utilizando un

escáner sonar. Se remolcó un EdgeTech, Inc. 4125 y se utilizó un EdgeTech Discover

software para tomar datos acústicos a 900 y 1250 kHz a lo largo de transectos espaciados.

51

Resumen en español

Las posiciones fueron proporcionadas por un Trimble AgGPS 132 con correcciones

diferenciales por radio de la U.S. Coast Guard. Los datos se preprocesaron con un

Chesapeake Technology Inc. SonarWiz5, se georeferenciaron y se exportaron las

imágenes georeferenciadas a ArcGIS 10.1 para su clasificación. Se delinearon las

praderas marinas manualmente en ArcGis tras examinar y verificar las zonas. Finalmente

se validó la interpretación de las imágenes y se verificó la extensión final de praderas de

zostera marina mediante un muestreo adicional en una serie de localizaciones aleatorias

elegidas al azar. Los flujos de agua subterránea y sus concentraciones asociadas de

nitratos fueron obtenidas en estudios previos (Kroeger et al., 2006; Ganju et al., 2012;

Hayn et al., 2014).

Finalmente, el análisis de los datos de campo reveló que todos los valores medios de las

principales variables de calidad de agua, tales como temperatura, salinidad, pH, oxígeno

disuelto y turbidez, fueron similares espacialmente, sin encontrarse grandes variaciones

a lo largo del estuario, excepto para la clorofila-a, que fue más alta en Snug Harbor. Las

medidas de clorofila-a sugirieron una mayor eutrofización en las zonas interiores de la

bahía, con un valor medio de 28 µg L-1, mientras que en la bahía exterior la concentración

media de clorofila-a fue de 6.5 µg L-1. En consecuencia, las medidas de irradiancia han

demostrado que Snug presentaba valores de PAR considerablemente menores que en

Outer. Hay que tener en cuenta además, que los datos de PAR de South Harbor no se

pudieron obtener debido a un mal funcionamiento del instrumento medidor. El coeficiente

de atenuación difuso de la luz en el agua, Kd, presentó un valor medio de 0.45 m-1 en

Outer y 0.86 m-1 en Snug. La distribución estadística de las medidas de clorofila-a y de

Kd en Outer y Snug confirma que el mayor coeficiente de atenuación de la luz en el agua

en Snug es consistente con la elevada concentración de clorofila-a (ver Figura 13).

Además, medidas anteriores confirman que el CDOM es espacialmente uniforme y

relativamente bajo en West Falmouth Harbor (M. Hayn, no publicado). Estas medidas

también indican que la turbidez es relativamente baja con diferencias espaciales mínimas.

52

Resumen en español

Por lo tanto, existe una gran relación entre la eutrofización y la atenuación de la luz en el

agua en la zona de estudio.

Figura 13. Histogramas de los datos de campo de clorofila y Kd en Outer y Snug Harbors.

4.2 Descripción resumida del modelo

Se ha integrado un modelo de irradiancia espectral (Gallegos et al., 2011) en un modelo

existente NPZD-biogeoquímico (Fennel et al., 2006) para calcular la penetración

espectral del PAR a través de la columna de agua. El modelo de irradiancia-

biogeoquímico resultante utiliza el PAR para calcular el crecimiento del fitoplancton.

Este modelo ha sido acoplado al Sistema de Modelado de Circulación Oceanográfica

Regional ROMS 3D (Haidvogel et al., 2008). Esto permite calcular la hidrodinámica, la

luz y las variables de calidad del agua tridimensionalmente en el mismo paso de tiempo.

El PAR y el Kd que se computan sirven de dato a un modelo bio-óptico (Zimmerman,

2003) que se ha conectado a los anteriores para calcular el balance de carbono de la

zostera marina en función de las condiciones lumínicas (Figura 14), el cual permite la

predicción de la presencia/ausencia de zostera marina y su potencial supervivencia.

Además, se definió con éxito un criterio para evaluar la futura evolución de las praderas

53

Resumen en español

de zostera marina en SECS eutróficos, basado únicamente en las condiciones de luz. Se

seleccionó como criterio el umbral de P/R=1 (P/R, Producción/Respiración), dado que

tanto los ecosistemas autótrofos como los heterótrofos tienden a aproximarse este valor a

lo largo del tiempo (Giddings and Eddlemon, 1978). Además, como el ecosistema bajo

estudio es autótrofo, se ha asumido que P/R>1 está asociado al éxito, supervivencia y

crecimiento de la zostera marina, mientras que P/R<1 da lugar a la desaparición de la

misma. Las capacidades fruto del enlace de estos modelos aporta una descripción integral

de las dinámicas físicas, ópticas y biológicas en la columna de agua.

54

Resumen en español

Figura 14. Diagrama de flujo del sistema de modelado y sus interacciones. En el panel inferior se

muestra que el ratio P/R>1 indica hábitat potencial de zostera marina, mientras que el P/R<1 indica

perdida potencial de hábitat de zostera marina. U y V son las velocidades, h la profundidad de la columna

de agua, T la temperatura del agua y η la variación de la superficie libre.

55

Resumen en español

4.3 Calibración del modelo

En esta sección se describe la calibración de los modelos biogeoquímico, de irradiancia

y bio-óptico, teniendo en cuenta que el modelo hidrodinámico y los flujos de agua dulce

se calibraron en un estudio previo (Ganju et al., 2012). Durante el proceso de calibración,

los parámetros del modelo se variaron para ajustarse a los valores medidos. Los

indicadores utilizados para calibrar los modelos biogeoquímico, de irradiancia y bio-

óptico fueron la concentración de clorofila-a, el coeficiente de atenuación de la luz en el

agua y la presencia/ausencia de zostera marina, respectivamente.

4.3.1 Calibración de los modelos biogeoquímicos y de irradiancia.

Se realizó un análisis de sensibilidad para evaluar la influencia de los parámetros

principales del modelo en los resultados de clorofila-a. Estos parámetros se fijaron a los

valores máximos y mínimos encontrados en la bibliografía (ver Tabla 5) obteniendo el

comportamiento observado en la Figura 15. Los parámetros que tienen mayor efecto en

los resultados de clorofila-a del modelo fueron μ0, α, gmax, mp, τ, y mz.

Figura 15. Análisis de sensibilidad del modelo biogeoquímico de Fennel et al. (2006) con un nuevo

módulo integrado de irradiancia espectral.

La calibración de la integración del modelo biogeoquímico con el de irradiancia se centró

en los tres puntos donde se obtuvieron las medidas de campo, cada uno representativo de

56

Resumen en español

Outer, Snug y South respectivamente (ver Figura 2). Por otra parte, los valores finales de

simulación de cada parámetro pueden verse en la Tabla 5, encontrándose todos ellos

dentro de los rangos de variación de la bibliografía.

Tabla 5. Parámetros principales del modelo biogeoquímico y de irradiancia y valor asignado.

SIMBOLO DEFINICIÓN VALOR

CALIBRADO

UNIDADES RANGO

µ0 Rango de crecimiento del

fitoplancton

3 d-1 0.62a-3.0b

KNO3 Concentración de semi-

saturación para el consumo de

NO3

0.1 Mmol N m-3 0.007-1.5c

KNH4 Concentración de semi-

saturación para el consumo de

NH4

1.5 Mmol N m-3 0.007-1.5 c

α Pendiente inicial de la curva P-I 0.13 Mol C gChl-1(Wm-2)-

1d-1

0.007-0.13d

gmax Ratio máximo de herbivorismo 0.6 (mmol N m-3)-1 d-1 0.5e-1.0f

Kp Concentración de semi-

saturación de ingestión del

fitoplancton

2 (mmol N m-3)2 0.56-3.5c

mp Mortalidad del fitoplancton 0.05 d-1 0.05-0.2g

Parámetro de agregación 0.005 (mmol N m-3)-1d-1 0.005-0.1c

Θmax Ratio máximo de clorofila-a a

fitoplancton

0.068 mgChl mg C-1 0.005-0.072d

mz Mortalidad del zooplancton 0.025 (mmol N m-3)-1 d-1 0.025-0.25c

RSD Ratio de re-mineralización de

detritos suspendidos

0.03 d-1 0.01-0.25h

RLD Ratio de re-mineralización de

grandes detritos

0.01 d-1 0.01-0.25h

Nmax Ratio máximo de nitrificación 0.05 d-1 0.05-0.1c

a(Taylor, 1988) b(Andersen et al., 1987) c(Lima and Doney, 2004) d(Geider et al., 1997) e(Wroblewski, 1989) f(Fasham, 1995) g(Taylor et al., 1991) h(Leonard et al., 1999).

57

Resumen en español

La comparación entre los promedios de los resultados del modelo y los datos de campo

en las estaciones de muestreo puede verse en la Tabla 6. En lo que respecta a la

concentración de clorofila-a y al coeficiente de atenuación de la luz en el agua se obtuvo

un BIAS cercano a cero en Snug y Outer. La similitud entre las desviaciones estándar

sugiere que el modelo describe apropiadamente la variabilidad del sistema en estas áreas.

Por el contrario, en South Cove, la diferencia entre los datos modelados y los resultados

de campo de clorofila-a es mayor que en los otros puntos de muestreo.

Tabla 6. Valores medios de clorofila-a y Kd, desviación estándar (Std), y BIAS para Outer, Snug y South.

Los datos de campo de clorofila-a y Kd fueron obtenidos procesando los datos de los sensores desplegados

durante el verano de 2012. Los resultados del modelo fueron obtenidos para el mismo periodo de tiempo.

CLOROFILA-A KD

Punto Media ± Std Modelo (µg/L)

Media ± Std Campo (µg/L)

BIAS (µg/L)

Media ± Std Modelo (1/m)

Media ± Std Campo (1/m)

BIAS (1/m)

Outer 6.9 ± 3.7 6.5 ± 2.8 0.41 0.45 ± 0.07 0.45 ± 0.30 -0.001

Snug 28 ± 12 28 ± 9.9 0.33 0.79 ± 0.19 0.86 ± 0.16 -0.077

South 6.3 ± 3.9 10 ± 9.3 -3.9 - - -

Finalmente, la media vertical promediada temporalmente de clorofila-a para el estuario

completo puede observarse en la Figura 16. Este cálculo se obtuvo con el procesado de

los resultados de concentración de clorofila-a tridimensionales durante todo el periodo de

estudio (Figura 16 a), para posteriormente calcular la concentración vertical promedio

(Figura 16 b). Los resultados de la simulación presentan mayores concentraciones de

eutrofización en las capas superiores (ver Figura 16 a), y también mayores niveles de

clorofila-a en Snug que en Outer y South (ver Figuras 16 a y b). Además, estos resultados

concuerdan con los datos de campo. Por lo tanto, el modelo es capaz de describir con

exactitud la concentración de clorofila-a con una alta resolución espacial.

58

Resumen en español

Figura 16. a) Variación de la concentración vertical de clorofila-a en tres capas del modelo; b) Promedio

temporal y vertical de clorofila-a en el estuario completo.

4.3.2 Calibración del modelo bio-óptico de zostera marina.

Los parámetros calibrados para el modelo bio-óptico fueron el ángulo de inclinación, la

máxima altura del manto de vegetación y la densidad de brotes. El ángulo de inclinación

seleccionado fue de 45º, representando el ángulo medio durante un ciclo mareal, y la

máxima altura de manto vegetal fue de 1 metro (Ackerman, 2002). La densidad elegida

fue de 525 brotes/m2, dado que la densidad observada de plantas varió entre 250-800

brotes/m2 en Outer (McGlathery, Marino, Hayn, y Howarth no publicado). El PAR

espectral proviene del modelo de irradiancia, y es propagado a través del manto de

vegetación hasta el fondo. Las propiedades de absorbancia y reflectancia del fondo

también fueron consideradas, siendo la composición de este una mezcla entre fango y

(a) (b)

59

Resumen en español

arena. La luz reflejada también se propagó hacia arriba a través del manto vegetal, por lo

que la producción primaria fue calculada con el total de luz absorbida. El hábitat potencial

fue evaluado en función de la distribución del ratio Producción/Respiración (P/R) (Figura

17 a) calculado a partir de los datos medios del verano completo. Se ha considerado el

ratio P/R obtenido como representativo de la estación del año bajo estudio. Además, se

ha asumido que para P/R>1 hay condiciones favorables para el crecimiento de zostera

marina, delimitando este umbral el hábitat potencial de z. marina en el estuario. Por el

contrario, para P/R≤1 se asumen condiciones desfavorables para la presencia de zostera

marina (Figura 17 b) ya que el crecimiento estaría limitado, siendo la respiración mayor

que la producción fotosintética en estas áreas. Basándose en este criterio, se ha obtenido

una correspondencia del 73.39 % entre los resultados del modelo y los datos de campo de

ausencia/presencia, teniendo en cuenta Outer y Snug (Figura 17 c), ya que en South se

estima que la zostera ha desaparecido debido a razones hidrodinámicas y perturbaciones

ambientales, y no debido a las condiciones de luz, como puede ser observado en la Figura

17 b. Del mismo modo en la Figura 17 se puede observar que la zostera no está presente

en zonas poco profundas del estuario. Esto se debe al efecto de inundación-secado

simulado por el modelo y el comportamiento sub-mareal de la zostera marina impuesto

en el modelo. Finalmente, destacar que la distribución de zostera obtenida por el criterio

P/R (Figura 17 d) está en correspondencia con la distribución de la profundidad crítica

obtenida cuando se aplica la ecuación propuesta por Duarte et al. (2007).

60

Resumen en español

Figura 17. a) Distribución del ratio Producción/Respiración; b) Distribución del ratio

Producción/Respiración aplicando el criterio de P/R>1, y comparación con los datos del campo (línea

negra continua); c) Detalle de la distribución de P/R en Snug y Outer para P/R>1 y comparación con

datos de campo (línea negra continua); d) Distribución de zostera marina obtenida con la ecuación de

limitación de profundidad (Duarte et al., 2007). El área blanca representa las zonas donde se descarta la

presencia de zostera marina (P/R<1), el área verde es el hábitat potencial de zostera marina (P/R>1) y la

línea negra continua delimita el área de presencia de zostera marina medida en la campaña de campo.

61

Resumen en español

4.4 Escenarios de carga de nitratos y subida del nivel del mar

El sistema de modelado propuesto se aplicó a West Falmouth Harbor para evaluar los

efectos de reducción de nitratos y de subida del nivel del mar en la concentración de

clorofila-a y en el hábitat potencial de zostera marina. El periodo de simulación fue de

dos meses, correspondientes a Julio y Agosto de 2012. Se simuló una disminución gradual

prevista de la carga de entrada de nitratos y una subida en el nivel del mar basada en las

predicciones del IPCC (IPCC, 2007) para los próximos cien años.

Primero, se analizaron los efectos de la reducción de nutrientes (NR) y de la subida del

nivel del mar (SLR) por separado, después se configuraron escenarios combinados (CS)

para evaluar el efecto simultáneo de ambos parámetros (Figura 18). Los resultados

apuntan a una potencial recuperación de la zostera marina en Snug basada en las

condiciones lumínicas, cuando la carga de nitratos se haya reducido un 50% (Figura 18;

NR 50). El ratio P/R mejora considerablemente en Snug con una reducción de nitratos

del 75% (NR 75). Por el contrario, la subida del nivel del mar provoca una reducción del

ratio P/R en áreas donde hay presencia de zostera marina, como puede verse en SLR

2112. No obstante, cuando ambos efectos (SLR y NR) se estudian en conjunto, se observa

una clara relación entre su comportamiento combinado (CS) y la respuesta del sistema a

la reducción de nitratos (NR). Por lo tanto, aunque la subida del nivel del mar incremente

el nivel de agua y reduzca la penetración de la luz, este factor no es tan significativo como

el de las cargas de nitrógeno. Esto es evidente comparando la variación temporal de P/R

debido al aumento del SLR y a la reducción de la carga de nitratos (Figura 19). En Snug,

la Figura 19 muestra una variación de P/R desde 0.97 hasta 1.14 debido a esta reducción,

mientras que el efecto de la atenuación de la luz debido al SLR hace decrecer P/R desde

0.97 hasta 0.94. Un efecto similar, debido al SLR, puede observarse en Outer, con una

variación de P/R desde 1.05 hasta 1.01 (Figura 19). Sin embargo, el efecto de la reducción

de nitratos es más bajo en Outer, con un P/R entre 1.05 y 1.10, debido a los menores

niveles de clorofila-a en este punto.

62

Resumen en español

Figura 18. Variación espacial de P/R debido a la reducción de nutrientes (NR), al aumento del nivel del

mar (SLR) y a los escenarios combinados (CS).

63

Resumen en español

Figura 19. Variación de P/R debida a la reducción de nutrientes y aumento del nivel del mar.

Un comportamiento similar se obtuvo en la concentración de clorofila-a, ya que la

reducción de nutrientes es en general el factor principal que provoca una disminución en

los niveles de clorofila-a en sistemas eutróficos, mientras que el aumento del nivel del

mar afecta a la eutrofización pero de una manera menos relevante (ver Figura 20). De

hecho, el descenso de clorofila-a promedio debido a la reducción de nutrientes en el

escenario final de NR es del 80%, mientras que esta disminución debida al aumento del

nivel del mar, en el escenario SLR final es solo del 24% (ver Figura 20). Este

comportamiento puede también observarse en la Figura 21, donde se presenta la

distribución espacial de clorofila-a para los diferentes escenarios (ver Table 4.5 Chapter

IV para una explicación de la nomenclatura de los escenarios). La figura presenta una

influencia dominante de la reducción de nutrientes en los escenarios combinados, y

también una relación inversamente proporcional entre la distribución de clorofila-a y la

presencia de zostera marina, que puede ser observada comparando las Figuras 18 y 20,

ya que la clorofila-a es uno de los principales factores que limitan la disponibilidad de

luz.

64

Resumen en español

Figura 20. Variación de clorofila-a debido a la reducción de nitratos y al SLR.

Adicionalmente, el efecto combinado del aumento del nivel del mar ha contribuido a una

disminución significativa de concentración de clorofila-a y Kd, especialmente en Snug,

con una reducción desde 27.77 µg L-1 hasta 2.79 µg L-1 y desde 0.79 m-1 hasta 0.36 m-1

respectivamente (Figura 22). En consecuencia, P/R en Snug crece desde 0.97 hasta 1.09,

proporcionando las condiciones adecuadas para el crecimiento de zostera marina a partir

de CS_3. Por otra parte, P/R en Outer incrementa ligeramente hasta CS_4, donde alcanza

un valor máximo de 1.07, y decrece hasta 1.05 en CS_5 debido a la influencia del aumento

del nivel del mar. También se obtuvo una evolución del área potencial de zostera marina

en el estuario para diferentes escenarios combinados (Figura 22). Se obtuvo un

incremento del 8% en CS_1, teniendo un crecimiento acumulado del 21% y del 34% en

CS_2 y CS_3 respectivamente. En el caso de CS_4 y CS_5 la influencia del aumento del

nivel del mar hace que la evolución sea más lenta, obteniéndose un incremento del área

desde CS_4 hasta CS_5 de solamente un 3%, teniendo en CS_5 un crecimiento

acumulado de área del 45% con respecto al escenario original (CS_0). Por lo tanto, los

resultados muestran que en este sistema las reducciones de carga de nitratos tendrán más

influencia que el aumento del nivel del mar. No obstante, en otros sistemas con baja carga

de nitratos el aumento del nivel del mar puede ser más relevante.

65

Resumen en español

Figura 21. Variación espacial de clorofila-a debido a la reducción de nutrientes (NR), aumento del nivel

del mar (SLR) y escenarios combinados (CS).

66

Resumen en español

Figura 22. Clorofila-a, Kd, P/R y variación del área de z. marina para los escenarios combinados.

4.5 Discusión

Los resultados de este estudio soportan la idea de que cuando llega una cantidad de luz

insuficiente a la superficie de la pradera marina la presencia y/o extensión de esta pradera

disminuye, lo cual está de acuerdo con otros estudios previos, por ejemplo: Dennison,

1987; Orth and Moore, 1988; Duarte, 1991; Duarte et al., 2007. Además, aunque hay

relativamente pocos ejemplos de praderas de zostera marina que se hayan recuperado

siguiendo una política únicamente de reducción de nitratos (Burkholder et al. 2007), los

resultados obtenidos en este estudio sugieren que con una reducción progresiva de

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

CS_0 CS_1 CS_2 CS_3 CS_4 CS_5

Kd

(m-1

)

Scenarios

Outer Snug

0.00

5.00

10.00

15.00

20.00

25.00

30.00


Chlo

rophy

ll-a

(µ

gL

-1)

Scenarios

Outer Snug

0.90

0.93

0.95

0.98

1.00

1.03

1.05

1.08

1.10

1.13


P/R

Scenarios

Outer Snug

0.00

10.00

20.00

30.00

40.00

50.00

60.00


Sea

gra

ss A

rea

Var

iati

on (

%)

Scenarios

Acumulated

Variation per scenario

67

Resumen en español

nitratos, Snug Harbor podría recuperarse desde una perspectiva estrictamente de mejora

de sus condiciones lumínicas, ya que la concentración de clorofila-a, debida

principalmente a la presencia de fitoplancton en sistemas con problemas de eutrofización,

es uno de los principales factores que atenúan la luz en el sistema. Otros factores como la

competición de las macroalgas y los cambios morfodinámicos también podrían tenerse

en cuenta. Por ejemplo, en South Cove, podrían estudiarse efectos como la competición

con macroalgas oportunistas, o la muerte de la zostera marina por perturbaciones

naturales.

La recuperación potencial de zostera marina en Snug Harbor podría ser posible debido al

hecho de que las presiones antropogénicas que afectan a esa zona principalmente se basan

en aportes de nitratos que podrían reducirse, y a que la cobertura de macroalgas es mínima

en esa zona. Además, también se obtuvo que la concentración de nutrientes y por lo tanto

la eutrofización son procesos que controlan la distribución de la zostera marina en el

estuario estudiado, debido a la atenuación de la luz producida por esta eutrofización, lo

cual está de acuerdo con Burkholder et al. (2007) y con Costa (1988). Sin embargo,

aunque la distribución de la zostera marina es altamente sensible a la eutrofización y a la

atenuación de la luz, también está afectada por otros factores, tales como la hipoxia, el

crecimiento de epifitos, el herbivorismo, y la interacción entre la hidrodinámica y la

vegetación, que no están incluidos en este modelo ya que tiene ciertas limitaciones.

La aproximación de modelado propuesta resuelve el patrón espacial del hábitat potencial

de la zostera marina desde el punto de vista de las condiciones lumínicas, y de la

influencia que la eutrofización tiene en las mismas. En contraste a la mayoría de los

modelos ecológicos existentes, esta implementación acoplada computa la atenuación

espectral de la luz como una función de diferentes sustancias atenuantes. Además, como

el modelo irradiancia-biogeoquímico que se ha acoplado está integrado en ROMS, que es

un sistema de modelado flexible de código abierto, esta técnica puede ser utilizada en un

amplio rango de aplicaciones. Por ejemplo, podría servir para estudiar la influencia de la

68

Resumen en español

resuspensión y del transporte de sedimentos en la disponibilidad de luz. Otra posible

aplicación sería analizar las variaciones de las condiciones lumínicas debidas al CDOM

producido por riadas y escorrentías. Todo esto es posible debido a que el ROMS está

acoplado al Sistema de Modelado de la Comunidad de Transporte de Sedimentos

(CSTMS) por lo que la interacción dinámica entre sedimentos y disponibilidad de luz

podrían ser modelados con esta implementación.

4.6 Conclusiones

En el presente estudio se ha desarrollado un nuevo sistema de modelado para SECS

eutróficos, que es capaz de describir la concentración de clorofila-a, la disponibilidad de

luz y la recuperación potencial de comunidades de zostera marina bajo escenarios futuros

de cargas de nutrientes y de subida del nivel del mar desde el punto de vista de las

condiciones lumínicas. Se ha evaluado el modelo en un SECS eutrófico, y se ha descrito

la variabilidad espacial de la clorofila-a, la atenuación de la luz, y la potencial

presencia/ausencia de zostera marina.

La implementación acoplada computa la atenuación espectral de la luz en el agua como

una función de diferentes sustancias atenuantes con alta resolución vertical y horizontal,

lo que permite la determinación con exactitud del ambiente lumínico en las praderas de

zostera marina. Se encontró que, en general, el aumento del nivel del mar reducirá la

disponibilidad de la luz en el agua, y se espera que esto tenga un impacto negativo en las

praderas de zostera marina; presentándose una reducción del 11.4% en el área de

presencia de zostera marina con un aumento del nivel del mar de 0.35 m. Sin embargo,

en el SECS estudiado, la reducción de la carga de nutrientes es el factor que mayor efecto

tiene en la mejora de la disponibilidad de luz, ya que la eutrofización debida a la carga de

nutrientes es uno de los principales problemas del sistema. De hecho, los resultados

obtenidos muestran que el hábitat de zostera marina se expande en un 42.3 % con un 94

% de reducción de carga de nitratos. Este estudio contribuye a los esfuerzos de modelado

69

Resumen en español

de eutrofización existentes aportando una nueva implementación que es capaz de evaluar

el hábitat potencial de zostera marina en términos de disponibilidad de luz con

formulaciones sencillas, computando resultados fácilmente interpretables con un alto

nivel de exactitud.

70

Resumen en español

Capítulo V. Conclusiones y futuras líneas de investigación

Según el objetivo general y los objetivos específicos establecidos, se desarrollaron dos

modelos para predecir el comportamiento de SECS hipereutróficos y eutróficos

respectivamente. Para evaluar las habilidades de las herramientas desarrolladas fueron

utilizados datos de campo procedentes de campañas de las zonas bajo estudio. Los

resultados obtenidos permiten la extracción de las siguientes conclusiones con respecto a

las características de los modelos, al análisis de sensibilidad, a las formulaciones de luz,

a la calibración y a los resultados obtenidos.

5.1 Conclusiones generales

Los modelos ecosistémicos simplificados con alta resolución espacial pueden

caracterizar la eutrofización en sistemas costeros semi-encerrados, debido a

la alta variabilidad que presenta la distribución del fitoplancton en estos

sistemas. Por lo tanto, la resolución espacial es uno de los factores principales

a tener en cuenta en el diseño de un modelo de eutrofización para SECS.

Los modelos ecosistémicos requieren un mayor nivel de complejidad para

describir SECS eutróficos que hipereutróficos, dado que en el primer caso

deben tenerse en cuenta un mayor número de procesos al tener estos una más

difícil simplificación. Por ello, los SECS hipereutróficos pueden ser descritos

por modelos con un nivel inferior de parametrización que en el caso de los

eutróficos.

Los modelos ecosistémicos simplificados con una demanda de datos no tan

alta como los tradicionales pueden ser herramientas eficaces para calcular

fácilmente resultados interpretables con un alto nivel de exactitud si las

71

Resumen en español

ecuaciones principales y consiguientes variables y parámetros están bien

seleccionados y definidos.

El nuevo modelo simplificado desarrollado en la presente Tesis fue capaz de

caracterizar la eutrofización en un SECS hipereutrófico fuertemente regulado,

que es uno de los casos existentes más complejos, describiendo con alta

resolución espacial y temporal la evolución de la concentración de clorofila-

a durante un año.

El acoplamiento de modelos es una herramienta eficaz y flexible para

describir procesos ecológicos específicos en sistemas complejos. De hecho,

este estudio contribuye a los esfuerzos de modelado existentes

proporcionando una nueva implementación, acoplamiento y conexión de

modelos que no sólo describe la concentración de clorofila-a, sino también el

hábitat potencial de zostera marina en términos de la disponibilidad de luz en

SECS eutróficos.

El modelo hidrodinámico es crítico para el correcto desempeño de los

modelos ecológicos y debe describir con precisión el comportamiento del

sistema. Una adecuada representación de la hidrodinámica es fundamental

para el análisis de SECS, debido a su singular conexión con los sistemas

adyacentes. Por lo tanto, los modelos y condiciones hidrodinámicas han sido

cuidadosamente seleccionadas en el presente estudio.

5.2 Conclusiones del análisis de sensibilidad

Los parámetros para los cuales la clorofila-a es altamente sensible en el

sistema hipereutrófico estudiado son la tasa de respiración endógena del

fitoplancton, la tasa de herbivorismo del zooplancton, la tasa de crecimiento

72

Resumen en español

de fitoplancton, el flujo de fósforo soluble reactivo del sedimento a la columna

de agua y la proporción de carbono-clorofila-a. Mientras que en el sistema

eutrófico fueron la tasa de crecimiento del fitoplancton, la pendiente inicial de

la curva que relaciona la fotosíntesis con la intensidad de luz para el

fitoplancton (curva PI, Photosynthesis-Light Intensity curve), la tasa de

herbivorismo del zooplancton, la muerte de fitoplancton, el parámetro de

agregación, y la tasa de mortalidad del zooplancton. Se puede concluir por

tanto, que en ambos casos los parámetros más influyentes son los relacionados

con el crecimiento y muerte del fitoplancton y del zooplancton, menos en el

estuario eutrófico, donde la pendiente inicial de la curva PI parece tener mayor

relevancia ya que los efectos de la luz siguen siendo importantes en los

sistemas eutróficos.

El análisis de sensibilidad ha demostrado a su vez ser una herramienta valiosa

para reducir el número de parámetros que deben ser ajustados en los modelos.

5.3 Conclusiones del modelado de luz

La disponibilidad de luz en SECS eutróficos e hipereutróficos está

influenciada principalmente por la concentración de clorofila-a.

Las condiciones lumínicas podrían mejorarse reduciendo la carga de

nutrientes en SECS eutróficos con eutrofización cultural, mientras que la

restauración de los sistemas hipereutróficos requiere un análisis más complejo

que involucra también las condiciones hidrodinámicas y los flujos de

sedimentos.

La implementación de la atenuación espectral de la luz en el agua en modelos

ecosistémicos en función de diversas sustancias, con alta resolución horizontal

73

Resumen en español

y vertical, permite una determinación de las condiciones lumínicas con mayor

exactitud que las formulaciones tradicionales. En este estudio, se han

implementado las formulaciones de atenuación espectral de la luz de Gallegos

et al. (2011) en el modelo de Fennel et al. (2006) que está acoplado al

ROMS. Esta configuración permite ser conectada con un modelo bio-óptico

de praderas marinas, como por ejemplo el modelo de Zimmerman (2003) ya

que este tipo de modelos necesita el PAR espectral como dato de entrada.

El presente estudio revela que un modelo tridimensional con formulaciones

de atenuación espectral de la luz es necesario en un SECS eutrófico para

describir con exactitud la heterogeneidad y el ambiente lumínico del sistema,

mientras que un modelo bidimensional con atenuación no-espectral fue

utilizado con éxito en un SECS hipereutrófico debido a la baja penetración de

la luz a través de la columna de agua en estos sistemas.

5.4 Conclusiones de calibración y resultados

La incertidumbre promedio de la predicción del modelo hipereutrófico fue de

menos del 6%, con dos coeficientes de correlación de Pearson de 0.93 y 0.92

para la calibración y validación respectivamente, y un coeficiente de eficacia

de Nash-Sutcliffe de 0.96, que son valores excelentes.

El sistema de modelado eutrófico reproduce con precisión la variabilidad

espacial de la clorofila-a y la atenuación de la luz con errores RMS de 3.72 µg

L-1 y de 0,07 m-1 respectivamente.

Se ha propuesto un criterio de hábitat potencial de zostera marina basado en

el ratio entre Producción/Respiración (P/R). Se ha asumido que para áreas

donde P/R>1 existe crecimiento de vegetación acuática sumergida y por lo

74

Resumen en español

tanto presencia potencial de la misma, delimitando este umbral el hábitat

potencial de zostera marina en el estuario. Sin embargo, para zonas donde

P/R≤1 asumimos que las condiciones son desfavorables para la presencia de

praderas zostera marina ya que el crecimiento sería limitado, siendo la

respiración más grande que la producción en esas áreas. Los resultados

obtenidos con el criterio de P/R seleccionado para evaluar el hábitat potencial

de zostera marina presentan una coincidencia de un 73.39% entre los

resultados modelados y los datos de campo.

Se ha contribuido a una mejor comprensión de los impactos del cambio

climático en SECS eutróficos. Se ha encontrado que, en general, el aumento

del nivel del mar reducirá la disponibilidad de luz y se espera que esto tenga

un impacto negativo en las praderas de zostera marina. En West Falmouth

Harbor el modelo muestra una reducción de 11.4% de hábitat potencial de

zostera marina con un aumento en el nivel del mar de 0.35 m.

Se ha concluido que la reducción de nutrientes puede ser el factor principal

necesario en la restauración de sistemas eutróficos, mientras que en los

hipereutróficos la solución es más compleja. En los SECS eutróficos

estudiados en esta Tesis, la reducción de la carga de nitratos es un factor

principal en la mejora de la disponibilidad de luz. La concentración de

clorofila-a se reduce un 89.3% mientras que el hábitat zostera marina se

expande en un 42.3%, con una reducción de un 94% de la carga de nitratos.

Según lo demostrado por el balance de masas calculado, en un SECS

hipereutrófico las cargas de entrada pueden ser superiores a las cargas de

salida, y la limitada conexión con el mar magnifica la eutrofización del

sistema. Además, el flujo SRP del sedimento a la columna de agua contribuye

75

Resumen en español

al mantenimiento de concentraciones elevadas de clorofila-a en el área

estudiada. Por lo tanto, en SECS hipereutróficos, la reducción de nutrientes

podría no tener un impacto significativo en la restauración del sistema si no

cambian las condiciones hidrodinámicas y las características de los

sedimentos del mismo.

Un modelo bio-óptico se ha conectado con éxito al modelo de eutrofización y

al ROMS, demostrando la flexibilidad de este sistema para describir procesos

específicos, como la relación entre la producción y la respiración de la

vegetación acuática sumergida.

5.5 Futuras líneas de investigación

Esta Tesis ha puesto de manifiesto algunas limitaciones en los modelos desarrollados que

abren nuevas líneas de estudio. A continuación se mencionan los aspectos más relevantes

de la Tesis que necesitan una futura investigación.

Los modelos desarrollados podrían aplicarse a otras zonas de estudio con

datos de campo disponibles. Esto sería útil para consolidar la eficacia y la

utilidad de las herramientas de modelado desarrolladas.

Del mismo modo, se plantea la conveniencia de diseñar y evaluar estrategias

de gestión factibles que puedan mejorar el estado de la Albufera de Valencia,

teniendo en cuenta tanto las características del sistema como los factores

económicos que indicen en la zona de estudio, tales como el cultivo del arroz,

las actividades industriales y el turismo entre otros.

También sería interesante implementar el modelo simplificado hipereutrófico

en el ROMS para expandir la técnica desarrollada y hacer el código abierto

76

Resumen en español

para la comunidad científica como se está haciendo con el sistema de

modelización eutrófico.

Además se podría continuar investigando la influencia de la concentración de

nitratos en el crecimiento del fitoplancton. Dicha influencia también podría

incluirse en el modelo simplificado para SECS hipereutróficos, para poder

extender su aplicación a sistemas donde el nitrógeno sea el nutriente limitante.

Convendría que futuros trabajos sobre el sistema de modelización eutrófico

incorporaran otras comunidades ecológicas como macroalgas y epifitos, para

ser capaces de estudiar la competición de las praderas marinas con macroalgas

oportunistas, y la influencia de los epifitos en la atenuación de la luz.

Otro factor que también podría ser de interés investigar en el sistema de

modelado eutrófico es la anoxia debida a la eutrofización. El contenido de

oxígeno en las praderas marinas es fuertemente dependiente de la fotosíntesis

y la respiración, que han sido computadas en el modelo como una función de

luz y temperatura. La concentración de oxígeno también podría ser calculada

por el modelo de eutrofización, y ser usada por el modelo bio-óptico, ya que

niveles bajos de oxígeno podrían limitar el crecimiento de la vegetación

acuática sumergida y su producción primaria. Por lo tanto, un término de

oxígeno debería ser incluido en el modelo bio-óptico para modificar la

formulación de la producción y respiración de las plantas basándose en la

concentración de oxígeno.

Se podría profundizar también en el estudio de la captura de partículas por

parte de la vegetación acuática sumergida y su efecto sobre las condiciones

lumínicas. Las praderas marinas disminuyen la velocidad del flujo y reducen

la turbulencia. Esto provoca la sedimentación de las partículas suspendidas en

77

Resumen en español

el agua en la zona de las praderas marinas, consiguiendo una mejora de las

condiciones lumínicas. Además, dado que en la presente Tesis se ha integrado

un modelo de irradiancia espectral en ROMS, que es un modelo de código

abierto flexible, podría evaluarse la influencia de la resuspensión y del

transporte horizontal del sedimento en la disponibilidad de luz. Esto es posible

debido a que el ROMS está acoplado al Sistema de Modelado de Transporte

de Sedimentos de la Comunidad Científica (CSTMS) por lo que la interacción

dinámica entre los sedimentos y la disponibilidad de luz podría ser modelada

con esta implementación.

Otra posible aplicación del sistema de modelado acoplado para SECS

eutróficos sería analizar la variación de las condiciones lumínicas debida a

cambios espaciales en el CDOM causados por aumentos en los flujos de los

ríos, escorrentía terrestre y/o procesos microbianos. Esto es de nuevo posible

debido al hecho de que ROMS está acoplado al CSTMS.

Otra mejora en el sistema de modelado podría ser el acoplamiento del modelo

de Zimmerman al resto del sistema para poder computar el hábitat potencial

de zostera marina al mismo tiempo que el PAR espectral y la clorofila-a, ya

que en el presente estudio el modelo de Zimmerman se ha conectado al resto

de los modelos acoplados, pero no ha llegado a acoplarse.

78

Resumen en español

79

Chapter I Introduction and research background

Chapter I

1 Chapter I. Introduction and research background

80


81


Chapter I. Introduction and research background

1.1 Motivation of the research

Semi-enclosed coastal systems (SECS) include coastal lagoons and transitional waters

(Newton et al., 2013) and are not only important ecological systems, but also have

considerable socio economic value (Lassere, 1979). In recent years, there has been

increasing recognition of the economic importance of SECS through their provision of

ecosystem services. However, these services are increasingly threatened as SECS are

among the most vulnerable coastal systems to both natural and human pressure

(Eisenreich, 2005). Moreover, the geomorphology of SECS makes them particularly

vulnerable to global changes, such as sea-level rise, increased temperatures, storminess,

droughts, floods and changes in sediment dynamics (Newton et al., 2013). Additionally,

human activities cause changes in demographics, urbanization, agriculture and land-use,

as well as industrial development and shipping that affect the structure and function of

these vulnerable and valuable coastal ecosystems. These systems are characterized by

their hydrodynamic exchange properties with the adjacent open sea, and can be classified

as open, leaky, restricted or choked (Kjerfve, 1994; Newton et al., 2013). Some examples

of different kind of SECS can be seen at Figure 1.1.

SECS are complex ecosystems characterized by a natural high spatial variability and high

productivity. They support a rich indigenous fauna and flora because they are sheltered

and, in most cases, shallow water systems of high productivity. Moreover, SECS are

important feeding and nesting sites for a multitude of bird species, as well as important

stop-over sites for bird migration (European Commission, 2009). The range of ecosystem

services provided by SECS is extensive and includes provisioning services, also known

as ecosystem goods, such as fish and shellfish; supporting services, such as oxygen

production from photosynthesis; and cultural services, such as recreation and ecotourism

(Millenium Ecosystem Assessment, 2005). The ecosystem functions of SECS provide

82


essential services including decomposition, nutrient cycling, and nutrient production.

They also function in the regulation of fluxes of nutrients, water, particles, and organisms

to and from land, rivers, and the ocean. SECS serve as buffers, sinks and transformers,

for example due to processes like sedimentation, transformation or de-nitrification.

Figure 1.1. Examples of different kind of SECS: a) Open (Bay of Wismar, Germany); b) Leaky (Venice

Lagoon, Italy); c) Restricted (Quanzhou Bay, China); d) Choked (Étang de Thau, France).

83


It is also important to remark that SECS help to protect the adjacent sea from

eutrophication and pollution, retaining riverine nutrients. However, these valuable

ecosystems are being subject to strong anthropogenic pressures due to tourism, heavy

shellfish or fish farming, wastewater discharges, pollution and urbanization. Moreover,

one of the consequences of the anthropogenic pressures on these systems has been the

loss of habitats such as seagrasses that can act as nurseries and trap particles. Therefore,

the ongoing eutrophication produced in these systems by the anthropogenic pressures and

climate change are a threat to future structure and function of SECS.

Eutrophication is the process by which a body of water acquires a high concentration

of nutrients, especially phosphates and nitrates. These typically promote excessive growth

of algae. As the algae die and decompose, the oxidation of this organic matter and the

respiration by the decomposing organisms can deplete the water of available dissolved

oxygen, causing the death of other organisms, such as fish (Art, 1993). This nutrient

enrichment may occur naturally or can be the result of human activity, then the process it

is called “cultural eutrophication”, and it is usually provoked by fertilizers runoff and

sewage discharge (Lawrence et al., 1998). Eutrophication is a natural slow-aging process

for a water body, but human activity sometimes greatly speeds up the process.

Many human activities result in water pollutants, with the main sources of eutrophication

being discharged from urban wastewater treatment and agricultural pollution. During the

last century, population growth, wastewater production and discharge from urban areas

(point sources) resulted in a marked increase in water pollution. On the one hand, in spite

of the improvements on wastewater treatment in recent years, and the optimization of

industrial production processes, pollution discharges are today sadly still related to

population growth and economic growth. On the other hand, agriculture is a key source

of diffuse pollution, as agricultural production is becoming increasingly intensive, with

high input of fertilizers and pesticides, in turn resulting in significant loads of pollutants

http://toxics.usgs.gov/definitions/nutrients.html

84


to the water environment through diffuse pollution and/or irrigation channels. All these

human activities can provoke different trophic status in an aquatic system.

Consequently, the study and determination of the trophic status of a coastal ecosystem is

an important issue due to the effects that this process has in semi-enclosed coastal

ecosystems, and because aquatic plants and animals react to changes in their environment

caused by changes in water quality. For example, it changes the function of the semi-

enclosed lagoons and embayments as buffer zones, and contribute to the proliferation of

macroalgal mats that outcompete with perennial benthic producers, such as seagrasses.

Therefore, in order to characterize the stage at which this process is at any given time in

a particular water body we have the ‘trophic status’ (see Figure 1.2). For this purpose, the

following terms are used (OCDE, 1982; Walmsley, 2000):

Oligotrophic: low nutrient levels and not productive in terms of aquatic animal

and plant life.

Mesotrophic: intermediate levels of nutrients, fairly productive in terms of

aquatic animal and plant life and showing emerging signs of water quality

problems.

Eutrophic: rich in nutrients, very productive in terms of aquatic animal and plant

life and showing increasing signs of water quality problems.

Hypertrophic: very high nutrient concentrations where plant growth is

determined and limited by physical factors. Water quality problems are serious

and almost continuous.

85


Figure 1.2. Trophic status.

Not surprisingly, the status that cause more concern are eutrophic and hypertrophic,

which are the ones that produce more severe consequences on the ecosystems.

Additionally, the processes that rule these systems have similarities and differences, and

a well understanding of these processes is important for the management and restoration

of their ecosystem services. In fact, one of the main differences between these systems is

the rule of light on the system’s productivity. In eutrophic systems the light can still

penetrate through the water column as the chlorophyll-a concentration still allows it, and

submerged aquatic vegetation such as seagrasses can live, although its growth could be

limited by light. However, in a hypertrophic system the phytoplankton and consequently

the chlorophyll-a concentration in the water surface is so high, and the system is so

impaired that the penetration of light through the water column is very limited. Therefore,

in hypertrophic systems the survival of seagrasses is unlikely as they have high light

requirements, and they are usually replaced by opportunistic macroalgae with lower light

requirements. Moreover, the overabundance of nutrients also produces a higher epiphytes

growth that also cause light attenuation (Figure 1.3).

86


Figure 1.3. Seagrass loss and pressures

Therefore, regarding seagrass growth it is important to take into account the spectral

character of light in order to accurately reproduce the photosynthesis process and the

biomass production. This is due to the fact that from the radiation that reaches the water

surface, only the ones of wavelengths between 400 and 700 nm can be used for

photosynthesis, and are called the Photosynthetic Active Radiation (PAR). However, in

the water column this light is limited first due to atmospheric processes such as ozone and

cloudiness that limit the amount of light that reaches the water surface, and then by the

water depth, chlorophyll-a and phytoplankton pigments, turbidity, and Coloured

Dissolved Organic Matter (CDOM) between others. Moreover these substances produce

light absorption, scattering and backscattering, changing the light field that reaches the

plant, and limiting their growth. Therefore, it is important to accurately determine the

light climate in systems with submerged aquatic vegetation where the habitat is

vulnerable to disappear as is the case of eutrophic systems, whereas in the case of

hypertrophic systems this processes are not so relevant due to the scarce influence of light

dick

Highlight

dick

Sticky Note

spelling: Photosynthetically

87


in the water column as most of it is absorbed at the water surface by the present high

levels of chlorophyll-a concentration and the lack of life in the bottom of these systems.

Moreover, the three-dimensional character that eutrophic systems, as they are usually

very shallow and the life on the bottom is very impaired. In these systems, the differences

between bottom and water column could be simplified assuming a well-mixed water

column, due to the fact that there is not usually stratification in the vertical direction in

these systems. However, in eutrophic systems there is still life in the bottom and complex

processes happened there, therefore it is important in this kind of system to accurately

describe the three-dimensional character of the system. However, both kind of systems

present a high horizontal variability as the horizontal gradients of the phytoplankton

distribution are usually very high as they are strongly dependent on the nutrient loading.

Another aspect that produces light attenuation in eutrophic systems is the sea-level rise

(SLR), which could modify the seagrass habitat in the long term (Duarte, 2002). When

the sea-level rises the light climate changes as the water depth increases. Therefore,

seagrass meadows distribution can change. In fact, the deeper parts of the system present

less light availability and seagrass will disappear from this areas, whereas in the shallower

areas new habitat could be created and some seagrass will migrate to those areas as can

be seen in Figure1.4. However, the new habitat creation will be limited in each system by

other factors such as physical barriers, unsuitability of the substrate, shoreline

constructions, increase of salinity, temperature and species competition (Hauxwell et al.,

2003). Therefore, although the light conditions would be good for seagrass growth, there

could be other parameters that could change the species distribution. As a conclusion,

seagrass will disappear due to sea-level rise as overall effect, although some migration is

possible.

dick

Highlight

dick

Sticky Note

there is plenty of 'life' (microbes, invertebrates etc.) in the benthos of these systems - even though they don't support SAV

dick

Highlight

dick

Sticky Note

incomplete sentence

88


Figure 1.4. Seagrass changes due to sea-level rise (SLR)

The complexity and spatial variability of these systems makes the study of the cause-

effect relationships between the different human actions and hydrographical,

hydrodynamical and ecological processes a difficult task, and complex models are widely

used for SECS independently of the level of eutrophication presented by the system (e.g:

Fulton et al., 2004; Zouiten, 2012). Therefore, there is a need of developing specific

simplified mathematical tools for the management of eutrophic and hypertrophic SECS,

taking into account the system hydrodynamics, the specific processes governing each

system and a proper space and time resolution. In recent years, coupled-linked models

have been demonstrated to be a good solution to integrate the main system processes,

simplifying the modelling approach, and optimizing computational resources (Mainardi

et al., 2015).

1.2 History of coupled-linked models in estuaries and SECS

Ecological models of aquatic systems can be traced back to the pioneering work of

Gordon A. Riley (1946, 1947). Riley’s major advancement was to formulate models in

89


which changes in daily growth and loss processes were functions of biomass, abundance

of prey and predators, and environmental variables such as temperature, light, and

nutrients. These independent models of phytoplankton (P) and zooplankton (Z) were then

combined along with a state variable for nutrients (N) to produce the first coupled NPZ

model of an aquatic system (Riley et al., 1949).

The next major phase in the application of ecological models to aquatic systems came

during the 1960s and 1970s with the development of models primarily for estuaries, shelf

ecosystems, and lakes. Early examples included Di Toro et al. (1971) model of the

Sacramento-San Joaquin Delta, Steele’s (1974) model of the North Sea, and Kremer and

Nixon’s (1978) model of Narragansett Bay. The use and complexity of models continued

to expand in the 1980s to systems such as Chesapeake Bay (HydroQual, 1987), the Baltic

Sea (Stigebrand and Wulff, 1987), and the Ems-Dollard Estuary (Baretta and Ruardij,

1988). These models added complexity, with dissolved oxygen as a state variable and

addition of dissolved organic and particulate forms of nutrients and implementation in a

relatively high resolution, three-dimensional domain. The application of models to

management began in this period, particularly related to anthropogenic nutrient

enrichment and the cultural eutrophication of estuaries (HydroQual, 1987).

Models continued to be developed in the 1990s with an increasing focus on management.

During this period, the developed ecological models included more mechanistic

relationships, a greater number of state variables, and frequently finer spatial scales, and

started to be coupled to other models (Ambrose et al., 1993; Doney et al., 1996; Bissett

et al., 1999a, b). Models also began to include benthic primary producers such as

macroalgae, seagrasses and their epiphytes, and tidal marshes (Bach et al., 1992;

Malmgren-Hansen and Warren, 1992; Vested et al., 1992; Madden and Kemp, 1996;

Solidoro et al., 1997; Buzzelli et al., 1999).

90


After 2000, models became essential management tools, and coupled models

considerably increased in order to better understand the interaction between

hydrodynamic and ecological processes in estuaries. These models included greater

complexity and processes (Fulton et al., 2004b), taking into account toxic chemical

materials (Tetra Tech, 1999), heavy metals (Hipsey et al., 2007), and submerged aquatic

vegetation (SAV) (Burd and Dunton, 2001).

As estuarine ecosystem models incorporated greater complexity and more numerous state

variables, modellers delved into the detail associated with various model components.

For example, the interaction between SAV and drag was investigated using a variety of

numerical approaches (Bouma et al., 2005; Bal et al., 2011), and is still an active area of

research. Another increasing field of interest nowadays are the benthic–pelagic coupled

ecosystem models to estimate hypoxia in estuaries, one of the most relevant recent

examples of these models is the model ECOHYM (Somha et al., 2008), which is a

coupled hydrodynamic-pelagic and benthic ecosystem model that succeeds on describing

the vertical DO profiles during hypoxic season in estuaries. For shellfish modelling,

particle uptake is a rate processes that is often a component of larger models (e.g. Cerco

and Noel, 2005, 2007; and Fulford et al., 2007). In such models, particle availability often

does not vary, although real hydrodynamic processes create varying concentrations.

Published models (Simpson et al., 2007; Petersen et al., 2013; Forsyth, 2014) have

incorporated particle variations into mussels and oyster uptake rates. Other shellfish

models that couple biology with physical processes include those simulating larval

transport (North et al., 2008; Bidegain et al., 2012), where simple rules describing larval

behaviour were shown to have a large effect on the spatial patterns of larval settlement.

All these coupled models have improved our knowledge in different processes in estuaries

in the last decades and continue under constant study and research. This is the case for

example of harmful algal blooms, although we may roughly attribute its occurrence to

excessive nutrients, there could be other mechanism that are under study as life cycle

91


(Hense, 2010), physiology (Hood et al., 2006) residence time and toxins models

(Ramanathan, 2010).

Despite the trend towards increasing model complexity and spatial resolution, there has

been a growing recognition of trade-offs among complexity, resolution, over

parameterization, and model uncertainty. Numerous investigators have addressed the role

of model complexity (Fulton et al., 2003, 2004a; Friedrichs et al., 2006; Ménesguen et al.

2007) and others have embraced the use of simpler, reduced and intermediate complexity,

and alternative modeling approaches (Rigler and Peters, 1995; NRC, 2000; Pace, 2001;

Duarte et al., 2003; Scavia et al., 2006; Swaney et al., 2008), going back to the initial

NPZD models (Fennel et al., 2006) due to its precision.

In order to have a better knowledge about the capabilities, limitations and complexity of

the most widely used models nowadays, a review with the description and main

applications of the most relevant existing models with coupled-linked proved capabilities

that can be used in SECS, estuaries, coastal areas and coastal lagoons, is presented in the

following section. In this section we have focused on hydrodynamic models, on water

quality and ecosystem models able to assess eutrophication and phytoplankton dynamics,

and on bio-optical irradiance models.

1.3 Review of hydrodynamic, water quality, ecosystem and bio-optical

irradiance models with coupled-linking capabilities: description

and applications.

Eutrophication in SECS is a complex environmental problem that can be produced by

different factors (hydrodynamic conditions, anthropogenic pressures, input loads etc.).

This problem can cause severe changes in the ecosystem such as submerged aquatic

vegetation disappearance between others, so there is a clear need for solutions for dealing

with this environmental problem. To achieve this goal, the use of models becomes

indispensable. Models are useful to provide a way to synthesize observations, theoretical

92


knowledge, and information about stocks, flows, rates, and interconnections of many

parts of the system. They are also a useful way to test scientific hypotheses, as each model

equation itself is a hypothesis about the way the system functions can be compared with

observations. Additionally, they allow for the generation of scenarios to be developed and

investigated, and reveal indirect pathways and connections between system components.

Finally, they are also useful tools for decision makers.

Due to the fact that primary producers play a relevant role in eutrophic and hypertrophic

systems, we give particular insight into the models that take into account phytoplankton

dynamics, and the interaction between light and primary producers. Due to the importance

of hydrodynamics in the transport and dispersion of pollutants a common type of model

used in coastal waters is a coupled physical hydrodynamic - water quality or ecological

model. Such models can be used to predict future conditions in a specific region to try to

fully understand the changes that have occurred or will occur, and the main human factors

and natural processes that can explain such changes. Therefore, in the next sections the

main characteristics of the physical hydrodynamic models which have coupled or linked

ecological models and the most important ecological models that could be applied to

eutrophic and hypertrophic semi-enclosed coastal systems are presented.

1.3.1 Hydrodynamic models

In aquatic systems, biological and chemical processes are, to a large degree, controlled

by physical processes. For example, rates of phytoplankton growth, zooplankton grazing,

and organic matter re-mineralization are temperature dependent, and species composition

of the biological community are affected by salinity and temperature. Recent open ocean

modeling studies have emphasized the importance of improving the representation of

physical processes and variability in order to improve the performance of biogeochemical

models (Oschlies and Garcon, 1999; Hood et al., 2003; Friedrichs et al., 2006). In fact, a

high spatial resolution and the capability of nesting or downscaling are important

93


characteristics in the selection of a physical model for its application to SECS. Due to the

importance of reproducing the correct physical conditions in a coupled physical-

biological model for modeling biogeochemical variability, the selection of the physical

model is fundamental issue in eutrophication modeling. Thus, we analyze the main

characteristics of the most widely used physical models with coupled or linked ecological

models in order to determine their suitability for its application to eutrophic and

hypertrophic SECS. The physical models we describe are ROMS, ECOM, FVCOM,

MIKE 3, MOHID, HAMSOM, DELFT3D FLOW, EFDC, ELCOM, TELEMAC, and

H2D/H3D.

1.3.1.1 Regional Ocean Modelling System (ROMS)

Description:

ROMS is a free-surface, three-dimensional, terrain-following, primitive equations model.

It includes accurate and efficient physical and numerical algorithms and several coupled

models for biogeochemical, sediment, and sea ice (Budgell, 2005) applications. The wet

and dry capabilities of ROMS have also been tested by Warner et al (2013), who proved

them to be successful. It also includes several vertical mixing schemes (Warner et al.,

2005a), and multiple levels of nesting and composed grids. ROMS is coupled to series of

biogeochemical models of increasing complexity, from simple 5-component or 7-

component NPZD models (Franks et al., 1986; Fennel et al., 2006; Gruber et al., 2006;

Powel et al., 2006;) to NEMURO, which is a lower trophic level model with 11 state

variables, PISCES with 24 components (Aumont, 2005), or ECOSIM (Bissett et al.,

1999a, 1999b) with 21 state equations and 115 parameters. It has also been successfully

linked to CE-QUAL-ICM (Kim et al., 2011), which is also a complex biogeochemical

model. Moreover, as ROMS is also coupled to the Community Sediment Transport Model

(Warner et al., 2005b; Blaas et al., 2007; Warner et al., 2008) it has also the capability to

take into account the effect of sediment dynamics on water quality .

http://www.myroms.org/includes/bib.php?id=Warner...05a


94


Applications:

The wide range of applications of ROMS coastal and ocean waters (e.g., Haidvogel et al.,

2000; Di Lorenzo, 2003; Dinniman et al., 2003; Marchesiello et al., 2003; Peliz et al.,

2003; Budgell, 2005; Warner et al., 2005a, b; Wilkin et al., 2005) and the number of

coupled systems (Franks et al., 1986; Fennel et al., 2006; Gruber et al., 2006; Powel et

al., 2006; Blaas et al., 2007; Warner et al., 2008; Kim et al., 2011) makes it to be a

pioneering force in ocean and coastal research modelling due to the continued innovative

improvements and new developments that are continuously made by its users.

1.3.1.2 Estuarine and Coastal Ocean Model (ECOM)

Description:

The ECOM is the 3-D finite-difference, primitive equation coastal ocean model

developed based on the original code of the Princeton Ocean Model (POM) (Blumberg

and Mellor, 1987). It was developed principally by Alan Blumberg of HydroQual. It has

incorporated a wet/dry technique. It can be used to simulate tides, tide-induced current,

and salinity distribution due to mixing between freshwater and oceanic water in estuaries.

Applications:

It is good for estuaries and bays, but it is inappropriate to be used for real-time simulation

and coastal management. In particular, ECOM fails to resolve areas with barriers, and

islands. It has been applied by HydroQual to Boston Harbor, New York Harbor and

Onondaga Lake (New York) between others, and it is not generally available as open

source.

1.3.1.3 Finite Volume Coastal Ocean Model (FVCOM)

Description:

FVCOM is an unstructured-grid, finite-volume, free-surface, 3-D primitive equation

coastal ocean circulation model developed by UMASSD-WHOI joint efforts (Chen et al.,

2006). The model consists of momentum, continuity, temperature, salinity and density

http://www.myroms.org/includes/bib.php?id=Haidvogel...00

http://www.myroms.org/includes/bib.php?id=Haidvogel...00

http://www.myroms.org/includes/bib.php?id=DiLorenzo03

http://www.myroms.org/includes/bib.php?id=Dinniman...03

http://www.myroms.org/includes/bib.php?id=Marchesiello...03

http://www.myroms.org/includes/bib.php?id=Peliz...03

http://www.myroms.org/includes/bib.php?id=Peliz...03

http://www.myroms.org/includes/bib.php?id=Budgell05


http://www.myroms.org/includes/bib.php?id=Warner...05b

http://www.myroms.org/includes/bib.php?id=Wilkin...05

http://www.hydroqual.com/

dick

Highlight

dick

Sticky Note

among

95


equations and is closed physically and mathematically using turbulence closure

submodels. The horizontal grid is comprised of unstructured triangular cells and the

irregular bottom is presented using generalized terrain-following coordinates. FVCOM is

solved numerically by a second-order accurate discrete flux calculation in the integral

form of the governing equations over an unstructured triangular grid. This approach

combines the best features of finite-element methods (grid flexibility) and finite-

difference methods (numerical efficiency and code simplicity) and provides a much better

numerical representation of both local and global momentum, mass, salt, heat, and tracer

conservation.

Applications:

The ability of FVCOM to accurately solve scalar conservation equations in addition to

the topological flexibility provided by unstructured meshes and the simplicity of the

coding structure has make FVCOM ideally suited for many coastal and interdisciplinary

scientific applications. The present version of FVCOM includes a water quality module

to simulate dissolved oxygen and other environmental indicators, a 3-D sediment

transport module for estuarine and near-shore applications, a flexible biological module

for the study of food web dynamics. FVCOM is only permitted for use in non-commercial

academic research and education. Moreover it has recently been applied to Tamar Estuary

(Uncles and Torres, 2011), and to the south-western coast of Korea (Lee et al., 2013).

1.3.1.4 MIKE 3

Description:

MIKE 3 is a 3D free surface model that is able to compute associated sediment and water

quality processes. The hydrodynamic module of MIKE 3 (MIKE HD) solves the

equations for the conservation of mass and momentum as well as for salinity and

temperature in response to a variety of forcing functions. MIKE 3 HD simulates the water

level variation and current velocities in response to a variety of forcing functions in lakes,

96


estuaries, bays and coastal areas. The water levels and currents are solved on either a

rectilinear grid, a curvilinear grid, a triangular element mesh or any combination covering

the area of interest. MIKE 3 HD includes formulations for the effects of: convective and

cross momentum, bottom shear stress, wind shear stress at the surface, barometric

pressure gradients, Coriolis forces, density effects, sources and sinks, evaporation and

precipitation, flooding and drying, and hydraulic structures.

Applications:

It is useful for a number of applications: assessment of hydrographic conditions for

design , coastal and oceanographic circulation studies, including fine sediment dynamics,

optimization of coastal outfalls, environmental impact assessment of marine

infrastructures, ecological modelling including optimization of aquaculture systems, lake

hydrodynamics and ecology, coastal and marine restoration projects, analysis and

optimization of cooling water recirculation and desalination. However, MIKE 3 is not

open source, and requires a license for its use.

1.3.1.5 Mohid Water Modelling System (MWMS)

Description:

MOHID is a three-dimensional water modelling system,

developed by MARETEC (Marine and Environmental Technology Research

Center) at Instituto Superior Técnico (IST) which belongs to Technical University of

Lisbon. Its development was started by Neves in 1985, and by that time the model was

bi-dimensional (Mohid 2D) and was used to study estuaries and coastal areas using a

classical finite difference approach. However, in the following years a eulerian and

lagrangian transport model were included in it, and also a turbulence model, a water

quality model, an ecosystem module and an oils spill model. The MOHID modelling

system allows the adoption of an integrated modelling, not only of processes (physical

and biogeochemical), but also of different scales (allowing the use of nested models) and

systems (estuaries and watersheds). MOHID has different tools integrated in tis modelling

http://194.65.82.103/maretec/index.htm

http://www.ist.utl.pt/

http://www.utl.pt/

http://www.utl.pt/

97


system, such as MOHID Water, MOHID Land and MOHID Soil, which can be used to

study the water cycle in an integrated approach. It has also coupled MOHID WQ and CE-

QUAL-W2 for water quality and ecological modelling, and it also has and Oil module for

oil spills tracking as mentioned before.

Applications:

Several different coastal and estuarine areas have been modelled with MOHID in the

framework of research and consulting projects. Along the Portuguese coast, different

environments have been studied, from river mouths, and estuaries (Martins et al., 2001;

Saraiva et al., 2007; Oliveira et al., 2015), to coastal lagoons (Vaz et al., 2007). The model

has also been adapted to simulate Galician Rías hydrodynamics, such as Ría de Vigo

(Montero et al., 1999; Montero, 1999) and Ría de Pontevedra (Taboada et al., 2000). Far

from the Atlantic coast of the Iberian Peninsula, some European estuaries have been

modelled, Western Scheldt (Holland), Gironde (France) (Cancino and Neves, 1999) and

Hellingford (Leitão, 1996). Regarding to open sea, MOHID has been applied to the

North-East Atlantic region where some processes including the Portuguese coastal

current (Coelho et al., 1994), the slope current along the shelf (Neves et al., 1998) and the

generation of internal tides (Neves et al., 1998) have been studied, and also to the

Mediterranean Sea to simulate the seasonal cycle (Taboada, 1999) or the circulation in

the Alboran Sea (Santos, 1995). Mohid has also been applied to the Portuguese Monte

Novo, Roxo and Alqueva reservoirs (Braunschweig, 2001).

1.3.1.6 HAMSOM

Description:

The development of the HAMSOM coding goes back to the mid-eighties where it

emerged from a co-operation between Backhaus and Maier-Reimer. From the very

beginning HAMSOM was designed with the intention to allow simulations of both

oceanic and coastal and shelf sea dynamics. It is a primitive equation model and it is

98


defined in Z co-ordinates on the Arakawa C-grid. It has been coupled with two eco-system

models (ECOHAM, ERSEM), an atmospheric model (REMO), and both Lagrangian and

Eulerian models for sediment transport. For polar applications HAMSOM was coupled

with a viscous-plastic thermo-hydrodynamic ice model of Hibler type. Since about 15

years in Hamburg, and overseas in more than 30 laboratories, HAMSOM is already being

in use as a community model.

Applications:

It has been applied to the Kara Sea (Harms, 1997) and to estuaries and bays, as it is the

case of the estuary Rio de la Plata (Meccia et al., unpublished) and to La ría de Vigo

(Souto et al., 2001).

1.3.1.7 Delft3D-Flow

Description:

Delft3D-FLOW is a multi-dimensional (2D/3D) hydrodynamic and transport simulation

program which calculates non-steady flow and transport phenomena that result from tidal

and meteorological forcing on a rectilinear or a curvilinear, boundary fitted grid. In 3D

simulations, the vertical grid is defined following the sigma coordinate approach.

The hydrodynamic conditions (velocities, water elevations, density, salinity, vertical eddy

viscosity and vertical eddy diffusivity) calculated in the Delft3D-FLOW module are used

as input to the other modules of Delft3D, which are: Delft3D-WAVE for short wave

propagation, Delft3D-SED for cohesive and non-cohesive sediment transport, Delft3D-WAQ

for far-field water quality, Delft3D-PART for mid-field water quality and particle tracking,

and Delft3D-ECO for ecological modeling. The hydrodynamics of Delft3D-FLOW are not

coupled to the other sub-models, being externally supplied to them.

Applications:

There are applications of Delft3D to SECS, lakes, rivers, estuaries and coastal

environments. Some examples of its application are to the Wadden Sea (Van Leeuwen et

99


al., 2010), to San Diego Bight (Van Dongeren, 2010), to Swansea Bay (Bent et al., 1991),

to the Gulf of Uraba (Escobar, 2011), to Nam Theun Reservoir (Chanudet et al., 2012)

and to the Bay of Cadiz (Zarzuelo, 2014).

1.3.1.8 Environmental Fluid Dynamics Code (EFDC)

Description:

The Environmental Fluid Dynamics Code (EFDC Hydro) is a hydrodynamic model that can

be used to simulate aquatic systems in one, two, and three dimensions. EFDC uses stretched

or sigma vertical coordinates and Cartesian or curvilinear, orthogonal horizontal coordinates

to represent the physical characteristics of a waterbody. It solves three-dimensional, vertically

hydrostatic, free surface, turbulent averaged equations of motion for a variable-density fluid.

Dynamically-coupled transport equations for turbulent kinetic energy, turbulent length scale,

salinity and temperature are also solved. The EFDC model allows for drying and wetting in

shallow areas by a mass conservation scheme. The physics of the EFDC model and many

aspects of the computational scheme are equivalent to the widely used Blumberg-Mellor

model (Blumberg and Mellor, 1987) and U. S. Army Corps of Engineers' Chesapeake Bay

model. The output of this model will provide all the necessary hydrodynamic inputs to the

Water Quality Analysis Simulation Program (WASP) (Di Toro et al., 1983; Connolly and

Winfield, 1984; Ambrose et al., 1988), QUAL2K from QUAL2E (Brown and Barnwell,

1987), AQUATOX (Park, 1990) and EPD-RIV1 (Martin and Wool, 2002).

Applications:

It has been applied to estuaries, coastal and ocean waters. It is remarkable its application to

St. Lucie Estuary, located on the east coast of south Florida by Ji et al. (2007) and also to

Morro Bay, California (Ji et al., 2001), both are semi-enclosed coastal systems.

100


1.3.1.9 Estuary and Lake Computer Model (ELCOM)

Description:

ELCOM (Estuary and Lake Computer Model) is a numerical modelling tool developed

at Centre for Water Research (CWR) of the University of Western Australia in 1996. It

is a 3D hydrodynamics, thermodynamics and transport model. It applies hydrodynamic

and thermodynamic models to simulate the temporal behaviour of stratified water bodies

with environmental forcing. ELCOM is designed to facilitate modelling studies of aquatic

systems over time scales extending to a few weeks, though the limit of computational

feasibility depends on the size and resolution requirements of an application and

computational resources. ELCOM is suited for comparative studies of the summer and

winter circulation patterns, spring versus neap tidal cycles, or dispersal conditions under

different flow regimes. ELCOM can be run either in isolation for hydrodynamic studies,

or coupled with the Computational Aquatic Ecosystem Dynamics (CAEDYM) for

simulation of biological and chemical processes.

Application:

It can be applied to semienclosed ecosystems (estuaries and coastal lagoons) but the most

important applications are to lakes, such as Lake Kinneret, Israel (Hodges et al., 2000).

However, as the code is not open source, it needs a licence to be used.

1.3.1.10 TELEMAC

Description:

TELEMAC-3D is a three-dimensional (3D) model that uses a horizontally unstructured

mesh of triangular elements with a finite-element or finite-volume computational method.

It can take into account the propagation of long waves with non-linear effects, bed

friction, Coriolis force, the influence of meteorological factors such as atmospheric

pressure and wind, turbulence, torrent and river flows, influence of horizontal

temperature or salinity gradients on density, Cartesian or spherical coordinates for large

101


domains, dry areas in the computational domain and diffusion of a tracer, with source and

sink terms, monitoring of floats and lagrangian drifts and singular point sources such as

pipes.

Applications:

It is an open source model that can be applied to bays, estuaries, coastal waters, seas and

oceans. Some examples of its application are to Atlanta Bay, to the English Channel, to

the beaches of Anglet and the Adour River (Brière et al., 2007), and wind induced

response of the Irish Sea (Jones and Davies, 2006). It is open-source since July 2010.

1.3.1.11 H2D/H3D

Description:

The H2D/H3D are two finite differences hydrodynamic models developed by the

Environmental Hydraulics Institute of the University of Cantabria (Castanedo, 2000).

They solve the two-dimensional/ three-dimensional hydrodynamic equations based on

the Reynolds Averaged Navier Stokes equations (RANS) for incompressible and

unsteady turbulent flows, including the effects of the earth’s rotation, bottom friction and

wind shear. These models involve the solution of the continuity, momentum and transport

equations for salinity and temperature dividing the study area into square cells. The

numerical computation should be carried out on a spatial domain that represents the entire

estuary through a finite-difference grid. The cell dimension is a function of the size of the

study area, and its resolution depends on the desired level of detail (García et al., 2010b).

H2D model has been linked to a medium complexity biogeochemical model (García et

al., 2010a) and to a very complex water quality model (EnvHydrEM) (Zouiten et al.,

2013).

Applications:

H2D has been successfully applied to coastal waters, estuaries and coastal lagoons. They

have been applied to estuaries such as Suances Estuary (Barcena et al., 2012), Urdaibai

102


(García et al., 2010a), and coastal lagoons such as Victoria (Spain) (Zouiten et al., 2013).

H3D has only been applied to Santander’s Bay (Castanedo, 2000).

1.3.1.12 Discussion

In Table 1.1 some of the most important physical models with application to SECS and

with coupled or linked ecological modeling capabilities are shown. All of them can be

used in both three-dimensional and two-dimensional configurations allowing the high

spatial and temporal resolution needed to accurately reproduce the hydrodynamic

processes, substances transport and biogeochemical and ecological processes in SECS.

However, some of them are not open-source, like MIKE 3, HAMSON, ELCOM, ECOM,

EFDC and H2D/H3D. Other models like the FVCOM have the code open only for

research or educational purposes, but not for commercial or management strategies. The

three-dimensional open source existing models are ROMS, TELEMAC, MOHID and

Delft3D, which are widely used models that have been successfully applied to many

different areas and problematics as could be seen in the previous section. These models

have therefore great future and their users are growing every day, and all of them could

be recommended to be used in many kinds of applications. Specifically, ROMS, is an

open source model since its creation, and it has been successfully coupled to several open-

source ecological models of different complexities. The great number of users of these

coupled configurations, the existing interest on improving them, and the flexibility of its

code makes it a good option for developing and testing new coupled or linked modeling

strategies. Moreover, ROMS is coupled to an open-source sediment transport model that

increase its capabilities. Additionally, at the IHCantabria we have been adding processes

and improving the H2D code, in order to have a simpler management tool to solve

different 2D situations. The knowledge of the code and its flexibility and accuracy in its

applications made it also a good tool for management strategies. In fact, its results are

easily linked to ecological models.

103


Therefore, at the present Thesis we have selected ROMS for the development of a three-

dimensional new hydrodynamical-ecological coupled system, and H2D for a two-

dimensional study where the third dimension could be omitted due to the characteristics

of the system under study. Although it is important to remark that nowadays there are

many different models that could be successfully applied in any of the cases mentioned

before.

Table 1.1. Characteristics of physical models with linked or coupled ecological models.

*Only for research or educational purposes

1.3.2 Water quality and ecosystem models

Water quality and ecosystem models are important tools for the study of semi-enclosed

coastal systems. On the one hand, water quality models deal with the basic aspects of

water quality of coastal areas influenced by human activities, namely, the oxygen

depletion as a result of nutrients, organic matter loadings, sediment interaction and fate

of pollutants. On the other hand, ecosystem models are essential tools for the better

knowledge of the coastal waters ecosystems and prediction of their evolution as well as

for their effective management (Jorgensen, 2011). Ecosystem models describe in general

phytoplankton and zooplankton dynamics, nutrient cycling, growth and distribution of

MODEL

CODE

LINKAGED OR COUPLED WATER QUALITY AND ECOSYSTEM MODELS

SEDIMENT TRANSPORT

MODULE

DOWNSCALING

ROMS

FREE EcoSim, PISCES, NEMURO, NPZD Fennel, NPZD Franks, NPZD Powell, NPZD Gruber, CE-QUAL-ICM

YES YES

FVCOM FREE* FBM(NPZ, NPZD), WQ YES YES

ECOM NO WQ YES YES

MIKE 3 NO WQ (ECO LAB) YES YES

MOHID FREE MOHID WATER QUALITY, CEQUALW2 YES YES

HAMSOM NO EcoHam1, EcoHam2, ERSEM NO NO

EFDC NO WASP, QUAL2K, AQUATOX,EPD-RIV1 YES YES

ELCOM NO CAEDYM NO NO

Delft3D FREE DELWAQ YES YES

TELEMAC-3D FREE DELWAQ YES YES

H2D/H3D NO T3DCOLI, T2D8, ENVYDREM NO YES

104


rooted vegetation, and macroalgae. In addition, they allow us to focus on phytoplankton

related processes: the primary production and the oxygen budget within the water column.

Water quality models can be effective tools to simulate and predict pollutant transport in

the water environment. Therefore, water quality models become an important tool to

identify water environmental pollution and the final fate and behaviours of pollutants in

the water environment. However, nowadays water quality models are usually coupled to

ecosystem models in order to take into account more processes. Therefore, their

capabilities have increased in the last decades and sometimes it is very difficult to

distinguish them from ecosystem models. Most of them started being only water quality

models and were afterwards coupled to ecological modules as mentioned before. A

simplified scheme of the main coupling and/or linking possibilities between the different

models can be seen in Figure 1.5. Additionally, some of the most important and widely

applied water quality models are: WASP, CE-QUAL-ICM, DELWAQ, WQ, MIKE3-

WQ.

Figure 1.5. Water quality and ecosystem models coupling and/or linkage possibilities.

105


Ecosystem models capture processes known to be physically and biologically important

determinants of phytoplankton dynamics, which are the driving process of eutrophication.

They also represent some of the most complex models available and so there is a lot of

scope for simplification. Ecosystem models explicitly include complex trophic webs,

nutrient dynamics and recycling. They can also be spatially resolved and include highly

detailed process formulations. The degree of detail employed in the formulation of any

one of these features and the data available for assessing them may have an important

impact upon model behaviour and performance. There are models that try to represent the

entire ecosystem by including all processes in the system, from physics to chemistry, and

plankton to fish. To achieve this, three types of models are usually coupled:

hydrodynamic models, lower trophic level (bacteria, phytoplankton and zooplankton)

models and complex higher trophic level (mainly fish species) models. Some examples

of higher trophic level models are: ERSEM, CAEDYM, PISCES, ECOPATH with

ECOSIM, and IGBEM, which are described in the next section. The complexity of higher

trophic level models due to the great number of state variables and parameters made them

difficult to use and interpret, so nowadays there is a need to build simpler models.

One example of simpler ecosystem models are biogeochemical models, which are

essentially lower trophic level models focusing in single species or a low number of

functional groups, such as NPZ and NPZD models. These models take into account only

the main processes of phytoplankton dynamics and are easier to calibrate and use, having

less parametrization errors. However, they have been equally used to eutrophic systems

without taking into account the level of eutrophication, and they usually only work with

nitrogen as the limitant nutrient of the system. Examples of lower trophic level models

are: ECO LAB, ECOHAM, FBM, MOHID Water Quality Module, CE-QUAL-W2,

NEUTRO, EnvHydrEM, NEMURO, Delft3D-ECO, PPBM, ICOLLS model, SEACOM,

P-Z, NPZ, NPZD, Fasham, Fennel, and SWEM.

106


Moreover, many more models have been published over the past few decades, all of

which cannot be described here. The ones below are a representative sample of the most

popular models, which have been applied equally to eutrophic and hypertrophic coastal

systems, and have been coupled or linked to hydrodynamic models.

1.3.2.1 Water Quality Analysis Simulation Program (WASP)

Description:

The Water Quality Analysis Simulation Program (WASP7), is an enhancement of the

original WASP (Di Toro et al., 1983; Connolly and Winfield, 1984; Ambrose et al.,

1988). This model helps users to interpret and predict water quality responses to natural

phenomena and man-made pollution for various pollution management decisions. WASP

is a dynamic compartment-modelling program for aquatic systems, including both the

water column and the underlying benthos. WASP allows the user to investigate 1, 2, and

3 dimensional systems, and a variety of pollutant types. The time varying processes of

advection, dispersion, point and diffuse mass loading and boundary exchange are

represented in the model. WASP also can be linked with hydrodynamic and sediment

transport models that can provide flows, depths velocities, temperature, salinity and

sediment fluxes. WASP has an ecosystem module called EUTRO that was specifically

designed for the assessment of processes impacting eutrophication and dissolved oxygen

dynamics. Five state variables are modelled in EUTRO: phytoplankton carbon, ammonia,

nitrate, carbonaceous biochemical oxygen demand, and dissolved oxygen.

Application:

WASP has been used to examine eutrophication of Tampa Bay (Sheng and Yassuda,

1995), phosphorus loading to Lake Okeechobee, eutrophication of the Neuse River

Estuary, eutrophication of Coosa River and its Reservoirs, PCB pollution of the Great

Lakes, eutrophication of the Potomac Estuary, volatile organic pollution of the Delaware

Estuary, heavy metal pollution of the Deep River, and mercury in the Savannah River.

107


1.3.2.2 CE-QUAL-ICM

Description:

CE-QUAL-ICM (Cerco and Cole, 1993) is a 3-D finite-volume water quality model

whose processes are based on WASP (Ambrose, et al., 1988) and was developed by the

Water Quality and Contaminant Modelling Branch at Waterways Experiment Station for

application to Chesapeake Bay. The model can link to hydrodynamic models of any

dimension or combination of dimensions using either structured or unstructured grids.

However, certain numerical solution techniques in the model complicate the linkage to

unstructured grid hydrodynamic models.

Application:

The model has been used by the U.S. Army Engineer Waterways Experiment Station

(WES) to evaluate eutrophication in Chesapeake Bay (Linker et al., 2001; Cerco and

Noel, 2013). It has also been used to address management issues in the New York Bight

(Hall and Dortch, 1994), Florida Bay (Cerco et al., 2000), San Juan Bay/Estuary (Bunch

et al., 2000), and many others.

1.3.2.3 DELWAQ

Description:

DELWAQ (Postma, 1988) is the engine of the D-Water Quality and D-Ecology programs

of the Delft3D suite (e.g: Delft3D-WAQ, Delft3D-ECO). It is based on a rich library from

which users and developers can pick relevant substances and processes to quickly put

water and sediment quality models together. It computes from basic tracers, dissolved

oxygen, nutrients, organic matter, inorganic suspended matter, heavy metals, bacteria and

organic micro-pollutants, to complex algae and macrophyte dynamics. High performance

solvers enable the simulation of long periods, often required to capture the full cycles of

the processes being modelled. The finite volume approach underlying DELWAQ allows

it to be coupled to both structured and flexible unstructured grid hydrodynamics.

http://oss.deltares.nl/web/delft3d/delwaq/-/message_boards/view_message/272768;jsessionid=650B26530617561847DFF88F49E1B983

108


Advanced features allow for increased flexibility in model configuration: hydrodynamics

can be aggregated in space and time, water and sediment layers may have a different

resolution and coarser sub-grids can be used to improve performance or facilitate input

specification.

The ecological module (Delft3D-ECO) includes processes related to algae growth and

mortality, mineralization of organic matter, nutrient uptake and release and oxygen

production and consumption. The Delft3D-ECO modelling instrument considers three

nutrient cycles: nitrogen, phosphorus and silica. The carbon cycle is partially modelled,

with a mass balance of all components containing organic carbon. Algal diversity are

represented in three species groups: diatoms, flagellates and green algae, and three genera

of cyanobacteria: Microcystis, Aphanizomenon and Planktothrix. To model variable

stoichiometry, each group is represented by three types defined by the physiological state

of the phytoplankton: phosphorus, nitrogen, or light limited. Different formulations are

available for the characterization of grazers, microphytobenthos, bottom sediment and

sediment–water exchange. Extinction of visible light is not spectrally computed, but is a

function of: inorganic suspended matter, yellow substances, detritus and phytoplankton

suspended particulate matter (SPM).

Applications:

It can be applied to marine, coastal waters and lakes (Los, 2009), estuaries (Blauw et al.,

2009), and SECS like the Berre Lagoon in France (Martin et al., 2010), and the Venice

lagoon (Runca et al., 1996) between many other applications.

1.3.2.4 WQ

Description:

The unstructured grid finite-volume water quality model (WQ), coupled to ECOM and

FVCOM, was developed based on the framework of the water quality analysis simulation

program (called WASP5) created by Ambrose et al. (1993). A Modification is made to

109


include the nutrient fluxes due to sediment resuspension via sedimentation process at the

bottom. In the water quality, the water quality variables include dissolved oxygen,

phytoplankton as carbon, carbonaceous biochemical oxygen demand, ammonium

nitrogen, nitrate and nitrite nitrogen, ortho-phosphorus or inorganic

phosphorus, organic nitrogen, and organic phosphorus.

Applications:

It has been used for estuarine, coastal and ocean applications. For example, it was applied

to Jiaozhou Bay, China (Zhang et al., 2012), and to industrial dumping and pollutant

dispersion (Wang et al., 2011), between other applications.

1.3.2.5 MIKE3-WQ (ECO LAB)

Description:

MIKE3 WQ is a water quality model that computes advection-dispersion, sediment

transport, eutrophication and heavy metals by using different templates. This templates

integrate the main biogeochemical processes associated to the phytoplankton

photosynthesis and respiration, oxidation due to the biochemical oxygen demand (BOD),

nitrification, sediment oxygen demand, phytoplankton, and bacterial respiration, and deal

with the basic aspects of water quality associated to oxygen depletion and the

transforming processes of the main biochemical compounds, namely, BOD and ammonia

levels resulting from the organic matter loadings. It is usually used with an ecosystem

template called ECO LAB that computes nutrient cycling, phytoplankton and

zooplankton growth, growth and distribution of rooted vegetation, macroalgae and

oxygen conditions; and also with a template called MIKE-EU for simulating algae growth

and primary production.

Application:

It has been widely applied by DHI (Danish Hydraulic Institute), for example to rivers

such as Yamuna River (India), to coral reefs, and to restore eelgrass.

http://fvcom.smast.umassd.edu/research_projects/Satilla/waterqualitymodel.html

110


1.3.2.6 European Regional Seas Ecosystem Model (ERSEM)

Description:

ERSEM (the European Regional Seas Ecosystem Model) is a plankton functional type

box model which was developed by Baretta et al. (1995) and improved by Baretta-Bekker

et al. (1997). It is related to NPZD type models but includes several refinements

necessaries to represent more accurately some processes of temperate shelf ecosystems:

plankton community complexity, the microbial loop, variable nutrient stoichiometry,

variable carbon, chlorophyll-a ratios and a comprehensive description of benthic

biochemical and ecological processes. It has 36 state variables, 20 pelagics and 16

benthics. The units of ERSEM are Carbon, Nitrogen, Phosphorus, Silica and Oxygen.

Applications:

It has been applied to the North Sea by Radach and Lenhart (1995) for assessing the

nutrient cycling. Additionally Ebenhoh et al. (1997) applied ERSEM to assess microbial

dynamics, and more studies showed that ERSEM was capable of continuously simulating

complex food webs (Baretta-Bekker et al., 1997). There were many applications to this

respect during the following years, and more recently ERSEM has been applied to test

scenarios of fisheries management strategies by Petihakis et al. (2007) for example.

1.3.2.7 Computational Aquatic Ecosystem Dynamics Model (CAEDYM)

Description:

CAEDYM (Hipsey et al., 2007) is a complex water quality model which can run in 1D

(coupled to DYRESM) or 3D (coupled to ELCOM) configurations, and it has 112 state

variables in total, taking into account inorganic particles, light, metals, phytoplankton

dynamics, zooplankton, bacteria, pathogens and microbial indicator organisms, higher

biology, carbon, nitrogen, phosphorus and silica, dissolved oxygen, sediments and

resuspension. However, for a simple water quality simulation there is a minimum

configuration of dissolved oxygen, extinction coefficient, Photosynthetically Active

111


Radiation (PAR), Particulate Organic Phosphorus, Particulate Organic Nitrogen,

Dissolved Organic Phosphorus, Dissolved Organic Nitrogen, Particulate Organic Carbon,

Dissolved Organic Carbon, Filterable Reactive Phosphorus, Ammonium, and Nitrate.

CAEDYM allows simulation of up to seven phytoplankton groups, five zooplankton

groups, three fish groups, four macroalgae groups, three invertebrate groups, three

clam/mussel groups, one seagrass/macrophyte group, one jellyfish group and three

pathogen/microbial indicator organism groups. However, although it computes PAR, it

does not compute the spectral character of PAR or the spectral character of the light

attenuation coefficient, which is fundamental for phytoplankton and seagrass growth and

primary production.

Applications:

It can be applied to SECS, lakes, estuaries, rivers and coastal waters but the 3D

configuration of ELCOM-CAEDYM is not open source. Some examples of its

application are to Swan River (Chan et al., 2002) and to Marmiom Marine Park coastal

lagoon in Western Australia (Machado and Imberger, 2012).

1.3.2.8 Pelagic Interaction Scheme for Carbon and Ecosystem Studies (PISCES)

Description:

PISCES (Pelagic Interaction Scheme for Carbon and Ecosystem Studies) was created by

Aumont and Bopp (2006) and is a medium complexity ecosystem model that has been

implemented in ROMS for its suitability for a wide range of applications. PISCES currently

has five modeled limiting nutrients for phytoplankton growth: Nitrate and Ammonium,

Phosphate, Silicate and Iron. It also has two phytoplankton size-classes/groups corresponding

to nanophytoplankton and diatoms, and two zooplankton size classes which are micro-

zooplankton and mesozooplankton. For phytoplankton, prognostic variables are total

biomass, the iron, chlorophyll-a and silicon contents. In addition to the ecosystem model,

PISCES also simulates dissolved inorganic carbon, total alkalinity and dissolved oxygen. The

112


dissolved oxygen formulation is also used to define the regions where oxic or anoxic

remineralization takes place.

Applications:

Some oceanic applications of PISCES can be seen at Tagliabue et al. (2014) for assessing

the impact of external sources of iron on the global carbon cycle, and some applications to

estuaries were made by Laws et al. (2000) and by Schlitzer (2000).

1.3.2.9 ECOPATH with ECOSIM (EwE)

Description:

ECOPATH (Christensen and Pauly, 1991) is an ecosystem mass balance model for food

webs, where functional groups are represented as biomasses, linked through their trophic

interactions. The model establishes mass balances by solving sets of linear equations that

describe the production and consumption of each group. ECOPATH has reasonably low

data requirements, and single mass balances give valuable insights into how energy is

transferred through the food web. Multiple balances are used for temporal or spatial

comparisons of system functioning. The time-dynamic module ECOSIM (Bissett et al.,

1999a, 1999b) applies differential equations to describe temporal variations of the flows

identified by ECOPATH mass balances and is mostly used to study the effects of

fisheries’ management policies in both marine and freshwater systems. In the EwE

software there is also a module called Ecospace, which is a spatial and temporal dynamic

module primarily designed for exploring impact and placement of protected areas. This

modelling system takes into account the spectral attenuation of light and spectral PAR for

phytoplankton growth kinetics, but has a great input parameter uncertainty.

Applications:

It is usually used to study management strategies of fisheries’ policies, and it can be

applied different to marine and coastal waters, lakes, lagoons, estuaries and SECS. One

example of its application to SECS is to Hudson Bay (Hoover et al., 2013), to estuarine

113


and coastal waters we can remark the case of Pearl River Delta (Duan et al., 2009), and

to lakes the application to Lake Victoria (Downing et al., 2012) and to Baoan Lake (Guo

et al., 2013). It is important to note that EcoSim model itself has also been applied to

SECS, as it is the case of Port Phillip Bay (Fulton and Smith, 2004), and also to the

Sargasso Sea (Bissett et al., 1999a, 1999b).

1.3.2.10 Integrated Generic Bay Ecosystem Model (IGBEM)

Description:

It is a box model developed by Fulton et al. (2004b). It is a coupled physical transport-

biogeochemical process box model constructed as a basis to explore the effects of model

structure and complexity. The formulations of the model are based on two existing

models, the European Regional Seas Ecosystem Model II (ERSEM II) and the Port Phillip

Bay Integrated Model (PPBIM). IGBEM provides a spatially and temporally resolved

model of nutrient cycles and population biomasses for enclosed temperate bays. The

model has 24 living components (groups), two dead, five nutrient, six physical and two

gaseous components. These components are linked through both biological and physical

interactions. The main processes taken into account in the water column are

Phytoplankton, Bacteria, Zooplankton, Fish, Nutrients and Detritus. In the epibenthic the

main processes are the Epifauna and the Phytobenthos, and in the sediment the Infauna,

Bacteria, Nutrients and Detritus. A daily time-step is utilized for standard runs. This is a

limitation as within the biological modules, a daily time-step may make the variables with

fast dynamics become unstable.

Application:

It was designed for semi-enclosed coastal systems, such as Port Phillip Bay (Fulton et al.,

2004b) where it was applied and tested.

114


1.3.2.11 Ecological North Sea Model, Hamburg (ECOHAM)

Description:

ECOHAM is a three-dimensional model for estimating the annual primary production in

the North Sea (Moll, 1998). The plankton system was represented by one phytoplankton

bulk variable, triggered by one nutrient, namely phosphate or dissolved inorganic

nitrogen. The mechanism for nutrient regeneration was represented by immediate

regeneration of part of the dead organic matter in the water and by a linear process for

regenerating the detritus at the bottom. Coupling of the benthic nutrient reservoir to the

water column was achieved by regeneration of inorganic nutrients from detritus and

transfer into the water column by bottom boundary conditions. There are two versions of

ECOHAM:

ECOHAM1

The ECOlogical North Sea Model, HAMburg, Version 1 (ECOHAM1) is a three-

dimensional model system that study nutrient and phytoplankton dynamics. It is the

German ecosystem model of the North Sea that has been developed at the Institute fur

Meereskunde, Hamburg, first, for the simple phosphorus cycle, and later extended for an

additional simple nitrogen cycle. It simulates the annual primary production under actual

circulation and solar radiation forcing (Moll, 1998) and describes concentrations and

fluxes of biologically important elements in space and time. Phytoplankton is represented

by one state variable and the model formulations are based on phosphorus and nitrogen

to limit the phytoplankton production. Grazing of phytoplankton by zooplankton is

treated dynamically due to a formulation according to Michaelis–Menten including a

threshold value below phytoplankton grazing ceases. The model is conceptualized for a

shelf sea including the shallow sea characteristic for the replenishment of the water

column with nutrients from the bottom. Two ordinary differential equations describe the

benthic detritus in terms of phosphorus and nitrogen pools. Underwater light is calculated

115


by a diagnostic ordinary differential equation that includes shelf shading due to

phytoplankton but it is not computed spectrally.

ECOHAM2

The “Ecological North Sea Model, Hamburg, Version 2” (ECOHAM2) predicts the

carbon and nitrogen cycle of the North Sea, the zooplankton total biomass dynamics and

the population dynamics of one single key copepod species (Pseudocalanus elongatus).

In contrast to the earlier version ECOHAM1, ECOHAM2 includes the carbon cycle in a

shelf sea and resolves the nitrogen cycle completely. It is a biogeochemical model with a

medium set of 13 “bulk” state variables to simulate the nutrient phytoplankton dynamics

in a three-dimensional physical frame for estimating bulk derived trophic state variables

for the functional units of phytoplankton, zooplankton, bacteria and additional detritus,

organic dissolved matter and nutrient concentrations including all fluxes between state

variables, especially the primary production, in shelf seas.

Applications:

Most remarkable applications of ECOHAM1 and ECOHAM2 were to the Bohai Sea (Wei

et al., 2004), and to the North Sea (Moll, 1998; Moll and Radach, 2003).

1.3.2.12 Flexible Biological Module (FBM)

Description:

The “Flexible Biological Module (FBM)” provides a platform that allows users to build

their own parallelized biological model from a discrete set of functions that is independent

of the physical model. This module can be run simultaneously coupled to FVCOM with

an unstructured-grid or an structured-grid for 3-D applications. Advection and diffusion

variables can be run separately by itself in 1-D applications. In the FBM code, the

biological module is an independent 1-D system that is self-maintained and upgraded

without linking to a physical model. It can be also converted to a Lagrangrian-based

biological model by linking it with the 3-D Lagrangian particle-tracking module.

116


FBM includes seven groups: nutrients, autotrophy, heterotrophy, detritus, dissolved

organic matter, and bacteria.

Application:

The most remarkable application of FBM was to build a NPZ model for the Gulf of Maine

(Tian and Chen, 2006).

1.3.2.13 Mohid Water Quality Module

Description:

The Water Quality Mohid model development started in 1985 by MARETEC, and

computes sinks and sources terms associate with the Carbon, Nitrogen and Phosphorous

cycle. The main dynamic processes that are computed by this model are: phytoplankton,

zooplankton and Nitrogen (Ammonia, Nitrate Nitrite, particulate organic nitrogen,

dissolved refractory and non-refractory organic nitrogen), Phosphorus inorganic and

organic, Dissolved Oxygen and BOD (Biochemical Demand of Oxygen). It can also solve

nutrients-dissolved oxygen-organic matter interactions, larvaes transport, selective

withdrawal from stratified reservoir outlets, hypolimnetic aeration, multiple algae,

epiphyton/periphyton, zooplankton, macrophyte, CBOD, and vegetative cover. Finally,

light penetration is not computed spectrally, but depends on suspended particulate matter

and phytoplankton.

Application:

It has been used for a wide range of applications such as for bathing water quality

management (Viegas and Nunes, 2009; Viegas et al., 2012), for assessing the quality of

the water of Madeira island (Portugal) (Campuzano et al., 2010).

http://www.mohid.com/WaterQuality.htm

117


1.3.2.14 CE-QUAL-W2

Description:

CE-QUAL-W2 (Cole and Wells, 2008) is an ecosystem model in 2D (longitudinal-

vertical). The model can simulate suspended solids, nutrient and organic matter groups,

residence time, derived variables such as TN, TKN, TOC, chlorophyll-a, as well as pH,

total dissolved gases and optional biotic groups, including multiple periphyton, multiple

phytoplankton, multiple zooplankton and multiple macrophyte groups interacting with

hydrodynamics (Berger and Wells, 2008). The model includes various vertical turbulence

closure, weirs/spillways, gates, pipes and pumps and re-aeration schemes for engineered

systems, which can be simulated depending on the nature of the water body. The model

is an open-source code written in FORTRAN.

Applications:

It has been used extensively throughout the United States for rivers, estuaries, lakes,

reservoirs and river basin systems (Deliman et al., 2002; Bowen and Hieronymous, 2003;

Debele et al., 2006) and elsewhere in the world (Chung and Oh, 2006; Kuo et al., 2006)

as a management and research tool. The model has also been used to drive models of food

web dynamics (Saito et al., 2001) and to support the studies of fish habitat (Sullivan et

al., 2003).

1.3.2.15 NEUTRO

Description:

NEUTRO is a 3-D eutrophication model, which is under development in the Physical

Oceanography Research Laboratory of Singapour (PORL) since 1997. Detailed

NEUTRO model description and eutrophication kinetics are given in Tkalich and

Sundarambal, 2003. It consists of a 3-D transport formulation, WASP eutrophication

kinetics and Silica cycles. NEUTRO takes into consideration tidal forcing, advection,

diffusion and settling of suspended particles, as well as chemical and physical kinetic

118


reactions of the diluted and suspended substances. It has been applied to coastal waters

within several commercial and research projects for environment impact assessment as

well as water quality management. The Eutrophication model (NEUTRO)

predicts water quality with respect to nutrients, plankton and dissolved oxygen,

suspended solids as well as bacteria decay. Seven interacting Systems (Cycles) are

included with benthic coupling for dissolved oxygen and nutrients: Nitrogen, Phosphorus,

Carbon and Silica cycles, Phytoplankton and Zooplankton Dynamics, and Dissolved

Oxygen Balance. The total of 13 state variables are considered: Ammonia Nitrogen,

Nitrate Nitrogen, Phosphate, Phytoplankton, Carbonaceous Biochemical Oxygen

Demand, Dissolved Oxygen, Organic Nitrogen, Organic Phosphorus, Zooplankton,

Bacteria, Total Solids, Available Dissolved Silica and Particulate Biogenic Silica.

Applications:

It has been used for Oceanic and Coastal waters of Southeast Asia, at Singapure area

(Sundarambal et al., 2010, 2012).

1.3.2.16 EnvHydrEM

Description:

EnvHydrEM (Zouiten et al., 2013) is a very complex computational time demanding 2D

model. The model considers a total of nineteen variables and 110 parameters. The

processes included are phytoplankton dynamics, total inorganic carbon, sediment carbon,

organic phosphorus, inorganic phosphorus, organic nitrogen, ammonia, nitrate, available

dissolved silica, particulate biogenic silica, dissolved oxygen, organic matter or

carbonaceous biochemical oxygen demand CBOD, zooplankton, bacterioplankton,

detritus, total iron and ferrous iron, total manganese and manganous ion. This model also

analyses a total of seventy-two interactions between the considered variables, including

the following sixteen processes: respiration, uptake, excretion, sedimentation, oxidation,

mineralization, nitrification, denitrification, photosynthesis, resuspension, grazing,

remineralization, predation, reaeration, sediment oxygen demand (SOD) and death. The

119


main limitations of the model are that it has a great input parameter uncertainty and it has

been explored to a limited extent.

Applications:

It has been applied to the Albufera of Valencia (Zouiten, 2012) and to Victoria coastal

lagoon, Spain (Zouiten et al., 2013).

1.3.2.17 Intermittently Closed and Open Lakes or Lagoons (ICOLLS) model

Description:

It is a spatially resolved, eleven-box ecological model for Australian Intermittently

Closed and Open Lakes or Lagoons (ICOLLs) (Everett et al., 2007). ICOLLs are

characterized by low flow from the catchment and a dynamic sand bar blocking oceanic

exchange, which creates two distinct phases, open and closed. The processes of the

ecological model are based on a combination of physical and physiological limits to the

processes of nutrient uptake, light capture by phytoplankton and predator-prey

interactions. An inverse model is used to calculate mixing coefficients from salinity

observations. The model is characterized by strong oscillations in phytoplankton and

zooplankton abundance, typical of predator-prey cycles. The model contains 17 state

variables (dissolved inorganic nitrogen, dissolved inorganic phosphorus, small

phytoplankton, large phytoplankton, small zooplankton, large zooplankton, epiphytic and

benthic microalgae, seagrass, refractory detritus, sediment dissolved inorganic nitrogen,

unflocculated phosphorus, flocculated phosphorus, sediment dissolved inorganic

phosphorus, unflocculated sediment phosphorus, flocculated sediment phosphorus,

unflocculated total suspended solids, flocculated total suspended solids). Moles of

nitrogen is the currency of the state variables, being one of them seagrasses. It is

remarkable that it takes into account seagrass nutrient uptake and constant light

attenuation. In this model, the photosynthetically active radiation in the water column is

assumed to be a constant value, the 43% of the surface irradiance, but it is attenuated not

spectrally by epiphytes and seagrasses between others.

120


Applications:

It was configured for Smiths Lake, NSW Australia and it has been explored to a very

limited extent (Everett et al., 2007).

1.3.2.18 SEACOM

Description:

It is a six-compartment ecological model designed to investigate the effects of

eutrophication on submerged vascular plants in coastal areas and bays. It was created by

Madden and Kemp (1996). The state variables include plant leaf biomass, leaves, plant

roots and rhizomes or roots, epiphytic algae attached to plant leaves, phytoplankton

(diatom group), phytoplankton (flagellate group), and labile sediment organic material.

Grazing by zooplankton is also taken into account but not as a state variable. In the

improved version of 2009 (Madden and McDonal, 2009) two seagrass species were taken

into account, Thalassia and Halodule. The attenuation of PAR is taken into account but

not spectrally. However, the attenuation constants of epiphytes, chlorophyll, water and

suspended particle matter are taken into account. The light attenuated by seagrasses is not

taken into account, the production is computed as a function of light reaching the top of

the canopy but attenuated by epiphytes.

Application:

It has been applied to coastal waters and bays, such as Chesapeake Bay (Madden and

Kemp, 1996), and Florida Keys (Madden and McDonald, 2009), but is potentially

applicable to SECS.

1.3.2.19 Phytoplankton-Zooplankton (P-Z) Models

Description:

These models are based on Lotka-Voltera predator-prey model. In 1926, the Italian

mathematician Vito Volterra proposed a differential equation model to explain the

observed increase in predator fish (and corresponding decrease in prey fish) in the

121


Adriatic Sea during World War I. At the same time in the United States, the equations

studied by Volterra were derived independently by Alfred Lotka (1925) to describe a

hypothetical chemical reaction in which the chemical concentrations oscillate. The Lotka-

Volterra model is the simplest model of predator-prey interactions. It is based on linear

per capita growth rates. Then, the equations were applied to phytoplankton and

zooplankton dynamics complicating them with a phytoplankton mortality term. In this

models production is limited by the value of P (Phytoplankton), which depends on the

values of Z (zooplankton).

Applications:

This kind of model was designed to be applied to semi-enclosed systems and bays, as it

is the case of Bay of Villefranche (Ross and Nival, 1976).

1.3.2.20 Nutrient-Phytoplankton-Zooplankton (NPZ) Model

Description:

This kind of model incorporates one of the simplest sets of dynamics that usefully

describe plankton dynamics. An NPZ model has, by definition, three state variables:

nutrients, phytoplankton and zooplankton. These are usually modelled in terms of

nitrogen content, since nitrogen is often limiting to primary production in both ocean and

coastal waters. In an NPZ model there are 5 transfer functions to consider: phytoplankton

response to light, phytoplankton nutrient uptake, zooplankton grazing, and phytoplankton

and zooplankton loss terms due to death, excretion and predation. The phytoplankton

response to irradiance in these kind of models can be formulated in different ways as

described by Franks (2002), but as far as we know, it has never been formulated

spectrally.

Some authors such as Denman and Gargett (1995) and McClain et al. (1996) started to

use NPZ models after the boom of complex models in the 70’s, and they have been using

since them, although there is always discussion between the degree of complexity needed

http://www.scholarpedia.org/article/Periodic_Orbit

122


to accurately describe the eutrophication in the system. One of the most used NPZ models

is NEMURO (Kishi, 1994; Kishi et al., 2011), which is coupled to ROMS, and has been

described in a separate section due to its importance. Another relevant NPZ model is Port

Phillip Bay Model (PPBM), which is also separately described. Depending on the model

they present different architectures from 1D (Denman and Gargett, 1995), 2D (Franks

and Walstad, 1997), to 3D (Franks and Chen, 2001; Kishi et al., 2011). NPZ models tend

to be very good at reproducing biomasses of nutrients and phytoplankton (Franks, 2002).

The level of detail that can be reproduced by the model depends on the physical model

employed. However, they have been usually used with low resolution physical models to

oceanic waters.

The NPZ model is a simplification of an extremely complex system, and it must be used

and applied carefully and appropriately. It is essential to make a clear statement about the

question being asked, so it could be extremely good for management purposes.

Applications:

Nowadays, one of the most common uses of NPZ models is for theoretical investigations.

There are many authors that have studied how does the model behaves if different transfer

functions are used (Sjoberg, 1977; Steele and Henderson, 1981, 1992; Franks et al., 1986;

Murray and Parslow, 1999; Ruan, 2001), if different parameters are used (Jeringan and

Tsokos, 1979, 1980; Hastings and Powell, 1991; Ruan, 1993; Abrams and Roth, 1994;

McCann and Yodzis, 1994; Truscott and Brindley, 1994; Edwards and Bridley, 1996,

1999; Edwards et al., 2000 a, b), or if different physical models are used (Franks, 1997).

Most of these studies showed a good behaviour under the use of different transfer

functions and parameterizations, as the little data requirement for its calibration made the

model easily obtain good and realistic results.

Other theoretical investigations have concentrated less on the model structure and

parameterization and more on the biological implications of the model. Evans (1978) and

123


Evans et al., (1977) coupled the simple NPZ model to a kinematic physical model of

vertical shear to explore how the interaction of vertical migration with vertical shear could

lead to patchiness of plankton. Steele and Frost (1977) used an elaboration of an NPZ

model to investigate the factors controlling the size structure of the phytoplankton. Kiefer

and Atkinson (1984) used an NPZ model to study nitrogen cycling efficiency in the

plankton. Evans and Parslow (1985) used their model to understand the factors

controlling the very different annual plankton cycles in the Atlantic and Pacific oceans.

In the 90’s, Marra and Ho (1993), Wroblewski et al. (1988) and McGillicuddy et al.

(1995a, b) between others used NPZ models to explore the spring phytoplankton bloom.

The NPZ model was usually the Franks et al. (1986) formulation, coupled to a range of

physical frameworks. Therefore, coupling the NPZ ecosystem model to a variety of

physical models has allowed exploration of a range of physical-biological interactions in

the ocean.

1.3.2.21 North Pacific Ecosystem Model for Understanding Regional

Oceanography (NEMURO)

It is known as the North Pacific Ecosystem Model for Understanding Regional

Oceanography. The model is a lower trophic ecosystem model that as most NPZ models

takes into account the nitrogen cycle. Nemuro was initially developed for the North

Pacific ecosystem and process formulations. Phytoplankton is represented by two

functional groups (small (flagellates) and large (diatoms)) and zooplankton by three

functional groups (small, large, and predatory). The unit of the model is nitrogen, and

many of the process equations are similar to other existing NPZ models.

Applications:

Nemuro model has been widely applied to the North Pacific (Kishi, 1994; Kishi et al.,

2007; Kishi et al., 2011).

124


1.3.2.22 Port Phillip Bay Model (PPBM)

Description:

It is a simple biogeochemical model created by Murray and Parslow in 1997 and detailed

by Murray and Parslow (1999). It is based on the biogeochemistry of only the lower

trophic levels, so it can no consider fisheries and eutrophication simultaneously. It is

locally designed for Port Phillip Bay (South Eastern Australia) biogeochemical processes.

It is a NPZ model with N as limitant nutrient; therefore, it is based on the nitrogen cycle.

The water column and the sediment are computed in different compartments, which are

linked by sedimentary flux of detritus from the water column to the sediment, and the net

flux of dissolved inorganic nitrogen from the sediment to the water column.

Application:

Semienclosed coastal systems such as Port Phillip Bay (Murray and Parslow, 1999;

Murray et al., 2001).

1.3.2.23 Nutrient-Phytoplankton-Zooplankton-Detritus (NPZD) Model

Description:

NPZD models simulate the interactions of the four variables nutrients (N), phytoplankton

(P), zooplankton (Z) and detritus (D). The mathematical formulation of the internal fluxes

varies in kind and complexity. For example, the growth of phytoplankton can be modelled

to be limited by nutrients only, to be limited additionally by light or even by more factors

(Denman and Peña, 1999; Oschlies and Garcon, 1999; Fennel et al., 2001; McCreary et

al., 2001; Kawamyko, 2002; Spitz et al., 2003; Losa et al., 2006; Pahlow et al., 2008).

There are well known NPZD models such as Fennel’s biological model (Fennel et al.,

2006), Fasham’s model (Fasham et al., 1990), Powell et al.’s model (2006), Gruber et

al.’s (2006) model, all of them coupled to ROMS. An others such as Doney et al. (1996),

Oschlies and Garcon (1999), Hood et al. (2001, 2003), and Hinckley and Dobbins (2004).

Fennel’s and Fasham models have been described in separate sections due to their

importance.

125


Applications:

NPZD models have been used for a wide range of study areas such as Middle Atlantic

Bight (Fennel et al., 2006), and Lake Hamana and Lake Sanaru (Ohno and Nakata, 2008)

between others.

1.3.2.24 Fasham

Description:

Biogeochemical model developed by Fasham et al. (1990) with seven compartments

(Phytoplankton, Zooplankton, Bacteria, Nitrate, Ammonium, Dissolved organic nitrogen

and Detritus), it has 7 state variables and 27 parameters. In this model ecosystem

seasonality was assumed to be driven by seasonal changes in incident Photosynthetically

Active Radiation (PAR) and mixed layer depth (Evans and Parslow, 1985). Nitrogen is

regarded as the limiting nutrient of primary production, therefore, the model is based on

the nitrogen cycle. The use of nitrogen as a model currency has the additional advantage

that the primary production can be partitioned into "new" production, fuelled by nitrate,

and "regenerated" production, fuelled mainly by ammonium (Dugdale and Goering,

1967; Eppley and Peterson, 1979). It is comprised of seven compartments:

Phytoplankton, Zooplankton, Bacteria, Nitrate nitrogen, Ammonium nitrogen, Labile

dissolved organic nitrogen and Detritus.

Applications:

It has been applied to oceanic waters. Some examples are its application to Bermuda area

(Fasham et al., 1990) and to the North Pacific (Fasham, 1995).

1.3.2.25 Fennel

Description:

It is a biogeochemical nutrient, phytoplankton, zooplankton, and detritus (NPZD) model

(Fennel et al., 2006) that is implemented into ROMS, and assumes nitrogen as the

controlling nutrient for primary production. Therefore, it is based on the nitrogen cycle,

126


and includes the source, sink, and biogeochemical transformation terms of seven state

variables: nitrate, ammonium, small and large detritus, phytoplankton, zooplankton, and

chlorophyll. The main biogeochemical model equations are described by Fennel et al.

(2006), who adapted them from the plankton dynamics model of Fasham et al. (1990). In

the Fennel implementation, phytoplankton growth is a function of temperature, nutrient

concentration, and the homogenously integrated PAR distribution.

Applications:

It has been applied for example to the Middle Atlantic Bight (MAB) (Fennel et al., 2006)

and to the Northwest North Atlantic (Fennel et al., 2008).

1.3.2.26 System Wide Eutrophication Model (SWEM)

Description:

An improved model, the System Wide Eutrophication Model (SWEM), has been

developed and tested by Hydroqual Inc. (1987) to simulate the biogeochemistry and

circulation in Long Island Sound (LIS) and adjacent waters. SWEM is a complex model

with many parameters that represent short-term variability in the rates of the processes

that influence the dissolved oxygen concentration. SWEM is coupled to a hydrodynamic

module that represents winds and the state of the ocean. The products of this module

(velocities and vertical eddy coefficients) are passed to the water quality module to

compute the evolution of nutrients, plankton, dissolved oxygen etc.

Application:

It has been applied to western Long Island Sound (LIS) (O’Donnell et al., 2009), and the

States of New York and Connecticut have developed a Comprehensive Conservation and

Management Plan (CCMP) using a computer model to assess the likely impact of

reductions in nitrogen discharged from water treatment plants and non-point sources.

127


1.3.2.27 Discussion

There are a high number of water quality and ecosystems models that vary from very high

complex ecosystem models such as ECOSIM (Bisset et al., 1999a, 1999b), ERSEM

(Baretta et al., 1995), IGBEM (Fulton et al., 2004b), CAEDYM (Hipsey et al., 2007),

CE-QUAL-ICM (Cerco and Cole, 1993), ICOLLs model (Everett et al., 2007) and

EnvHydrEM (Zouiten et al., 2013) to medium complexity models such as PPBM and

PPBIM (Murray and Parslow, 1997), ECOHAM (Moll, 1998), NEUTRO

(Tkalich and Sundarambal, 2003), WASP (Di Toro et al., 1983; Connolly and Winfield,

1984; Ambrose et al., 1988), to simple models such as P-Z (Ross and Nival, 1976), NPZ

(Denman and Gargett, 1995; McClain et al., 1996), NPDZ (Fasham et al., 1990; Fennel

et al., 2006) (see Table 1.2). All these models solve the eutrophication taking into account

a different number of processes and are used equally for systems with different trophic

levels (oligotrophic, mesotrophic, eutrophic, hypertrophic). The trophic level and the

main governing processes of each system should be taken into account when deciding the

model to use. In fact, complex water quality and ecosystem models are many times used

without taking into account the level of eutrophication and therefore the main processes

that rule the ecosystem. Moreover, the large parametrization of complex models with

numerous equations, the complexity of getting data for calibrating them, and the

indiscriminate use that the scientific community makes of them justifies the increasing

development of simpler models with less data requirements, especially for decision

makers and management strategies. Besides, these models could be coupled to others if

necessary, depending on the processes needed to accurately define the system under

study.

One of the most important simple models are PZ, NPZ and NPDZ. Although they have

been traditionally used for testing hypothesis, after our analysis we support that they can

be used for management strategies, due to their accurate results and its design. As

mentioned in their corresponding section, they are focused on the most important

128


variables and parameters of eutrophic systems and give a quick and certain answer to

relevant questions. We support the use of these kind of models to evaluate eutrophication

in SECS due to the fact that they take into account the phytoplankton, the limitant nutrient

cycle, and they generally have the possibility of taking into account sediment interaction

which could be very relevant in these kind of systems. However, most of the

biogeochemical models suitable for SECS are codified with the nitrogen cycle, especially

those that are open source; whereas in some SECS phosphorus is the limitant nutrient.

There is another important aspect to remark about the analyzed models, which is the light

formulation. Only one of the analyzed models take into account the spectral behavior of

irradiance (see Table 1.2) and its influence on phytoplankton growth. Additionally, only

few of them have the possibility to model submerged aquatic vegetation (SAV), for which

light formulation is essential. In fact, SAV growth is based on the spectral irradiance that

reaches the canopy, and on the photosynthesis and respiration carbon balance. This could

be solved by coupling or linking the analyzed models to a bio-optical irradiance model.

Therefore, we conclude that there is a need of using different simplified biogeochemical

models for studying eutrophication in SECS, if possible NPZ or NPDZ type, with

different parameters and formulations depending on the trophic level of the studied area,

as in the case of eutrophic and hypertrophic SECS. Additionally, for studying the

submerged aquatic vegetation dynamics the coupling or linkage of specific irradiance and

bio-optical models is recommended.

129


Table 1.2. Some relevant characteristics of the main water quality and ecosystem models.

MODEL COMPLEXITY SAV MODULE LIGHT ATTENUATION

SPECTRAL LIGHT ATTENUATION

SPATIAL RESOLUTION

WASP MEDIUM NO YES NO HIGH

CE-QUAL-ICM COMPLEX NO NO NO LOW

DELWAQ COMPLEX NO YES NO HIGH

WQ MEDIUM NO YES NO HIGH

MIKE WQ (ECO LAB) MEDIUM YES YES NO HIGH

ERSEM COMPLEX NO *Added in 2008

NO *Added in 2008

NO LOW

CAEDYM COMPLEX YES YES NO LOW & HIGH

PISCES MEDIUM NO YES NO HIGH

ECOPATH with ECOSIM COMPLEX NO YES YES HIGH

IGBEM COMPLEX YES YES NO LOW

ECOHAM MEDIUM NO YES NO LOW

FBM SIMPLE NO YES NO LOW

MOHID WQ MEDIUM YES YES NO HIGH

CE-QUAL-W2 COMPLEX NO YES NO MEDIUM

NEUTRO MEDIUM NO NO NO HIGH

EnvHydrEM COMPLEX NO YES NO HIGH

ICOLLS COMPLEX YES YES NO LOW

SEACOM MEDIUM YES YES NO LOW

P-Z models SIMPLE NO YES NO LOW

NPZ models SIMPLE NO NO NO LOW

NEMURO SIMPLE NO NO NO LOW

PPBM MEDIUM YES YES NO LOW

NPZD models SIMPLE NO NO NO LOW

FASHAM MEDIUM NO YES NO LOW

FENNEL SIMPLE NO YES NO LOW

SWEM COMPLEX NO YES NO HIGH

1.3.3 Bio-optical irradiance models with coupling or linking

capabilities

Light is essential for photosynthetic plants and algae. However, due to the rapid

attenuation in water, light is often a limiting factor in primary production in the aquatic

environment (Fisher et al., 1999). The degree of light attenuation also varies

130


tremendously in SECS due to the variable presence of “chromophoric” organic matter,

such as phytoplankton, CDOM and detritus. Therefore, reproducing the correct

underwater light field is a key problem in modelling the biogeochemical processes in

these systems.

Because the incoming Photosynthetically Active Radiation (PAR) at the air-water

interface can be measured or calculated quite accurately (Fisher et al., 2006), the main

issue in calculating the underwater light field is to have correct estimate of the vertical

diffuse attenuation coefficient (Kd). The diffuse attenuation coefficient is referred to as

an Apparent Optical Property (AOP) because its value depends on the ambient

underwater light field. For monochromatic light the vertical light attenuation can be

decomposed as a set of partial attenuation coefficients, each characterizing absorption

and scattering by a different waterborne material. Strictly speaking, a complete spectrum

of Kd (Kd (λ)) is needed to obtain the average Kd for the whole photosynthetic waveband

and for each narrow band it is necessary to know the wavelength-specific absorption and

scattering coefficients for each waterborne material. Spectral bio-optical models have

been developed and applied to different kinds of water bodies (Platt and Sathyendranath,

1988; Gallegos et al., 1990; Arrigo and Sullivan, 1994). However, due to the optical

complexity of estuarine waters, and specifically of eutrophic SECS, and the complexity

of obtaining good quality light data most of the ecological models do not take into account

the spectral character of light (see Table 1.2), which is one of the functions of bio-optical

models. The main bio-optical models that take into account the spectral underwater light

climate are described below.

1.3.3.1 Hydrolight-Ecolight v5 (HE5)

Description:

Hydrolight was developed in 1992 and Ecolight in 2008. They were coupled by Mobley

and Sundman in 2008. It is a modelling system composed by Inherent Optical Properties

(IOP) models with different formulations for: particle absorption, CDOM absorption, and

dick

Highlight

dick

Sticky Note

more accurately on the geometric distribution of the underwater light field

131


scattering. In fact, it has specific absorption and scattering models, bottom reflectance

models, sky radiance and irradiance models, and inelastic-scattering models. It is a

complex modelling system that allows obtaining all the inherent and apparent optical

properties of the water and present substances, and the light climate. It is not open source

but its free use is allowed by the developers for scientific research with a license.

However, Ecolight was coupled to EcoSim (a complex ecosystem model detailed in

section 1.3.2.2). The complexity of the coupling of the two models makes them difficult

to use and interpret.

Applications:

It can be applied to all kind of water bodies and ecosystems, and especially by NASA to

obtain information from satellite data. It has been recently applied to assess the light

climate of different ocean ecosystems (Mobley, 2011).

1.3.3.2 Fuji et al.’s model

Description:

Fuji et al. (2007) developed an optical and radiative transfer model that explicitly

represented and spectrally resolved the Inherent Optical Properties (IOPs, e.g. absorption,

scattering, and attenuation) from the ecosystem, and a radiative-transfer model to obtain

the Apparent Optical Properties (AOPs), such as diffuse attenuation, and radiometric

quantities, Photosynthetically Active Radiation (PAR) and remotely sensed reflectance

(ocean colour). The absorption coefficient is determined from the sum of the absorption

coefficients of seawater, picoplankton (plankton composed by cells between 0.2 and 2

μm), and diatoms (based on their chlorophyll-a content), non-algal particles (NAP), and

coloured dissolved organic matter (CDOM). The absorption spectra of different

phytoplankton functional groups are also taken into account. This model was coupled to

a modelling system developed by Fuji et al., (2007) that consists of three individual

models: a physical-ecosystem model (simulating the dynamics of different ecosystem

components in time and space), a photo-acclimation model (specifying the chlorophyll-a

dick

Highlight

dick

Sticky Note

allows one to obtain

dick

Highlight

dick

Sticky Note

no, the IOPs have to be provided by the user.

dick

Highlight

dick

Sticky Note

only if you buy Mobley a new kayak. Otherwise it's $10K per copy.

132


to carbon ratio of phytoplankton), an optical (converting ecosystem state variables into

inherent optical properties) and a radiative transfer (calculating the underwater light field

and the ocean colour) model. It is a complex model that is not open source.

Application:

It can be applied to all kind of aquatic systems but as far as we know it has only been

calibrated and validated by Fuji et al. (2007).

1.3.3.3 Gallegos et al.’s model

Description:

It is a simplified model design for management purposes and decision makers (Gallegos

et al., 2011). It includes the spectral attenuation effects of water, CDOM, phytoplankton,

and non-algal particulates (e.g., detritus, minerals, bacteria). Because the model is

spectrally based, it can be used to calculate the attenuation of either Photosynthetically

Active Radiation (PAR, equally weighted quanta from 400 nm to 700 nm) or

Photosynthetically Usable Radiation (PUR, the integral of the quantum spectrum

weighted by the pigment absorption spectrum of SAV). PUR is a more accurate

measurement of light that can be absorbed by SAV and it is more strongly affected by

phytoplankton chlorophyll-a in the water column than is PAR. In this model the empirical

descriptor of the light available at a depth in terms of that available at the surface is the

diffuse attenuation coefficient of downward propagating irradiance, Kd.

Application:

It has been applied on its actual and early form to the Rhode River (Gallegos et al., 1990),

Chesapeake Bay (Gallegos et al., 1990; Gallegos et al., 2011), the Indian River Lagoon

(Gallegos and Kenworthy, 1996; Gallegos et al., 2009), Carrie Bow Cay, Belize (Gallegos

et al., 2009), and in Bahia Almirante, Panama (Gallegos et al., 2009).

133


1.3.3.4 Zimmerman model

Description:

The model developed by Zimmerman (2003) consists of three different modules: a

module that simulates the seagrass relative biomass and architecture including leaf

geometry, an irradiance module that calculates the light absorption and scattering through

the canopy, and a photosynthesis module that calculates the carbon balance (production

and respiration) of the submerged plant canopy. The model simulates the light

environment of a submerged canopy at a fixed horizontal point (1D model). It allows the

description of the light environment by dividing the seagrass canopy volume, including

the leaves and the water column, into a series of horizontal sections of finite thickness.

The optical properties of each section are based on the architecture of the canopy, the

orientation and optical properties of the leaves, and the optical properties of the dissolved

materials and suspended particles in the water column. Given the spectral PAR at canopy

height from an irradiance model, it is able to compute the seagrass Photosynthetically

Usable Radiation (PUR) by computing the spectral absorption, reflection, and diffraction

of the downwelling and upwelling photosynthetically active irradiance through the

seagrass canopy. Finally, the model calculates the canopy production and respiration.

Application:

It has been applied to populations near Lee Stocking Island, Bahamas (Zimmerman,

2003) and to eelgrass from California, USA (Zimmerman, 2003).

1.3.3.5 Discussion

There are different characteristics to take into account when selecting a Bio-optical

model. Hydrolight-Ecolight is a complex model used mainly to obtain the apparent

optical properties of different substances in the water column by means of two coupled

models. Another complex bio-optical model is Fuji’s et al. (2007) model, which has only

134


been calibrated in one area in 1D. On the other hand Gallegos et al. (2011) is a simple

bio-optical model designed for decision makers and management strategies, which in

spite of its detailed formulations it is easy to use and interpret, it can be used for eutrophic

systems and SAV (although it does not take into account the canopy architecture), and

gives us the spectral light climate and the Kd (λ). Finally Zimmerman’s model is a simple

bio-optical model focused on SAV. It takes into account the architecture of the canopy,

but needs the spectral PAR at canopy height to start computing. Therefore, a combination

of Zimmerman’s model with another model (see Table 1.3) could be a good solution for

assessing not only the spectral underwater light climate in a eutrophic SEC, but also the

SAV potential habitat and photosynthesis-respiration carbon balance.

Therefore, in terms of simplicity and accuracy, a combination between Gallegos’ et al.

(2011) and Zimmerman’s models can be a good way to predict light climate in both water

column and canopy. Gallegos’ et al. (2011) could be used to obtain the spectral PAR at

canopy height and Zimmerman to propagate light through the canopy, and calculate the

production-respiration ratio and potential SAV habitat based on light climate.

Table 1.3. Some relevant characteristics of the main bio-optical models.

MODEL COMPLEXITY SPECTRAL PAR SAV FORMULATION

HYDROLIGHT-ECOLIGHT COMPLEX YES NO

FUJI ET AL. (2007) COMPLEX YES NO

GALLEGOS ET AL. (2011) SIMPLE YES NO

ZIMMERMAN (2003) SIMPLE NO YES

1.4 Balancing spatial and temporal resolution

Because computer resources are limited, one inevitable problem faced in all numerical

modelling studies is resolution. Even with today’s powerful computers, it is still

impossible to resolve all processes at all relevant scales in a model.

135


In this section, the trade-offs involved in the balance between spatial and temporal

resolution in coupled-linked ecological models are highlighted by presenting a series of

model case studies. It is clear that a variety of modelling frameworks have developed over

time that operate over rather different time and space scales. Each of these model types

are guided by different questions, the answer to which influences the chosen spatial and

temporal aggregation used in the model. Below, a series of examples of model space-time

complexity is discussed.

1.4.1 Low spatial, low temporal

As the sophistication of modelling has increased over time, few models currently exist

that include both low spatial and low temporal resolution, although they were historically

developed (e.g., Riley, 1946). The remaining models that include both low spatial and

temporal resolution are, in general, empirical models that describe the relationship

between a variable of interest (chlorophyll-a, hypoxic volume) and one or more

controlling variables (e.g., Cole and Cloern, 1987; Lee et. al., 2013), or semi-empirical

approaches based on first principles (Scavia et al., 2006; Testa and Kemp, 2008). In

general, these models predict variables at a single place or over a defined area, and are

primarily used to understand the dominant controls on a variable of interest. Such models

do not include mechanisms explicitly, but mechanisms are often inferred based on our

knowledge of the system derived from experimental and observational effort. The fact

that mechanisms are often inferred from empirical models allows these tools to be

excellent hypothesis generators; that is, when significant statistical relationships emerge

from an empirical model, the speculated underlying mechanisms that drive these

relationships can be tested with more resolved mechanistic models. For example, such

statistical models have been used to predict the spatial extent of hypoxic water in many

coastal ecosystems (e.g., Greene et al., 2009; Lee et al., 2013). These models have

generally included multiple linear regressions, which suggest that multiple controlling

136


variables, operating during specific seasons, ultimately control the spatial extent of

hypoxic volume in temperate ecosystems. In reality, however, the controlling variables

influence the biology and physics of the system at time and space scales that are much

finer than the seasonal or regional mean values used in the statistical model, so in many

cases, the hypotheses generated from statistics can be readily tested using hydrodynamic-

ecosystem models.

1.4.2 High spatial, low temporal

Some ecological models focus on the description of ecological processes that does not

change in a small period of time, but are spatially variable. One example of these

processes are those that involve vegetation presence/absence which are seasonally

variable and could be predicted by modelling strategies. In that line models with high

spatial, low temporal resolution try to answer questions like the influence of different

pressures on the seagrass presence absence, the effects of eutrophication, pollutants and

macroalgae competition on seagrass biomass distribution (Giusti et al., 2010), the

influence of phytoplankton and epiphytes on the underwater light environment

(Cunningham, 2002), seagrass shoot densities (Bearling et al., 1999), and the conditions

necessary for the restoration of submerged vegetation (Duarte et al., 2013).

High spatial resolution has increased in modelling techniques as shoot density and roots

of seagrasses are very variable with sediment type and light availability. In fact, depth

and type of sediments are quite variable in estuaries, and physiological and morphological

responses of seagrass are quite sensible to these parameters. The role of time and the

importance of temporal scale have received considerable less attention than issues on

spatial scale in recent years, although understanding complex ecological systems requires

linking space with time over the appropriate spatial and temporal scales (O’Neill et al.,

1986). Time resolution could be low (days, months, seasons, years) regarding some

ecological processes such as submerged aquatic vegetation dynamics due to low growth

137


kinetics and the biological time scales and seasonal variability of the different species

(Alexandre et al., 2008).

1.4.3 Low spatial, high temporal

Some examples of models with low spatial, but high temporal resolution include simple

sediment biogeochemical models, 1-D water-column process models, and box models.

These models can be useful for single point phenomena or processes which are

homogeneous over large areas. However, there is a considerable risk of missing adjacent

factors with high spatial variability that could affect the description of the main processes

of the system. In this respect, SWEM is a complex model that has been applied to the

western Long Island Sound with low spatial and high temporal resolution. However, the

low resolution of the model diminishes its value to the management of water quality in

embayments like Hempstead Harbour, Smithtown Bay, etc, but due to the complexity of

SWEM model higher spatial resolution will need many computational resources, data,

time and costs. Anyway, the high spatial, low temporal resolution can be many times

determined based on field data spatial and temporal variations.

1.4.4 High spatial, high temporal

Individual-based models (IBMs) represent another class of models developed in estuarine

systems that combine biological and physical processes, in this case using high resolution

spatial scales and temporal resolution that varies with the relevant life history

characteristic being modelled. They cannot be classified as in a lower or higher trophic

level group in general as it depends on the individual been studied in each case, for

example a type of larvae, as larvae transport should be simulated at a very high temporal

resolution. Miller (2007) provides a review of IBMs describing larval recruitment using

2-D and 3-D hydrodynamic simulation platforms and a variety of approaches to describe

particle transport and larval behaviour.

138


Other examples of models with both high spatial and temporal resolution include some

of the coupled hydrodynamic water quality and ecosystem models explained in section

1.3.2, such as particle transport models, and multi-layer ecosystem models used to assess

eutrophication and phytoplankton and nutrients dynamics. However, these models

(especially ecosystems models) usually have a relatively high complexity (as can be seen

in Table 1.2), so they are not commonly used with high temporal and spatial resolution

because of the computational requirements.

1.5 Modelling Tradeoffs

Due to their complexity, many coupled or linked hydrodynamic-ecosystem or even bio-

optical models are not used with high spatial or temporal resolution because of the

computational times and resources, and data requirements. This makes more difficult to

obtain quality and reliable modelling results, and limits the application of models to

research activities as they are slow and hard to understand and are not useful for decision

makers who need a quick answer to their questions.

Due to biological complexity, Levins (1966) supported that when building a model a

trade-off between generality, precision or realism should be chosen (Figure 1.6). There is

no best all-purpose model and there is no perfect strategy associated simultaneously for

maximizing the three aspects. In fact, in some cases models should prioritize precision

and realism at the expenses of generality. Whereas, in some other cases models misprize

realism for generality and precision. There could be also useful models that build on

generality and realism at the expense of precision in some other cases. In the water quality

and ecological modelling community, there was a tendency to make more and more

complex models to ensure that all relevant processes were included in the model.

However, nowadays there is a need in reducing model complexity due to the difficulties

in expressing all the relevant processes by a mathematical equation and finding the values

for all the induced parameters, which can end in more parametrization errors. Although

dick

Highlight

dick

Sticky Note

you probably mean "sacrifice". I have never heard of the word "misprize"

139


ecological modelling is a branch of science that has been strongly developed in the last

40 years, there is not yet a conclusion about the most appropriate trade-off in each case

(Fath et al., 2011), and all approaches are still implemented and applied for any ecological

problem or study site.

Therefore, the increasing complexity of ecological models is a growing concern in the

modelling community. Ecological models are used to integrate process knowledge from

different parts of the system, and in doing so allow us to test system understanding and

generate hypotheses about how the system will respond to particular actions via virtual

experiments. However, as we strive to make our models more ‘realistic’, the more

parameters and processes we include. With increased model complexity we are less able

to manage and understand model behaviour. We need enough complexity to realistically

model a process, but not so much that we ourselves cannot handle. In fact, a recent

biogeochemical model intercomparison study has shown that increasing model

complexity may not lead to increased skill or predictive ability (Friedrichs et al., 2006).

It is though that the ecosystem, the problematic, the available data and the objective of

the study (management or research) should determine the complexity and the balance

between generality, precision and realism of the model (Figure 1.6). In fact, a simpler

integrated approach, ease of use can give the reliable results needed to meet management

goals to protect marine resources and support their sustainable exploitation, and to help

decision makers in their strategies. Therefore it is essential to develop models for

addressing the complex impact of drivers and ecosystem responses, being essential

numerical models which can simulate and predict changes in the state of marine

ecosystem in response to different drivers and management scenarios, and who can

support the decision-making processes. Simplifications can be made depending on the

characteristics of the area under study and its problematic. Therefore, a detailed

knowledge of the study area is a fundamental issue in building or choosing the model.

dick

Sticky Note

Try this instead: However, as we strive to make our models more ‘realistic’, complexity increases as we include more processes that can be difficult to paramaterize accurately.

dick

Highlight

140


Figure 1.6. Model tradeoff between generality, precision, and realism, adapted from Levins (1966).

1.6 Discussion

The need to develop models for ecological and environmental management has been

increasing since the late 1970s. Indeed, models provide both a keener understanding of

causal relationships driving ecological functioning, and the quantitative knowledge which

is required for evaluation, at ecological and economic levels, of consequences of the

implementation of possible alternative scenarios of environmental policy options in

SECS. Many mathematical models have been developed with the main aim of gaining

insights into given biogeochemical and ecological processes, regardless of the relevance

of the models for management purposes. In most cases, high spatial variability has not

been taken into consideration, the proper reproduction of the hydrodynamics that rule

these complex systems has not had a fundamental role, and the complexity of these

models made them hard to use for decision makers while a tradeoff between generality,

realism and precision should have been done. Besides, most of the models that have been

applied to SECS and lead with eutrophication in recent years are complex (e.g. Zouiten

et al., 2013) and difficult to apply and interpret. Additionally, despite the complexity of

the existing models there is a lack of models leading processes like the effects of

141


eutrophication on underwater light climate and submerged aquatic vegetation between

others. The phytoplankton blooms produced by nutrient overloading, turbidity, CDOM

concentrations and increase in water depth due to sea-level rise are fundamental issues

that produce light attenuation in the water column and limit the photosynthesis process

and submerged aquatic vegetation growth. The study of this process from a spectral light

modelling perspective with high spatial resolution has scarcely been done for

management strategies, as the existing models that could lead with these characteristics

are complex and difficult to use. In fact, the degree of simplification and the selection of

the processes to be taken into account in a model is a difficult task that depends on the

system and requires a complete understanding of the processes that control it.

Additionally, we support that coupling and linking simple models could be an adequate

solution for many problematics when defining the model complexity and processes

needed. In fact, complex eutrophication models are usually been used equally for

hypertrophic and eutrophic systems, although the problematic and processes governing

each one are different. Moreover, the great number of equations and parameters of these

models make their calibration difficult and data demanding, and therefore expensive for

management purposes.

Consequently, there is a need of developing new ecological modelling tools, arising the

objectives of the present Thesis.

142


1.7 Objectives

As described above, the main objective of this Thesis is to develop novel ecological

modelling tools for assessing and describing the behaviour of semi-enclosed eutrophic

and hypertrophic coastal systems. Thus, the present Thesis provides a deeper knowledge

about the behaviour of semi-enclosed coastal systems such as the Albufera of Valencia,

and West Falmouth Harbour which are complex hypertrophic and eutrophic systems

difficult to study. The specific objectives of this Thesis are:

To analyse the differences between eutrophic and hypertrophic systems and the

limitations of the existing models.

To design, link and implement a simplified water quality model for a hypertrophic

heavily regulated semi-enclosed coastal system.

To design, couple and implement a system of models to describe the behaviour of

a semi-enclosed coastal eutrophic system and the influence of eutrophication on

light attenuation and on submerged aquatic vegetation.

To assess the models sensitivity, calibrate them with field data and apply them to

semi-enclosed coastal systems with cultural eutrophication problems and socio-

economic and environmental importance.

To analyse different factors, such as the effects of future nutrient reduction and

sea level rise in a eutrophic semi-enclosed coastal system, and the input and output

loads in a hypertrophic system performing a mass balance.

To carry out an analysis of the developed models limitations and the future

research needed.

143


1.8 Layout of Thesis

The Thesis is organized in five chapters that address the objectives of the study. Chapter

I includes an introduction and the research background, Chapter II the description and

problematic of the study sites, and Chapters III and IV are edited versions of published

articles in SCI journals that describe the models development and its applications. Finally,

Chapter V reveals the conclusions and proposes future research lines. The structure of

the Thesis and a brief description of the contents of each chapter can be seen as follows:

Chapter I: Introduction and research background

In the present chapter are explained: the motivations of the research, the socio-economic

importance of SECS, the different behaviour and characteristics of eutrophic and

hypertrophic systems, and the state of the art with the history, classification, description

and applications of the main existing models. Other important aspects such as an

evaluation of cases with different temporal and spatial resolution, and the modelling

trade-off between generality, precision and realism needed when building a model have

also been analysed. At the end of this chapter the specific objectives designed to solve

the needs raised, and the structure of the present Thesis are also described.

Chapter II: Study sites.

In Chapter II, a detailed description of the study sites is presented. The problematic and

importance of the Albufera of Valencia (Spain) and West Falmouth Harbour (USA),

which are hypertrophic and a eutrophic SECs respectively, is explained. This chapter also

includes a comparison between the two sites, assessing their differences, similarities and

modelling strategies.

144


Chapter III - Development and implementation of a simplified model for semi-enclosed

hypertrophic coastal systems. Application to a heavily regulated coastal lagoon.

A simplified ecological model was linked to a hydrodynamic model to assess the

eutrophication in hypertrophic heavily regulated coastal lagoons. This chapter presents

the model description, the numerical techniques, the calibration and validation, the

results, discussion and conclusions.

Chapter IV - Development and implementation of a coupled ecological modelling system

for semi-enclosed eutrophic coastal systems. Application to a groundwater-fed estuary

with submerged aquatic vegetation.

We have developed a modelling system to assess eutrophication and potential seagrass

habitat in eutrophic semi-enclosed coastal systems. We have coupled a hydrodynamic

model, a simple biogeochemical model, and a spectral irradiance model to describe the

eutrophication and its effect on the light climate of SECS. We have linked a bio-optical

model to the system coupled before, in order to have a reliable tool to assess the potential

habitat of submerged aquatic vegetation in SECS based on a light perspective.

The chapter describes the observational methods and results, the description and

assessment of the modelling system, the results, discussion and conclusions.

Chapter V - Conclusions and future research

In this chapter, the general and specific conclusions that arise from this study are

described. Future research lines and the impact and dissemination of the present Thesis

are also summarized.

145

Chapter II Study sites

Chapter II

2 Chapter II. Study sites

146


147


Chapter II. Study sites

2.1 Introduction

The presented study has been done in two SECS with different characteristics and some

similarities, the Albufera of Valencia (Spain) and West Falmouth Harbor (USA) (see

Figure 2.1). Both were selected for being SECS with the adequate characteristics for

assessing the developed models, and for having environmental problems for which

modeling strategies can be efficient tools to help solving them. The Albufera of Valencia

is a hypertrophic SECS with regulated connection with the sea, whereas West Falmouth

Harbor is a eutrophic SECS whose connection with the sea is limited but not regulated.

The singularity and problematic of each site, a comparison between them and the

modelling strategy are presented in the next sections.

Figure 2.1. a) Aerial view of the Albufera of Valencia Natural Park (abc.es) and b) West Falmouth

Harbor (fineartamerica.com).

148


2.2 Albufera of Valencia

The Albufera of Valencia is an oligohaline coastal lagoon situated on the Mediterranean

coast 12 km south of the city of Valencia, Spain (see Figure 2.2).

Figure 2.2. Albufera of Valencia, irrigation channels and sampling stations location.

Its average depth is approximately 0.90 m (TYPSA, 2005; Soria, 2006) and it is located

within the Natural Park of the Albufera of Valencia, which is mainly formed by rice fields.

The Natural Park of the Albufera of Valencia was included in the List of Wetlands of

International Importance of the Ramsar Conference in 1990 (Soria, 2006). It was

designated as SPAS (Special Protection Area for Birds) based on Directive 94/24/EC on

the 8th of June 1994 for the conservation of Wild Birds. Furthermore the Albufera of

Valencia has also been declared a Site of Community Importance (SCI).

149


Despite its environmental interest, the future of the lagoon is threaten by contamination

and silting (Canet et al., 2003). Currently the Albufera of Valencia can be considered a

highly modified water body with “bad” water quality according to the Water Framework

Directive (WFD) (Romo et al., 2008). This situation is produced by the contributions of

nutrients from the watershed and by its hydrological regulation.

Although in 1991 a partial removal of wastewaters in the lagoon commenced, after which,

as described by Romo et al. (2005), a reduction of 77% of phosphorus load was achieved,

the Albufera of Valencia has been hypertrophic since the 1970s, as it is subjected to many

different environmental pressures such as wastewaters coming from wastewater treatment

plants, chemical and manufacturing industries, scrapyards, and agriculture activities

among others (see Figure 2.3).

Figure 2.3. Some of the pollutant pressures that surround the Albufera of Valencia.

150


One of the most important pressures are the irrigation waters of the 14000 ha. of rice

fields that surround the Albufera, which are fed into the lagoon through irrigation

channels and gullies (see Figure 2.2 and Figure 2.4). Additionally, the communication of

the Albufera with the sea is carried out through three artificial channels called golas

(Pujol, Perellonet and Perelló) (see Figure 2.2 and Figure 2.5) whose water flow is

regulated by sluice gates that keep the lake level at the appropriate values for rice

cultivation (Roselló, 1979) and limit the hydrodynamic flow (García Alba et al., 2014).

Figure 2.4. Rice fields and irrigation channels surrounding the Albufera of Valencia.

151


Figure 2.5. Albufera of Valencia connection with the sea (“golas”), a) Gola Pujol, b) Gola Perellonet, c)

Gola Perelló.

In addition to this, waste water laden with phosphorus and ammonia come into the lagoon

through numerous ditches. In fact, combined sewer overflows (CSOs) as well as water

from sewage treatment plants from a population of 300,000 inhabitants and industrial

waste water are dumped into the Albufera. Between 1980-1988, the stress suffered by the

lagoon, due to the contributions of P and N, reached an average value of chlorophyll-a of

300 µg L-1 and a density of phytoplankton of 10 5 -10 6 ind mL-1, corresponding to a

biomass of 30-300mg L-1 (fresh weight) (Vicente and Miracle, 1992). Nine years after

the partial remove of wastewaters which started in 1991, the annual mean chlorophyll-a

composition decreased to 180 µg L-1 (Romo et al., 2005). However, despite these actions

the lagoon still presents serious hypertrophic conditions, as the situation has not

significantly improved since then (Usaquen et al., 2012). The evolution of water quality

in the lagoon over the years causes great concern and makes it necessary to take measures

for the management and recovery of the system due to its ecological and environmental

relevance.

152


An aspect that should not be forgotten is that in the Albufera the phytoplankton is

dominated by filamentous cyanobacteria (over 80%) (Miracle et al., 1984). These are

characterized by their ability to fix nitrogen and to produce cyanotoxins. There are other

groups of phytoplankton in the lagoon, such as diatoms and chlorophyta, but to a much

lesser extent. As far as the zooplankton is concerned, during most of the year the dominant

species is Brachionus angularis, a detritivore filter feeder that belongs to the rotifers and

their predator Acantocyclops belonging to the vernalis copepods (Vicente and Miracle,

1992). However, around February and March a “clear-water phase”, usually takes place

in the Albufera because of the substitution of the cyanobacteria plankton by micro algae

and the increase of the zooplankton grazing activity. As described by Romo et al. (2005),

during the clear-water phase, phytoplankton is dominated by chlorophytes and diatoms,

which replace the filamentous cyanobacteria. These authors support the idea that during

this period zooplankton grazing increases and Secchi depth reaches the lagoon bottom.

Moreover, the development of the zooplanktonic specie Daphnia magna occurs, which is

a Cladocera planktonic crustacean that disappears in late March and is responsible for

much of the consumption of phytoplankton in this period (Romo et al., 2008; Sahuquillo

et al., 2007). Furthermore, clear-water phases were observed in the lake with values under

10 µg L-1 of chlorophyll-a (Romo et al., 2008), which are characterized by the increase

of the zooplankton grazing (Romo et al., 2005) and the communication of the system with

the sea. However, this phase is produced only once a year during one month, and the rest

of the time the lagoon is considerably impaired. The economic consequences of its

contamination is also important due to the activities that are disappearing such as fishing,

which has dropped considerably in the area. Moreover, some fauna and flora species have

disappeared also due to high levels of pollutants. In fact, there used to be submerged

aquatic vegetation (potamogetun pectinatus) in the bottom that have disappeared (Blanco

and Romo, 2006). Figure 2.6 shows the kind of submerged aquatic vegetation that the

Albufera used to have and the aspect of the water column and bottom nowadays at

different areas of the lagoon. Nevertheless, as rice fields give to Valencia high economic

153


profits, and they need so many fertilizers and pesticides, the solution to this problem is

complex, and before taking any action having a model able to properly describe the

complex behaviour of the system in a simple but accurate way is the first step.

Figure 2.6. a) Submerged aquatic vegetation that used to be at the Albufera of Valencia (potamogetun

pectinatus); b,c,d) water column and bottom of the Albufera nowadays at different areas of the lagoon.

154


2.3 West Falmouth Harbor

West Falmouth Harbor is a eutrophic groundwater-fed estuary situated on the western

shore of upper Cape Cod, Massachusetts, USA (Figure 2.7). Tidal range at the harbor

entrance is 1.9 m during spring tides and 0.7 m during neap tides (Ganju et al., 2012). The

average depth is approximately 1 m, and the surface area is 0.7 km2. The Harbor is

connected to Buzzards Bay and ultimately the Atlantic Ocean through a 3 m deep, 150 m

wide channel constrained by rock jetties on both sides (Ganju et al., 2012) (see Figure

2.8). The Harbor is comprised of different sub-embayments (Outer Harbor, South Cove,

and Snug Harbor).

Figure 2.7. West Falmouth Harbor, site locations and input loads.

155


(a) (b)

Figure 2.8. West Falmouth Harbor connection with the sea (a) and view from the marked point (b)

Regarding the socio-economic importance of West Falmouth Harbor, it used to be a

shellfish recreational area. However, this activity has been forbidden due to bacterial

pollution (see Figure 2.9). This pollution is especially important in the harbor during the

summer, due to the increasing number of boats (see Figure 2.10) and population in the

area as it is located in a very touristic area due to their beaches and its short distance to

Boston. Moreover, due to this pollution a number of species have also disappeared, and

not only shellfish but also fishing activity and submerged aquatic vegetation has been

affected. Moreover, the eutrophication of the area makes the light penetration through the

water column difficult, so submerged aquatic vegetation dies. This has provoked the

accelerated disappearance of eelgrass meadows in the last decades. This has caused great

concern as seagrasses are considered one of the major primary producers in shallow

waters, they provide habitat and food for a variety of organisms, act as nursery for many

species, stabilize bottom sediments and baffle currents, and improve water quality by

filtering and trapping suspended matter.

156


Figure 2.9. West Falmouth Harbor closed shellfish activity due to pollution.

Figure 2.10. West Falmouth Harbor sailing activity.

157


Therefore, the presence of eelgrass (Zostera marina), fish, and shellfish communities in

the Harbor is particularly important from a habitat perspective. However, Costello and

Kenworthy (2011) showed that there has been an ecologically significant alteration of

eelgrass distribution in West Falmouth Harbor within the past decades. In 1981, eelgrass

meadows were found throughout the Harbor, with beds in Outer Harbor, South Cove, and

Snug Harbor (Costa, 1988) (see Figure 2.11).

Figure 2.11. West Falmouth Harbor seagrass (area delimited by red line) disappearance. Adapted from

http://buzzardsbay.org/historical-eelgrass-west-falmouth.htm.

Seagrasses disappeared first from South Cove in very little time. It is thought that this

disappearance was mainly due to a big storm that happened at the late 80’s, which

158


devastated that area. Then seagrasses started to disappear from Snug Harbor (see Figure

2.11) and although this disappearance was slower it has not stopped yet. At present,

seagrass meadows have been lost from the landward basins, and eelgrass beds are only

found in the Outer Harbor (see Figure 2.12 and Figure 2.13). In Snug Harbor, eelgrass

beds died off as of mid-summer of 2010 remaining only patches (Hayn, 2012). Moreover,

grasses present during the previous several seasons had very high epiphyte loads on their

blades and showed signs of considerable physiological stress (Hayn, 2012; Howarth et al.

2014).

Figure 2.12. Outer and Snug Harbors seagrass presence in 2010 (green area) and 2012 (blue area).

Adapted from Hayn (2012).

As of 2012, no eelgrass beds were present in Snug Harbor (see Figure 2.12). This pattern

of eelgrass loss from the landward portions of the Harbor expanding toward the seaward

regions is due to the excess nitrate loading to the Harbor, which has contributed to

eutrophication (Hayn et al. 2014; Howarth et al. 2014). This high nitrate load comes

mainly from groundwater which naturally flows from the Sagamore lens of the Cape Cod

aquifer; nitrate loads are high due to input from the Town of Falmouth Wastewater

Treatment Plant (FWTP). The FWTP was constructed in the mid 1980's and is located

159


landward of the Harbor, at a distance of 1 km east, with an average elevation of 30 m

above sea level (Howes et al., 2006). Since 2005, nitrate input to groundwater from the

FWTP has been substantially reduced due to an improvement in the sewage treatment.

Nevertheless, given the groundwater travel time between the FWTP and West Falmouth

Harbor (up to 10 years) (Kroeger et al., 2006), the effects of the nitrate loading reduction

were still not apparent as of 2012 (Hayn et al., 2014). Moreover, surveys indicate that

both the inner and outer basins would be capable of supporting eelgrass when the

watershed nitrogen loading rates reached the levels of 1979-1985 (Howes et al., 2006).

Therefore, it is thought that lowering nitrogen inputs to this system should provide the

possibility of recovering seagrass communities and benthic habitats.

Figure 2.13. West Falmouth Harbor seagrass meadows at outer harbor (a and b), seabed covered by

macroalgae at south cove (c) and with some Ulva lactuca at Snug Harbor (d) in 2012.

However, due to the complexity of the system, and its environmental and economic

relevance, a model able to describe the behaviour of this eutrophic SEC and to predict

160


future scenarios of seagrass presence-absence, and its relationship with eutrophication

and sea-level rise is needed.

2.4 Study sites comparison and modeling strategies

The Albufera of Valencia and West Falmouth Harbor are two SECS with both similarities

and differences. On the one hand, both are SECS with a cultural eutrophication problem,

the systems mean depth are approximately the same, and both are close to a population

area, having a big economic and environmental importance. On the other hand, their

specific characteristics are different as can be seen in Table 2.1.The Albufera of Valencia

is damaged in a bigger extent than West Falmouth Harbor, as it is close to a big city

(Valencia, Spain), it is a SECS whose connection with the sea is only a few times a year

by three artificial channels, and it is also surrounded by rice fields what makes the nutrient

input load excessive and uncontrolled. These factors have provoked the hypertrophic

status of the system, which has led to the disappearance of many species of fauna and

flora such as submerged aquatic vegetation. One of the main reasons of this disappearance

is the limited light penetration through the water column, which has led into a

disappearance of submerged aquatic vegetation and therefore in a decrease of oxygen

levels in the lagoon that have provoked also the disappearance of many kinds of fishes.

On the contrary, West Falmouth Harbor is a eutrophic SECS close to the Town of

Falmouth (Massachusetts, USA), with input loads coming from a wastewater treatment

plant through underground waters with long travel times (up to 10 years). This harbor is

a restricted SECS, but this restriction is provoked by two jetties so the connection with

the sea is constant. In this harbor, the submerged aquatic vegetation has been disappearing

during the years due to the limitation of the light penetration through the water column.

In fact, recent studies support the idea that the less light penetrates the water column in

this area, the more light requirements seagrasses develop (Kenworthy et al., 2014). All

these factors have been demonstrated to provoke the disappearance of seagrasses, and

161


there is also concern with the consequences of sea-level rise in this area due to the constant

although restricted connection with the sea.

Table 2.1. Albufera of Valencia and West Falmouth Harbor main characteristics comparison.

CHARACTERISTICS ALBUFERA OF VALENCIA WEST FALMOUTH HARBOR

SURFACE (Km2) 23.2 0.7

DEPTH (m) 0.9 1

POLLUTION SOURCES -WASTEWATER TREATMENT PLANTS AND SEPTIC SYSTEMS

- FERTILIZERS AND PESTICIDES FROM RICE FIELDS

-INDUSTRIAL WASTEWATER

-WASTEWATER TREATMENT PLANT

-SEPTIC SYSTEMS

TROPHIC STATUS HYPERTROPHIC EUTROPHIC

LIMITANT NUTRIENT P N

TYPE OF SEC CHOKED RESTRICTED

CONNECTION WITH THE SEA REGULATED WITH 3 CHANNELS AND GATES LIMITED BY 2 JETTIES

SEAGRASS PRESENCE NO YES

Focusing on the use of simple and accurate reliable models, the differences between them

make the modeling strategy to be different although both are eutrophic SECS. For a

hypertrophic model a simple NPZ model could be used, as phytoplankton blooms due to

the overabundance of nutrients, and zooplankton grazing are the main processes

governing the system. However, in a eutrophic system were light penetration through

water column could be still enough for seagrass survival, not only a simple

biogeochemical model would be needed, but also the use of bio-optical models, especially

if one of the main points of concern is seagrass disappearance.

Additionally, the hydrodynamic conditions are quite different in both SECS, being

necessary for the Albufera to take into account the anthropogenic regulation with the sea,

whereas in West Falmouth Harbor the tidal forcing could determine in a bigger extent the

system behavior. In fact, as West Falmouth Harbor is a shallow SEC whose connection

with the sea is not regulated, sea-level rise could have consequences in light attenuation

and therefore in seagrass habitats in future years. Finally, the three-dimensional character

of West Falmouth Harbor due to the hydrodynamics, chlorophyll-a concentration,

162


seagrass presence and light climate, versus the two-dimensionality of the Albufera of

Valencia where the communication with the sea is very limited and light does not

penetrate in the water column provoking no changes through it contributes to justify the

use of different modeling strategies, and consequently of developing two different models

applicable to SECS. The selection of each model depends on the system characteristics,

such as the connection with the sea, the level of eutrophication and the main important

physical and biological processes.

163

Chapter III Development and implementation of a simplified model for semi-enclosed hypertrophic

coastal systems. Application to a heavily regulated coastal lagoon.

Chapter III

3 Chapter III. Development and implementation of a

simplified model for semi-enclosed

hypertrophic coastal systems. Application

to a heavily regulated coastal lagoon.

164



165



Chapter III. Development and implementation of a simplified

model for semi-enclosed hypertrophic coastal systems.

Application to a heavily regulated coastal lagoon.

This chapter is an edited version of two published Articles:

“A model for describing the eutrophication in a heavily regulated coastal lagoon.

Application to the Albufera of Valencia (Spain).” by Pilar Del Barrio Fernández, Andrés

García Gómez, Javier García Alba, César Álvarez Díaz, José Antonio Revilla Cortezón.

Journal of Enviromental Management. 2012.

DOI: 10.1016/j.jenvman.2012.08.019

“Hydrodynamic modelling of a regulated Mediterranean coastal lagoon, the Albufera of

Valencia (Spain)” by Javier García Alba, Aina G. Gómez, Pilar del Barrio Fernández,

Andrés García Gómez, César Álvarez Díaz, Journal of Hydroinformatics. 2014

DOI: 10.2166/hydro.2014.071

Abstract

A simplified two-dimensional eutrophication model was developed to simulate temporal

and spatial variations of chlorophyll-a in semi-enclosed hypertrophic coastal systems.

This model considers the hydrodynamics of a coastal system with a heavily regulated

connection with the sea, taking into account the whole study area, the variability of the

input and output nutrient loads, the flux from the sediments to the water column, the

phytoplankton growth and mortality kinetics, and the zooplankton grazing (see Figure

3.1). The model was calibrated and validated by applying it to the Albufera of Valencia,

a hypertrophic SECS whose connection to the sea is strongly regulated by a system of

166



sluice-gates. The calibration and validation results presented a significant agreement

between the model and the data obtained in several surveys. The accuracy was evaluated

using a quantitative analysis, in which the average uncertainty of the model prediction

was less than 6%. The results confirmed an expected phytoplankton bloom in April and

October, achieving mean maximum values around 250 µg L-1 of chlorophyll-a. A mass

balance revealed that the eutrophication process is magnified by the input loads of

nutrients, mainly from the sediments, as well as by the limited connection of the lagoon

with the sea. This study has shown that the developed model is an efficient tool to manage

the eutrophication problem in semi-enclosed hypertrophic coastal systems.

Figure 3.1. Graphical Abstract

167



3.1 Introduction

Eutrophication is a widespread phenomenon in inhabited areas of the planet, being

considered one of the major threats to the health of marine and coastal ecosystems (Nixon,

1995). This phenomenon produces a very large increase of biomass in the system, a

serious impoverishment of the diversity, and a decline in the quality of the affected water

body (Chau and Haisheng, 1998). Due to the importance, complexity, and variability of

eutrophicated systems, mathematical models are essential tools to represent the degree of

eutrophication of natural water bodies (Fan et al., 2009; Chao et al., 2010). The

complexity of the models that describe eutrophication in aquatic systems ranges from

simple NPZ (Denman and Gargett, 1995; McClain et al., 1996), or NPDZ (Oschlies and

Garcon, 1999; Hood et al., 2003), to multi-nutrient, multi-species and size-structured

ecosystem models (Lima and Doney, 2004; Lopes et al., 2009; Sundarambal et al., 2010).

In fact, most of the available models assessing water quality in variable and high

productive environments like coastal lagoons (Ferrarin and Umgiesser, 2005; Everett et

al., 2007; Ohno and Nakata, 2008) are complex. Moreover, large data requirements and

high computational costs make them time consuming and expensive to develop (Lawrie

and Hearne, 2007). In addition, more complexity in an ecosystem model does not

necessarily improve model performance (Hood et al., 2003; Friedrichs et al., 2006). On

the contrary, models that use simpler formulations have lower computational demands

and can be easier to parameterize and interpret (Fulton, 2001). These mathematical tools

are usually coupled to physical models that range from 1-box models (Li et al., 1999;

Usaquen et al., 2012), which do not represent the heterogeneity of the entire system, to

full models (Skogen et al., 1995; Lima and Doney, 2004), which usually have fine

resolution grids and sophisticated numerical schemes to describe the system

hydrodynamics. However, in spite of the fact that spatial resolution and heterogeneity are

crucial characteristics in model performance (Fulton, 2001) the significant increase of the

use of low spatial resolution models (Baird et al., 2003; Everett et al., 2007) is common

168



in order to avoid high time consumption and computational demands. Hence, the

combination of a fine resolution grid, a sophisticated numerical scheme, and a simple

ecological model, could give a reasonable description of the observed phenomena in

complex aquatic systems.

Due to the complexity of describing eutrophication processes in semienclosed

hypertrophic coastal systems the use of mathematical models is needed. These numerical

tools are based on different assumptions and formulations to characterize the system. On

the one hand, there are models that present complex formulations (Fulton et al., 2004b ;

Everett et al., 2007;), which use a large number of parameters (ranging from 41 to 775).

These models focus on giving a detailed description of the biogeochemical processes and

interactions of the system. However, these complex models are usually applied with low

spatial resolution grids in order to reduce their computational costs (Martin, 1998; Fulton

et al., 2004a,b; Everett et al., 2007). On the other hand, there are models with simple

formulations that assume the system is a well-mixed box (Murray and Parslow, 1999;

Baird et al., 2003; Bastón, 2008;). This reduces computational demands but can only

provide a rough picture of the transport and distribution of the chlorophyll-a in the system.

Additionally, some models with complex formulations and high resolution grids (Lonin

and Tuchkovenko, 2001) do not have an expression to describe variable connections of

the system with the sea, which is a critical factor that must be implemented to successfully

describe the behaviour of the system. In fact, regulated connections are one of the most

complex scenarios in SECS and are usually strongly linked to economic and social

interests and high levels of eutrophication. There are also three-dimensional models such

as NEUTRO (Sundarambal et al., 2010), CE-QUAL-ICM (Cerco and Cole, 1993),

DELWAQ (Postma, 1988), and MOHID (Neves, 1985) which can be applied to these

systems. However, the shallowness of hypertrophic SECS justifies the deep-averaged

well-mixed assumption in order to simplify the model. In fact, most of the existing models

that operate with high resolution or three-dimensional formulations have a long execution

169



time, require a large amount of data for calibration and validation, and involve many

parameters that are usually difficult to measure and whose choice can affect model

outputs. This situation can improve substantially when simplifying the model equations.

In addition, modelling the regulated connection with the sea would help on the

understanding of these systems, as they are one of the most complex and less studied

cases. Therefore, in order to manage and predict the chlorophyll-a concentration in

semienclosed hypertrophic coastal systems, a simplified, high resolution model, which

takes into account regulated hydrodynamics, is needed.

The aim of this chapter is to develop a simple two-dimensional depth-averaged

eutrophication model, for semienclosed hypertrophic coastal systems. The model presents

high spatial and temporal resolution, takes into account the regulated connection with the

sea, considers the hydrodynamics of the whole study area, the variability of the input and

output loads of the system, the flux of nutrients from the sediment to the water column,

the phytoplankton growth and mortality, and the zooplankton grazing. Moreover, it is

able to accurately describe the chlorophyll-a distribution in these systems, with simple

formulations, 3 state variables, 14 parameters, and low computational demands. In this

study, we present the sensitivity analysis of the key model parameters, the calibration and

validation, and the application of the model to a hypertrophic SECS, the Albufera of

Valencia, a heavily regulated coastal lagoon, which was described in Chapter II.

3.2 Materials and methods

3.2.1 The hydrodynamic model

The definition and implementation of a hydrodynamic model for a hypertrophic SECS

shall be based on a deep knowledge of the system under study. The hydrodynamic

behaviour of the lagoon is controlled by several factors, such as the fresh water inputs

from the irrigation channels, the balance between precipitation and evaporation, the wind,

170



and the outflows through the golas. The last one is strongly dependent of the opening

regime of the sluice gates and the water level difference between the lagoon and the sea.

In order to assess these processes two different finite hydrodynamic models (a long wave

model and a wind model) were used in this work. The first one, a two-dimensional depth-

averaged model, was used to characterize the water circulation in the Albufera, the water

flows entering the lagoon through the irrigation channels, and those occurring through

the three golas to the Mediterranean Sea. The second one, a quasi three-dimensional

model, was applied to calculate the wind induced currents in the system. Both models

consider the whole domain (irrigation channels, lagoon, golas, sea), being the effect on

the water circulation of the outflows through the golas specifically included in the long

wave model.

3.2.1.1 The long wave model

The complex relations between the irrigation channels, the lagoon, the golas and the

Mediterranean Sea were analyzed using a long wave model which has been proved to

provide good results in shallow coastal and estuarine areas (García et al., 2010a; Barcena

et al., 2012). The model solves the depth-averaged three-dimensional Reynolds Averaged

Navier-Stokes equations, dividing the study area into rectangular cells to calculate

velocity and water surface elevation. Governing motion equations are expressed as

follows:

St

H

x

VH

x

UH

(Eq. 1)

bxsxyx

y

UHN

x

UHN

x

gH

xgHfVH

y

UVH

x

HU

t

UH

0

22

0

0

22

1

2 (Eq. 2)

171



bysy

o

yx

o

o

y

VHN

x

VHN

y

gH

ygHfUH

y

HV

x

UVH

t

VH

1)()(

.2

)()(

2

2

2

2

22

(Eq. 3)

Where U and V are the components of depth-averaged velocities in the x and y directions,

H is the water depth, g is gravity, S is the balance between precipitation and evaporation

(P-E), is the water surface elevation from mean sea level, f is the Coriolis parameter,

Nx and Ny are the horizontal eddy viscosity coefficients, 0 is the averaged density, sx,sy

are the friction terms in the water surface, and bx,by are the friction terms in the bed.

Water surface friction terms are expressed as a function of wind as:

22)(yxx

o

aa

o

sz WWWC

(Eq. 4)

22)(yxy

o

aa

o

sz WWWC

(Eq. 5)

Where Ca is a drag coefficient, a is the air density, and Wx and Wy are the wind velocities

in the x and y directions.

The bed friction terms are given by the following expressions:

HC

VUgUh

o

bx

2

22)(

(Eq. 6)

HC

VUgVh

o

by

2

22)(

(Eq. 7)

Where C is the Chezy loss friction coefficient which can adopt variable values depending

on water depth as follows:

172



K

HC

12log18 (Eq. 8)

where K is the Nikuradse roughness. Besides this formulation, the model also allows

specifying a constant Chezy friction coefficient.

As far as the water circulation is concerned, one important aspect should be mentioned,

which is how the connection between the lagoon and the sea through the golas was

modelled, as this connection is controlled by several gates. The effect of these hydraulic

structures on the lagoon discharge was included in the long wave model, using the

following weir discharge equation:

2/32

11

2/1

2/3

23

2

g

UhgbCQ g

(Eq. 9)

Where Q is the discharged flow, Cg is a discharge coefficient which was calibrated, b is

the weir width, g is the gravity constant, U1 is the upstream velocity, and h1 is the free

surface height over the weir. The discharged flow is correlated with the number of open

sluice gates through the discharge coefficient Cg. This coefficient determines the effect

of the flow through the golas on the lagoon velocity field calculated in the long wave

model.

3.2.1.2 The wind model

A quasi three-dimensional wind model, which takes into account the different structure

over the depth of horizontal velocities due to wind action, was used. This model provided

good results in shallow coastal areas (García et al., 2010b). Its governing equations are

the following:

0

t

H

x

VH

x

UH (Eq. 10)

173



HH

U

H

fVx

gy

UaV

x

UaU

y

UV

x

UU

t

U

sxsxsx

yx

000

5.018.0

402.0

402.0

(Eq. 11)

HH

V

H

fUx

gy

VaV

x

VaU

y

VV

x

VU

t

V

sysysy

yx

000

5.018.0

402.0

402.0

(Eq. 12)

Where

0

6.16

sxxa (Eq. 13)

0

6.16

sy

ya (Eq. 14)

Where sx,sy are the friction terms on the water surface.

3.2.2 Eutrophication model

The mathematical model proposed is a two-dimensional simplified numerical model

which solves the depth-averaged advection-dispersion equation for each water quality

variable selected. The shallowness that usually characterizes hypertrophic SECS along

with their low vertical variation justifies the depth averaged simplification (Calero et al.,

2003), so complete vertical mixing has been assumed. The developed eutrophication

model simulates water quality with respect to phytoplankton and soluble reactive

phosphorus in the water column. In this regard, phytoplankton is an indicator of primary

biomass producers and of chlorophyll-a present in the lagoon, and soluble reactive

174



phosphorus is the limiting nutrient in the system that usually controls the phytoplankton

growth when the main input flows come from fresh water sources, as is the case of many

hypertrophic SECS like the Albufera (Soria et al., 1987; Martín, 1998) being its N/P ratio

19.6. Based on this approach, phytoplankton growth was calculated as a function of the

soluble reactive phosphorus, temperature, and light intensity in the water column;

whereas its consumption was mainly focused on the endogenous respiration and the

zooplankton grazing activity.

As far as the phosphorus cycle is concerned, soluble reactive phosphorus (SRP) is utilized

by phytoplankton for growth, and is incorporated into phytoplankton biomass. As

observed by Chao et al. (2006), the various forms of organic phosphorus undergo settling,

hydrolysis, and mineralization, and are converted to inorganic phosphorus at temperature

dependent rates. Furthermore, phosphorus may interact with sediments through the

processes of adsorption, desorption, and bed release. It is important to mention that only

soluble reactive phosphorus (SRP) has been defined in the current simplified model

because it is considered the limiting nutrient of the system. Moreover, the initial

conditions of the model are the phytoplankton concentration (Cf), soluble reactive

phosphorus concentration (Cp), water temperature (T), solar radiation (Iom), and the light

extinction coefficient (Ke), all of which were initially assigned the mean measured values

for the first simulation period. Figure 3.2 describes the flow chart of the developed model,

and the main biological processes considered.

175



Figure 3.2. Model flow chart

176



3.2.2.1 Transport equation

The water quality model is coupled to the hydrodynamic model through the depth-

averaged transport equation Eq. 15, which integrates the advection and the diffusion

properties of the flow, as well as the main processes occurring in the water column.

i

i

y

i

x

iii Ry

CHD

yx

CHD

xy

vHC

x

uHC

t

HC

(Eq. 15)

Where Ci is the depth-averaged concentration of the substance i (soluble reactive

phosphorus and phytoplankton); H is the depth of the water column that is given by

, η being the free surface elevation and h the mean water depth; Dx and Dy are

the diffusion coefficients; u and v are the current velocity in the x and y directions; Ri

represents the chemical and biological transformations of the substance i.

3.2.2.2 Chemical and biological interactions

A system of two differential equations describes the main chemical and biological

transformations for soluble reactive phosphorus (Cp) and phytoplankton (Cf):

(Eq. 16)

f

fCDG

dt

dC (Eq. 17)

where Cp is the concentration of soluble reactive phosphorus (SRP) (g m-3); Cf is the

concentration of phytoplankton in (g m-3); Fs is soluble reactive phosphorus released from

the sediment; G is the growth rate of phytoplankton; D is the death rate of phytoplankton,

and apc is the phosphorus to carbon ratio in phytoplankton.

hH

fpc

spCGa

H

F

dt

dC

177



3.2.2.3 Phytoplankton growth

Phytoplankton growth is described by a first-order kinetic expression where the net

growth rate is defined as the difference between the growth (G) and the death (D) rates.

The proposed model considers the population as a whole, using the total biomass of the

phytoplankton present.

The growth rate of phytoplankton (G) in a natural environment is a complex function of

the phytoplankton present and its differing reactions to solar radiation, temperature, and

the balance between nutrient availability and phytoplankton requirements. The growth

rate of phytoplankton (G) depends on four basic components: the maximum growth rate

at 20 º C (Gmax), the temperature correction factor (GT), the light limitation factor, and the

nutrient effect (GN), as shown in the following expression:

(Eq. 18)

One of the factors that affect phytoplankton growth is temperature. The variation and

relationship between growth rate and temperature is described by Eppley (1972). This

relationship is determined by the temperature correction factor (GT). In order to define

this factor, the reference temperature has been fixed at 20 º C. This factor is expressed as:

(Eq. 19)

Where θ is a temperature coefficient.

Moreover, the degree of penetration of sunlight into the water column has a significant

effect on phytoplankton growth, as phytoplankton needs sunlight to carry out its

photosynthetic function. The light limitation factor, GI, allows for photosynthesis to

increase with light levels up to a maximum, after which further increases in light result in

photo-inhibition (Tkalich and Sundarambal, 2003). The highest productivity occurs under

NITmáxGGGGG

20 T

TG

178



conditions of constant temperature and nutrients for a given light intensity known as

optimal intensity. Additionally, in hypertrophic systems, the overabundance of

phytoplankton, the big light attenuation, and the consequent absence of submerged

vegetation justifies the simplification of using non-spectral radiation.

The penetration of incoming solar radiation is described by the Lambert-Beer equation:

𝐼𝑧 = 𝐼0 𝑒𝑥𝑝 (−𝐾𝑒𝑍) (Eq. 20)

Where Iz is the light intensity at depth z (ly day-1) calculated from the light surface

intensity (I0) and the light extinction coefficient Ke (m-1).

The light limitation function of Steele and Baird (Thomann and Mueller, 1987) is used in

the model. The vertically averaged light limitation factor (GI) over a given water depth is

integrated as:

01

7182 expexp

HK

.G

e

I (Eq. 21)

Where

HKI

Ie

S

T exp1 (Eq. 22)

S

T

I

I0

(Eq. 23)

IT being the photosynthetically active solar radiation at the water surface (ly day-1), and

IS the saturating light intensity of phytoplankton (ly day-1). The photosynthetically

179



active solar radiation at the water surface (IT) is calculated using the equation given by

Kremer and Nixon (1978):

bmT C.II 71010 (Eq. 24)

Where Iom (ly day-1) is half the incident solar radiation, and Cb is the cloudiness (oktas).

The final effect on the growth that must be evaluated is the impact of varying nutrient

levels on the growth rate of the phytoplankton. The nutrient limitation factor (GN) which

describes the nutrient effect on the growth of the phytoplankton is expressed as a function

of dissolved inorganic phosphorus (Cp) in the form (Chapra, 1997):

pmp

p

NCK

CG (Eq. 25)

Where the constant, Kmp, is the Michaelis or half-saturation constant of phosphorus.

3.2.2.4 Phytoplankton death

Phytoplankton mortality is described as the sum of the phytoplankton endogenous

respiration and the zooplankton grazing. Therefore the mortality rate of phytoplankton

(D) can be expressed as the sum of two components (Thomann and Mueller, 1987):

zr DTKD (Eq. 26)

Where Kr (T) is the endogenous respiration of phytoplankton as a function of temperature

and Dz is the death rate due to zooplankton grazing.

The endogenous respiration of phytoplankton represents the processes by which the

phytoplankton oxidizes its organic carbon into CO2. The endogenous respiration rate of

phytoplankton (Kr) varies with temperature as follows (Di Toro and Matystik, 1980):

180



20

T

rrr TK (Eq. 27)

Where μr is the endogenous respiration rate of phytoplankton at 20 °C, and θr is a

temperature coefficient for phytoplankton respiration.

The loss of phytoplankton due to zooplankton grazing by herbivorous zooplankton is

proportional to the concentration of zooplankton present in the environment. Therefore

the mortality rate due to grazing (DZ) can be expressed as (Thomann and Mueller, 1987):

(Eq. 28)

where Cg is the grazing (filtering) rate of zooplankton (L mgC-1 day-1), which is the rate

at which the zooplankton feed on the phytoplankton, and Z is the zooplankton

concentration in equivalent carbon units (mgC L-1).

3.2.2.5 Chlorophyll-a concentration

The concentration of chlorophyll-a (CCHL-a) expressed in μgChl-a L-1 is a good indicator

of the eutrophication level of an aquatic system and can be obtained from the

concentration of phytoplankton by the following expression (Thomann and Mueller,

1987):

(Eq. 29)

where Cf is the concentration of phytoplanktonic carbon (mgC L-1) and aC / CHL-a is the

carbon to chlorophyll-a ratio (mgC / mgChl-a).

ZCDgz

f

aCHLC

aCHL Ca

C

/

1000

181



3.3 Numerical techniques

The transport equation Eq. 15 was solved using the eulerian technique, with an explicit

finite-difference discretization scheme based on the split operator approach, in which

advection and diffusion processes are computed independently for each time-step (García

et al., 2010b). Hence, different numerical methods were used to solve each process.

Advective transport was computed using an upwind scheme, whereas diffusion was

described through a centred scheme. Moreover, the system of differential equations that

describes chemical and biological interactions in the transport equation was solved by the

4th order Runge-Kutta integrator, with a relative and absolute tolerance of 10-8. The model

also includes an algorithm which adjusts the time-step based on two numerical stability

criteria, Courant (Courant and Hilbert, 1962) and Peclet (Fletcher, 1991). The time-step

for the model was 20 seconds, in order to ensure stable conditions, and a constant

horizontal diffusion coefficient of 0.4 m2 s-1 was used. The model was coded in Fortran

90 (non-parallel codification), and it is able to run a 12-month simulation with high spatial

(345x300 cells) and temporal resolution (time-step 20 s) in 4.46 hours using an Intel Core

i7 2.3 GHz with 8 Gb RAM.

3.3.1 The numerical grid

Both hydrodynamic and eutrophication models were applied over the same numerical

grid. This grid consists of 345x300 square cells with a cell dimension of 50.0 m, to

provide the high spatial resolution of the model. The bathymetric data was taken from

detailed topographic works developed for the entire Albufera Park by the Polytechnic

University of Valencia and TYPSA (2005), and the Navigation Charts of the State Naval

Hydrographical Institute (numbers 47, 48, 474, 481, 482 and 791). This numerical grid

includes the irrigation channels that discharge into the lagoon, the lagoon itself, its

connection to the sea through the three golas, and the coastal area. Due to the complexity

of the system and to the high number of discharge points in the Albufera, a simplification

182



was introduced in the numerical grid in order to limit the complexity of the model. The

sewage and irrigation water has been distributed into thirteen main irrigation channels.

3.3.2 Field data and model set up

This study was carried out for the hydrological year 2005/2006, where the available data

was enough for calibrating and validating the model. The main input data of the model

were measured by the Entidad Pública de Saneamiento de Aguas de Valencia (EPSAR)

between October 2005 and September 2006, at seven sampling stations distributed within

the lagoon. The collected data included chlorophyll-a, temperature, SRP and Secchi

depth, which were sampled monthly, taking 12 samples for each variable and station

throughout the hydrological year, making a total of 336 samples. The location of the

sampling stations, the irrigation channels, and the golas, can be observed in Figure 2.1.

Furthermore, SRP measurements from the main irrigation ditches were sampled during

each month of the study period, in order to describe the nutrient concentrations flowing

into the lagoon, making a total of 156 SRP samples. The maximum values of SRP were

found in the north irrigation channels (Alfafar to Beniparrell), due to the origin of this

kind of nutrient, which is mainly urban and industrial.

Chlorophyll-a levels were not measured for each individual phytoplankton species, since

in the Albufera the phytoplankton is strongly dominated by filamentous cyanobacteria

(over 80%) (Miracle et al., 1984). Instead, the total chlorophyll-a level was measured.

This simplification has been used in recent studies (He et al., 2011). The mean

concentration of chlorophyll-a is 115.7 μg l-1, meaning that the Albufera of Valencia is a

hypertrophic system. The mean Secchi depth varies between 0.12 and 0.36 metres, which

means that the light penetration is blocked in the water column. Consequently, the light

extinction coefficient values are considerable, in agreement with Martin (1998). With the

observed data of chlorophyll-a and Secchi disc depth in seven sampling stations

distributed around the lagoon, an expression to describe the variation of the light

183



extinction coefficient (Ke) with the chlorophyll-a concentration was obtained as can be

seen in Eq. 30 and Figure 3.3.

74.0

0457.04245.3

2

r

CK aCHLe (Eq. 30)

where CCHL-a is the chlorophyll-a concentration in µg L-1.

Figure 3.3. Variation of the light extinction coefficient (Ke) with the chlorophyll-a concentration.

The calculated regression coefficient (r2) in Eq. 30 shows a positive correlation between

the concentration data and the obtained light extinction coefficient (Ke) values. The good

obtained value of r2 enables us to use this expression to describe the relationship between

the chlorophyll-a concentration (CCHL-a) and the light extinction coefficient (Ke).

Therefore the light extinction coefficient (Ke) was calculated as a function of the

chlorophyll-a concentration (see Eq. 30).

184



Climatological data for the hydrological year 2005/2006, such as cloudiness and wind

parameters, were obtained for each hour from the Valencia Viveros Meteorological

Station (10 km from the Albufera). The cloudiness was measured by The Meteorological

State Agency (Agencia Estatal de Meteorología) on a scale of 0-8 oktas, 0 oktas being

the minimum cloud coverage and 8 oktas the maximum. The cloudiness evolution at the

Albufera of Valencia during the hydrological year 2005/2006 can be seen in Figure 3.4.

Figure 3.4. Average cloudiness variation during the hydrological year 2005/2006.

The most frequent wind was from the southeast, being in general of low intensity.

However, in some cases, wind events reached intensities above 3 m s-1, thus being able

to transport the pollutants from the banks of the lagoon to its centre, due to the

shallowness of the lagoon. Meteorological data, water temperature, and nutrient loads

were introduced into the model as input data per hour, showing the temporal resolution

of the model.

Another important input parameter that has been taken into account in this work is the

flux of soluble reactive phosphorus from the sediment to the water column. This

185



parameter was obtained from a specific one day field survey, in order to characterize the

influence of the SRP flux from sediments. Data were sampled once at 17 stations in order

to describe the spatial distribution of the SRP flux in the lagoon.

3.4 Calibration and validation

In the present study the calibration and validation of the water quality model is described.

Moreover, the hydrodynamic model was also calibrated taking into account the discharge

flow through the golas, which is characterised by a weir discharge coefficient (Cg) (see

Eq. 9). The hydrodynamic calibration was conducted through the comparison of the

predicted model results with water lagoon levels measured at the Pujol gola during the

simulation period (see Figure 3.5). The values of the discharge coefficient Cg varied from

0.2 to 0.5 as a function of the number of opened sluice gates.

Figure 3.5. Comparison between calculated and observed lagoon water surface during the period October

2005-September 2006 in a point of the lagoon located in front of gola Pujol.

As far as the eutrophication model is concerned, the calibration of this model involves

adjusting the parameter rates so that the model output fits the measured values in some

periods. In addition, a sensitivity analysis which describes the effect of model parameters

on the model output (Van Griensven et al., 2006) was carried out. Once calibrated, the

186



model was validated for a hydrological year, in order to confirm the values of the

parameters previously set.

3.4.1 Sensitivity analysis

Global sensitivity analysis is used as an initial screening tool to identify the most

influential model parameters (Arhonditsis and Brett, 2005). The ranges shown in Table

3.1 for each of the major model parameters were used to set acceptable parameter limits

for the model calibration. All of the parameter ranges were assigned based on published

literature values (Di Toro and Matystik, 1980; Thomann and Mueller, 1987; Ambrose,

1988; Chau and Haisheng, 1998; Martín, 1998), or sampling surveys.

Table 3.1. Range of variation and assigned calibration value of the main eutrophication parameters of the

model.

PARAMETER DESCRIPTION UNITS ASSIGNED RANGE

ASSIGNED VALUE

apc Phosphorus to carbon ratio gP gC-1 0.011-0.025a,d 0.011***, d

Kr Endogenous phytoplankton respiration rate

day -1 0.05-0.5b,a,d 0.12***, g, h, d

Kmp Half-saturation constant of phosphorus

mgP L-1 0.001-0.005c 0.0027***, d

aC/CHL-a Carbon to chlorophyll-a ratio mgC mgChl-a-1 50 – 133d 88***, d

Gmax Maximum phytoplankton growth rate

day-1 1.5-2.5c 1.5***, i

Is Saturating light intensity of phytoplankton

Ly day-1 100-400c * (see Eq.33)

Fs Factor of soluble reactive phosphorus from the sediment

mgP m-2 day-1 5-50e 20.72 **

Cg Grazing (filtering) rate of zooplankton

LmgC-1day-1 0.05-0.3f 0.3 ***, f

Source: a(Ambrose, 1988), b(Di Toro and Matystik, 1980), c(Thomann and Mueller, 1987), d(Martín, 1998), e field data, f (Chau and Haisheng, 1998); g (Lindenschmidt, 2006); h (Ambrose et al., 1993); i (Parslow et al., 1999).

*The values were assigned by empiricism;**The values were field data; ***The values were verified by calibration and literature.

For the sensitivity analysis, eight adjustable parameters were fixed at the mean of their

defined range, given in Table 3.1. Each simulation was performed with one of the

187



parameters fixed to its minimum or maximum value, and the remaining free parameters

set to the medium value of their assigned range. This process was repeated for each

parameter. The variation of mean concentration of chlorophyll-a over the whole lagoon

was used for the evaluation of the sensitivity.

The histogram in Figure 3.6 reveals that the parameters for which chlorophyll-a was

highly sensitive are Kr, Cg, Gmax, Fs, and acp. It is important to mention that Kr, Cg and

Gmax directly alter growth rates, whereas Fs and acp affect the possibility to take in or

utilise phosphorus. This is in good agreement with the results obtained in other studies

(Fasham et al., 1990; Schladow and Hamilton, 1997; Wu et al., 2009), in which Kr was

the sensitivity parameter that most affects chlorophyll-a concentration and phytoplankton

growth.

Figure 3.6. Mean chlorophyll-a (Chl-a) concentration calculated for the period of the sensitivity analysis

in the Albufera of Valencia for the minimum and maximum calibration parameter values.

188



3.4.2 Model calibration

The model free parameters were calibrated by trial-and-error adjustment to give the best

match with trends in the measured field data over four calibration periods, which were

selected because they usually present the minimum or maximum values of chlorophyll-a.

The selected periods were October, February, May, and July. In October and May a

phytoplankton bloom usually occurs, and during February and July the concentration of

chlorophyll-a is usually low in comparison with the rest of the year. The parameters can

be varied within a certain range (see Table 3.1) to match the model results with the

measurements of the most critical variable regarding water quality in the system (Fasham

et al., 1990; García et al., 2010b); being, in this case, chlorophyll-a concentration. This

methodology is in agreement with Lonin and Tuchkovenko (2001), Martin (1998) and

Garcia et al. (2010b). The comparison between measurements and model predictions was

performed with the help of different types of errors calculated between the observed and

predicted values for a given variable. The errors calculated were absolute error (AE),

relative error (RE), mean relative error (MRE), root mean square error (RMSE),

normalised root mean squared error (PRMSE), mean absolute error (MAE), normalised

mean absolute error (NMAE), and BIAS. The formulation of these errors can be seen in

Table 3.2.

189



Table 3.2. Error formulations applied in the calibration process. Фi is the calculated concentration in cell

i, Фiobs is the observed concentration in cell i and N is the number of cells analyzed.

DESCRIPTION FORMULATION

Absolute error

Relative error

Mean relative error

Root mean square error

Normalised root mean squared error

Mean absolute error

Normalised mean absolute error

BIAS

The model calibration was carried out taking into account the most influential parameters

from the sensitivity analysis, and the seasonal parameters. Despite the complex nature

and high variability of the Albufera, the assigned value of most of the parameters used in

the model were kept constant over all the periods. However, different values of the light

intensity saturation (Is) have been used for distinct simulation periods. Light intensity

saturation (Is) changes depending on the temperature and the season, being minimal in

winter and maximal in summer (Macedo et al., 2001). Moreover, Is is positively correlated

with temperature and can be described as a function of temperature, using an Arrhenius

iobsiAE

100(%)1

N

i iobs

iobsiRE

1001

(%)1

N

i iobs

iobsi

NMRE

N

i

iobsi

NRMSE

1

2

100obs

RMSEPRMSE

N

i

iobsi

NMAE

1

N

i

iobs

iobsi

NNMAE

1

N

i

iobsi

NBIAS

1

190



equation (Falkowski and Raven, 1997). A linear equation describing Is as a function of

temperature was obtained (see Eq. 31), based on the maximum and minimum

temperatures measured at the lagoon and the Is range described by Thoman and Mueller,

1987.

(Eq. 31)

Where Is is the saturation light intensity (ly day-1) and T is the water temperature (º C).

Equation 31 was applied to obtain a value of Is for the rest of the periods, using the mean

water temperature of the lagoon for each month. As Is changes with temperature following

an Arrhenius equation, the data obtained was approximated to Eq.32.

ln(𝐼𝑠) = 6.557 − 18.7741

𝑇 (Eq. 32)

Finally, this logarithmical equation was solved, and was included in the model as follows

(see Eq. 33):

𝐼𝑠 = 704.156𝑒−18.774

𝑇 (Eq. 33)

Another important input parameter that has been calibrated in this work is the flux of

soluble reactive phosphorus from the sediment to the water column. This parameter data

was obtained through a specific field survey developed to characterise the influence of

the SRP flux from sediments. Data were sampled at 17 stations in order to represent the

spatial distribution of the SRP flux in the lagoon. Sampling data were interpolated

following the Kriging Gridding Interpolation Method (Kitanidis, 1997). The results,

measured data, and location of the sampling stations, are presented in Figure 3.7. As can

be seen in Figure 3.7, the maximum diffusive flux is found in the north of the lagoon, due

561.47516.15 TI s

191



to the high discharges of SRP coming from the northern ditches. In addition, a mean SRP

flux value of 20.72 mgPm-2day-1 was found in the lagoon.

Figure 3.7. Distribution of soluble reactive phosphorus (SRP) flux from the sediment to the water column

into the lagoon.

Additionally, a simulation process was carried out in order to fix some of the most

significant free parameters of the model, including Kr, acp, Kmp, ac/chl-a, Gmax, and Cg, which

were adjusted to achieve the lowest error between observed data and model results. The

assigned values of these parameters corresponding to the best fit are described in Table

3.1, and are in good agreement with those of the published literature (Ambrose et al.,

1993; Chau and Haisheng, 1998; Martín, 1998; Parslow et al., 1999; Lindenschmidt,

2006). Moreover, due to the model sensitivity to the endogenous phytoplankton

respiration rate, special care has been taken in order to set the assigned value, which is

that commonly recommended by published literature (Ambrose et al., 1993;

Lindenschmidt, 2006).

192



As can be observed in Table 3.3, there are differences between the values of the errors

calculated at the selected calibration periods where the light intensity saturation was

adjusted. The errors summarised in Table 3.3 were calculated using the observed data at

the seven sampling stations for each month compared with the model output obtained at

the same points. In fact, the month with the lowest mean relative error is February,

whereas the highest mean relative error occurs in May. The root-mean-square error

(RMSE) and the mean absolute error (MAE) have also been calculated. Each of these

measures is “dimensioned”, meaning that it expresses average model-prediction error in

the units of the variable of interest, in this case the chlorophyll-a concentration in μg L-1.

As each period has different levels of chlorophyll-a, depending on the inputs and the

weather conditions, the percent mean square error (PRMSE) has also been calculated for

every calibration period, and has been compared with the normalised mean absolute error

(NMAE). As a result, the lowest values of both errors have been found in the month of

May, being PRMSE 7.30% and NMAE 8.50%.

Table 3.3. Errors obtained with the chlorophyll-a values of all the sampling stations in each calibration

period.

Period MRE (%) RMSE(μg L-1) PRMSE (%) MAE(μg L-1) NMAE (%) BIAS(μg L-1)

October 2.52 26.94 11.95 23.10 10.70 2.41

February 1.75 10.46 19.26 8.52 16.13 -1.22

May -6.56 16.12 7.30 12.45 8.50 8.36

July -3.62 6.23 14.76 4.84 14.01 -0.28

The BIAS has also been calculated, and July is the month where the BIAS is closest to

zero. In fact, the values calculated by the model were about 0.28 μg L-1 lower than the

observed ones. Afterwards, the overall errors for each month were calculated comparing

the mean chlorophyll-a concentration given by the model for the whole system with the

averaged data measured at the sampling stations. This methodology is in agreement with

193



Haggard et al. (1999). Simulations showed that for October, February, May and July, the

relative errors obtained comparing the mean observed data with the mean model results

are between 3-5%. The absolute errors are between 1.86 and 7.15 µg L-1, as can be seen

in Table 3.4.

Table 3.4. Global errors obtained with the mean chlorophyll-a concentration of each calibration period in

the whole lagoon.

Period Фi (μg L-1 ) Фiobs (μg L-1) Фi - Фiobs (μg L-1) EA (μg L-1) ER (%)

October 232.59 225.44 7.15 7.15 3.17

February 52.47 54.33 -1.86 1.86 -3.42

May 174.33 167.20 7.13 7.13 4.27

July 39.99 42.23 -2.23 2.23 -5.28

As can be seen in Figure 3.8, the comparison between the results given by the numerical

model and the measured data for the best fit is adequate. In addition, Figure 3.8 shows

the evolution of chlorophyll-a concentration at the seven sampling stations and the lagoon

averaged concentration for the whole lagoon in the calibration periods. Concerning

chlorophyll-a, the best adjustment was found at station A2, and the global relative error

for the whole lagoon is less than 6%. The oscillation of the calculated values describes

the variation produced between day and night.

194



Figure 3.8. Comparison between results obtained by the model (solid lines) and the observed data (black

dots) in the sampling stations and the whole lagoon for each calibration period.

3.4.3 Model validation

Once calibrated, the numerical model was validated for the remaining months of the year

during which the calibration was not carried out. These are November, December,

January, March, April, June, August, and September. Model validation implies verifying

that the parameter values assigned in the calibration process are the best to describe the

degree of eutrophication of the Albufera of Valencia. Therefore, errors have been

calculated at each station for the validation periods (see Table 3.5). As can be observed

195



in Table 3.5 and Figure 3.9, the station that best fits the chlorophyll-a concentration is

C2, with a mean relative error of 1.02%, whereas the worst adjustment occurs at A1, with

a mean relative error of 47.74%. Additionally, all the stations except for A1, A2, and C1

have mean relative errors of under 16%, which is an acceptable value for the spatial

validation.

Table 3.5. Errors of the different sampling stations obtained with the validation periods.

MRE (%) RMSE(μg l-1) PRMSE (%) MAE(μg l-1) NMAE (%) BIAS(μg l-1)

A1 47.74 40.72 38.24 29.35 59.92 14.67

A2 29.20 31.28 28.74 21.80 30.95 18.85

A3 14.21 20.82 18.21 16.67 20.72 6.77

B1 -3.69 30.92 25.55 24.31 19.22 -13.57

B2 15.65 34.35 27.83 24.07 24.14 4.51

C1 22.70 37.76 31.01 28.05 39.11 -1.83

C2 1.02 12.63 10.63 11.41 13.56 -3.23

196



Figure 3.9. Evolution of the simulated chlorophyll-a concentration (“full line”, calculated) and the

observed data (“black dots”, observed) in each sampling station, for the validation periods of the

hydrological year 2005/2006.

As far as the temporal validation is concerned, Figure 3.10 shows the comparison between

lagoon-averaged calculated and observed data for the hydrologic year 2005/2006, in the

months where the validation has been carried out. The overall relative error obtained is

5.81 %, indicating that the model is able to reproduce the observed data values and trends.

197



Figure 3.10. Evolution of the lagoon-averaged chlorophyll-a concentration calculated by the model (“full

line”, calculated) for the hydrological year 2005/2006 and the observed values (“black dot”, observed) for

the validation periods.

3.5 Results and discussion

Our results provide compelling evidence that the proposed model is effective in

describing the chlorophyll-a distribution in the Albufera with high spatial resolution (see

Figure 3.11). In addition, the spatial distribution is in accordance with the temporal

evolution as shown in Figures 3.10 and 3.11, where October and April are the months

that present higher chlorophyll-a concentrations, whereas February and March are the

ones that present lower concentration values of chlorophyll-a. These findings are in good

agreement with those of Romo et al. (2008), achieving the lowest levels of chlorophyll-a

in February and March and the highest levels in October and April. Additionally, the

chlorophyll-a concentration values obtained in the present study are consistent with

results obtained by Romo et al. (2005) with maximum mean values between 200 and 250

µg L-1. The increase of the chlorophyll-a concentration in the month of October is caused

by the warm temperature of the water (23 ºC), the hydrodynamism, and the large input

loads of nutrients during that period. Between September and October the harvest of the

rice fields takes place (Villena and Romo, 2003) and there is a large quantity of nutrients

198



that go into the lagoon, increasing the eutrophication process. Similar behaviour was

observed by Menendez et al. (2002) in the Buda lagoon (Spain), which is also a

Mediterranean coastal lagoon where the fresh water comes from the irrigation of rice

fields. This lagoon also shows a phytoplankton bloom in the month of October, as the

input loads are determinated by the rice cultivation. In April and May a large number of

nutrients arrive at the Albufera from the fertilizers and pesticides used in the preparation

of the surrounding rice fields. February and March, on the other hand, present

significantly lower chlorophyll-a concentrations. In February most of the gates of the

output channels are open, so the hydrodynamic of the lagoon increases considerably. In

March there is also a renewal of water, and the input loads of nutrients are lower than in

February. This nutrient reduction leads to an increase of the zooplanktonic species

Daphnia magna, which is a Cladocera planktonic crustacean that is the responsible for

much of the consumption of phytoplankton in this period (Romo et al., 2005). As a

consequence of this, zooplankton grazing effect increases, so a clear water phase is

produced in this period, and the chlorophyll-a concentration is the lowest of the year. In

view of these results, the connection of the lagoon with the sea, the zooplanktonic grazing,

and the nutrient loads directly affect the eutrophication of this heavily regulated coastal

lagoon. In order to evaluate the limiting nutrient influence on the chlorophyll-a evolution,

a mass balance has been carried out to determine the input and output SRP loads of the

lagoon.

199



Figure 3.11. Chlorophyll-a spatial distribution in the Albufera of Valencia in the hydrological year

2005/2006.

200



The net mass balance analysis of pollutant inputs to an ecosystem minus outputs from

that system provides a measure of how it is coupled with adjacent systems as, in this case,

on the coast of Valencia. On the one hand, the input loads taken into account in the mass

balance are the sewage water and nutrients that come from the surrounding townships and

from the irrigation water streams, particularly from 14,000 ha. of rice fields that surround

the Albufera. On the other hand, the output load includes the SRP flux that goes to the

sea, which depends on the sluice gates opening regime. In addition, the mass balance

reveals that most of the SRP that comes into the lagoon remains in it. The total SRP load

that enters the Albufera is 26.4 t y-1, and the output load is 5.8 t y-1. These results are

similar to those of Burger et al (2008) in Lake Rotorua, which is a eutrophic lake that has

an SRP input load of 27.5 t y-1.

Using the method described above, we found that April was one of the months with the

highest SRP input load (see Figure 3.12). In April, the chlorophyll-a concentration

increases mainly due to the high SRP input load, the warm temperature, and the light

intensity in the water column. During autumn the load of SRP that comes into the lagoon

is also high (see Figure 3.12). In October both SRP concentration and water flow in the

irrigation channels increase, making it one of the months with the highest input load of

SRP. As can be seen in Figure 3.12, in all the months of the 2005/2006 hydrological year

the SRP input load was considerably higher than the output load, especially in October

and April, where there was no outflow because the golas are almost completely closed.

In these months the input SRP load was considerably higher than that of the other months,

and is equal to the net load accumulated in the Albufera lagoon, which produces the main

eutrophication problems. Additionally, it is important to note that in the Albufera of

Valencia the SRP input load coming from the irrigation channels is only approximately

35 percent of the total SRP water column input load, whereas the sediment SRP flux to

the water column constitutes approximately 65 percent (IHCantabria, 2009). Once in the

water column, the SRP can be assimilated by phytoplankton or can become adsorbed to

201



sediment as particulate phosphorus, which can be dissolved again, returning to the water

column. Hence, we can conclude that the phytoplankton blooms are directly affected not

only by the temperature, but also by the input load of nutrients, the SRP flux to the water

column, and the connection of the lagoon with the sea.

Figure 3.12. Mass balance of the soluble reactive phosphorus loads that comes in and out of the Albufera

of Valencia.

In order to assess the accuracy of the model and to make it suitable for other similar

situations, statistical model evaluation techniques were applied. The simulated

chlorophyll-a values were positively correlated to the measured values with a Pearson

correlation coefficient of 0.933 for the calibration periods and of 0.917 for the validation

periods (see Figure 3.13). The Nash-Sutcliffe efficiency coefficient (Moriasi et al., 2007)

has also been calculated, resulting in a value of 0.96, which is considered as excellent

according to Usaquen et al. (2012). Finally, the error indices were used to quantify the

deviation between measured and calculated data. In this study, the calculated deviation

was about 5.81%. Due to the mathematical simplicity formulation, the high resolution

capacities and the accuracy of the developed model, it could be extended to other heavily

regulated aquatic systems.

202



Figure 3.13. Comparison between calculated and observed chlorophyll-a calibration and validation data

As far as the sensitivity analysis is concerned, the results obtained are in agreement with

those presented by Wu et al. (2009) applied to the Fuchunjiang Reservoir. The results of

this sensitivity analysis show that the parameters that strongly influence simulated

chlorophyll-a concentration are the endogenous respiration, the grazing rate of

zooplankton, the maximum growth rate and the sediment SRP flux.

3.6 Conclusions

In this study a simple two-dimensional eutrophication model, for hypertrophic SECS was

developed. This model was successfully tested in the Albufera of Valencia, a

hypertrophic system whose connection to the sea is strongly regulated. The chlorophyll-

a concentration in the Albufera of Valencia was used to calibrate and validate the

proposed model, as well as to assess its sensitivity.

203



Better knowledge of the influence that the main sensitivity parameters have on the model

was achieved. These parameters were proved to be the main ones for the model

calibration. Therefore the sensitivity analysis permits a reduction in the number of

parameters to be adjusted. We can conclude that the parameters for which chlorophyll-a

is highly sensitive are Kr, Cg, Gmax, Fs and acp. The endogenous respiration rate, Kr, is the

dominant parameter affecting the chlorophyll-a concentration, as it directly alters the

phytoplankton growth.

The model gave us a better understanding of the system. In fact, the results of the

modelling concluded that there were phytoplankton “blooms” in April and October, due

not only to the temperature, but also to the high nutrient loads and the lagoon-sea

connection characteristics. Nevertheless, the results confirmed that a “clear water phase”

took place around the month of March, mainly due to the nutrient reduction and the

zooplankton grazing effect. Moreover, the zooplankton species Daphnia magna is

primarily responsible for the predation on phytoplankton during the “clear water phase”.

Furthermore, a quantitative statistical analysis was applied to determine modelling

uncertainties between the measured and calculated data. The average uncertainty of the

model prediction for this study was less than 6%, which is an acceptable limit, with two

Pearson correlation coefficients of 0.933 and 0.917 for calibration and validation

respectively and a Nash-Sutcliffe efficiency coefficient of 0.96, which are excellent

values. Therefore, the modelled results demonstrated that a simplified model can

characterise eutrophication in heavily regulated SECS.

As demonstrated by the calculated mass balance, the input loads in the lagoon are higher

than the output loads, so the limited connection of the lagoon with the sea magnifies the

eutrophication of the system. Furthermore, the SRP flux from the sediment to the water

column contributes to maintaining high chlorophyll-a concentrations.

204



The results confirmed that the model constitutes a valuable tool for the eutrophication

management in heavily regulated hypertrophic SECS like the Albufera of Valencia, being

able to describe, with high temporal and spatial resolution, the chlorophyll-a

concentration evolution during a whole year.

205

Chapter IV Development and implementation of a coupled ecological modelling system for semi-enclosed

eutrophic coastal systems. Application to a groundwater-fed estuary with submerged aquatic vegetation.

Chapter IV

4 Chapter IV. Development and implementation of a

coupled ecological modelling system for

semi-enclosed eutrophic coastal systems.

Application to a groundwater-fed estuary

with submerged aquatic vegetation.

206



207



Chapter IV. Development and implementation of a coupled

ecological modelling system for semienclosed eutrophic coastal

systems. Application to a groundwater-fed estuary with

submerged aquatic vegetation.

This chapter is an edited version of a published Article:

“Modeling future scenarios of light attenuation and potential seagrass success in a

eutrophic estuary.” by Pilar Del Barrio, Neil K. Ganju, Alfredo L. Aretxabaleta, Melanie

Hayn, Andrés García, Robert W. Howarth. Estuarine, Coastal and Shelf Science. 2014.

DOI: 10.1016/j.ecss.2014.07.005

This work has also been successfully presented to the scientific community at the

following events:

Workshop: Linking hydrodynamic and ecological models in estuaries: a workshop to

discuss recent advances and approaches. Presentation: “A modeling approach to

assess light availability and potential seagrass success under nitrate loading and sea

level rise scenarios”. Authors: Pilar del Barrio, Neil K. Ganju, Alfredo L. Aretxabaleta,

Melanie Hayn, Andrés García, Robert W. Howarth. Woods Hole, Massachusetts, USA.

September 10-11, 2013.

MABPOM2012 Symposium. Poster presentation: "A seagrass and light attenuation

model for a eutrophic estuary: Calibration, validation and predictions under nitrogen

loading scenarios". Authors: Pilar del Barrio, Neil K. Ganju, Alfredo L. Aretxabaleta.

Groton, Connecticut, USA. Nov. 2012.

208



Abstract

Eutrophication in SECS has led to numerous ecological changes, including loss of

seagrass beds. One potential cause of these losses is a reduction in light availability due

to increased attenuation by phytoplankton. Future sea level rise will also tend to reduce

light penetration and modify seagrass habitat. In the present study, we integrate a spectral

irradiance model into a biogeochemical model coupled to the Regional Ocean Model

System (ROMS). It is linked to a bio-optical seagrass model to assess potential seagrass

habitat in a eutrophic SECS under future nitrate loading and sea-level rise scenarios. The

model was applied to West Falmouth Harbor, a shallow semienclosed estuary located on

Cape Cod (Massachusetts) where nitrate from groundwater has led to eutrophication and

seagrass loss in landward portions of the estuary. An schematic description of the model

and the processes can be seen in Figure 4.1. Measurements of chlorophyll-a, turbidity,

light attenuation, and seagrass coverage were used to assess the model accuracy. Mean

chlorophyll-a measurements varied from 28 µg L-1 at the landward-most site to 6.5 µg L-

1 at the seaward site, while light attenuation ranged from 0.86 to 0.45 m-1. The model

reproduced the spatial variability in chlorophyll-a and light attenuation with RMS errors

of 3.72 µg L-1 and 0.07 m-1 respectively. Scenarios of future nitrate reduction and sea-

level rise suggest an improvement in light climate in the landward basin with a 75%

reduction in nitrate loading. This coupled model can be useful to assess chlorophyll-a

variation and seagrass habitat availability changes from a light perspective. It also fully

considers spatial variability on the tidal timescale in eutrophic SECS.

209

Chapter IV Development and implementation of a coupled ecological modelling system for semi-enclosed eutrophic coastal systems.

Application to a groundwater-fed estuary with submerged aquatic vegetation.

Figure 4.1. Graphical Abstract

210



4.1 Introduction

In eutrophic SECS light may penetrate through the water column and can be the limiting

factor controlling the presence of submerged aquatic vegetation (SAV) such as

seagrasses. Seagrass meadows are found in many coastal areas around the world and are

regarded as key indicators of ecosystem health (Dennison et al., 1993). They are among

the most productive plant communities, and represent one of the major sources of primary

production in shallow waters worldwide (Hemminga and Duarte, 2000). These plants

serve as a nursery for many species, providing habitat and food for a variety of marine

organisms (Orth et al., 2006). They also trap nutrients, thereby improving water

transparency and filtering substantial quantities of both N and P from estuarine waters,

serving as a buffer between land-based pollution sources and adjacent estuaries (Nixon et

al. 2001; Short and Short 2004; McGlathery et al., 2007; Hayn et al., 2014). Consequently,

the increasing loss of seagrass beds raises concern because of a potential reduction in

coastal ecosystem productivity, a decrease in water quality, and a decline in fishing

resources. Additionally, in a report prepared for the European Union, Terrados and Borum

(2004) estimate the value of ecosystem services provided by seagrasses as two orders of

magnitude higher than productive agricultural lands.

Despite the ecological and economic value of seagrass meadows, their disappearance has

accelerated in the last decades (Short and Wyllie-Echeverria, 1996; Waycotta et al.,

2009). The causes of decline range from natural disturbances (e.g., storms) to

anthropogenic pressures (e.g., nutrient loading). In temperate estuaries and coastal areas,

one of the dominant factors for seagrass loss is eutrophication (Short and Neckles, 1999;

Orth et al., 2006). In eutrophic SECS, there is an overabundance of nutrients that leads to

phytoplankton blooms, an increase in epiphytes growing on seagrass tissues, and

subsequent light reduction (Burkholder et al., 2007). This reduction can impede seagrass

growth and its ability to assimilate nitrogen, as they are vascular benthic autotrophs that

dick

Highlight

dick

Sticky Note

Waycott

211



require clear water and high levels of Photosynthetically Active Radiation (PAR). In fact,

minimum light requirements of seagrasses (2-37% of surface irradiance, SI) are much

higher than those of macroalgae and phytoplankton (about 1-3% of SI) (Dennison et al.,

1993; Lee et al., 2007). Therefore, seagrass photosynthesis, and thereby their growth,

survival, and depth distribution, are directly linked to PAR reaching the plant surface

(Cabello-Pasini et al., 2003). The spatial variation in light availability in eutrophic SECS

can cause changes in the spatial distribution of seagrass on the order of meters. Another

aspect that should be taken into account is that the allocation and abundance of seagrasses

have changed over evolutionary time in response to sea-level rise (SLR) (Orth et al.,

2006). In areas where the tidal range increases, plants at the lower edge of the bed will

receive less light at high tide, which increases plant stress, reduces photosynthesis, and

therefore decreases the growth and survival of the vegetation (Short and Neckles, 1999;

Titus et al., 2009). The complexity and variability of eutrophic SECS with seagrass

meadows highlights the need for a spatially explicit model that can resolve spatial

distributions of chlorophyll, turbidity, colored dissolved organic matter (CDOM), and

ultimately light attenuation. There are relatively few coupled hydrodynamic-light models

that calculate light attenuation as a function of different attenuating substances apart from

chlorophyll and water (Everett et al., 2007; Hipsey and Hamilton, 2008), and even fewer

take into account spectral underwater irradiance (Bissett et al., 1999a, b).

In the present study, we develop a new tool to assess eutrophication and potential seagrass

habitat in eutrophic SECS. We have used a three-dimensional circulation model

(Regional Ocean Model System, ROMS) coupled to a Nutrient Phytoplankton

Zooplankton Detritus (NPZD) eutrophication model (Fennel et al., 2006), where we have

integrated a spectral light attenuation formulation (Gallegos et al., 2011). We describe the

model and the linkage of this tool with a benthic seagrass model (Zimmerman, 2003),

which calculates seagrass distribution. We apply the model to West Falmouth Harbor, a

212



eutrophic SECS where we assess the effects of future nitrate loading and sea level rise

scenarios on seagrass habitat and eutrophication. In the sections that follow we describe:

the observational methods and results, the numerical model and skill assessment, and

future scenarios of nitrate loading and sea-level rise. Finally, we discuss the utility and

limitations of the approach and future directions.

4.2 Observational methods

We deployed instrumentation in West Falmouth Harbor (described in Chapter II) to

measure meteorological, hydrodynamic, water quality, and light conditions during

summer 2012 (See Figure 4.2 ).

Figure 4.2. Field survey stations and equipment

Meteorological data were measured at 1 minute intervals by an Onset weather station

from 28 June 2012 to 11 September 2012. Parameters included wind direction, wind

speed, atmospheric pressure, relative humidity, shortwave radiation, PAR, and air

temperature. The subaqueous instrument platform consisted of a Nortek Aquadopp

ADCP (water velocity), a SeaBird SeaCat (pressure), a YSI 6600 multisonde (salinity,

temperature, chlorophyll-a, turbidity, dissolved oxygen), and a pair of WetLabs ECO-

213



PARSB sensors (PAR). All sensors were located 0.3 mab except for the upper PAR sensor

located at 0.8 mab. The PAR sensors were equipped with wipers to prevent bio-fouling.

The chlorophyll-a values were obtained by a YSI 6025 sensor located in the YSI 6600

multisonde. This sensor uses a light source with a peak wavelength of 470 nm which

provokes the chlorophyll-a emission of light between 650 – 700 nm. The output of the

sensor is automatically processed via the sonde software, which provides the chlorophyll-

a (µg/L) readings. Measurements were collected at 5 min intervals from 3 to 19 July 2012

in Outer Harbor, from 19 July 2012 to 9 August 2012 in South Cove, and from 9 to 27

August 2012 in Snug Harbor. Due to the fact that no large intra-seasonal changes were

observed between July and August during a previous field survey in 2010 (Ganju et al.,

2011), the data collected in each location were considered representative of the season.

Following Gallegos et al. (2011), we calculated the diffuse attenuation coefficient of

downward propagating irradiance, Kd, as:

upper

lowerd

PAR

PAR

zK ln

1 (Eq. 34)

where Kd is the light attenuation coefficient, z is the distance between the two sensors

(0.5 m), PAR lower and PAR upper are the PAR measurements near the bottom and just

below the water surface respectively during daylight.

To determine the areal extent of seagrass beds in West Falmouth Harbor we conducted

surveys during early June 2012, using side scan sonar. We used an EdgeTech, Inc. 4125

towfish and EdgeTech Discover software to collect acoustic data at 900 and 1250 kHz

along survey transects spaced to provide 200% bottom coverage in the survey area, and

horizontal positions were provided by a Trimble AgGPS 132 with U.S.Coast Guard

beacon differential corrections. We pre-processed the data in Chesapeake Technology

Inc. SonarWiz5 to adjust for signal attenuation through the water, georeference the data,

and export georeferenced imagery to ArcGIS 10.1 for classification. We manually

214



delineated the seagrass beds in ArcGIS after examining ground-truthed locations to

calibrate our image interpretation, and verified the final extent by surveying a random set

of locations using a combination of surface and underwater observational techniques.

Groundwater fluxes and associated nitrate concentrations (Figure 4.3) were obtained in

previous studies (Kroeger et al., 2006; Ganju et al., 2012; Hayn et al., 2014).

Figure 4.3. Groundwater fluxes and nitrate concentrations at West Falmouth Harbor. Arrows indicate

main contributions from Falmouth Wastewater Treatment Plant (FWTP).

4.3 Observational results

Mean values of water column properties (temperature, salinity, pH, dissolved oxygen and

turbidity) were overall spatially similar except for chlorophyll-a, which was highest at

site Snug (Table 4.1).

215



Table 4.1. Mean values and standard deviation (Std) of measurements.

FIELD

MEASUREMENT

UNITS

MEAN ± STD

OUTER SNUG SOUTH

Chlorophyll-a µgL-1 6.46 ± 2.75 27.50 ± 9.90 10.22 ± 9.26

Turbidity NTU N/A* 4.26 ± 1.49 3.41 ± 1.39

Temperature °C 24.64 ± 0.84 25.84 ± 0.91 24.37 ± 0.81

Salinity psu 30.94 ± 0.31 29.74 ± 0.65 30.51 ± 0.56

pH - 8.01 ± 0.10 8.06 ± 0.49 7.97 ± 0.32

Dissolved oxygen mgL-1 7.17 ± 1.17 7.54 ± 1.44 6.75 ± 1.22

* Turbidity at site Outer was compromised by reflective copper tape (for anti-fouling) accidentally placed

near the optical window. The tape tarnished within two weeks and did not affect subsequent measurements

at sites Snug or South.

Chlorophyll-a measurements suggested more eutrophication at landward ends of the

harbor, with a mean value of 28 µg L-1, whereas in the outer harbor the mean

concentration was 6.5 µg L-1. Accordingly, site Snug demonstrated considerably lower

PAR values than site Outer (Table 4.2). PAR data from site South were not obtained due

to instrument malfunction.

Table 4.2. Mean values, standard deviation (Std) and percentile 84 of measured optical data during

daylight hours.

FIELD

MEASUREMENT

UNITS

MEAN ± STD PERCENTILE 84

OUTER SNUG OUTER SNUG

PARupper µE/m2s 504 ± 387 301 ± 300 945 545

PARlower µE/m2s 416 ± 327 198 ± 209 795 363

It is important to note that for the Kd calculation we only considered PAR values over the

84th percentile of the distribution, which corresponds to the hours of highest light

incidence, usually around noon. These values were selected because when a beam of light

impacts the water surface perpendicularly or with low angles measured from the vertical,

216



most of the light penetrates the water column, and the scattering on the water surface is

minimal. However, the larger the incident angle, the less light penetrates the water column

and the less accurate the PAR measurements become. We have used PAR values during

times when sunbeams impact the water surface with low incidence angles to minimize

this effect.

The mean diffuse light attenuation coefficient Kd was 0.45 m-1 at site Outer and 0.86 m-1

at site Snug. Statistical distribution of chlorophyll-a measurements and Kd between the

two sites confirms that the larger light attenuation coefficient present at site Snug is

consistent with elevated chlorophyll-a concentration (Figure 4.4). Prior measurements

showed that CDOM is spatially uniform and relatively low in West Falmouth Harbor

(absorbance at 440 nm < 0.01 m-1; M. Hayn, unpublished). These measurements also

indicate that turbidity is relatively low with minimal spatial differences. Therefore there

is a strong relationship between eutrophication (and ensuing chlorophyll-a levels) and

light attenuation in the study area.

Figure 4.4. Chlorophyll-a and Kd field data histograms in Outer and Snug Harbors. Data collected from

sensors deployed during summer 2012 with a 5 minutes sampling interval.

217



4.4 Model description

We integrated a spectral irradiance model (Gallegos et al., 2011) into an existing NPZD-

biogeochemical model (Fennel et al., 2006) to compute the spectral penetration of PAR

through the water column. The coupled optical-biogeochemical model uses the PAR from

the irradiance model to calculate phytoplankton growth. This model was integrated in the

ROMS 3D circulation model (Haidvogel et al., 2008) that simulates the three-dimensional

hydrodynamics. The computed PAR and Kd are provided to a bio-optical model

(Zimmerman, 2003) that calculates the seagrass carbon balance under the estimated light

climate (Figure 4.5). The carbon balance allows the prediction of seagrass

presence/absence and its potential survival. The capabilities of the linkage of these models

provide an integral description of the physical, optical, and biological dynamics of the

water column. This allows us to define a success criterion for assessing seagrass future

evolution in a eutrophic SECS, based on light climate alone. In this section, we present a

brief description of the main processes of each different model, although further

information can be found in their respective references.

218



Figure 4.5. Modeling system flowchart and interactions. In the bottom panel, P/R ratio > 1 indicates

potential seagrass habitat; P/R < 1 indicates potential loss of seagrass habitat. U and V are the velocities,

h the water depth, T the water temperature and η the water surface variation.

219



4.4.1 Physical model

The circulation model used is the Regional Ocean Modeling System (ROMS)

(Marchesiello et al., 2003; Shchepetkin and McWilliams, 2005; Warner et al., 2005a;

Haidvogel et al., 2008). ROMS is a three-dimensional, free-surface, terrain-following

numerical model that solves the Reynolds-averaged Navier-Stokes equations using the

hydrostatic and Boussinesq assumptions (Haidvogel et al., 2008). The main equations and

parameters of the hydrodynamical core of this model can be seen at Table 4.3.

In the physical configuration adopted, the outer domain is 2 km in the north-south

direction centered on 41.6° latitude and 1.5 km in the east-west direction centered on -

70.64° longitude (Figure 2.7) and includes the entire West Falmouth Harbor estuarine

system. The horizontal grid spacing is 10 m (150x200 grid points). The grid has 10

vertical levels using an evenly spaced vertical stretching. The spatial discretization of

such a model allows for the representation of the spatial heterogeneity of the estuary in

terms of light climate.

The model was forced at the western boundary (Figure 2.7) with tidal free surface

elevation, velocity, salinity and temperature. Additionally, the atmospheric forcing

included wind velocities, atmospheric pressure, shortwave radiation, surface air

temperature and relative humidity. These data were obtained from the weather station

located in the study area (Figure 4.2) and were applied as surface forcing in the entire

computational domain. Groundwater fluxes, nitrogen loads, and fresh water temperature

were given to the model as point sources (Figure 4.3). These fluxes were quantified from

velocity and salinity measurements and a Total Exchange Flow (TEF) methodology

(Ganju et al., 2012).

220



Table 4.3. Physical model main equations and parameters

ROMS HYDRODYNAMIC CORE (SHCHEPETKIN AND MCWILLIAMS, 2005; WARNER ET AL., 2005a; HAIDVOGEL ET AL., 2008)

𝒖, 𝒗, and w: components of velocity in the horizontal (x and y) and vertical (scaled sigma coordinate, s) directions respectively. ζ: wave-averaged free-surface elevation. h: depth of the sea floor below mean sea level. Hz: vertical level thickness. f: Coriolis parameter. p: Pressure. ρ and ρ0: total and reference densities. g: acceleration due to gravity. 𝐯 𝐚𝐧𝐝 𝐯𝟎: Molecular viscosity and diffusivity. C: tracer quantity. Csource: tracer source/sink terms. KM: eddy viscosity for momentum. KH: eddy diffusivity for tracers. Note: An over-bar represents a time average, and a prime (‘) represents turbulent fluctuations.

Reynolds-averaged Navier Stokes equations

𝜕(𝐻𝑧𝑢)

𝜕𝑡+

𝜕(𝑢𝐻𝑧𝑢)

𝜕𝑥+

𝜕(𝑣𝐻𝑧𝑢)

𝜕𝑦+

𝜕(𝑤𝐻𝑧𝑢)

𝜕𝑠− 𝑓𝐻𝑧𝑣 = −

𝐻𝑧

𝜌0

𝜕𝑝

𝜕𝑥− 𝐻𝑧𝑔

𝜕𝜁

𝜕𝑥−

𝜕

𝜕𝑠(𝑢′𝑤′ −

v

𝐻𝑧

𝜕𝑢

𝜕𝑠)

𝜕(𝐻𝑧𝑣)

𝜕𝑡+

𝜕(𝑢𝐻𝑧𝑣)

𝜕𝑥+

𝜕(𝑣𝐻𝑧𝑣)

𝜕𝑦+

𝜕(𝑤𝐻𝑧𝑣)

𝜕𝑠− 𝑓𝐻𝑧𝑢 = −

𝐻𝑧

𝜌0

𝜕𝑝

𝜕𝑦− 𝐻𝑧𝑔

𝜕𝜁

𝜕𝑦−

𝜕

𝜕𝑠(𝑣′𝑤′ −

𝑣

𝐻𝑧

𝜕𝑣

𝜕𝑠)

0 = −1

𝜌0

𝜕𝑝

𝜕𝑠−

𝑔

𝜌0

𝐻𝑧𝜌

The continuity equation:

𝜕𝜁

𝜕𝑡+

𝜕(𝐻𝑧𝑢)

𝜕𝑥+

𝜕(𝐻𝑧𝑣)

𝜕𝑦+

𝜕(𝐻𝑧𝑤)

𝜕𝑠= 0

Scalar transport: 𝜕(𝐻𝑧𝐶)

𝜕𝑡+

𝜕(𝑢𝐻𝑧𝐶)

𝜕𝑥+

𝜕(𝑣𝐻𝑧𝐶)

𝜕𝑦+

𝜕(𝑤𝐻𝑧𝐶)

𝜕𝑠= −

𝜕

𝜕𝑠(𝑐′𝑤′ −

𝑣0

𝐻𝑧

𝜕𝐶

𝜕𝑠) + 𝐶𝑠𝑜𝑢𝑟𝑐𝑒

State equation: 𝜌 = 𝑓(𝐶, 𝑝)

Reynolds stresses and turbulent tracer fluxes

parametrization: 𝑢′𝑤′ = −𝐾𝑀

𝜕𝑢

𝜕𝑧

𝑣′𝑤′ = −𝐾𝑀

𝜕𝑣

𝜕𝑧

𝑐′𝑤′ = −𝐾𝐻

𝜕𝜌

𝜕𝑧

Variable Grid Location 𝒖 X 𝒗 C, Csource, ρ, ρ0

w, KM, KH

221



4.4.2 Biogeochemical model

The phytoplankton dynamics are simulated using a biogeochemical nutrient,

phytoplankton, zooplankton, and detritus (NPZD) model (Fennel et al., 2006). This model

is implemented into ROMS, and assumes nitrogen as the controlling nutrient for primary

production. Therefore, it is based on the nitrogen cycle (Figure 4.5), and includes the

source, sink, and biogeochemical transformation terms of seven state variables: nitrate

(NO3), ammonium (NH4), small and large detritus (SDet and LDet), phytoplankton

(Phy), zooplankton (Zoo), and chlorophyll-a (Chl). We added the effect of seagrass

nutrient uptake in order to account for its influence in nutrient cycling. This is represented

in Figure 4.5 by the arrow that goes from NH4 to the sediment. We assumed that the

uptake decreases with depth as seagrass biomass and production are strongly related with

light availability (Cunningham, 2002). The mean nitrogen uptake by seagrasses in the

bottom layer of the model varies between 0 and 10 mmolNm-2day-1 (Hemminga et al.,

1991; Risgaard-Petersen et al., 1998; Hansen et al., 2000; Risgaard-Petersen and Ottosen,

2000), describing the nitrogen removal due to seagrass.

The main biogeochemical model equations were described by Fennel et. al (2006), who

adapted them from the plankton dynamics model of Fasham et al. (1990). In the Fennel

implementation, phytoplankton growth is a function of temperature, nutrient

concentration, and the homogenously integrated PAR distribution. The main equations

and parameters of this model can be seen at Table 4.4.

222



Table 4.4. Irradiance model main equations and parameters

BIOGEOCHEMICAL MODEL (FENNEL ET AL., 2006)

mp: Phytoplankton mortality.

: Aggregation parameter. KNO3: Half-saturation concentration for uptake of NO3 KNH4: Half-saturation concentration for uptake of NH4 μr: Phytoplankton growth rate at reference temperature (Tref) T: Temperature α: Initial slope of the P-I curve I: Photosynthetically available radiation lBM: excretion rate due to basal metabolism lE: maximum rate of assimilation related excretion. β: Assimilation efficiency. mz: Zooplankton mortality. gmax: Maximum grazing rate kp: Half-saturation concentration of phytoplankton ingestion. Θmax: Maximum chlorophyll-a to phytoplankton ratio. rSD: Remineralization rate of suspended detritus. rLD: Remineralization rate of large detritus. wPhy: Sinking velocity of phytoplankton wSDet: Sinking velocity of suspended detritus wLDet: Sinking velocity of larger particles. nmax: Maximum nitrification rate. I0: Threshold for light-inhibition of nitrification KI: half-saturated light intensity of nitrification inhibition.

Phytoplankton balance: 𝜕𝑃ℎ𝑦

𝜕𝑡= 𝜇𝑃ℎ𝑦 − 𝑔𝑍𝑜𝑜 − 𝑚𝑝𝑃ℎ𝑦 − 𝜏(𝑆𝐷𝑒𝑡 + 𝑃ℎ𝑦)𝑃ℎ𝑦 − 𝑤𝑃ℎ𝑦

𝜕𝑃ℎ𝑦

𝜕𝑧

Growth rate of phytoplankton: 𝜇 = 𝜇𝑚𝑎𝑥𝑓(𝐼) (𝑁𝑂3

𝐾𝑁𝑂3+𝑁𝑂3∙

1

1+𝑁𝐻4

𝐾𝑁𝐻4

+𝑁𝐻4

𝐾𝑁𝐻4+𝑁𝐻4)

Maximum phytoplankton growth rate:

𝜇𝑚𝑎𝑥 = 𝜇𝑟1.066𝑇−𝑇𝑟𝑒𝑓

Photosynthesis-light (PI) relationship:

𝑓(𝐼) =𝛼𝐼

√𝜇𝑚𝑎𝑥2 + 𝛼2𝐼2

Photosynthesis available radiation:

𝐼 = 𝐼0 ∙ 0.43 ∙ 𝑒𝑥𝑝{−𝑧𝐾𝑑}

Light attenuation coefficient:

𝐾𝑑 = 𝐾𝑤 + 𝐾𝑐ℎ𝑙 ∫ 𝐶ℎ𝑙(𝑧)𝑑𝑧0

𝑧

Zooplankton balance: 𝜕𝑍𝑜𝑜

𝜕𝑡= 𝑔𝛽𝑍𝑜𝑜 − 𝑙𝐵𝑀𝑍𝑜𝑜 − 𝑙𝐸

𝑃ℎ𝑦2

𝐾𝑃+𝑃ℎ𝑦2 𝛽𝑍𝑜𝑜 − 𝑚𝑧𝑍𝑜𝑜2

Chlorophyll balance: 𝜕𝐶ℎ𝑙

𝜕𝑡= 𝜌𝐶ℎ𝑙𝜇𝐶ℎ𝑙 − 𝑔𝑍𝑜𝑜

𝐶ℎ𝑙

𝑃ℎ𝑦− 𝑚𝑝𝐶ℎ𝑙 − 𝜏(𝑆𝐷𝑒𝑡 + 𝑃ℎ𝑦)𝐶ℎ𝑙

Relationship between chl-a and phytoplankton:

𝜌𝐶ℎ𝑙 =𝜃𝑚𝑎𝑥𝜇𝑃ℎ𝑦

𝛼𝐼𝐶ℎ𝑙

Rate of phytoplankton grazing by zooplankton:

𝑔 = 𝑔𝑚𝑎𝑥

𝑃ℎ𝑦2

𝐾𝑝 + 𝑃ℎ𝑦2

Small Detritus balance:

𝜕𝑆𝐷𝑒𝑡

𝜕𝑡= 𝑔(1 − 𝛽)𝑍𝑜𝑜 + 𝑚𝑧𝑍𝑜𝑜2 + 𝑚𝑝𝑃ℎ𝑦 − 𝜏(𝑆𝐷𝑒𝑡 + 𝑃ℎ𝑦)𝑆𝐷𝑒𝑡 − 𝑟𝑆𝐷𝑆𝐷𝑒𝑡 − 𝑤𝑆𝐷𝑒𝑡

𝜕𝑆𝐷𝑒𝑡

𝜕𝑧

Large Detritus balance: 𝜕𝐿𝐷𝑒𝑡

𝜕𝑡= 𝜏(𝑆𝐷𝑒𝑡 + 𝑃ℎ𝑦)2 − 𝑟𝐿𝐷𝐿𝐷𝑒𝑡 − 𝑤𝐿𝐷𝑒𝑡

𝜕𝐿𝐷𝑒𝑡

𝜕𝑧

Nitrate balance 𝜕𝑁𝑂3

𝜕𝑡= −𝜇𝑚𝑎𝑥𝑓(𝐼) (

𝑁𝑂3

𝐾𝑁𝑂3+𝑁𝑂3∙

1

1+𝑁𝐻4

𝐾𝑁𝐻4

) 𝑃ℎ𝑦 + 𝑛𝑁𝐻4

Nitrification rate: 𝑛 = 𝑛𝑚𝑎𝑥 (1 − 𝑚𝑎𝑥 [0,𝐼−𝐼0

𝑘𝐼+𝐼−𝐼0])

Ammonium balance:

𝜕𝑁𝐻4

𝜕𝑡= −𝜇𝑚𝑎𝑥𝑓(𝐼) (

𝑁𝐻4

𝐾𝑁𝐻4 + 𝑁𝐻4) 𝑃ℎ𝑦 − 𝑛𝑁𝐻4 + 𝑙𝐵𝑀𝑍𝑜𝑜 + 𝑙𝐸

𝑃ℎ𝑦2

𝑘𝑝 + 𝑃ℎ𝑦2 𝛽𝑍𝑜𝑜 + 𝑟𝑆𝐷𝑆𝐷𝑒𝑡 + 𝑟𝐿𝐷𝐿𝐷𝑒𝑡

223



4.4.3 Irradiance model

Phytoplankton and seagrass growth are intrinsically dependent not only on light quantity

but also on light quality. The basic irradiance formulation included in Fennel et al. (2006)

did not account for spectral effects in considering light attenuation by water and

chlorophyll-a (see Table 4.4). To better approximate light behavior, the spectral

irradiance model used by Gallegos et al. (2011) was implemented. The atmospheric

evolution of the spectral irradiance was formulated following Gregg and Carder (1990),

which included absorption and scattering by ozone, oxygen, water vapor, and marine

aerosols and also reflectance at the air-sea interface. In the current experiment, the

observed PAR at the weather station was imposed at the water surface while enforcing

the spectral shape given by the Gregg and Carder (1990) formulation for that time.

The implemented spectral attenuation in the water column from Gallegos et al. (2011)

included the effects of water, CDOM, phytoplankton, and non-algal particulates (e.g.,

detritus, minerals, bacteria). In our simulations, the attenuation due to CDOM was

assumed to be minimal because CDOM concentrations were negligible in the area under

study (<3 Quinine Sulfate Units, QSU; <0.01 m-1 absorbance at 400 nm). The water

absorption and backscattering was assumed to follow the spectral characteristics of pure

water. The light absorption and scattering by phytoplankton was represented as

proportional to the chlorophyll-a concentration given by the Fennel et al. (2006)

implementation. Meanwhile, the non-algal component of the spectral attenuation was

taken as proportional to the total suspended solids concentration, which was considered

constant in the present study. The spectral shape of the attenuation by each component

followed the description in Gallegos et al. (2011). The PAR distribution across the entire

spectrum was integrated and used for the calculation of phytoplankton growth. The

spectral PAR was used to determine seagrass growth as part of the Zimmerman (2003)

224



bio-optical model. The main equations and parameters of Gallegos et al. (2011) model

can be seen at Table 4.5.

Table 4.5. Irradiance model main equations and parameters

SPECTRAL IRRADIANCE MODEL (GALLEGOS ET AL, 2011)

λ: Wavelenght z: Depth within the canopy E0(λ): Surface incident downwelling irradiance θ0: Solar incidence angle aw(λ): Absorption due to water aCDOM(λ): Absorption due to CDOM aNAP(λ): Absorption due to nonalgal particulates aChla(λ): Absorption due to chlorophyll-a bbw(λ): Backscattering due to water bbNAP(λ): Backscattering due to nonalgal particulates bbPhyto(λ): Backscattering due to phytoplankton

Spectral diffuse attenuation coefficient

𝐾𝑑(𝜆) = (1 + 0.005𝜃0)𝑎(𝜆) + 4.18{1 − 0.52𝑒𝑥𝑝[−10.8𝑎(𝜆)]}𝑏𝑏(𝜆)

Total light absorption 𝑎(𝜆) = 𝑎𝑤(𝜆) + 𝑎𝐶𝐷𝑂𝑀(𝜆) + 𝑎𝑁𝐴𝑃(𝜆) + 𝑎𝐶ℎ𝑙𝑎(𝜆)

Total light backscattering 𝑏𝑏(𝜆) = 𝑏𝑏𝑤(𝜆) + 𝑏𝑏𝑝ℎ𝑦𝑡𝑜(𝜆) + 𝑏𝑏𝑁𝐴𝑃(𝜆)

Downwelling spectral irradiance

𝐸𝑑(𝑧, 𝜆) = 𝐸0(𝜆)𝑒𝑥𝑝[−𝐾𝑑(𝜆)𝑧]

4.4.4 Bio-optical seagrass model

We linked a bio-optical model (Figure 4.5) to compute the carbon balance based on light

conditions and photosynthesis. The model developed by Zimmerman (2003) consists of

three different modules: a module that simulates the seagrass relative biomass and

architecture including leaf geometry, an irradiance module that calculates the light

225



absorption and scattering through the canopy, and a photosynthesis module that calculates

the carbon balance (primary production and respiration) of the submerged plant canopy.

The original 1D model simulates the light environment of a submerged canopy at a fixed

horizontal point. However, we have applied the model to the entire domain. We have

assumed initial seagrass presence in the entire system allowing the description of the light

environment by dividing the canopy volume, including the leaves and the water column,

into a series of horizontal sections of finite thickness. The optical properties of each

section are based on the architecture of the canopy, the orientation and optical properties

of the leaves, and the optical properties of the dissolved materials and suspended particles

in the water column. Given the spectral PAR at canopy height from the irradiance model,

it computes the seagrass Photosynthetically Usable Radiation (PUR) by computing the

spectral absorption and scattering of the downwelling and upwelling photosynthetically

active irradiance through the seagrass canopy. Finally, the model calculates the canopy

carbon balance, by computing the photosynthesis/respiration ratio. This ratio was used to

assess the seagrass presence/absence and survival under different scenarios and

conditions. The threshold of P/R=1 was chosen, as both autotrophic and heterotrophic

ecosystems tend to approach P/R=1 over time (Giddings and Eddlemon, 1978).

Moreover, as the ecosystem under study is autotrophic, we have assumed that P/R>1 is

associated with seagrass success and growth, while P/R<1 leads to seagrass

disappearance. The main equations and parameters of this model can be seen at Table 4.6.

226



Table 4.6. Bio-optical seagrass model main equations and parameters

BIO-OPTICAL SEAGRASS MODEL (ZIMMERMAN, 2003)

λ: Wavelength z: Depth within the canopy β: Bending angle of the seagrass canopy hm: Maximum canopy height h(z): Height above the seabed s: Shape factor for biomass distribution Rd (λ,z): Canopy reflectance of downwelling irradiance. �̅�𝒅(𝒛): Average cosine of downwelling irradiance. 𝑲𝒅(𝝀, 𝒛): Water column attenuation of downwelling irradiance. 𝒂𝑳(𝝀): Leaf absorption coefficient Rd (λ,z): Canopy reflectance of upwnwelling irradiance. 𝑲𝒖(𝝀, 𝒛): Water column attenuation of upwelling irradiance. Ap (λ): Photosynthetically absorptance DT: Daylenght ∅𝒑: Light-use efficiency

FR: Fraction of total plant biomass represented by root and rhizome. RR: Respiration rate of below-ground tissue. FL: Fraction of total plant biomass represented by leaves.

RL: Leaf respiration

Module I: vertical canopy architecture and leaf geometry

Horizontally projected leaf area at depth z 𝑙𝑝(𝑧) = 𝑙(𝑧)𝑠𝑖𝑛𝛽

Leaf area index at depth z 𝑙(𝑧) = 𝐿 × 𝐵(𝑧)

Canopy leaf area index 𝐿 = 𝑠ℎ𝑜𝑜𝑡 𝑑𝑒𝑛𝑠𝑖𝑡𝑦 × 𝐿𝑠

Leaf area per shoot 𝐿𝑠 = 0.0063ℎ𝑐 + 0.019ℎ𝑐2

Realized canopy height ℎ𝑐 = ℎ𝑚 cos 𝛽

Biomass fraction in layer z 𝐵(𝑧) = 𝜓 (1 + [

ℎ(𝑧)

𝐼]

𝑠

)⁄

Percent of canopy biomas at the seabed 𝜓 = 2.51ℎ𝑐−0.79

Intermediate height of biomas distribution 𝐼 = 0.588[1 − 𝑒𝑥𝑝(−1.12ℎ𝑐)]

Module II: two-flow irradiance distribution

Downwelling plane irradiance transmitted through layer z

𝐸𝑑(𝜆, 𝑧) = 𝐸𝑑(𝜆, 𝑧 − 1)[1 − 𝑅𝑑(𝜆, 𝑧)] × 𝑒𝑥𝑝 [−𝑎𝐿(𝜆)𝑡𝐿𝑙𝑝(𝑧)

�̅�𝑑(𝑧)− 𝐾𝑑(𝜆, 𝑧)∆𝑧]

Upwelling plane irradiance transmitted through layer z

𝐸𝑢(𝜆, 𝑧) = {[𝐸𝑑(𝜆, 𝑧)𝑅𝑏(𝜆, 𝑧 + 1)] + 𝐸𝑢(𝜆, 𝑧 + 1)} × [1 − 𝑅𝑢(𝜆, 𝑧)] × 𝑒𝑥𝑝 [−𝑎𝐿(𝜆)𝑡𝐿𝑙𝑝(𝑧)

�̅�𝑢− 𝐾𝑢(𝜆)∆𝑧]

Module III: Canopy photosynthesis

Photosynthetically used irradiance in layer z

𝑃𝑈𝑅(𝑧) = ∑ 𝐴𝑝(𝜆)𝑙𝑝(𝑧) [𝐸𝑑(𝜆, 𝑧 − 1)

�̅�𝑑(𝑧 − 1)+

𝐸𝑢(𝜆, 𝑧 + 1)

�̅�𝑢

]

𝜆

Daily integrated biomass-specific photosynthesis 𝑃 = ∑ 𝐵(𝑧){1 − 𝑒𝑥𝑝[−0.67∅𝑝𝑃𝑈𝑅(𝑧)]}𝐷𝑇

𝑧

Daily plant respiration (adapted from Zimmerman et al. 1995)

𝑅 = 24 × 𝑅𝐿 × 𝐹𝐿 + 𝐷𝑇 × 𝑅𝑅 × 𝐹𝑅 + (24 − 𝐷𝑇)× 0.65𝑅𝑅 × 𝐹𝑅

227



4.5 Model skill assessment

In the present section the assessments of the biogeochemical, irradiance and bio-optical

models are described. The hydrodynamics and freshwater fluxes were assessed in a

previous study (Ganju et al., 2012). During the calibration process, all model parameters

were adjusted to match measured values. The indicators used to calibrate the

biogeochemical, irradiance and bio-optical models were chlorophyll-a concentration,

light attenuation coefficient, and seagrass presence/absence, respectively.

4.5.1 Biogeochemical and irradiance model assessment

A sensitivity analysis was conducted to assess the influence of the main model parameters

on chlorophyll-a results. These parameters were fixed to their minimum and maximum

literature value (see Table 4.7) obtaining the behavior observed in Figure 4.6 The

parameters that have a major effect on the chlorophyll-a model results were μr, α, gmax,

mp, τ, and mz, which is in agreement with Fasham et al. (1990) and Fennel et al. (2011).

Figure 4.6. Comparison of hourly model and field data values of chlorophyll-a at the sampling stations.

The calibration of the biogeochemical and irradiance model was focused on the three sites

where field measurements were obtained, each one representative of Outer, Snug and

South, respectively. The values of the model parameters were chosen within the range

228



found in the literature (Andersen et al., 1987; Taylor, 1988; Wroblewski, 1989; Taylor et

al., 1991; Fasham, 1995; Geider et al., 1997; Leonard et al., 1999; Lima and Doney, 2004)

to maximize the agreement between model results and field data (Table 4.7). We achieved

the highest skill by adjusting values of some of the main influential parameters according

to the results of the sensitivity analysis (see Figure 4.6) setting the rest of the parameters

with the same value as Fennel et al. 2006. The final simulation values can be seen in Table

4.7, being all of them in agreement with the literature variation ranges.

Table 4.7. Main biogeochemical and irradiance model parameters and chosen value.

SYMBOL DEFINITION CALIBRATED VALUE

UNITS RANGE

µr Phytoplankton growth rate at

reference temperature

3 d-1 0.62a-3.0b

KNO3 Half-saturation concentration for

uptake of NO3

0.1 Mmol N m-3 0.007-1.5c

KNH4 Half-saturation concentration for

uptake of NH4

1.5 Mmol N m-3 0.007-1.5 c

α Initial slope of the P-I curve 0.13 Mol C gChl-1(Wm-2)-1d-1 0.007-0.13d

gmax Maximum grazing rate 0.6 (mmol N m-3)-1 d-1 0.5e-1.0f

Kp Half-saturation concentration of

phytoplankton ingestion

2 (mmol N m-3)2 0.56-3.5c

mp Phytoplankton mortality 0.05 d-1 0.05-0.2g

Aggregation parameter 0.005 (mmol N m-3)-1d-1 0.005-0.1c

Θmax Maximum chlorophyll-a to

phytoplankton ratio

0.068 mgChl mg C-1 0.005-0.072d

mz Zooplankton mortality 0.025 (mmol N m-3)-1 d-1 0.025-0.25c

RSD Remineralization rate of

suspended detritus

0.03 d-1 0.01-0.25h

RLD Remineralization rate of large

detritus

0.01 d-1 0.01-0.25h

Nmax Maximum nitrification rate 0.05 d-1 0.05-0.1c a(Taylor, 1988) b(Andersen et al., 1987) c(Lima and Doney, 2004) d(Geider et al., 1997) e(Wroblewski, 1989) f(Fasham, 1995) g(Taylor et al., 1991) h(Leonard et al., 1999).

Calibration results show that in general, hourly temporal series of modelled chlorophyll-

a follows the field data behaviour as can be seen in Figure 4.7, finding the biggest

deviations at South Cove.

229

Chapter IV Development and implementation of a coupled ecological modelling system for semi-enclosed eutrophic coastal systems.

Application to a groundwater-fed estuary with submerged aquatic vegetation.

Figure 4.7. Comparison of hourly model and field data values of chlorophyll-a at the sampling stations.

230



A comparison between mean modeled and field data at the sampling stations allows a

better understanding of the model results (see Table 4.8 and Figure 4.8). In Snug and

Outer harbors, with regards to the chlorophyll-a concentration and the light attenuation

coefficient we achieved a BIAS close to zero. The similarity between the standard

deviations suggests that the model properly describes the variability of the system in those

areas. By contrast, in South Cove the difference between modeled and observed

chlorophyll-a is larger.

Table 4.8. Mean values of chlorophyll-a and Kd, standard deviation (Std), and BIAS for Outer, Snug and

South Harbors. Field values of chlorophyll-a and Kd were obtained processing data from sensors deployed

during summer 2012. Model results were obtained for the same time-period.

CHLOROPHYLL-A KD

Site Mean ± Std

Model (µg/L)

Mean ± Std

Field (µg/L)

BIAS

(µg/L)

Mean ± Std

Model (1/m)

Mean ± Std

Field (1/m)

BIAS (1/m)

Outer 6.9 ± 3.7 6.5 ± 2.8 0.41 0.45 ± 0.07 0.45 ± 0.30 -0.001

Snug 28 ± 12 28 ± 9.9 0.33 0.79 ± 0.19 0.86 ± 0.16 -0.077

South 6.3 ± 3.9 10 ± 9.3 -3.9 - - -

Figure 4.8. Comparison of mean model and field data values of Chlorophyll-a and Kd at the sampling

stations.

231



Additionally, a spectral analysis was performed in order to assess the tidal and diurnal

influence in chlorophyll-a model results at the sampling stations (see Figure 4.9). The

spectral analysis shows that chlorophyll-a in the Outer and South basins is mainly

influenced by tidal transport from landward to award (12 h peak in Figure 4.9) while Snug

Harbor is strongly influenced by solar radiation (24 h peak in Figure 4.9). This is

consistent with the fact that Snug presents the highest phytoplankton concentration being

its variation strongly correlated with light availability, whereas Outer and South have a

stronger tidal influence due to the lower phytoplankton presence.

Figure 4.9. Spectral analysis of chlorophyll-a based on model results at the sampling stations

Finally, the mean vertical and time averaged chlorophyll-a concentration for the whole

estuary can be seen in Figure 4.10. This calculation was obtained by computing the

chlorophyll-a concentration three-dimensionally during the complete study period

(Figure 4.10 a) and then calculating the time-averaged and mean vertical concentration

(Figure 4.10b). The simulation results show higher eutrophication levels in the upper

layers (see Figure 4.10a), and also higher chlorophyll-a levels in Snug harbor than in

232



Outer Harbor and South Cove (see Figure 4.10 a and b). Moreover, these results are in

agreement with field data as can be seen in Figure 4.8. Therefore, the model is able to

accurately reproduce chlorophyll-a concentration behavior with high spatial resolution.

Figure 4.10. a) Vertical chlorophyll-a variation in three layers of the model; b) Time-averaged and mean

vertical chlorophyll-a concentration for the whole estuary.

4.5.2 Seagrass bio-optical model assessment

The calibrated parameters for the bio-optical model were the bending angle, the maximum

canopy height, and the shoot density. The bending angle selected was 45 º representing

the average angle over a tidal cycle, and the maximum canopy height was 1 meter

(Ackerman, 2002). The chosen density was 525 shoots/m2, which is the mean observed

plant density, as it varies between 250-800 shoots/m2 in Outer Harbor (McGlathery,

Marino, Hayn, and Howarth unpublished). The spectral PAR comes from the irradiance

233



model, and is propagated through the canopy to the seafloor. The seafloor absorbance and

reflectance properties were also considered with the composition of the bottom being a

mixture between mud (representing organic detritus) and sand. The reflected light was

also propagated upward through the canopy, so the primary production was calculated

with the total light absorbed. The potential habitat was evaluated as a function of the

Photosynthesis/Respiration (P/R) ratio distribution (Figure 4.11a) obtained using the

mean data of the entire summer. We have considered the P/R ratio obtained as

representative of the season in the year under study. We assumed that for P/R >1 there is

seagrass growth, and therefore presence, delimiting this threshold as the potential

seagrass in the estuary. However, for P/R≤1 we assumed conditions are unfavorable to

seagrass presence (Figure 4.11b) as growth would be limited, being respiration larger than

photosynthesis in those areas. Based on this criterion, we have obtained an agreement of

73.39 % between modeled and field presence/absence data, taking into account Outer and

Snug Harbor (Figure 4.11c), as in South Cove seagrass is thought to have disappeared

due to hydrodynamic reasons, and not due to the light conditions as can be seen in Figure

11b. Additionally, in Figure 4.11 we can see that seagrass is not present in the shallower

areas of the estuary. This is due to the wetting and drying effect simulated by the model

and the subtidal behaviour of zostera marina imposed in the model. The seagrass

distribution obtained by the selected P/R criterion (Figure 4.11d) was in agreement with

the critical depth distribution obtained applying the depth-limitation equation proposed

by Duarte et al. (2007).

234



Figure 4.11. a) Photosynthesis/Respiration ratio distribution; b) Photosynthesis/Respiration ratio

distribution applying the P/R>1 criterion and comparison with field data (black solid line); c) Detail of

Snug and Outer P/R distribution with P/R> 1, and comparison with field data (black solid line) ; d)

Seagrass distribution obtained with the depth-limited equation (Duarte et al., 2007). The white area

represents where seagrass presence is discouraged (P/R<1), the light green area the potential seagrass

habitat (P/R >1), and the black solid line delimits the seagrass presence area measured in the field survey.

235



4.6 Nitrate loading and sea-level rise scenarios

The coupled modeling system was applied to West Falmouth Harbor to assess the effects

of nitrate reduction and sea-level rise on chlorophyll-a concentration and potential

seagrass habitat during the summer. The simulations time period was of two months

corresponding with July and August. Different nitrate loading and sea-level rise scenarios

were conducted corresponding to anticipated future summer scenarios (Table 4.9). We

implemented a gradual decrease of the nitrate input load and an increase in sea-level rise

based on the IPCC predictions (IPCC, 2007) for the next one-hundred years.

Table 4.9. Nitrate reduction and sea level rise scenarios, being CS_0/ NR_0/ SLR_2012 the initial

scenario.

COMBINED SCENARIO (CS)

NITRATE REDUCTION

SCENARIOS (NR)

NITRATE INPUT LOAD REDUCTION

(%)

SEA-LEVEL RISE SCENARIOS

(SLR)

SEA LEVEL RISE (m)


NR_0 NR_10 NR_25 NR_50 NR_75 NR_94

0 10 25 50 75 94

SLR_2012 SLR_2015 SLR_2022 SLR_2037 SLR_2062 SLR_2112

0 0.02 0.04 0.09 0.18 0.35

First, we analyzed the effects of nitrate reduction (NR) and sea-level rise (SLR)

separately, and then we have configured combined scenarios (CS) to evaluate the

simultaneous effect of both parameters (Table 4.9; Figure 4.12). Not surprisingly, our

results support the idea that improvements in light conditions for seagrass, and

consequently higher P/R ratios, are achieved with decreases in nitrate loading (Figure

4.12). The results point to a potential recovery of seagrass in Snug Harbor area when

nitrate loading is reduced by 50% (Figure 4.12; NR 50). The P/R ratio improves

considerably in Snug Harbor with a 75% nitrate reduction (NR 75). On the contrary, sea-

level rise provokes a P/R ratio decline in areas where there is currently seagrass presence,

as can be seen in SLR 2112. However, when both effects (SLR and NR) were studied

together, a clear relationship between their combined behavior (CS) and the system

236



response to nitrate reduction (NR) was observed. Hence, although sea-level rise increases

water level and reduces light penetration, the light attenuation change is not as significant

as the nitrate loading effect. This is evident from comparing the temporal variation of P/R

due to SLR vs. nitrate loading (Figure 4.13). In Snug, the Figure shows a variation of P/R

from 0.97 to 1.14 due to nitrate reduction, whereas light attenuation due to sea-level rise

decreases P/R from 0.97 to 0.94. A similar effect, due to sea-level rise, can be observed

in Outer, with a P/R variation from 1.05 to 1.01 (Figure 4.13). However, the nitrate

reduction effect is lower in Outer, ranging from 1.05 to 1.10, due to the lower chlorophyll-

a levels at this point.

237



Figure 4.12. P/R spatial variation under nitrate reduction (NR), sea-level rise (SLR) and combined (CS)

scenarios. See Table 4.9 for an explanation of the scenarios nomenclature.

238



Figure 4.13. P/R variation due to nitrate reduction and sea-level rise

A similar behavior was obtained on chlorophyll-a concentration, as nutrient reduction is

the main factor that provokes a decrease in chlorophyll-a levels, whereas sea level rise

also affects to eutrophication but in a lower extent (see Figure 4.14). In fact, the mean

chlorophyll-a decrease due to nutrient reduction at the final NR scenario is 80 %, and due

to sea level rise at the final SLR scenario is 24 % (see Figure 4.14). This behavior can

also be seen in Figure 4.15, where the spatial chlorophyll-a distribution for the different

scenarios is presented. The figure shows a dominant influence of nutrient reduction

239



effects on the combined scenarios, and also an inverse relationship between chlorophyll-

a spatial distribution and seagrass presence, which can be seen comparing Figures 4.12

and 4.15, as chlorophyll-a is one of the main factors that limit light availability. In fact,

in general, seagrass presence is strongly related with eutrophication, as its presence is

inversely related to chlorophyll-a concentration due to its light limitation effect. South

cove was not represented in Figure 4.12 as the seagrass absence in that area is not related

to light availability but to a natural disturbance. As a consequence of this, the bio-optical

model is not able to represent its absence on this area. On the contrary, eutrophication

distribution can be correctly described by the model for the whole estuary as presented in

Figure 4.15. Additionally, measured chlorophyll-a data was available at South Cove,

whereas an instrument malfunction occurred at the light sensor for that area, so the

calibration of light attenuation at South could not have been properly done anyway.

240



Figure 4.14. Chlorophyll-a variation due to nitrate reduction and sea-level rise

241



Figure 4.15. Time-averaged and mean vertical chlorophyll-a spatial variation under nitrate reduction

(NR), sea-level rise (SLR) and combined (CS) scenarios. See Table 4.9 for an explanation of the

scenarios nomenclature.

242



Moreover, the combined effect of sea-level rise and nutrient reduction led to a significant

decrease of chlorophyll-a concentration and Kd, especially in Snug Harbor, with a

reduction from 27.77 µg L-1 to 2.79 µg L-1 and 0.79 m-1 to 0.36 m-1 respectively (Figure

4.16). Consequently, P/R in Snug Harbor increases from 0.97 to 1.09, providing adequate

conditions for seagrass growth from CS_2. P/R in Outer Harbor slightly increased until

CS_4, where it reached a maximum value of 1.07, and decreased to 1.05 in CS_5 due to

sea-level rise influence. We have also obtained the evolution of potential seagrass area

on the estuary for the different combined scenarios (Figure 4.16). We obtained an 8%

increase at CS_1, having an accumulated growth of 21 % and 34% at CS_2 and CS_3

respectively. In the case of CS_4 and CS_5 the influence of sea level rise makes the

evolution slower, obtaining an area increase from CS_4 to CS_5 of only 3 %, having

CS_5 an accumulated area growth of 45% with respect to the original scenario (CS_0).

Therefore, our results show that in this system, potential reductions in nitrate loading will

be more important than sea-level rise. However, in other systems with low nitrate loading,

sea-level rise may be more relevant.

243



Figure 4.16. Chlorophyll-a, Kd, P/R and seagrass area variation for the combined scenarios (CS). See

Table 4.9 for an explanation of the scenarios nomenclature.

4.7 Discussion

Phytoplankton bloom intensity can significantly depress the light climate in the bottom

of the water column, and therefore, light sensitive biogeochemical processes such as

photosynthesis and photo-oxidation. This has significant effects on seagrass distribution,

as light is one of the main factors for seagrass growth and primary production. Our results

support the idea that when insufficient light reaches the canopy seagrass presence

diminishes, which is in agreement with previous studies (e.g. Dennison, 1987; Orth and

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90


Kd

(m-1

)

Scenarios

Outer Snug

0.00

5.00

10.00

15.00

20.00

25.00

30.00


Ch

loro

ph

yll

-a (

µg

L-1

)

Scenarios

Outer Snug

0.90

0.93

0.95

0.98

1.00

1.03

1.05

1.08

1.10

1.13

CS_0 CS_1 CS_2 CS_3 CS_4 CS_5P

/RScenarios

Outer Snug

0.00

10.00

20.00

30.00

40.00

50.00

60.00


Sea

gra

ss A

rea

Var

iati

on

(%

)

Scenarios

Acumulated

Variation per scenario

244



Moore, 1988; Duarte, 1991; Duarte et al., 2007). Additionally, although there are

relatively few known examples of seagrass meadow recovery following nitrate reductions

(Burkholder et al., 2007) our results suggest that with progressive nitrate removal, Snug

Harbor may recover from strictly a light perspective, as chlorophyll-a concentration,

which is one of the main light attenuation factors, will decrease in the system. Other

factors such as macroalgae competition and morphodynamic changes should also be

taken into account. For example, in South Cove, effects like competition with

opportunistic macroalgae or seagrass death due to natural disturbances should be studied.

However, the potential recovery of seagrass in Snug Harbor could be achieved due to the

fact that the anthropogenic pressures that affect that area mainly consist of nitrate loading

while macroalgal coverage is minimal. Nevertheless, the future effects of nitrate

reduction in West Falmouth Harbor are hard to predict or evaluate with conventional

techniques, as the transit time of the enriched water in the aquifer to the estuary is as much

as 10 years (Kroeger et al., 2006). Moreover, we also obtained that nutrient concentration

and therefore eutrophication are the processes that control seagrass distribution in the

studied semienclosed microtidal shallow estuary, due to the light attenuation produced by

them, which is in agreement with Burkholder et al. (2007) and Costa (1988). However,

although seagrass distribution is strongly sensitive to eutrophication and light attenuation,

it is also affected by other factors such as hypoxia, epiphyte growth, grazing, and

hydrodynamic feedback, which are not included in this model as it has some limitations.

However, further work could include these formulations. One of the factors that could

also be considered in the modeling system is anoxia due to eutrophication, as the plant

oxygen content is strongly dependent on photosynthesis and respiration (Greve et al.,

2003), which have been computed in the model as a function of light and temperature.

In fact, low oxygen levels could cause anoxia in the meristem that could also limit

seagrass growth and primary production. The maintenance of oxic conditions in

meristematic and belowground tissues is important for support seagrass growth, nutrient

245



uptake by roots, and translocation of nutrients between roots and leaves (Zimmerman and

Alberte, 1996; Greve et al., 2003). We have also neglected the role of epiphytes.

Epiphytes attenuate light and expand around leaves limiting uptake of oxygen, inorganic

carbon, and nutrients (Hauxwell et al., 2001). It is also important to include the effect of

grazers: small invertebrate grazers generally have minimal negative impacts on seagrass

growth and biomass, and may have important positive functions by controlling epiphyte

growth; however, large grazers can impact seagrass meadows significantly (Stoner et al.,

1995). Moreover, although currents and wave action can play an important role in

seagrass distribution, the interactions between hydrodynamics and seagrasses were

considered negligible in the present study, as this estuary is microtidal with limited fetch.

Most of the recent modeling advances with respect to seagrasses are based on flow motion

(Maza et al., 2013), morphodynamic changes (Bouma et al., 2008), and particle trapping

(Hendriks et al., 2008). However, our modeling approach resolves the spatial pattern of

seagrass habitat quality from a light perspective. In contrast to most of the existing

ecological models, our coupled implementation computes spatially varying spectral light

attenuation as a function of different attenuating substances with high vertical and

horizontal discretization, allowing for delineation of the light climate for seagrass

meadows. Moreover, as the irradiance model has been integrated into ROMS, which is

an open-source flexible modeling system, our technique could be used in a wide range of

applications. Moreover, the simplicity of its formulation makes it a non-high data

demanding tool that is able to easily compute interpretable results. For example, the

influence of sediment re-suspension and of the horizontal sediment transport on light

availability could be assessed with this tool. Another possible application would be to

analyze light climate variations due to spatial changes in CDOM caused by rivers flows,

terrestrial runoff and/or microbial processes. This is possible due to the fact that ROMS

is coupled to the Community Sediment Transport Modeling System (CSTMS) so the

246



dynamical interaction between sediments and light availability can be modeled with this

implementation.

4.8 Conclusions

In the present study, we have developed a new ecological modelling system for eutrophic

SECS, which is able to describe chlorophyll-a concentration, light availability and the

potential recovery of seagrass communities under future nitrate loading and sea-level-rise

scenarios from a light perspective. We have assessed the model in a shallow temperate

estuary, and capture the spatial variability of chlorophyll-a, light attenuation, and seagrass

presence/absence. The coupled implementation computes spectral light attenuation as a

function of different attenuating substances with high vertical and horizontal resolution,

which allows the accurate determination of the light climate in the seagrass meadow. We

find that, in general, increased sea-level will reduce light availability and is expected to

negatively impact seagrasses, with a 11.4% reduction in presence/absence area with a

0.35 m increase in sea level. However, in the estuary studied here, reduction of nitrate

loading is a larger factor in improvement of light availability, as eutrophication due to

nutrient loading is one of the main problems of the system. In fact, our results show that

seagrass habitat is expanded by 42.3 % with a 94 % reduction in nitrate loading. This

study contributes to existing eutrophication modeling efforts by providing a new linked

implementation which is able to assess seagrass potential habitat in terms of light

availability with simple formulations. This linkage of simple ecological models is a

powerful non-high data demanding tool, which easily computes interpretable results with

a high level of accuracy. Future work should incorporate other ecological communities

(macroalgae, epiphytes, grazers) as well as the effects of oxygen stress and hydrodynamic

drag caused by vegetation.

247

Chapter V Conclusions and future research

Chapter V

5 Chapter V. Conclusions and future research

248


249


Chapter V. Conclusions and future research

5.1 Introduction

Semi-enclosed coastal systems (SECS) are complex and vulnerable areas with relevant

economical and environmental importance. The increasing impact of eutrophication and

its consequences is causing great concern, especially in systems with the most severe

trophic levels, eutrophic and hypertrophic. The complexity of these systems makes the

study of the involved processes a difficult task. Complex modelling tools are widely used

for this purpose, but the high number of parameters and variables make them difficult to

calibrate and interpret, and require important computational resources. Therefore, there

is a need of developing simplified mathematical tools that take into account the most

relevant processes for eutrophic and hypertrophic SECS. In addition, the technique of

coupling and linking models seems to be a good solution to integrate the main system

processes, simplifying the modelling approach.

In this study, the main hydrodynamic, ecological and irradiance models with coupled-

linking capabilities were reviewed. The complexity and formulations of these models

confirmed the need of developing novel simplified ecological modelling tools with

coupling and linking capabilities in order to assess eutrophic and hypertrophic SECS. As

a consequence, the general aim of this Thesis was to develop novel ecological modelling

tools for assessing and describing the behaviour of eutrophic and hypertrophic SECS.

In order to accomplish this general objective, two models were developed to predict the

behaviour of hypertrophic and eutrophic SECS respectively. The first one was applied to

the Albufera of Valencia, a hypertrophic system with regulated connection with the sea,

and the second to West Falmouth Harbor, a eutrophic system with an additional issue of

seagrass meadows disappearance. Field data was successfully used to assess the skills of

the developed tools. The results obtained permit the extraction of the conclusions

described at the following section, regarding the characteristics of the models, the

sensitivity analysis, the light formulations, calibration and results.

250


5.2 Conclusions

5.1.1 General conclusions

Simplified modelling tools with high spatial resolution can characterize

eutrophication in semi-enclosed coastal systems due to the high spatial variability

of phytoplankton distribution. Therefore, spatial resolution is one of the main

factors to consider when designing an ecosystem modeling tool for SECS.

Ecosystem models require one more level of complexity for describing eutrophic

SECS than for hypertrophic SECS, as more ecological processes should be taken

into account in the first case. In fact, hypertrophic SECS can be described by

models with a lower level of parametrization than eutrophic ones.

Non-high data demanding models can be powerful tools, which can easily

compute interpretable results with a high level of accuracy if the main equations

and consequent variables and parameters are well selected and defined.

Simplified tools can be very useful to accurately describe eutrophication in

hypertrophic SECS. The new developed simplified model was able to characterise

eutrophication in hypertrophic heavily regulated SECS, describing with high

temporal and spatial resolution, the chlorophyll-a concentration evolution during

a whole year.

The coupling and/or linkage of models is an efficient and flexible tool to describe

specific ecological processes in complex systems. In fact, this study contributes

to existing modeling efforts by providing a new implementation that not only

describes chlorophyll-a concentration, but also seagrass potential habitat in terms

of light availability in eutrophic SECS.

251


The hydrodynamic model is critical for the performance of the ecological models.

An adequate representation of the hydrodynamics is a fundamental factor for the

analysis of SECS, due to their singular connection with the adjacent systems.

Therefore, the hydrodynamic models and conditions have been carefully selected

at the present study.

5.1.2 Sensitivity analysis conclusions

Better knowledge of the most influential processes that affect eutrophication in

semi-enclosed systems was achieved. We can conclude that the parameters for

which chlorophyll-a is highly sensitive in the hypertrophic studied area are

phytoplankton respiration rate, grazing rate, phytoplankton growth rate, flux of

soluble reactive phosphorus from sediment to the water column and chlorophyll-

a phytoplankton carbon ratio. Whereas in the eutrophic studied area were

phytoplankton growth rate, initial slope of Photosynthesis-Light Intensity curve

(PI curve), grazing rate, phytoplankton death, aggregation parameter, and

zooplankton mortality rate. Therefore, we can conclude that in both cases the most

influential parameters are those related to phytoplankton and zooplankton growth,

and death, unless in the eutrophic estuary, where the initial slope of the PI curve

appears to have more relevance as light effects are still important in eutrophic

systems.

The sensitivity analysis has been also proved to be a valuable tool to reduce the

number of parameters to be adjusted in both models.

5.1.3 Light modelling conclusions

Light availability in eutrophic and hypertrophic SECS is mainly influenced by

chlorophyll-a concentration.

252


Light climate could be improved by reducing nutrient loading in eutrophic SECS

with cultural eutrophication, whereas restoration of hypertrophic systems require

a more complex analysis involving also hydrodynamic conditions and sediment

fluxes.

The implementation of spectral light attenuation in ecological models as a

function of different attenuating substances with high vertical and horizontal

resolution, allows an accurate determination of the light climate. In this study, we

have implemented Gallegos et al (2011) formulations into Fennel et al (2006)

model. This configuration allowed the linkage of a bio-optical seagrass model that

required spectral PAR as input data.

The present study reveals that a three dimensional model with spectral light

attenuation formulations is required in a eutrophic SECS to accurately describe

the heterogeneity and light climate of the system, whereas a two-dimensional

model with non-spectral attenuation was successfully used in hypertrophic SECS

due to the low light penetration through the water column in these systems.

5.1.4 Conclusions of calibration and results

The average uncertainty of the hypertrophic model prediction was less than 6%,

with two Pearson correlation coefficients of 0.933 and 0.917 for calibration and

validation respectively and a Nash-Sutcliffe efficiency coefficient of 0.96, which

are excellent values.

The eutrophic modeling system accurately reproduced the spatial variability in

chlorophyll-a and light attenuation with RMS errors of 3.72 µg L-1 and 0.07 m-1

respectively.

253


A potential seagrass habitat criteria based on Production/Respiration (P/R) ratio

was proposed. We assumed that for areas where P/R >1 there is potential seagrass

growth, and therefore presence, delimiting this threshold as the potential seagrass

habitat in the estuary. However, for P/R≤1 we assumed conditions are unfavorable

to seagrass presence as growth would be limited, being respiration larger than

photosynthesis in those areas. The results obtained with the P/R criterion selected

to assess seagrass potential habitat present an agreement of 73.39 % between

modeled and field data.

We contribute to a better understanding of the impacts of climate change on

eutrophic SECS. We find that, in general, increased sea-level will reduce light

availability and is expected to negatively impact seagrasses. In West Falmouth

Harbor the model showed an 11.4% reduction in seagrass presence area with a

0.35 m increase in sea level.

We conclude that nutrient reduction can led to system restoration in eutrophic

systems, whereas in hypertrophic ones the solution is more complex. In the

eutrophic SECS studied in this Thesis, the reduction in nitrate loading is a larger

factor in improvement of light availability. Chlorophyll-a concentration is reduced

a 89.3% whereas seagrass habitat is expanded by 42.3 %, with a 94 % reduction

in nitrate loading.

As demonstrated by the calculated mass balance, in hypertrophic SECS the input

loads can be higher than the output loads, so the limited connection with the sea

magnifies the eutrophication of the system. Furthermore, the SRP flux from the

sediment to the water column contributes to maintain high chlorophyll-a

concentrations in the studied area. Therefore, in hypertrophic SECS, nutrient

reduction could not have a significant impact on the system restoration if the

hydrodynamic and sediment conditions do not change.

254


A bio-optical model has been successfully linked to the eutrophication model,

demonstrating the flexibility of this approach to describe singular processes

occurring in SECS, such as the ratio between production and respiration.

5.2 Future research

This Thesis has revealed some limitations in the developed models that open new research

lines. These issues have been analyzed in detail in each chapter at their corresponding

discussion section. Here, the most relevant aspects of the Thesis needing future research

are mentioned.

The developed models should be applied to other locations with available field

data. This would be helpful in order to consolidate the effectiveness and utility of

the modeling tools.

Evaluate realistic management strategies at the Albufera of Valencia in order to

find solutions to its ecological problem, taking into account the economical factors

that have an impact on the study site, such as the rice cultivation, manufacturing

industry and tourism.

It would also be interesting to implement the hypertrophic simplified model into

ROMS in order to expand our technique and make it open source for the scientific

community as we are doing with the eutrophic modelling system.

Nitrogen influence on phytoplankton growth could also be included in the

hypertrophic simplified model, in order to allow its application to systems where

nitrogen is the limiting nutrient.

255


Future work on the eutrophic modelling system could incorporate other ecological

communities such as macroalgae and epiphytes. For example, effects like

competition of seagrass with opportunistic macroalgae, and epiphytes influence

on light attenuation should be studied.

One of the factors that could also be considered in the eutrophic modeling system

is anoxia due to eutrophication. The seagrass oxygen content is strongly

dependent on photosynthesis and respiration, which have been computed in the

model as a function of light and temperature. Oxygen concentration could be also

computed by the eutrophication model and be used by the seagrass model, as low

oxygen levels could limit seagrass growth and primary production. Therefore, an

oxygen term shall be included to the seagrass model formulation in order to

modify the calculation of the plant respiration and production based on the water

oxygen concentration.

Particle trapping and its effect on light climate could also be included in the

eutrophic modeling system. Seagrass canopies decrease flow velocity and reduce

turbulence. This promotes sedimentation within a seagrass meadow and reduces

resuspension of particles improving light climate. Moreover, as in the present

Thesis an spectral irradiance model has been integrated into ROMS, which is an

open-source flexible modeling system, the influence of sediment re-suspension

and of the horizontal sediment transport on light availability could be assessed.

This is possible due to the fact that ROMS is coupled to the Community Sediment

Transport Modeling System (CSTMS) so the dynamical interaction between

sediments and light availability can be modeled with this implementation.

Another possible application of the coupled modelling system for eutrophic SECS

would be to analyze light climate variations due to spatial changes in CDOM

256


caused by rivers flows, terrestrial runoff and/or microbial processes. This is again

possible due to the fact that ROMS is coupled to the CSTMS.

As we have linked Zimmerman’s model to Fennel’s model in which we integrated

the spectral light formulation, and Fennel’s is coupled into ROMS, the next step

would be to couple Zimmerman’s model in order to compute the seagrass

potential habitat at the same time than chlorophyll-a and spectral PAR.

257


5.3 Thesis impact and dissemination

5.3.1 Research articles

“A model for describing the eutrophication in a heavily regulated coastal lagoon.

Application to the Albufera of Valencia (Spain).” by Pilar Del Barrio, Andrés García

Gómez, Javier García Alba, César Álvarez Díaz, José Antonio Revilla Cortezón. Journal

of Enviromental Management. 2012.

DOI: 10.1016/j.jenvman.2012.08.019

“Modeling future scenarios of light attenuation and potential seagrass success in a

eutrophic estuary.” by Pilar Del Barrio, Neil K. Ganju, Alfredo L. Aretxabaleta,

Melanie Hayn, Andrés García, Robert W. Howarth. Estuarine, Coastal and Shelf Science.

2014.

DOI: 10.1016/j.ecss.2014.07.005

“Hydrodynamic modelling of a regulated Mediterranean coastal lagoon, the Albufera of

Valencia (Spain)” by Javier García Alba, Aina G. Gómez, Pilar del Barrio, Andrés

García Gómez, César Álvarez Díaz. Journal of Hydroinformatics. 2014

DOI: 10.2166/hydro.2014.071

“Progress and challenges in coupled hydrodynamic-ecological estuarine modeling” by

Neil K. Ganju, Mark J. Brush, Brenda Rashleigh, Alfredo L. Aretxabaleta, Pilar del

Barrio, Melinda Forsyth, Jason S. Grear, Lora A. Harris, Samuel J. Lake, Grant

McCardell, James O’Donnell, David K. Ralston, Richard P. Signell1, Jeremy M. Testa,

and Jamie M.P. Vaudrey. Estuaries and Coasts. In press.

258


5.3.2 Communications in conferences and workshops

This work has also been successfully presented with great acceptation to the scientific

community at the following events:

MABPOM2012 Symposium. Poster presentation: "A seagrass and light attenuation

model for a eutrophic estuary: Calibration, validation and predictions under nitrogen

loading scenarios". Authors: Pilar del Barrio, Neil K. Ganju, Alfredo L. Aretxabaleta.

Groton, Connecticut, USA. Nov. 2012.

Linking hydrodynamic and ecological models in estuaries: a workshop to discuss recent

advances and approaches. Presentation: “A modeling approach to assess light availability

and potential seagrass success under nitrate loading and sea level rise scenarios”.

Authors: Pilar del Barrio, Neil K. Ganju, Alfredo L. Aretxabaleta, Melanie Hayn,

Andrés García, Robert W. Howarth. Woods Hole, Massachusetts, USA. September 10-

11, 2013.

259

References

References

260

References

261

References

A

Abrams, P. A., and Roth, J. D. 1994. The effects of enrichment of three-species food

chains with nonlinear functional responses. Ecology, 75: 1118–1130.

Ackerman, J.D. 2002. Diffusivity in a marine macrophyte canopy: Implications for

submarine pollination and dispersal. American Journal of Botany, 89: 1119-1127.

Alexandre, A., Silva, J., Bouma, T.J., and Santos, R. 2008. Inorganic nitrogen uptake

kinetics and whole plant nitrogen budget in the seagrass Zostera noltii. Journal of

Experimental Marine Biology and Ecology, 401: 7-12.

Ambrose, R.B., 1988. WASP4, A Hydrodynamic and Water Quality Model. Model

Theory, User's Manual, and Programmer's Guide. U.S. Environmental Protection

Agency.

Ambrose, R.B., Wool, T.A., and Martin, J.L., 1993. The Water Quality Analysis

Simulation Program, WASP5, Part A: Model Documentation. U.S. EPA Center for

Exposure Assessment Modeling, Athens, GA.

Andersen, V., P. Nival, and R.P. Harris. 1987. Modelling of a planktonic ecosystem in an

enclosed water column. Journal - Marine Biological Association, 67: 407-430.

Arhonditsis, G.B., and Brett, M.T., 2005. Eutrophication model for Lake Washington

(USA): Part I. Model description and sensitivity analysis. Ecological Modelling, 187:

140-178.

Arrigo, K.R., and Sullivan, C.W. 1994. A high resolution bio-optical model of microalgal

growth: Tests using sea ice algal community time series data. Limnology and

Oceanography, 39: 609-631.

Art, H.W. 1993. A dictionary of ecology and environmental science (1st ed.): New York,

New York, Henry Holt and Company, 196 pp.

262

References

Aumont, O., and Bopp, L. 2006. Globalizing results from ocean in situ iron fertilization

studies, Glob. Biogeochem. Cy., 20 (2): 1944-9224.

Aumont, O. 2005. PISCES biogeochemical model description. Internal report, 36 pp.

B

Bach, H., Malmgren-Hansen, A., and Birklund, J. 1992. Modelling of eutrophication

effects on coastal ecosystems with eelgrass as the dominating macrophyte. In

Vollenweider, R.A., R. Marchetti, and R. Viviani (eds.), Marine Coastal Eutrophication.

Elsevier, Amsterdam, 993-999 pp.

Baird, M.E., Walker, S.J., Wallace, B.B., Webster, I.T., and Parslow, J.S., 2003. The use

of mechanistic descriptions of algal growth and zooplankton grazing in an estuarine

eutrophication model. Estuarine, Coastal and Shelf Science, 56: 685-695.

Bal, K.D., Bouma, T.J., Buis, K., Struyf, E., Jonas, S., Backx, H., and Meire, P. 2011.

Trade-off between drag reduction and light interception of macrophytes: comparing five

aquatic plants with contrasting morphology. Functional Ecology, 25: 1197-1205.

Baretta, J., and P. Ruardij. 1988. Tidal Flat Estuaries: Simulation and Analysis of the Ems

Estuary. Springer-Verlag, Berlin.

Baretta, J.W., Ebenhoh, W., and Ruardij, P. 1995. The European regional seas ecosystem

model, a complex marine ecosystem model. Netherlands Journal of Sea Research, 33 (3-

4): 233-246.

Baretta-Bekker, J.G., Baretta, J.W., and Ebenhoh, W. 1997. Microbial dynamics in the

marine ecosystem model ERSEM II with decoupled carbon assimilation and nutrient

uptake. Journal of Sea Research, 38: 195-211.

Barcena, J.F., Garcia, A., Garcia, J., Alvarez, C., and Revilla, J.A. 2012. Surface analysis

of free surface and velocity to changes in river flow and tidal amplitude on a shallow

263

References

mesotidal estuary: An application to Suances Estuary (Nothern Spain). Journal of

Hydrology 420-421, 301-318.

Bastón, S., 2008. Propagation of long waves in semi-enclosed bays, Department of

Science and Techniques of Water and Environment. University of Cantabria, Santander.

Ph.D. Thesis.

Bearlin, A.R., Burgman, M.A., Regan, H.M. 1999. A stochastic model for seagrass

(Zostera muelleri) in Port Phillip Bay, Victoria, Australia. Ecological Modeling, 118:

131-148.

Bent, E., Postma, L., Roelfzema, A., and Stive, R. 1991. Hydrodynamic and dispersion

modelling of Swansea Bay. Environmental Hydraulics, 1: 865-870.

Berger, C. J., and Wells, S. A. 2008. Modeling the Effects of Macrophytes on

Hydrodynamics, Journal of Environmental Engineering, 134 (9): 777-788.

Bidegain, G., Bárcena, J.F., García, A., and Juanes, J.A. 2012. Modelling dispersal and

potential recruitment patterns of Ruditapes phillippinarum and Ruditapes decussatus

based on particle tracking, habitat suitability and larval dynamics. 50th ECSA Conference.

Venice. Italy.

Bissett, W.P., K.L. Carder, J.J. Walsh, and D.A. Dieterle. 1999a. Carbon cycling in the

upper waters of the Sargasso Sea: II. Numerical simulation of apparent and inherent

optical properties. Deep-Sea Research Part I: Oceanographic Research Papers, 46: 271-

317.

Bissett, W.P., J.J. Walsh, D.A. Dieterle, and K.L. Carder. 1999b. Carbon cycling in the

upper waters of the Sargasso Sea: I. Numerical simulation of differential carbon and

nitrogen fluxes. Deep-Sea Research Part I: Oceanographic Research Papers, 46: 205-269.

264

References

Blaas, M., Dong, C., Marchesiello, P., McWilliams, J.C., and Stolzenbach, K.D. 2007.

Sediment-trasnport modeling on Southern Californian shelves: A ROMS case study.

Continental Shelf Research, 27: 832-853.

Blanco, S. and Romo, S. 2006. Ictiofauna del lago de la Albufera de Valencia: evolución

histórica y situación actual. Boletín de la Real Sociedad Española de Historia Natural.

Sección Biología, 101 (1-4), 45-56.

Blauw, A.N., Los, H.F.J., Bokhorst, M. and Erftemeijer, P.L.A. 2009. GEM: a generic

ecological model for estuaries and coastal waters. Hydrobiologia, 618 (1): 175–198.

Blumberg, A. F. and Mellor, G.L. 1987. A description of a three-dimensional coastal

ocean circulation model. In Three-Dimensional Coastal Ocean Models. Coastal and

Estuarine Series. Ed. Moores, 4: 1-16 pp.

Bouma, T.J., De Vries, M.B., Low, E., Peralta, G., Tánczos, I.C., van de Koppel, J.,and

Herman, P.M.J. 2005. Trade-offs related to ecosystem-engineering: a case study on

stiffness of emerging macrophytes. Ecology, 86:2187–2199

Bouma, T.J., Friedrichs, M., Van Wesenbeeck, B.K., Brun, F.G., Temmerman, S., De

Vries, M.B., Graf, G., and Herman, P.M.J. 2008. Plant growth strategies directly affect

biogeomorphology of estuaries. In River, Coastal and Estuarine Morphodynamics:

RCEM 2007. 285-292.

Bowen, J. D., and Hieronymus, J. W. 2003. A CE-QUAL-W2 Model of Neuse Estuary

for Total Maximum Daily Load Development. ASCE Journal of Water Resources

Planning and Management, 129 (4): 283-294.

Braunschweig, F. 2001. Generalização de um modelo de circulação costeira para

albufeiras, MSc. Thesis, Instituto Superior Técnico, Technical University of Lisbon.

265

References

Briere, C., Abadie, S., Bretel, P., and Lang, P. 2007. Assessment of TELEMAC system

performances, a hydrodynamic case study of Anglet, France. Coastal Engineering, 54:

345-356.

Bronw, L.C., and Barnwell, T.O. 1987. The Enhanced Stream Water Quality Models:

QUAL2E and QUAL2E-UNCAS: Documentation and User Model. EPA/600/3-87/007.

Budgell, W.P. 2005. Numerical simulation of ice-ocean variability in the Barents Sea

region, Ocean Dynamics, DOI 10.1007/s10236-005-0008-3.

Bunch, B. W., Cerco, C. F., Dortch, M. S., Johnson, B. H., and Kim, K. W. 2000.

Hydrodynamic and Water Quality Model Study of San Juan Bay Estuary, ERDC TR-00-

1, U.S. Army Engineer Research and Development Center, Vicksburg, MS.

Burd, A.B., and Dunton, K. H. 2001. Field verification of a light-driven model of biomass

changes in the seagrass Halodule wrightii. Marine Ecology Progress Series 209, 85-98.

Burger, D.F., Hamilton, D.P., and Pilditch, C.A. 2008. Modelling the relative importance

of internal and external nutrient loads on water column nutrient concentrations and

phytoplankton biomass in a shallow polymictic lake. Ecological Modelling 211, 411-423.

Burkholder, J.M., Tomasko, D.A., and Touchette, B.W. 2007. Seagrasses and

eutrophication. Journal of Experimental Marine Biology and Ecology 350: 46-72.

Buzzelli, C.P., Wetzel, R.L., and Meyers, M.B. 1999. A linked physical and biological

framework to assess biogeochemical dynamics in a shallow estuarine ecosystem.

Estuarine, Coastal and Shelf Science, 49: 829-850.

266

References

C

Cabello-Pasini, A., R. Muñiz-Salazar, and D.H. Ward. 2003. Annual variations of

biomass and photosynthesis in Zostera marina at its southern end of distribution in the

North Pacific. Aquatic Botany, 76: 31-47.

Calero, L.A., Tuchkovenko, Y.S., and Lonin, S.A. 2003. Modelo numérico de calidad de

aguas para la Ciénaga Grande de Santa Marta. Ingeniería del Agua, 10 (4): 479-492.

Campuzano F., Nunes S., Malhadas M., Neves R. 2010. Modelling hydrodynamics and

water quality of Madeira Island (Portugal). GLOBEC International Newsletter, 16 (1):

40-42.

Cancino, L., and Neves, R. 1999. Hydrodynamic and sediment suspension modelling in

estuarine systems. Part I: Description of the numerical models. Journal of Marine

Systems, 22: 105-116.

Canet, R., Chaves, C., Pomares, F., and Albiach, R. 2003. Agricultural use of sediments

from the Albufera Lake (eastern Spain). Agr Ecosyst Environ, 95: 29-36.

Castanedo, S. 2000. Desarrollo de un modelo hidrodinámico tridimensional para el

estudio de la propagación de ondas largas en estuarios y zonas someras. Ph.D.

Dissertation. University of Cantabria. Spain.

Cerco, C.F., and Cole, T. 1993. Three-dimensional eutrophication model of Chesapeake

Bay. Journal of Environmental Engineering, 119: 1006-1025.

Cerco, C. F., Bunch, B. W., Teeter, A. M., and Dortch, M. S. 2000. Water Quality Model

of Florida Bay, Report no. ERDC/EL TR-00-10, US Army Engineer Research and

Development Center, Vicksburg, MS.

Cerco, C. F. and Noel, M.R. 2005. Evaluating ecosystem effects of oyster restoration in

Chesapeake Bay. A Report to the Maryland Department of Natural Resources.

267

References

Cerco, C.F., and Noel, M.R. 2007. Can Oyster Restoration Reverse Cultural

Eutrophication in Chesapeake Bay? Estuaries and Coasts, 30 (2): 331-343.

Cerco, C.F. and M.R. Noel, 2013. Twenty-One-Year Simulation of Chesapeake Bay

Water Quality. Using the CE-QUAL-ICM Eutrophication Model. Journal of the

American Water Resources Association (JAWRA) 49(5): 1119-1133.

Chan, T.U., Hamilton, D.P., Robson, B.J., Hodges, B.R., and Dallimore, C. 2002. Impacts

of hydrological changes on phytoplankton succession in the Swan River, Western

Australia. Estuaries, 25 (6B): 1406-1415.

Chanudet, V., Fabre, V., and Van Der Kaaij, T., 2012. Application of a three-dimensional

hydrodynamic model to the Nam Theun 2 Reservoir (Lao PDR). Journal of Great Lakes

Research, 38 (2): 260-269.

Chao, X., Jia, Y., M. Cooper, C., Shields, J., F. Douglas, and Wang, S., S. Y., 2006.

Development and Application of a Phosphorus Model for a Shallow Oxbow Lake. Journal

of Environmental Engineering, 132: 1498-1507.

Chao, X., Jia, Y., Shields Jr., F.D., Wang, S.S.Y., and Cooper, C.M., 2010. Three-

dimensional numerical simulation of water quality and sediment-associated processes

with application to a Mississippi Delta lake. Journal of Environmental Management,

91:1456-1466.

Chapra, S.T., 1997. Surface water-quality modeling. McGraw-Hill International Editions,

Singapore.

Chau, K.W., and Haisheng, J., 1998. Eutrophication Model for a Coastal Bay in Hong

Kong. Journal of Environmental Engineering, 124: 628-638.

Chen, C., Beardsley, R.C., and Cowles, G. 2006. An Unstructured Grid, Finite-Volume

Coastal Ocean Model (FVCOM) System. Oceanography, 19 (1): 78-89.

268

References

Christensen, V., and Pauly, D. 1991. A guide to ECOPATH II Software System (version

2.0). ICLARM Software, 6: 72 pp.

Chung S.W., and Oh, J.K. 2006. Calibration of CE-QUAL-W2 for a monomictic reservoir

in a monsoon climate area. Water Sci Technol, 54: 29–37

Coelho, H., Santos, A., Rosa, T. L. and Neves, R. 1994. Modelling the wind driven flow

off Iberian Peninsula, GAIA, 8: 71-78.

Cole, B.E., and Cloern, J.E. 1987. An empirical model for estimating phytoplankton

productivity in estuaries. Marine Ecology Progress Series, 36: 299-305.

Cole, T.M., and Wells, S.A. 2008. CE-QUAL-W2- A two-dimensional, laterally

averaged, hydrodynamic and water-quality model, version 3.6: U.S. Army Corps of

Engineers, Instruction Report EL-08-1 [variously paged].

Connolly, J.P., and Winfield, R. 1984. A User’s Guide for WASTOX, a Framework for

Modeling the Fate of Toxic Chemicals in Aquatic Environments. Part 1: Exposure

Concentration. EPA-600/3-84-077, U.S. Environmental Protection Agency, Gulf Breeze,

FL.

Costa, J.E. 1988. Eelgrass in Buzzards Bay: distribution, production and historical

changes in abundance. 223 pp. Boston, Massachusetts: United States Environmental

Protection Agency.

Costello, C.T., and Kenworthy, W.J. 2011. Twelve-Year Mapping and Change Analysis

of Eelgrass (Zostera marina) Areal Abundance in Massachusetts (USA) Identifies

Statewide Declines. Estuaries and Coasts, 34: 232-242.

Courant, R., and Hilbert, D. 1962. Methods of Mathematical Physics. Interscience

Publishers, New York.

Cunningham, I. 2002. Modelling Nitrogen Fluxes in Seagrass, The University of Western

Australia. 109 pp.

269

References

D

Debele, B., Srinivasan, R., and Yves Parlange, J. 2006. Coupling upland watershed and

downstream waterbody hydrodynamic and water quality models (SWAT and CE‐QUAL‐

W2) for better water resources management in complex river basins. Environ. Model. and

Assess, 13 (1): 135‐153.

Deliman, P., Gerald, N., and Jeffrey, A. 2002. Application of the two-dimensional

hydrothermal and water quality model, CE-QUAL-W2, to the Chesapeake Bay -

Conowingo Reservoir. Lake and Reservoir Management, 18 (1): 10-19.

Dennison, W.C. 1987. Effects of light on seagrass photosynthesis, growth and depth

distribution. Aquatic Botany, 27: 15-26.

Dennison, W.C., R.J. Orth, K.A. Moore, J.C. Stevenson, V. Carter, S. Kollar, P.W.

Bergstrom, and R.A. Batiuk. 1993. Assessing Water Quality with Submersed Aquatic

Vegetation. BioScience, 43: 86-94.

Denman, K.L., and Gargett, A.E., 1995. Biological-physical interactions in the upper

ocean: The role of vertical and small scale transport processes. Annual Review of Fluid

Mechanics, 27: 225-255.

Denman, K.L., and Peña, M.A. 1999. A coupled 1-D biological/physical model of the

northeast subarctic Pacific Ocean with iron limitation. Deep-Sea Res Part II, 46: 2877–

2908.

Di Lorenzo, E. 2003. Seasonal dynamics of the surface circulation in the southern

California Current System, Deep-Sea Res., Part II, 50: 2371-2388.

Dinniman, M. S., J. M. Klinck, and W. O. Smith Jr. 2003. Cross shelf exchange in a model

of the Ross Sea circulation and biogeochemistry, Deep-Sea Res., Part II, 50: 3103-3120.

270

References

Di Toro, D.M., O’Connor, D.J., and Thomann R.V. 1971. A dynamic model of the

phytoplankton population in the Sacramento-San Joaquin Delta. In Hem, J.D. (ed.),

Nonequilibrium Systems in Natural Water Chemistry, Advances in Chemistry Series 106.

Washington: American Chemical Society, 131-180 pp.

Di Toro, D.M., and Matystik, W.F. 1980. Mathematical models of water quality in large

lakes. Part 1: Lake Huron and Saginaw Bay.Ecol.Res.Ser., in: Agency, U.S.E.P. (Ed.),

EPA. EPA-600/3-80-056, 28-30 pp.

Di Toro, D.M., Fitzpatrick, J.J., and Thomann, R.V. 1983. Water Quality Analysis

Simulation Program (WASP) and Model Verification Program (MVP) – Documentation.

Contract No. 68-01-3872, Hydroscience Inc., Westwood, NY, for U.S. EPA, Duluth, MN.

Doney, S.C., Glover, D.M., and Najjar, R.G. 1996. A new coupled, one-dimensional

biological-physical model for the upper ocean: Application to the JGOFS Bermuda

Atlantic Time-series Study (BATS) site. Deep-Sea Research II 43 (2-3), 591-624.

Downing, A., van Nes, E., Janse, J., Witte, F., Cornelissen, I., Scheffer, M., and Mooij,

W. 2012. Collapse and reorganization of a food web of Mwanza Gulf, Lake Victoria.

Ecological Applications, 22: 229-239.

Duan, L.J., Li, S.Y., Liu, Y., Moreau, V., and Christensen. 2009. Modeling changes in

the coastal ecosystem of the Pearl River Estuary from 1981 to 1998. Ecological

Modelling, 220: 2801-2818.

Duarte, C.M. 1991. Seagrass depth limits. Aquatic Botany, 40: 363-377.

Duarte, C.M. 2002. The future of seagrass meadows. Environmental Conservation, 29:

192-206.

Duarte, C.M., Amthor, J., DeAngelis, D.L., Joyce, L.A., Maranger, R.J., Pace, M.L.,

Pastor, J., and Running, S. 2003. The limits to models in ecology. In Models in Ecosystem

271

References

Science. (Eds.) C.D. Canham, J.J. Cole, and W.K. Lauenroth. 437-451 pp. Princeton

University Press. Princeton. New Jersey.

Duarte, C.M.; Marba, N., Krause-Jensen, D., and Sánchez-Camacho M. 2007. Testing the

Predictive Power of Seagrass Depth Limit Models. Estuaries and Coasts, 30: 652-656

Duarte, C.M., Sintes, T. and Marba, N. 2013. Assessing the CO2 capture potential of

seagrass restoration projects. Journal of Applied Ecology, 50: 1341–1349.

Dugdale, R. C., and Goering, J. J. 1967. Uptake of new and regenerated forms of nitrogen

in primary productivity. Limnology and Oceanography, 12: 196-206.

E

Ebenhoh, W., Baretta-Bekker, J.G., and Baretta, J.W. 1997. The primary producer

module in the marine ecosystem model ERSEM II, with emphasis on the light forcing.

J. Sea Res., 38: 173-193.

Edwards, A. M., and Brindley, J. 1996. Oscillatory behavior in a three-component

plankton population model. Dyn. Stabil. Syst., 11: 347–370.

Edwards, A. M., and Brindley, J. 1999. Zooplankton mortality and the dynamical

behaviour of plankton population models. Bull. Math. Biol., 61: 303–339.

Edwards, C. A., Batchelder, H. P., and Powell, T. M. 2000a: Modeling microzooplankton

and macrozooplankton dynamics within a coastal upwelling system. J. Plankton Res., 22:

1619–1648.

Edwards, C. A., Powell, T. A. and Batchelder, H. P. 2000b. The stability of an NPZ model

subject to realistic levels of vertical mixing. J. Mar. Res., 58: 37–60.

Eisenreich, S.J. 2005. Climate Change and the Water Dimension. JRS, Ispra, Italy. EU

Report Nº 21553. pp 253.

272

References

Eppley, R.W. 1972. Temperature and phytoplankton growth in the sea. Fish. Bull, 70:

1063-1085.

Eppley, R.W., and Peterson, B.J. 1979. Particulate organic matter flux and planktonic

new production in the deep ocean. Nature, 282 (5740): 677–680.

Escobar S., C.A. 2011. Relevance of coastal processes on the hydrodynamics of the Gulf

of Urabá (Colombian Caribbean Sea). Boletin de Investigaciones Marinas y Costeras-

INVEMAR, 40 (2).

European Commission, The N2K GROUP, 2009. Estuaries and Coastal Zones Within

The Context of The Birds And Habitats Directives. Technical Supporting Document 124

pp.

Evans, G. T., Steele, J. H., and Kullenberg, G. E. B. 1977. A preliminary model of shear

diffusion and plankton populations. Scottish Fisheries Res. Proj. Report nº9.

Evans, G. T. 1978. Biological effects of vertical-horizontal interactions. In Spatial

Patterns in Plankton Communities, ed. by J. H. Steele, Plenum Press, New York. 157–

179 pp.

Evans, G. T., and Parslow, J. S. 1985. A model of annual plankton cycles. Biol.

Oceanogr., 3: 327- 347.

Everett, J.D., Baird, M.E., and Suthers, I.M. 2007. Nutrient and plankton dynamics in an

intermittently closed/open lagoon, Smiths Lake, south-eastern Australia: An ecological

model. Estuarine, Coastal and Shelf Science, 72: 690-702.

F

Falkowoski, P.G., and Raven, J.A. 1997. Aquatic Photosynthesis. Blackwell Science. 375

pp. ISBN 0 86542 387 3.

http://www.nature.com/nature/journal/v282/n5740/abs/282677a0.html

http://www.nature.com/nature/journal/v282/n5740/abs/282677a0.html

http://en.wikipedia.org/wiki/Nature_(journal)

273

References

Fath, B.D., Jorgersen, S.E., and Scharler, U.M. 2011. Ecological Modeling in

Environmental Management: History and Applications. Treatise On Estuarine and

Coastal Science, 9: 23-33.

Fan, C., Ko, C.-H., and Wang, W.-S., 2009. An innovative modeling approach using

Qual2K and HEC-RAS integration to assess the impact of tidal effect on River Water

quality simulation. Journal of Environmental Management, 90: 1824-1832.

Fasham, M.J.R., H.W. Ducklow, and S.M. McKelvie. 1990. A nitrogen-based model of

plankton dynamics in the oceanic mixed layer. Journal of Marine Research, 48: 591-639.

Fasham, M.J.R. 1995. Variations in the seasonal cycle of biological production in

subarctic oceans: A model sensitivity analysis. Deep-Sea Research Part I: Oceanographic

Research Papers, 42: 1111-1149.

Fennel, K., Losch, M., Schröter, J., and Wenzel, M. 2001. Testing a marine ecosystem

model: sensitivity analysis and parameter optimization. J Marine Syst, 28: 45–63.

Fennel, K., Wilkin, J., Levin, J., Moisan, J., O'Reilly, J., and Haidvogel, D. 2006.

Nitrogen cycling in the Middle Atlantic Bight: Results from a three-dimensional model

and implications for the North Atlantic nitrogen budget. Global Biogeochemical Cycles,

20.

Fennel, K., Wilkin, J., Previdi, M., and Najjar, R. 2008. Denitrification effects on air-sea

CO2 flux in the coastal ocean: Simulations for the Northwest North Atlantic. Geophysical

Research Letters, 35, L24608.

Fennel, K., Hetland, R., Feng, Y., and DiMarco, S. 2011. A coupled physical-biological

model of the Northern Gulf of Mexico shelf: model description, validation and analysis

of phytoplankton variability. Biogeosciences, 8: 1881-1899.

Ferrarin, C., and Umgiesser, G., 2005. Hydrodynamic modeling of a coastal lagoon: The

Cabras lagoon in Sardinia, Italy. Ecological Modelling, 188: 340-357.

274

References

Fisher, T.R., Gustafson, A.B., Sellner, K., Lacouture, R., Haas, L.W., Wetzel, R.L.,

Magnien, R., Everitt, D., Michaels, B., and Karrh, R. 1999. Spatial and temporal

variation of resource limitation in Chesapeake Bay. Marine Biology, 133: 763-778.

Fisher, T. R., Hagy III, J. D., Boynton, W. R., and Williams, M. R. 2006. Cultural

eutrophication in the Choptank and Patuxent estuaries of Chesapeake Bay. Limnol.

Oceanogr., 51: 435-447.

Fletcher, C.A.J., 1991. Computational Techniques for Fluid Dynamics., Second Edition

ed, New York, Springer, 1, 401 pp.

Forsyth, M. 2014. Investigations of the effects of oyster allometry and reef morphology

on filtration rate and particle capture using numerical models. Master of Science Thesis.

University of Maryland. 134 pp.

Franks, P.J.S., Wroblewski, J.S., and Flierl, G.R. 1986. Behavoiur of a simple plankton

model with food-level acclimation by herbivores. Marine Biology, 91:121-129.

Franks, P. J. S. 1997. New models for the exploration of biological processes at fronts.

I.C.E.S. J. Mar. Sci., 54: 161– 167.

Franks, P.J.S., and Walstad, L.J. 1997. Phytoplankton patches at fronts: a model of

formation and response to transient wind events. Journal of Marine Research, 55: 1-30.

Franks, P.J.S., and Chen, C. 2001. A 3-D prognostic numerical model study of the

Georges Bank ecosystem. Part II: biological-physical model. Deep-Sea Research II, 48:

457-482.

Franks, P.J.S. 2002. NPZ models of plankton dynamics: Their construction, coupling to

physics, and application. Journal of Oceanography, 58: 379-387.

Friedrichs, M.A.M., Hood, R.R., and Wiggert, J.D., 2006. Ecosystem model complexity

versus physical forcing: Quantification of their relative impact with assimilated Arabian

Sea data. Deep Sea Research Part II: Topical Studies in Oceanography, 53: 576-600.

275

References

Fujii, M., Boss, E., and Chai, F. 2007. The value of adding optics to ecosystem models:

a case study. Biogeosciences Discussions, European Geosciences Union (EGU), 4 (3):

1585-1631.

Fulford, R.S., Breitburg, D.L., Newell, R.I.E., Kemp, W. M., and Luckenbach, M. 2007.

Effects of oyster population restoration strategies on phytoplankton biomass in

Chesapeake Bay: A flexible modeling approach. Marine Ecology Progress Series, 336:

43-61.

Fulton, E.A., 2001. The effects of model structure and complexity on the behaviour and

performance of marine ecosystem models, School of Zoology. University of Tasmania,

Hobart, Australia, 427 pp.

Fulton, E.A., Smith, A.D.M., and Johnson, C.R. 2003. Effect of complexity on marine

ecosystem models. Marine Ecology Progress Series, 253: 1-16.

Fulton, E.A., and Smith, A.D.M. 2004. Lessons learnt from the comparison of three

ecosystem models for POrt Phillip Bay, Ausatralia. African Journal of Marine Science,

26: 219-243.

Fulton, E.A., Smith, A.D.M., and Johnson, C.R. 2004a. Effects of spatial resolution on

the performance and interpretation of marine ecosystem models. Ecological Modelling,

176 (1-2): 27-42.

Fulton, E.A., Smith, A.D.M., and Johnson, C.R., 2004b. Biogeochemical marine

ecosystem models I: IGBEM - A model of marine bay ecosystems. Ecological Modelling,

174: 267-307.

G

Gallegos, C.L., Correll, D.L., and Pierce, J.W. 1990. Modeling spectral diffuse

attenuation, absorption, and scattering coefficients in a turbid estuary. Limnology and

Oceanography, 35: 1486-1502.

276

References

Gallegos, C.L., and Kenworthy, W.J. 1996. Seagrass depth limits in the Indian River

Lagoon (Florida, USA): application of an optical water quality model. Estuarine, Coastal

and Shelf Science, 42: 267-288.

Gallegos, C.L., Kenworthy, W.J., Biber, P.D. and Wolfe, B.S. 2009. Underwater spectral

energy distribution and seagrass depth limits along an optical water quality gradient.

Smithsonian contributions to the marine sciences, 38: 359-367.

Gallegos, C.L., Werdell, P.J., and McClain, C.R. 2011. Long-term changes in light

scattering in Chesapeake Bay inferred from Secchi depth, light attenuation, and remote

sensing measurements. Journal of Geophysical Research, 116.

Ganju, N., Dickhudt, P., Thomas, J., Borden, J., Sherwood, C., Montgomery, E.,

Twomey, E., and Martini, M. 2011. Summary of oceanographic and water-quality

measurements in West Falmouth Harbor and Buzzards Bay, Massachusetts, 2009-2010:

U.S. Geological Survey Open-File Report 2011-1113.

Ganju, N.K., Hayn, M., Chen, S.N., Howarth, R.W., Dickhudt, P.J., Aretxabaleta, A.L.

and Marino, R. 2012. Tidal and Groundwater Fluxes to a Shallow, Microtidal Estuary:

Constraining Inputs Through Field Observations and Hydrodynamic Modeling. Estuaries

and Coasts, 35: 1285-1298.

García Alba, J., G. Gómez, A., del Barrio, P., García, A. and Álvarez C. 2014.

Hydrodynamic modelling of a regulated Mediterranean coastal lagoon, the Albufera of

Valencia (Spain). Journal of Hydroinformatics, 16.5: 1062-1076.

García, A., Juanes, J.A., Alvarez, C., Revilla, J.A., and Medina, R. 2010a. Assesment of

the response of a shallow macrotidal estuary to changes in hydrological and wastewater

inputs through numerical modelling. Ecological Modelling, 221: 1194-1208.

García, A., Sámano, M.L., Juanes, J.A., Medina, R., Revilla, J.A., and Álvarez, C. 2010b.

Assessment of the effects of a port expansion on algae appearence in a coastal bay through

277

References

mathematical modelling. Application to San Lorenzo Bay (North Spain). Ecological

Modelling, 221: 1413-1426.

Geider, R.J., MacIntyre, H.L. and Kana, T.M. 1997. Dynamic model of phytoplankton

growth and acclimation: Responses of the balanced growth rate and the chlorophyll-a:

carbon ratio to light, nutrient-limitation and temperature. Marine Ecology Progress

Series, 148: 187-200.

Giddings, J., and Eddlemon, G.K. 1978. Photosynthesis/respiration ratios in aquatic

microcosms under arsenic stress. Water, Air and Soil Pollution 9: Issue 2, 207-212

Giusti, E., Marsili-Libellia, S., Renzib, M., and Focardi, S. 2010. Assessment of spatial

distribution of submerged vegetation in the Orbetello lagoon by means of a mathematical

model. Ecological modelling 221, 1484–1493.

Greene, R. M., Lehrter, J.C., and Hagy, J.D. 2009. Multiple regression models for

hindcasting and forecasting midsummer hypoxia in the Gulf of Mexico. Ecological

Applications, 19: 1161-1175.

Gregg, W.W., and Carder, K.L. 1990. A simple spectral solar irradiance model for

cloudless maritime atmospheres. Limnology & Oceanography, 35: 1657-1675.

Greve, T.M., Borum, J., and Pedersen, O. 2003. Meristematic oxygen variability in

eelgrass (Zostera marina). Limnology & Oceanography, 48: 210-216.

Gruber, N., Frenzel, H., Doney, S. C., Marchesiello, P., McWilliams, J. C., Moisan, J. R.,

Oram, J. J., Plattner, G.-K., and Stolzenbach, K. D. 2006. Eddy-resolving simulation of

plankton ecosystem dynamics in the California Current System, Deep Sea Res. I,

53:1483–1516.

Guo, C., Ye, S., Lek, S., Liu, J., Zhang, T., Yuan, J., and Li, Z. 2013. The need for

improved fishery management in a shallow macrophytic lake in the Yangtze River basin:

278

References

Evidence from the food web structure and ecosystem analysis. Ecological Modelling,

267: 138-147.

H

Haggard, B.E., Moorre, P.A., Daniel, T.C., and Edwards, D.R. 1999. Trophic conditions

and gradients of the headwater reaches of Beaver Lake. Arkansas. Proc Okla Acad Sci,

79: 73–84.

Haidvogel, D. B., Arango, H. G., Hedstrom, K., Beckmann, A., Malanotte-Rizzoli, P.

and Shchepetkin, A. F. 2000. Model evaluation experiments in the North Atlantic Basin:

Simulations in nonlinear terrain-following coordinates, Dyn. Atmos. Oceans, 32: 239-

281.

Haidvogel, D.B., Arango, H., Budgell, W.P., Cornuelle, B.D., Curchitser, E., Di Lorenzo,

E., Fennel, K., Geyer, W.R., Hermann, A.J., Lanerolle, L., Levin, J., McWilliams, J.C.,

Miller, A.J., Moore, A.M., Powell, T.M., Shchepetkin, A.F., Sherwood, C.R., Signell,

R.P., Warner, J.C. and Wilkin, J. 2008. Ocean forecasting in terrain-following

coordinates: Formulation and skill assessment of the Regional Ocean Modeling System.

Journal of Computational Physics, 227: 3595-3624.

Hall, R. W., and Dortch, M. S. 1994. New York Bight Study, Report 2: Feasibility Study

of the Development and Application of a General Water Quality Model of the New York

Bight, Technical Report CERC-94-4, U.S. Army Engineer Waterways Experiment

Station, Vicksburg, MS.

Hansen, J.W., Pedersen, A.-G.U., Berntsen, J., Ronbog, I.S., Hansen, L.S. and Lomstein,

B.A. 2000. Photosynthesis, respiration, and nitrogen uptake by different compartments of

a Zostera marina community. Aquatic Botany, 66: 281-295.

Harms, I.H. 1997. Modelling the dispersion of 137Cs and 239Pu released from dumped

waste in the Kara Sea. Journal of Marine Systems, 13: 1-19.

279

References

Hastings, A., and Powell, T. 1991. Chaos in a 3-species food chain. Ecology, 72, 896–

903.

Hauxwell, J., Cebrian, J., Furlong, C., and Valiela, I. 2001. Macroalgal canopies

contribute to eelgrass (Zostera Marina) decline in temperate estuarine ecosystems.

Ecology, 82: 1007-1022.

Hauxwell, J., Cebrian, J., and Valiela, I. 2003. Eelgrass Zostera marina loss in temperate

estuaries relationship to land-derived nitrogen loads and effect of light limitation imposed

by algae. Marine Ecology Progress Series, 247: 59-73.

Hayn M. 2012. Exchange of nitrogen and phosphorus between a shallow estuary and

coastal waters. Department of Natural Resources, Cornell University. MS Thesis. 73pp.

Hayn, M., Howarth, R., Marino, R., Ganju, N., Berg, P., Foreman, K. and Giblin, A. 2014.

Exchange of nitrogen and phosphorus between a shallow lagoon and coastal waters.

Estuaries and Coasts, 37 (1).

He, G., Fang, H., Bai, S., Liu, X., Chen, M., and Bai, J. 2011. Application of a three-

dimensional eutrophication model for the Beijing Guanting Reservoir, China. Ecological

Modelling 222, 1491-1501.

Hemminga, M.A., Harrison, P.G., and Van Lent, F. 1991. The balance of nutrient losses

and gains in seagrass meadows. Marine Ecology Progress Series, 71: 85-96.

Hemminga, M., and Duarte, C. 2000. Seagrass ecology: Cambridge University Press.

Hendriks, I.E., Sintes, T., Bouma, T.J. and Duarte, C.M. 2008. Experimental assessment

and modeling evaluation of the effects of the seagrass Posidonia oceanica on flow and

particle trapping. Marine Ecology Progress Series, 356: 163-173.

Hense, I. 2010. Approaches to model life cycle of harmful algae. Journal of Marine

Systems, 83 (3-4): 108-114.

280

References

Hinckley, S. and Dobbins, E.L. 2004. Investigating Gulf of Alaska data sets using an

NPZD model. PICES 13th Annual Meeting, Honolulu, Hawai, 22 October.

Hipsey, M.R., Brookes, J.D., and Aldridge, K. 2007. Setup and validation of a coupled

3D hydrodynamic-biogeochemistry model for the River Murray Lower Lakes. Report

prepared for Land and Water Australia.

Hipsey, M.R., and Hamilton, D.P. 2008. Computational aquatic ecosystem dynamic

model: CAEDYM v3 Science Manual: Centre for Water Research.

Hodges, B.R., Imberger, J., Saggio, A., and Winters, K.B. 2000. Modeling basins-sclae

internal waves in a stratified lake. Limnol. Oceanogr., 45 (7): 1603-1620.

Hood, R.R., Capone, D.G., and Olson, D.B., 2001. Modeling seasonal and interanual

biogeochemical and N2 cycles at BATS. Deep-Sea Res. II, 48: 1609–1648.

Hood, R.R., Kohler, K.E., McCreary, J.P., and Smith, S.L., 2003. A four-dimensional

validation of a coupled physical-biological model of the Arabian Sea. Deep-Sea Research

Part II: Topical Studies in Oceanography, 50: 2917-2945.

Hood, R.R., Laws, E.A., Armstrong, R.A., Bates, N.R., Brown, C.W., Carlson, C.A.,

Chai, F., Doney, S.C., Falkowski, P.G., Feely, R.A., Friedrichs, M.A.M., Landry, M.R.,

Moore, J.K., Nelson, D.M., Richardson, T.L., Salihoglu, B., Schartau, M., Toole, D.A.

and Wiggert, J.D. 2006. Pelagic functional group modeling: Progress, challenges and

prospects. Deep Sea Research Part II: Topical Studies in Oceanography, 53 (5): 459-512.

Hoover, C., Pitcher, T., and Christensen, V. 2013. Effects of hunting, fishing and climate

change on the Hudson Bay marine ecosystem II Ecosystem model future projections.

Ecological Modelling, 264:143-156.

Howarth, R.W., Hayn, M., Marino, R.M., Ganju, N., Foreman, K., McGlathery, K.,

Giblin, A.E., Berg, P. and Walker, J.D. 2014. Metabolism of a nitrogen-enriched coastal

marine lagoon during the summertime. Biogeochemistry, 118: 1-20.

281

References

Howes, B.L., Kelley, S.W., Ramsey, J.S., Samimy, R. and Schlezinger, D. 2006. Linked

watershed-embayment model to determine critical nitrogen loading thresholds for West

Falmouth Harbor, Falmouth, Massachusetts. Massachusetts Estuaries Project., 161 pp.

Boston, Massachusetts: Massachusetts Department of Environmental Protection.

HydroQual, Inc. 1987. A Steady-State Coupled Hydrodynamic/Water Quality Model of

the Eutrophication and Anoxia Process in Chesapeake Bay. Mahwah: HydroQual Inc.

I

IHCantabria, 2009. Estudio de los efectos del saneamiento sobre las aguas de la Albufera.

Environmental Hydraulics Institute of Cantabria. University of Cantabria, Santander.

IPCC. 2007. Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution

of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on

Climate Change. Parry, M.L., Canziani, O.F., Palutikof, J.P., van der Linden, P.J. and

C.E., ed. E. Hanson, Cambridge University Press, 976. Cambridge, UK.

J

Jernigan, R. W., and Tsokos, C. P. 1979. Phytoplankton modeling involving random rate

constants. Part I: Deterministic setting. Intern. J. Env. Std., 14: 97–105.

Jernigan, R. W., and Tsokos, C. P. 1980. Phytoplankton modeling involving random rate

constants. Part II: stochastic formulation. Intern. J. Env. Std., 15: 217–227.

Ji, Z-G, Morton, M.R., and Hamrick, J.M. 2001. Wetting and Drying Simulation of

Estuarine Processes. Estuarine, Coastal and Shelf Science, 53: 683-700.

Ji, Z-G, Hu, G., Shen, J., and Wan, Y. 2007. Three-dimensional modeling of

hydrodynamic processes in the St. Lucie Estuary. Estuarine Coastal and Shelf Science,

73: 188-200.

282

References

Jones, J.E., and Davies, A.M. 2006. Application of a finite element model (TELEMAC)

to computing the wind induced response of the Irish Sea. Continental Shelf Research, 26:

1519-1541.

Jorgensen, S.E. 2011. Handbook of Ecological Models used in Ecosystem and

Environmental Management. Applied Ecology and Environmental Management Series.

CRC Press. 636 pp.

K

Kawamyko, M. 2002. Numerical model approaches to address recent problems on pelagic

ecosystems. Journal of Oceanography, 58: 365–378.

Kenworthy, W.J., Gallegos, C.L., Costello, C., Field, D., and di Carlo, G. 2014.

Dependence of eelgrass (Zostera marina) light requierements on sediment organic matter

in Massachsetts coastal bays: implications for remediation and restoration. Marine

Pollution Bulletin, 83 (2): 446-57.

Kjerfve, B. 1994. Coastal Lagoon Processes. Elsevier Oceanography Series, 60:1-8.

Kiefer, D. A., and Atkinson, C. A. 1984. Cycling of nitrogen by phytoplankton: a

hypothetical description based upon efficiency of energy conversion. J. Mar. Res., 42:

655–675.

Kim, C.S., Lim, H.S. and Cerco, C.F. 2011. Three-dimensional water quality modeling

for tidal lake and coastal waters with ROMS-ICM. Journal of Coastal Research, SI64:

1068-1072.

Kishi, M.J. 1994. Prediction of phytoplankton growth in a Warm-Core Ring using three

dimensional ecosystem model. Journal of Oceanography, 50: 489-498.

Kishi, M.J., Kashiwai, M., Ware, D.M., Megrey, B.A., Eslinger, D.L., Werner, F.E., Aita,

M.N., Azumaya, T., Fujii, M., Hashimoto, S., Huang, D., Iizumi, H., Ishida, Y., Kang, S.,

283

References

Kantakov, G.A., Kim, H-C., Komatsu, K., Navrotsky, V.V., Smith, L.S., Tadokoro, K.,

Tsuda, A., Yamamura, O., Yamanaka, Y., Yokouchi, K., Yoshie, N., Zhang, J., Zuenko,

Y.I., Zvalinsky, V.I. 2007. NEMURO-Introduction to a lower trophic level model for the

North Pacific marine ecosystem. Ecol. Model, 202: 12–25.

Kishi, M.J., Ito, S., Megrey, B.A., Rose, K.A., and Werner, F.E. 2011. A review of the

NEMURO and NEMURO. FISH models and their application to marine ecosystem

investigations, Journal of Oceanography, 67: 3-16.

Kitanidis, P.K., 1997. Introduction to geostatistics: applications in hydrogeology.

Cambridge University Press.

Kremer, J.N., and Nixon, S.W., 1978. A Coastal Marine Ecosystem: Simulation and

Analysis. New York. Ecological Studies. Springer-Verlag.

Kroeger, K.D., Cole, M.L., York, J.K., and Valiela, I. 2006. Nitrogen loads to estuaries

from waste water plumes: Modeling and isotopic approaches. Ground Water, 44: 188-

200.

Kuo, J.T., Lung, W.S., Yang, C.P., Liu, W.C., Yang, M.D., and Tang, T.S. 2006

Eutrophication modelling of reservoirs in Taiwan. Environ Model Softw., 21: 829-844

L

Lassere, P., 1979. Coastal Lagoons: Sanctuary Ecosystems, Cradles of Culture, Targets

for Economic Growth. Nature and Resources, 15 (4): 2-21.

Lawrence, E., Jackson, A.R.W., and Jackson, J.M., 1998. Longman Dictionary of

Environmental Science: London, England, Addison Wesley Longman Limited, 144-145.

Lawrie, J., and Hearne, J. 2007. Reducing model complexity via output sensitivity.

Ecological Modelling, 207: 137-144.

284

References

Laws, E. A., Falkowski, P. G., W. O. S. Jr., Ducklow, H. and McCarthy, J. J. 2000.

Temperature effects on export production in the open ocean, Global Biogeochem. Cycles,

14: 1231–1246.

Lee, K.S., Park, S.R., and Kim, Y.K. 2007. Effects of irradiance, temperature, and

nutrients on growth dynamics of seagrasses: A review. Journal of Experimental Marine

Biology and Ecology, 350: 144-175.

Lee, Y.J., Boynton, W.R., Li, M., and Li, Y. 2013. Role of late winter-spring wind

influencing summer hypoxia in Chesapeake Bay. Estuaries and Coasts, 36: 683-96.

Lee, M.E., Oh, N.S., Kim, G., and Kang, J.W. 2013. Modeling tidal current around

Mokpo the south western coastal zone of Korea. Proceedings of the 7th International

Conference on Asian and Pacific Coasts (APAC 2013), Bali, Indonesia.

Leitão, P. C. 1996. Modelo de Dispersão Lagrangeano Tridimensional. Ms. Sc. Thesis,

Universidade Técnica de Lisboa, Instituto Superior Técnico.

Leonard, C.L., C.R. McClain, R. Murtugudde, E.E. Hofmann, and L.W. Harding Jr. 1999.

An iron-based ecosystem model of the central equatorial Pacific. Journal of Geophysical

Research C: Oceans, 104: 1325-1341.

Levins, R., 1966. The strategy of model building in population biology. American

Scientist, 54 (4): 421-431.

Li, M., Gargett, A., and Denman, K., 1999. Seasonal and interannual variability of

estuarine circulation in a box model of the Strait of Georgia and Juan de Fuca Strait.

Atmosphere-Ocean, 37: 1-19.

Lima, I.D., and S.C. Doney. 2004. A three-dimensional, multinutrient, and size-structured

ecosystem model for the North Atlantic. Global Biogeochemical Cycles, 18 (GB3019):

3011-3021.

285

References

Lindenschmidt, K.E., 2006. The effect of complexity on parameter sensitivity and model

uncertainty in river water quality modelling. Ecological Modelling, 190: 72-86.

Linker, L.C., Shenk, G., Wang, P., Hopkins, K.J., and Pokharel, S. 2001. A short history

of Chesapeake Bay Modeling and the next generation of watershed and estuarine models.

Proceedings of the Water Environmental Federation.

Lonin, S.A., and Tuchkovenko, Y.S. 2001. Water quality modelling for the ecosystem of

the Ciénaga de Tesca coastal lagoon. Ecological Modelling, 144: 279-293.

Lopes, J.F., Cardoso, A.C., Moita, M.T., Rocha, A.C., and Ferreira, J.A. 2009. Modelling

the temperature and the phytoplankton distributions at the Aveiro near coastal zone,

Portugal. Ecological Modelling, 220: 940-961.

Los, H. 2009. Eco-hydrodynamic modelling of primary production in coastal waters and

lakes using BLOOM. Thesis Wageningen University.

Losa, S.N., Vézina, A., Wright, D., Lu, Y., Thompson, K., and Dowd, M. 2006. 3D

Ecosystem modelling in the North Atlantic: relative impacts of physical and biological

parametrizations. Journal of Marine System, 61: 230–245.

M

Macedo, M.F., Duarte, P., Mendes, P., and Ferreira, J.G., 2001. Annual variation of

environmental variables, phytoplankton species composition and photosynthetic

parameters in a coastal lagoon. Journal of Plankton Research, 23: 719-732.

Machado, D.A., and Imberger, J. 2012. Managing wastewater effluent to enhance aquatic

receiving ecosystem productivity: A coastal lagoon in Western Australia. Journal of

Environmental Management, 99: 52-60.

286

References

Madden, C.J., and Kemp, W.M., 1996. Ecosystem model of an estuarine submersed plant

community: calibration and simulation of eutrophication responses. Estuaries, 19: 457-

474.

Madden, C.J., and McDonald, A.A. 2009. Florida Bay SEACOM: Seagrass Ecological

Assessment and Community Organization Model. Model documentation, version II.

South Florida Water Management District Coastal Ecosystems Division.

Mainardi F., F., Santos F., A., Collischonn, W., Ames, D.P., and Rigo, D., 2015. Large-

scale analytical water quality model coupled with GIS for simulation of point sourced

pollutant discharges. Environmental Modelling & Software, 64: 58-71.

Malmgren-Hansen, A., and Warren, R. 1992. Management of urban, industrial and

agricultural pollution to control coastal eutrophication. In Vollenweider, R.A., R.

Marchetti, and R. Viviani (eds.), Marine Coastal Eutrophication. Elsevier, Amsterdam,

pp. 1115-1132.

Marchesiello, P., J.C. McWilliams, and A. Shchepetkin. 2003. Equilibrium structure and

dynamics of the California current system. Journal of Physical Oceanography, 33: 753-

783.

Marra, J., and Ho, C. 1993. Initiation of the spring bloom in the northeast Atlantic (47°N,

20°W): a numerical simulation. Deep-Sea Res. II, 40: 55–73

Martín, M., 1998. Modelación de la calidad en aguas superficiales. Aplicación al caso de

la Albufera de Valencia. Tesis doctoral, Departamento de Ingeniería Química.

Universidad de Valencia, Valencia, 398 pp.

Martin, J.L. and Wool, T. 2002. A dynamic one dimensional model of hydrodynamics

and water quality EPD-RIV1, Version 1.0. User’s manual, AScI Cooperation, Athens,

Georgia.

287

References

Martin, L., Gouze, E., Razafindrakoto, E., Hervouet, J.M., Durand, N. and Pham, C.T.

2010. 3D coupled modeling of hydrodynamics and water quality in the Berre Lagoon

(France). Environmental Hydraulics, Vol. 2, Chapter 76. Christodoulou & Stamou (Eds).

Chatou, France.

Martins, F., Neves, R.J., Leitao, P.C. and Silva, A. 2001. 3D modelling in the Sado

estuary using a new generic coordinate approach. Oceanologica Acta, 24: S51-S62.

Maza, M., J.L. Lara, and I.J. Losada. 2013. A coupled model of submerged vegetation

under oscillatory flow using Navier–Stokes equations. Coastal Engineering, 80: 16-34.

McCann, K., and Yodzis, P. 1994. Nonlinear dynamics and population disappearances.

Am. Nat., 144: 873–879.

McClain, C.R., Arrigo, K., Tai, K.S., and Turk, D., 1996. Observations and simulations

of physical and biological processes at ocean weather station P, 1951-1980. Journal of

Geophysical Research C: Oceans, 101: 3697-3713.

McCreary, J.P., Kohler, K.E., Hood, R.R., Smith, S., Kindle, J., Fischer, A.S. and Weller,

R.A. 2001. Influences of diurnal and intraseasonal forcing on mixed-layer and biological

variability in the central Arabian Sea. Journal of Geophysical Research, 106: 7139-7155.

McGillicuddy, D. J., McCarthy, J. J., and Robinson, A. R. 1995a. Coupled physical and

biological modeling of the spring bloom in the North-Atlantic. 1. Model formulation and

one dimensional bloom processes. Deep-Sea Res. I, 42: 1313– 1357.

McGillicuddy, D. J., Robinson, A. R., and McCarthy, J. J. 1995b. Coupled physical and

biological modeling of the spring bloom in the North-Atlantic. 1. 3-dimensional bloom

and post-bloom processes. Deep-Sea Res. I, 42: 1359–1398.

McGlathery, K.J., K. Sundback, and I.C. Anderson. 2007. Eutrophication in shallow

coastal bays and lagoons: the role of plants in the coastal filter. Marine Ecology Progress

Series, 348:1-18.

288

References

Meccia, V., Simionato, C., Dragani, W. y Núñez, M. Circulación estacional y corredores

de flujo en el estuario del Río de la Plata sobre la base del modelo HamSOM/CIMA.

Simposio FREPLATA. Resumen extendido 4 pp.

Menéndez, M., Hernandez, O., Comin, F.A. 2002. Spatial distribution and

ecophysiological characteristics of macrophytes in a Mediterranean coastal lagoon.


Ménesguen, A., Cugier, P., Loyer, S., Vanhoutte-Brunier, A., Hoch, T., Guillaud, J.F.,

and Gohin, F. 2007. Two-or three-layered box-models versus fine 3D models for coastal

ecological modelling? A comparative study in the English Channel (Western Europe).

Journal of Marine Systems, 64 (1): 47-65.

Millenium Ecosystem Assessment, 2005. Ecosystems and Human Well-Being.:

Synthesis. Island Press, Washington, DC. 155 pp.

Miller, T.J., 2007. Contribution of individual-based coupled physical-biological models

to understanding recruitment in marine fish populations. Marine Ecology Progress Series,

347: 127-138.

Miracle, M.R., García, M.P., and Vicente, E., 1984. Heterogeneidad espacial de las

comunidades fitoplanctónicas de la Albufera de Valencia. Limnetica, 1: 20-31.

Mobley, C. D., and Sundman, L. K. 2008. Hydrolight 5 Ecolight 5 Technical

Documentation, 52 pp., Sequoia Sci., Inc., Bellevue, Wash.

Mobley, C.D. 2011. Fast light calculations for ocean ecosystem and inverse models.

Optics Express, 19 (20): 18927-18944.

Moll, A. 1998. Regional distribution of primary production in the North Sea simulated

by a three-dimensional model. Journal of Marine Systems, 16: 1-2.

289

References

Moll, A., and Radach, G. 2003. Review of three-dimensional ecological modelling related

to the North Sea shelf system: Part 1. Models and their results. Progress in Oceanography,

57 (2): 175-217.

Montero, P. 1999. Estudio de la hidrodinámica de la Ría de Vigo mediante un modelo de

volúmenes finitos (Study of the hydrodynamics of the Ría de Vigo by means of a finite

volume model), Ph.D. Dissertation, Universidad de Santiago de Compostela, in Spanish.

Montero, P., Gómez-Gesteira, M., Taboada, J.J., Ruiz-Villarreal., M., Santos, A.P.,

Neves, R. J. J., Prego, R. and Pérez-Villar, V. 1999. On residual circulation of Vigo Ría

using a 3D baroclinic model, Boletín Instituto Español de Oceanografía, nº15.

SUPLEMENTO-1.

Moriasi, D.N., Arnold, J.G., Van Liew, M.W., Bingner, R.L., Harmel, R.D., and Veith,

T.L., 2007. Model evaluation guidelines for systematic quantification of accuracy in

watershed simulations. Transactions of the ASABE, 50: 885-900.

Murray, A. G., and Parslow, J. S. 1997. Port Phillip Bay Integrated model: Final Report.

Port Phillip Bay Environment Study Technical Report no. 44, Melbourne.

Murray, A.G., and Parslow, J.S., 1999. The analysis of alternative formulations in a

simple model of a coastal ecosystem. Ecological Modelling, 119: 149-166.

A.G. Murray, Parslow, J.S. and Walker, S.J. 2001. Modelling treated waste disposal in

Port Phillip Bay and Bass Strait: biogeochemical and physical removal. Environment

International, 27: 249-255.

N

Neves, R. J. J. 1985. Étude Experimentale et Modélisation des Circulations Trasitoire et

Résiduelle dans l’Estuaire du Sado. Ph. D. Thesis, Universidade de Liège

290

References

Neves, R., Coelho, H., Leitão, P., Martins, H., and Santos, A. 1998. A numerical

investigation of the slope current along the western European margin. In: Burgano V.,

Karatzas G., Payatakas A., Brebbia C., Gray W. and Pinder G. (Ed.), Computational

Methods in Water Resources XII (2): 369-376.

Newton, A., Icely, J., Cristina, S. Brito, A., Cardoso, A.C., Colijn, F., Dalla Riva, S.,

Gertz, F., Hansen, J.W., Holmer, M., Ivanova, K., Leppäkoski, E., Canu, D.M.,

Mocennim, C., Mudge, S., Murray, N., Pejrup, M., Razinkovas, A., Reizopoulou, S.,

Pérez-Ruzafa, A., Schernewski, G., Schubert, H., Carr, L., Solidoro, C., Viaroli, P., and

Zaldívar, J.M. 2013. An overview of ecological status, vulnerability and future

perspectives of European large shallow, semi-enclosed coastal systems, lagoons and

transitional waters. Estuarine, Coastal and Shelf Science, 140: 95-122.

Nixon, S.W., 1995. Coastal marine eutrophication: a definition, social causes, and future

concerns. OPHELIA, 41: 199-219.

Nixon, S.W., B. Buckley, S. Granger, and J. Bintz. 2001. Responses of very shallow

marine ecosystems to nutrient enrichment. Human Ecological Risk Assessment, 7: 1457-

1481.

North, E. W., Schlag, Z., Hood, R. R., Li, M., Zhong, L., Gross, T., and Kennedy, V. S.

2008. Vertical swimming behavior influences the dispersal of simulated oyster larvae in

a coupled particle-tracking and hydrodynamic model of Chesapeake Bay. Marine

Ecology Progress Series, 359: 99-115.

NRC. 2000. Clean Coastal Waters: Understanding and Reducing the Effects of Nutrient

Pollution. Washington: National Research Council, National Academy Press.

O

291

References

OCDE, 1982. Eutrophisation des eaux. Métodes de surveillance, d'evaluation et de lutte.

Paris, 164 pp.

O’Donnell, J., Dam, H., McCardell, G., and Fake, T. 2009. Simulation of Long Island

Sound with the System-Wide Eutrophication Model (SWEM): Inter-annual Variability

and Sensitivity. Long Island Sound Study EPA Assistance Award Final Report. EPA

Grant Number: LI-97127101.

Ohno, S., and Nakata, K., 2008. Ecosystem model application to lake Hamana-Sanaru,

coastal lagoon system, OCEANS 2008.

Oliveira, A.P., Mateus, M.D., Cabecadas, G., and Neves, R. 2015. Water-air CO2 fluxes

in the Tagus estuary plume (Portugal) during two distinct winter episodes. Carbon

Balance and Management,10 (1): 2.

Onandia, G., Miracle, M.R., Blasco, C., and Vicente, E. 2014. Diel and seasonal

variations in bacterial production in a hypertrophic shallow lagoon. Aquat. Microb.

Ecol., 72: 255-267.

Onandia, G., Gudimov, A., Miracle, M.R., and Arhonditis, G. 2015. Towards the

development of a biogeochemical model for addressing the eutrophication problems in

the shallow hypertrophic lagoon of Albufera de Valencia, Spain. Ecological Informatics,

26 (3): 70-89.

O’Neill, R.V., Deangelis, D.L., Waide, J.B. and Allen, G.E. 1986. A Hierarchical Concept

of Ecosystems. Princeton University Press. ISBN: 9780691084374. 262 pp.

Orth, R.J. and K.A. Moore. 1988. Distribution of Zostera marina L. and Ruppia maritima

L. sensu lato along depth gradients in the lower Chesapeake Bay, U.S.A. Aquatic Botany,

32: 291-305.

Orth, R.J., T.J.B. Carruthers, W.C. Dennison, C.M. Duarte, J.W. Fourqurean, K.L. Heck

Jr, A.R. Hughes, G.A. Kendrick, W.J. Kenworthy, S. Olyarnik, F.T. Short, M. Waycott,

292

References

and S.L. Williams. 2006. A global crisis for seagrass ecosystems. BioScience, 56: 987-

996.

Oschlies, A., and Garcon, V., 1999. An eddy-permitting coupled physical-biological

model of the North Atlantic. 1. Sensitivity to advection numerics and mixed layer physics.

Global Biogeochemical Cycles, 13: 135-160.

P

Pace, M.L. 2001. Prediction and the aquatic sciences. Canadian Journal of Fisheries and

Aquatic Sciences, 58 (1): 63-72.

Pahlow, M., Vezina, A.F., Casaul, B., Maass, H., Malloch, L., Wright, D., and Lu, Y.

2008. Adaptive model of plankton dynamics for the North Atlantic. Progress in

Oceanography, 76: 151–191.

Park, R.A., 1990. AQUATOX, A Modular Toxic Effects Model for Aquatic Ecosystems.

U.S. Environmental Protection Agency, Corvallis, Oregon.

Parslow, J., Hunter, J., and Davidson, A., 1999. Estuarine Eutrophicaton Models, in:

Australia, L.a.w. (Ed.).

Peliz, A., Dubert, J., Haidvogel, D. B., and Le Cann, B. 2003. Generation and unstable

evolution of a density-driven Eastern Poleward Current: The Iberian Poleward Current, J.

Geophys. Res., 108 (C8), 3268.

Petersen, J.K., Marr, M., Ysebaert, T., and Herman, P.M.J. 2013. Near-bed gradients in

particles and nutrients above a mussel bed in the Limfjorden: Influence of physical mixing

and mussel filtration. Marine Ecology Progress Series, 490: 137-146.

Petihakis, G., Smith, C.J., Triantafyllou, G., Sourlantzis, G., Papadopoulou, K.-N.,

Pollani, A., and Korres, G. 2007. Scenario testing of fisheries management strategies

293

References

using a high resolution ERSEM-POM ecosystem model. ICES Journal of Marine Science,

64 (9): 1627-1640.

Platt, T., and Sathyendranath, S. 1988. Oceanic primary production: estimation by remote

sensing at local and regional scales. Science, 241: 1613-1620.

Postma, L. 1988. DELWAQ Users manual, Version 3.0., Delft Hydraulics, The

Netherlands, 190 pp.

Powell, E.N., Hoffman, E.E., Klinck, J.M., and Ray, S.M. 1992. Modeling oyster

populations I. A commentary on filtration rate. Is faster always better? Journal of Shellfish

Research, 11: 387-398.

Powell T. P., C. V. W. Lewis, E. N. Curchitser, D. B. Haidvogel, A. J. Hermann, and E.

L. Dobbins. 2006. Results from a three-dimensional, nested biological-physical model of

the California Current System and comparisons with statistics from satellite imagery, J.

Geophys. Res., 111, C07018.

R

Radach, G., and Lenhart, H.J. 1995. Nutrient dynamics in the North Sea-Fluxes and

budgets in the water column derived from ERSEM. Neth. J. Sea Res., 33: 301-335.

Ramanathan, V. 2010. Climate Impacts of Black Carbon. Testimony before the U.S.

House of Representatives, Select Committee on Energy Independence and Global

Warming. Washington, DC: March 16. Accessed April 10, 2011.

Rigler, F.H., and R.H. Peters. 1995. Science and Limnology. Book 6 In Kinne, O. (ed.),

Excellence in Ecology. Oldendorf/Luhe: International Ecology Institute.

Riley, G.A. 1946. Factors controlling phytoplankton populations on Georges Bank.

Journal of Marine Research, 6: 54-73.

294

References

Riley, G.A. 1947. A theoretical analysis of the zooplankton population of Georges Bank.

Journal of Marine Research, 6: 104-113.

Riley, G.A., H. Stommel, and D.F. Bumpus. 1949. Quantitative ecology of the plankton

of the Western North Atlantic. Bulletin of the Bingham Oceanographic Collection, 12: 1-

169.

Risgaard-Petersen, N., T. Dalsgaard, S. Rysgaard, P.B. Christensen, J. Borum, K.

McGlathery, and L.P. Nielsen. 1998. Nitrogen balance of a temperate eelgrass Zostera

marina bed. Marine Ecology Progress Series, 174: 281-291.

Risgaard-Petersen, N., and L.D.M. Ottosen. 2000. Nitrogen cycling in two temperate

Zostera marina beds: Seasonal variation. Marine Ecology Progress Series, 198: 93-107.

Romo, S., Villena, M.J., Sahuquillo, M., Soria, J.M., Giménez, M., Alfonso, T., Vicente,

E., and Miracle, M.R., 2005. Response of a shallow Mediterranean lake to nutrient

diversion:does it follow similar patterns as in northen shallow lakes? Freshwater Biology,

50: 1706-1717.

Romo, S., García-Murcia, A., Villena, M.J., Sánchez, V., and Ballester, A., 2008.

Tendencias del fitoplancton en el lago de la Albufera de Valencia e implicaciones para su

ecología, gestión y recuperación. Limnetica, 27: 11-28.

Ross, C. G., and Nival, P. 1976. Plankton modelling in the bay of Villefranche. J. Theor.

Biol., 156: 381–389.

Roselló, V.M., 1979. Els espais albuferencs del País Valencia. Acta Geológica Hispánica,

14: 487-493.

Ruan, S. G. 1993. Persistence and coexistence in zooplanktonphytoplankton-nutrient

models with instantaneous nutrient cycling. J. Math. Biol., 31: 633–654.

295

References

Ruan, S. G. 2001. Oscillations in plankton models with nutrient recycling. J. Theor. Biol.,

208 (1): 15–26.

Runca, W., Bernstein, A.G., Postma, L. and Di Silvio, G. 1996. The framework of

analysis to evaluate environmental measures in the Venice Lagoon. Quademi Trimestrali

(monographic supplement to Year III, No. 4 and Year IV No. 1): 55-70.

S

Sahuquillo, M., Melao, M. G. G. and Miracle, M. R. 2007. Low filtering rates of Daphnia

magna in a hypertrophic lake: laboratory and in situ experiments using synthetic

microspheres. Hydrobiologia, 594: 141-152.

Saito, L., Johnson, B.M., Bartholow, J., and Hanna, R.B. 2001. Assessing ecosystem

effects of reservoir operations using food web-energy transfer and water quality

models. Ecosystems, 4: 105-125.

Santos, A. J. 1995. Modelo Hidrodinâmico Tridimensional de Circulação Oceânica e

Estuarina. Ph. D, Thesis, Universidade Técnica de Lisboa, Instituto Superior Técnico.

Saraiva, S., Pina, P., Martins, F., Santos, M. Braunschweig, F., and Neves, R. 2007.

Modelling the influence of nutrient loads on Portuguese estuaries. Hydrobiologia, 587: 5-

18.

Scavia D., Kelly, E.L.A., and Hagy, J.D. 2006. A simple model for forecasting the effects

of nitrogen loads on Chesapeake Bay hypoxia. Estuaries and Coasts, 29: 674-84.

Shchepetkin, A.F., and McWilliams, J.C. 2005. The regional oceanic modeling system

(ROMS): a split-explicit, free-surface, topography-following-coordinate oceanic model.

Ocean Modelling, 9: 347-404.

296

References

Schladow, S.G., and Hamilton, D.P. 1997. Prediction of water quality in lakes and

reservoirs: Part II - Model calibration, sensitivity analysis and application. Ecological

Modelling, 96: 111-123.

Schlitzer, R. 2000. Applying the adjoint method for biogeochemical modeling: Export

of particulate organic matter in the World Ocean, in Inverse methods in biogeochemical

cycles, AGU Monograph, Vol. 114, edited by P. Kasibhata, 107–124 pp.

Sheng, Y.P., and Yassuda, E.A. 1995. Application of a three-dimensional circulation

model to Tampa Bay to Support Water Quality Modeling. Tampa Bay National Estuary

Program. Tampa Bay Regional Planning Council. University of Florida. Gainesville,

Florida. 60 pp.

Short, F.T., and Wyllie-Echeverria, S. 1996. Natural and human-induced disturbance of

seagrasses. Environmental Conservation, 23: 17-27.

Short, F.T., and Neckles, H.A. 1999. The effects of global climate change on seagrasses.

Aquatic Botany, 63: 169-196.

Short, F.T., and Short, C.A. 2004. Seagrasses of the Western North Atlantic. In: Green

EP and Short FT (eds) World Atlas of Seagrasses. University of California Press.

Berkeley, CA. pp. 207-215.

Simpson, J.H., Berx, B., Gascoigne, J., and Saurel, C. 2007. The interaction of tidal

advection, diffusion and mussel filtration in a tidal channel. Journal of Marine Systems,

68: 556-568.

Sjoberg, S. 1977. Are pelagic ecosystems inherently unstable? Ecological Modelling, 3:

17-37.

Skogen, M.D., Svendsen, E., Berntsen, J., Aksnes, D., and Ulvestad, K.B., 1995.

Modelling the primary production in the North Sea using a coupled three-dimensional

297

References

physical-chemical-biological ocean model. Estuarine, Coastal and Shelf Science, 41: 545-

565.

Solidoro, C., Brando, V.E., Dejak, C., Franco, D., Pastres, R., and Pecenik, G. 1997.

Long term simulations of population dynamics of Ulva r. in the lagoon of Venice.

Ecological Modelling, 102: 259-272.

Somha, A., Sekiguchi, Y., Kuwae, T., and Nakamura, Y. 2008. A benthic-pelagic coupled

ecosystem model to estimate the hypoxic estuary including tidal flat-Model description

and validation of seasonal/daily dynamics. Ecological Modelling, 215: 10-39.

Soria, J.M., Miracle, M.R., and Vicente, E., 1987. Aporte de nutrientes y eutrofización de

la Albufera de Valencia. Limnetica, 3: 227-242.

Soria, J.M., 2006. Past, present and future of the Albufera of Valencia Natural Park.

Limnetica, 25: 135-142.

Souto C., Fariña, L., Alvarez, E., and Rodríguez, I. 2001. Wind and tide current prediction

using a 3D finite difference model in the Ría de Vigo (NW Spain), Sci. Mar., 65 (Suppl.

1): 269-276.

Spitz, Y.H., Newberger, P.A., Allen, J.S. 2003. Ecosystem response to upwelling off the

Oregon coast: behavior of three nitrogen-based models. Journal of Geophysical Research,

108 (C3): 3062.

Steele, J.H. 1974. The Structure of Marine Ecosystems. Cambridge: Harvard University

Press

Steele, J. H., and Frost, B.W. 1977. The structure of plankton communities. Phil. Trans.

R. Soc. Lond., 280: 485–534.

Steele. J. H., and Henderson, E. W. 1981. A simple plankton model. Am. Nat., 117: 676-

691.

298

References

Steele, J. H., and Henderson, E. W. 1992. The role of predation in plankton models.

Journal of Plankton Research, 14: 157–172.

Stigebrandt, A., and Wulff, F. 1987. A model for the dynamics of nutrients and oxygen

in the Baltic proper. Journal of Marine Research, 45: 729-759.

Stoner, A. W., Ray, M., and Waite, J.M. 1995. Effects of a large herbivorous gastropod

on macro-fauna communities in tropical seagrass meadows. Marine Ecology Progress

Series, 121: 125-137.

Sullivan, A.B., Jager, H.I., and Mayers, R. 2003. Modeling white sturgeon movement in

a reservoir: the effect of water quality and sturgeon density. Ecological Modelling, 167

(1–2): 97–114.

Sundarambal, P., Tkalich, P., and Balasubramanian, R., 2010. Impact of biomass burning

on surface water quality in Southeast Asia through atmospheric deposition:

Eutrophication modeling. Atmospheric Chemistry and Physics Discussions, 10: 7797-

7818.

Sundarambal, P., Balasubramanian, R., and Tkalich, P. 2012. A 3-D Model on the

Possible Role of Atmospheric Deposition in Tropical Coastal Eutrophication. A

commemorative volume 2012: Contributions to Marine Science, Editor: Tan Koh-Siang,

TMSI St. John's Island 10th anniversary, NUS, Singapore, 11- 21 pp.

Swaney, D.P., Scavia, D., Howarth, R.W., and Marino, R.M. 2008. Estuarine

classification and response to nitrogen loading: insights from simple ecological models.


299

References

T

Taboada, J.J. 1999. Aplicación de modelos numéricos al estudio de la hidrodinámica y

del flujo de partículas en el Mar Mediterráneo (Application of numerical models for the

study of hydro-dynamics and particle fluxes in the Mediterranean Sea), Ph. D.

Dissertation, Universidad de Santiago de Compostela. In Spanish.

Taboada, J.J., Ruíz-Villarreal, M., Gómez-Gesteira, M., Montero, P., Santos, A. P., Pérez-

Villar, V. and Prego, R. 2000. Estudio del transporte en la Ría de Pontevedra (NO España)

mediante un modelo 3D: Resultados preliminares, In: Estudos de Biogeoquímica na zona

costeira ibérica, Eds. A. Da Costa, C. Vale and R. Prego, Servicio de Publicaciones da

Universidade de Aveiro.

Tagliabue, A., Aumont, O., and Bopp, L. 2014. The impact of different external sources

of iron on the global carbon cycle. Geophys. Res. Lett., 41 (3): 920-926.

Taylor, A.H. 1988. Characteristic properties of models for the vertical distribution of

phytoplankton under stratification. Ecological Modelling, 40: 175-199.

Taylor, A.H., Watson, A.J., Ainsworth, M., Robertson, J.E. and Turner, D.R. 1991. A

modelling investigation of the role of phytoplankton in the balance of carbon at the

surface of the North Atlantic. Global Biogeochemical Cycles, 5: 151-171.

Terrados, J., and J. Borum. 2004. Why are seagrasses important?-Goods and services

provided by seagrass meadows. Pages 8-10. In Borum, J., C.M. Duarte, D. Krause-Jensen,

and T.M. Greve. 2006. European seagrasses: an introduction to monitoring and

management. 88 p.

Testa, J.M. and Kemp, W.M. 2008. Variability of biogeochemical processes and physical

transport in a partially stratified estuary: a box-modeling analysis. Marine Ecology

Progress Series, 356: 63-79.

300

References

Tetra Tech. 1999. A Theoretical Description of Toxic Contaminant Transport

Formulations Used in the EFDC Model. Tech. Memo TT-EFDC-99-2, Tetra Tech, Inc.,

Fairfax, VA.

Thomann, R.V., and Mueller, J.A., 1987. Principles of Surface water Quality Modeling

and Control. HarperCollinsPublishers, New York.

Tian, R., and Chen, C. 2006. Influence of model geometrical fitting and turbulence

parameterization on phytoplankton simulation in the Gulf of Maine. Deep-sea research

Part II, 53: 2808-2832.

Titus, J.G., K.E. Andreson, D.R. Cahoon, D.B. Gesch, S.K. Gill, B.T. Gutierrez, E.R.

Thieler, and S.J. Williams. 2009. Coastal Sensitivity to Sea-Level Rise: A focus on the

Mid-Atlantic Region, 320 pp.: U.S. Climate Change Science Program.

Tkalich, P., and Sundarambal, P. 2003. Eutrophication modelling for Singapore waters.

Singapore Maritime & Port Journal, 122-136.

Truscott, J. E., and Brindley, J. 1994. Equilibria, stability and excitability in a general

class of plankton population models. Phil. Trans. R. Soc. Lond., A347: 703–718.

TYPSA, 2005. Estudio para el desarrollo sostenible de L'Albufera de Valencia.

Confederación Hidrográfica del Jucar, in: Jucar, C.H.d. (Ed.). Ministerio de Medio

Ambiente. Dirección General de Obras Hidráulicas y Calidad de las Aguas.

Confederación Hidrográfica del Júcar.

U

Uncles, R.J., and Torres, R. 2011. Estimating dispersion and flushing time-scales in a

coastal zone: application to the Plymouth area. Ocean and Coastal Management, 72: 3-

12.

301

References

Usaquen Perilla, O.L., Gómez, A.G., Gómez, A.G., Díaz, C.A., and Cortezón, J.A.R.,

2012. Methodology to assess sustainable management of water resources in coastal

lagoons with agricultural uses: An application to the Albufera lagoon of Valencia (Eastern

Spain). Ecological Indicators, 13: 129-143.

V

Van Dongeren, A.P. 2010. Delft3D-Modelling of Hydrodynamic and Water Quality

Processes in San Diego Bight. Deltares. Award number: N00014-08-1-1016.

Van Griensven, A., Meixner, T., Grunwald, S., Bishop, T., Diluzio, M., and Srinivasan,

R., 2006. A global sensitivity analysis tool for the parameters of multi-variable catchment

models. Journal of Hydrology, 324: 10-23.

Van Leeuwen, B., Augustijn, D.C.M., van Wesenbeeck, B.K., Hulscher, S.J.M.H., and

de Vries, M.B., 2010. Modeling the influence of a young mussel bed on fine sediment

synamics on an intertidal flat in the Wadden Sea. Ecological Engineering, 36 (2): 145-

153.

Vaz, N., Dias, J.M., Leitao, P.C., and Nolasco, R. 2007. Application of the Mohid-2D

model to a mesotidal temperate coastal lagoon. Computers and Geosciences, 33: 1204-

1209.

Vested, H.J., Jensen, O.K., Ellegaard, A.C., Bach, H.K. and Rasmussen, E.K. 1992.

Circulation modelling and water quality prediction. In Spaulding, M.L., K. Bedford, A.

Blumberg, R. Cheng, and C. Swanson (eds.), Estuarine and Coastal Modeling,

Proceedings of the 2nd International Conference. American Society of Civil Engineers,

New York, 317-331 pp.

Vicente, E. and Miracle, M. R. 1992. The Coastal Lagoon Albufera of Valencia: An

Ecosystem Under Stress. Limnetica 8: 87-100.

302

References

Viegas, C.N., and Nunes, S. 2009. Streams contribution on bathing water quality after

rainfall events in Costa do Estoril - a tool to implement an alert system for bathing water

quality. Journal of Coastal Research, SI (56): 1691-1695.

Viegas, C., Neves, R., Fernandes, R., and Mateus, M. 2012. Modelling tools to support

an early alert system for bathing water quality. Environmental Engineering and

Management Journal, 11 (5): 907-918.

Villena, M.J., and Romo, S., 2003. Phytoplankton changes in a shallow Mediterranean

lake (Albufera of Valencia, Spain) after sewage diversion. Hydrobiologia, 506-509, 281-

287.

W

Walmsley, R.D. 2000. Perspectives on Eutrophication of surface waters: Policy/research

needs in South Africa. WRC Research Report No. KV 129/00. Water Research

Commission, Pretoria, South Africa.

Wang, Z., Mu, D., Li, Y., Cao, Y., and Zhang, Y. 2011. Recent eutrophication and human

disturbance in Daya Bay, the South China Sea: dinoflagellate cyst and geochemical

evidence. Estuarine, Coastal and Shelf Science, 92: 403–414.

Warner, J.C., Sherwood, C.R., Arango, H.G., and Signell, R.P. 2005a. Performance of

four turbulence closure models implemented using a generic length scale method. Ocean

modelling, 8: 81-113.

Warner, J. C., Geyer, W. R., and Lerczak, J. A. 2005b. Numerical modeling of an estuary:

A comprehensive skill assessment, J. Geophys. Res., 110, C05001.

Warner, J.C., Sherwood, C.R., Signell, R.P., Harris, C.K., and Arango, H.G. 2008.

Development of a three-dimensional, regional, coupled wave, current, and sediment-

trasnport model. Computers & Geosciences, 34: 1284-1306.

303

References

Warner, J., Defne, Z., Haas, K., and Arango, H. 2013. A wetting and drying scheme for

ROMS. Computers & Geosciences, 58: 54-61.

Waycotta, M., Duarte, C.M., Carruthers, T.J.B., Orth, R.J. , Dennison, W.C., Olyarnik,

S., Calladine, A., Fourqurean, J.W., Heck Jr, K.L., Hughes, A.R., Kendrick, G.A.,

Kenworthy, W.J., Short, F.T. and Williams, S.L. 2009. Accelerating loss of seagrasses

across the globe threatens coastal ecosystems. Proceedings of the National Academy of

Sciences of the United States of America, 106: 12377-12381.

Wei, H., Sun, J., Moll, A. and Zhao, L. 2004. Phytoplankton dynamics in the Bohai Sea-

observations and modelling. Journal of Marine Systems, 44 (3-4): 233-251.

Wilkin, J.L., H.G. Arango, D.B. Haidvogel, C.S. Lichtenwalner, S.M. Durski, and K.S.

Hedstrom, 2005: A regional Ocean Modeling System for the Long-term Ecosystem

Observatory. J. Geophys. Res., 110, C06S91.

Wroblewski, J. S., Sarmiento, J. L., and Flierl, G. R. 1988. An ocean basin scale model

of plankton dynamics in the North Atlantic 1. Solutions for the climatological

oceanographic conditions in May. Global Biogeochem. Cycles, 2: 199–218.

Wroblewski, J.S. 1989. A model of the spring bloom in the North Atlantic and its impact

on ocean optics. Limnology & Oceanography, 34: 1563-1571.

Wu, T., Luo, L., Qin, B., Cui, G., Yu, Z., and Yao, Z., 2009. A vertically integrated

eutrophication model and its application to a river-style reservoir - Fuchunjiang, China.

Journal of Environmental Sciences, 21: 319-327.

Z

Zarzuelo, C., 2014. Modelado hidrodinámico de la Bahia de Cádiz. Proyecto Fin de

Máster. Universidad de Granada. Instituto Interuniversitario del Sistema Tierra Andalucía

(IISTA), Grupo de Dinámica de Flujos Ambientales., 18 pp.

304

References

Zhang, Y., Sun, Y., Yu, J., Yuan, D., and Zhan, R. 2012. A three-Dimensional Water

Quality Model and Its Application to Jiaozhou Bay, China. China Ocean Eng., 26 (4):

669-684.

Zimmerman, R.C., Reguzzon, J.L., and Randall, S.A. 1995. Eelgrass (Zostera marina L.)

transplants in San Francisco Bay: Role of light availability on metabolism, growth and

survival. Aquatic Botany, 51: 67-86.

Zimmerman, R.C., and Alberte, R.S. 1996. Effect of light/dark transition on carbon

translocation in eelgrass Zostera marina seedlings. Marine Ecology Progress Series, 136:

305-309.

Zimmerman, R.C. 2003. A biooptical model of irradiance distribution and photosynthesis

in seagrass canopies. Limnology and Oceanography, 48: 568-585.

Zouiten, H. 2012. Análisis Mediante Modelado Avanzado de Procesos de Eutrofización

en Lagunas Litorales: Aplicación a Masas de Agua Atlánticas y Mediterráneas.

Universidad de Cantabria. Tesis Doctoral. 371 pp.

Zouiten, H., Álvarez Díaz, C., García Gómez, A., and Revilla Cortezón, J.A. 2013. An

advanced tool for eutrophication modeling in coastal lagoons: Application to the Victoria

lagoon in the north of Spain. Ecological Modelling, 265: 99-113.

305

References

Documents

TESIS DOCTORAL - ccpo.odu.eduklinck/Reprints/PDF/delbarrioPHD2015.pdf · Resumen en español 17 Capítulo I: Introducción y antecedentes de la investigación 19 1.1 Motivación de