Embed Size (px)
Citation preview
Lessons learnt –what can be modelled?
EstMar meeting Sigulda 2009-11-27
Martin Isæus
Karl Florén
+46 8161007
Why model distributions? Blue mussel distribution based on UV-video transects
Scale 1:2 000 000 – 1:500 000
1:300 000 -1:100 000
1:50 000 -1:25 000
1:10 000 –1:5000
Tourism ?????? ??????
Detailed study
(habour, cable)
Scoping map Site study
Aquaculture Suitability assessment
Farm siting (ecological impact)
Farm construction
Renewable energy
Regional assessment
Site study (ecological impact
Site study (development/
Scales in mangement
energy assessment (ecological impact (development/ construction)
Aggregate /dredging
Regional assessment
Site study (licensing)
Coastal MPA Regional Sea scoping
National/ subregional scoping
Management map
Higher Sea MPA designation
Regional Sea scoping
Fisheries Resource assessment
Management map
Draft by Jacques Populus, Ifremer
Scales in survey methods and modelling
Scale 1:2 000 000 -1:500 000
1:300 000 -1:100 000
1:50 000 -1:25 000
1:10 000 - 1:5 000
Bathymetry from nauticalcharts x xBathymetry from digitalized oldmeasurments
x x x
Bathymetry from multi-beam x x x xMarine geologiy Regional quality xMarine geology, detailedquality x xquality x xInterpreted back-scatter
x x x xInterpreted side scan sonar
x xSediment modeling basedon multi-beam bathymetry x x x
Wave exposure (SWM)x x x
Predictive modeling, national scale xMarine landsscape(BALANCE ) xMPredictive modeling, pilot areas x
Spatial modelering
Abundans
Miljövariabel
1. Statistical
analysisField data
Model
3. Prediction
Model performance
Prediction accuracy
4. External validation
2. Cross validation
Area under curve (AUC)
AUC-value Quality
0,9 – 1,0 Excellent
0,8 – 0,9 Good
0,7 - 0,8 Intermideate
Pro
port
ion o
f corr
ect
cla
ssific
ations
0,7 - 0,8 Intermideate
0,5 - 0,7 Poor
Proportion of incorrect
classifications
Pro
port
ion o
f corr
ect
cla
ssific
ations
Fucus vesiculosus
Zostera marina
Forsmark
area,
Southern
Bothnian
Sea (SKB)
Carnivorous fish, biomass
Forsmark
area,
Southern
Bothnian
Sea (SKB)
Zooplankton-eating fish, biomass
Prediction of Bladder wrack in MopoDeco
Probability of presence
Prevalence 50% in training data
AUC training data: 0.924
AUC validation data PO: 0.914
AUC validation data PA: 0.520
Prediction
Probability >50%
Variable contribution:
depth 74.7%
salinity 15.9
sand 7.8
fetch 1.6fetch 1.6
New Finnish
bathymetry
Old bathymetry New bathymetry
New salinity in
Lithuania and
Kaliningrad
Old salinity New salinity
Prediction
Probability of presence
Prevalence 50% in training data
AUC training data: AUC training data: 0.946
AUC validation data PO: 0.944
AUC validation data PA: 0.616
2009/2010 – Modeling Östergötland and Västernorrland Counties
• Digitalized bathymetry
• Oceanographic model, salinity, temp, sediment
• Complementing drop-video survey• Complementing drop-video survey
• Validation of bathy, sed and bio data
• Detailed species modeling
• Coastal zone managers involved
Previous SWMwave exposure calculations
SWM Estonian coast
Simplified Wave Model, SWM
The full method description is found in thesis:
Isæus, Martin. 2004 ”Factors structuring Fucus communities at open and complex coastlines in the Baltic Sea”, Dept. Of Botany, Stockholm University, Sweden, ISBN 91-7265-846-0, p40.
Simplified refers to the fact that bathymtry is not used, only the coastline. The reason for that is that hight resolution bathymetry is rarely available. is rarely available.
Calculations are based on
• 10 years mean wind from coastal stations
• Best available shoreline
Fetch calculations, Diffraction
• A diffraction algorith is used during the fetch calculations which make the waves spread
• Callibrated using aerial photograps on wave crests
Islandwave crests
(i, J)
(i+1, J-2) (i+1, J-1) (i+1, J+1) (i+1, J+2)
BioEx
R2 = 39.6
SWM
R2 = 55.2
Comparison with other wave models. Correlations with Biological exposure index (BEI) based on shore organisms
distributions.
STWAVE
R2 = 36.2
Oceanographic
method, uses
bathymetry
FWM
R2 = 48.9
SWM vs. STWAVE wave characteristicsin modeling kelp (Laminaria hyperborea) in Møre, Norway
Model/parameters AIC cvROC
SWM 424 0.78
STWAVE – orbital velocity 474 0.73
STWAVE – bed stress 483 0.72STWAVE – bed stress 483 0.72
STWAVE – wave height 467 0.73
Conclusions = lessons learnt
Modelled species/habitat maps are wanted!
It is possible to make good models! (indicated by high model performance values)
To be able to make good regional predictions the environmental layers must be improved environmental layers must be improved considerably! (poor external validation values)
SWM Wave exposure –robust and relevant
Map quality should meet 1:50 000 scale for most management purposes
Thank you!
Martin Isæus
+46 8161011
www.aquabiota.se
