LINKING SPECIES RICHNESS, LITTER CHEMICAL DIVERSITY AND SOIL CARBON DYNAMICS IN THE ATLANTIC FOREST, BAHIA, BRAZIL
By
KIMBERLY Y. EPPS
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
2009
1
© 2009 Kimberly Y. Epps
2
To Vera Jean and Omar Jamal, in whose hearts no bitterness could take root
3
ACKNOWLEDGMENTS
Little can be accomplished in solitude. From its inception to its completion, this work is
beholden to a cast of thousands. Foremost, I thank my mother, Yvonne D. Epps, for her
steadfast encouragement and unconditional love, my family consisting of the Shareef-Epps-
Minter-Perkins-Justice-Runge Units, and its drafted members Joyce Dow, Toni Greene-Garner,
and Edith Yamanoha.
My life of science began at my kindergarten discovery that stars were fiery balls of gas.
Since then I have been fortunate to have been influenced by phenomenal teachers and mentors
along the years. I express sincere gratitude to Mrs. Rosenthal (3rd grade) and Mrs. Millet (4th
grade) of P.S. 95Q; Mrs. Proimos (6th grade) and Mr. Ytuarte (7th grade) of the Immaculate
Conception School of Jamaica; Mme. Gerton (French), Mr. Gordon (Physics), Mr. Greenberg
(Geometry), and Mr. Kelly (English) of Bronx Science; Prof. Shirlynn Spacapan (Psychology)
and Prof. Art Benjamin (“Fun Math”) of HMC; Dr. Paul Marcotte (International Agricultural
Development) , Dr. Bill Rains (Agronomy), and Dr. Kate Scow (Soil Microbiology) of U.C.
Davis. For training my eyes, ears, head and hands, I thank Sheila Pinkel (Photography,
Pomona), Marti Kanin and Julie Hochman (‘Cello). For unparalleled training in field grit and
economy, I am indebted to Dr. Eliska Rejmankova of U.C. Davis.
Here at my current home in Gator Country, I thank my doctoral committee for expanding
my horizons while tightening my scientific investigation skills: Wendell Cropper for introducing
me to the wonders of open source code and to the genius of Ramon Margalef; Willie Harris for
being a model thinker and professor; Quintino Araújo for my immersion in the scientific and
artistic culture of cacao; and P.K. Nair and Nigel Smith for grounding my inquiries in application
and human need. Adding to the embarrassment of riches have been my “behind-the-scenes”
advisors, Jim Reeves (USDA); Nairam Barros (UFV), Jack Ewel (UF); Al Medvitz (UCD);
4
Mark Van Horn (UCD); Arlicélio de Queiroz Paiva (UESC); Rosana Higa (EMBRAPA
Floresta), George Sodrê (CEPLAC); Yuncong Li (UF); and Virupax Baligar (USDA) with whom
it would be a pleasure and honor to consider future research plans.
For bringing me up to speed on soil microbiology and the wonders of Biolog, I thank Cory
Krediet and Max Teplitski. For my crash course in the Atlantic Rain Forest I am grateful for the
technical orientation of André M. Amorim and José Limão da Paixão of the CEPLAC Herbarium
as well as to my dear friend, right hand, and mateiro superior, Agnaldo Oliveira da Silva. I
acknowledge the support of the community of the Assentamento Frei Vantuy and in particular,
the leadership of Maïsa Fontana and Marlene Rebouças. The expertise and daily efforts of Dr.
Helena Serôdio, José Nelson Machado, Celso, Claudionou, Osmario of CEPLAC-CEPEC;
Edilson Fernando da Silva, “Seu” José Alberto, Luis Claudio de Almeida Barbosa, Antônio Lelis
Pinheiro, Verónica, and Aldemir at the Federal University of Viçosa were indispensible to my
work, as were the time and tutelage of the inimitable DRIFTS-Squad, Barry Francis and Tanesha
Simmons of USDA-EMBUL.
For their willingness to brave dangers ranging from vigilante wasps and busted boots to
siever’s lung, pipettor’s thumb and dishpan hands, I thank my Brothers and Sisters in Soil past
and present: Dalton Abdala, Kahlil Apuzen-Ito, Elena Azuaje, Christine Bliss, Lizette Borges
Gómez, Oldair Vinhas Costa, Eric Carvalho, Joy Futrell, Antônio and Emanuela Gama-
Rodrigues, Sarah Johnson, Francisco Lópes Neto, Isabel Lopez-Zamora, Melissa Martin, Edson
Márcio Matiello, Marina Morales, Deoyani Sarkhot, Shinjiro Sato, Sharon Schnabel, Lauren
Serra, Lisa Stanley, Aja Stoppe, Dashuai Sun, Ken van Rees, Marcela Quintero, and especially,
Barbara Cade-Menun for leading the way down the dirt path.
5
For reminding me that life is not a one-note samba, but rather a frevo, forró and at times a
choro, I thank Abby Lóces; Ana Amelia Guimarães Araujo; Anthony Heric; Arceli Trindade da
Silva; Ayesha Nibbe; Carol Souza; Charlotte Norris; David Giles; David Healy; Flora Piasentin;
Francisco “Chico” de Vale; Glória Fidelis da Paixão; Jina Kim; Justine Di Fiore; Katie Painter;
Kipp Sutton; Larry Dieterich; the “Rainha das Aguas”, Maria Apareçida; Maria Mariano; Maura
Barros; Michelle Young; Morena Maia; Neyde Alice Pereira; “ O Paulozão”; Prashanth Ak;
Rozeli Shiroma; Scott Looney; and Sylvia Ward. I am grateful for those who kept my heart and
belly fed – Renata Santiago, Judite de Ilhéus; Regina Alvez Ferreira; Jean Rissman; Karen van
Epen; Danielle Calin; Paty Guerra; Gustavo Dantas; Ana Elisa Del’Arco; Cecilia and Wilton
Carvalho; Halter Maia; Ivani de Brito; Sônia Sampaio, and Gina Gonçalves dos Santos.
This work was performed under the authorization of CNPq and IBAMA Permit No. RMC -
004/05. I am grateful to the kind attentiveness of Dr. Francisco Guerra who led us along the way
to authorization. For their financial support, I thank the generous contributions of the University
of Florida Alumni Fellowship; the Graduate Women in Science-Nell I. Mondy Award; the Ford
Foundation Pre-Doctoral Diversity Fellowship and the Epps-Tadross-Minter-Painter Student
Emergency Funds.
Above all, I welcome the opportunity to once again express my deep gratitude for the
otherworldly patience, good humor, cheerleadership and translation abilities of my advisor and
role model, Nicholas B. Comerford.
Lastly, for giving me the home of my spirit, a second mother tongue, and a very, very fine
dog, I thank Bahia, herself.
6
TABLE OF CONTENTS
page
ACKNOWLEDGMENTS ...............................................................................................................4
LIST OF TABLES...........................................................................................................................9
LIST OF FIGURES .......................................................................................................................11
ABSTRACT...................................................................................................................................13
CHAPTER
1 INTRODUCTION ..................................................................................................................15
2 CHEMICAL DIVERSITY – HIGHLIGHTING A SPECIES RICHNESS AND ECOSYSTEM FUNCTION DISCONNECT.........................................................................17
Introduction.............................................................................................................................17 Materials and Methods ...........................................................................................................22
Characterization of Leaf Tissue.......................................................................................22 Mathematical Treatment of Spectral Data.......................................................................23
Results and Discussion ...........................................................................................................24 Experimental Ramifications of Chemical Diversity........................................................27
Conclusion ..............................................................................................................................30
3 THE EFFECT OF CHEMICAL IDENTITY AND CHEMICAL DIVERSITY ON THE DECOMPOSITION OF TROPICAL LEAF MIXTURES.....................................................38
Introduction.............................................................................................................................38 Materials and Methods ...........................................................................................................41
Leaf Material Collection and Characterization ...............................................................41 Calculation of Chemical Diversity ..................................................................................42 Incubation Study..............................................................................................................43 Microbial Community-Level Physiological Profiling.....................................................45 Statistics...........................................................................................................................47
Results.....................................................................................................................................49 Species Identity on Carbon Mineralization of Mixtures .................................................49 Chemical Identity on Carbon Mineralization of Mixtures ..............................................50 Identification of Functional Chemical Traits ..................................................................52 Microbial Functional Diversity and Non-additive Effects ..............................................53
Discussion...............................................................................................................................54 Importance of Species Identity and Chemical Identity on Carbon Mineralization .........54 Which Traits? ..................................................................................................................57 Microbial Underpinnings of Non-Additive Effects.........................................................59
Conclusion ..............................................................................................................................60
7
4 TREE SPECIES DIVERSITY AND NUTRIENT CYCLING POTENTIALS OF SMALL-SCALE CACAO PRODUCTION SYSTEMS AND SECONDARY FOREST FRAGMENTS IN SOUTHERN BAHIA, BRAZIL ..............................................................91
Introduction.............................................................................................................................91 Materials and Methods ...........................................................................................................94
Site Selection and Characterization.................................................................................94 Species Inventory and Tree Diversity .............................................................................95 Forest Floor .....................................................................................................................96 Litterfall Production ........................................................................................................96 Soil Nutrient Status..........................................................................................................96 Statistical Analysis ..........................................................................................................97
Results and Discussion ...........................................................................................................98 Tree Density and Species Inventory................................................................................98 Litter Production, Standing Biomass.............................................................................102 Nutrient Inflows through Litterfall ................................................................................103 Soil Nutrient Status........................................................................................................104
Conclusion ............................................................................................................................106
5 OVERVIEW AND SYNTHESIS.........................................................................................127
APPENDIX
A USING INFRARED SPECTROSCOPY TO DETERMINE THE RELATIVE CONTRIBUTION OF ORGANIC INPUTS TO SURFACE SOIL ORGANIC MATTER..............................................................................................................................132
Introduction...........................................................................................................................132 Materials and Methods .........................................................................................................133
Study Site and Sampling ...............................................................................................133 Single-species stands..............................................................................................133 Mixed-species stands..............................................................................................134
Sample Preparation........................................................................................................135 Spectral Analysis and Isotopic Carbon Analysis...........................................................135 Statistical Analysis ........................................................................................................135
Results and Discussion .........................................................................................................136 Relative Importance of Above- and Belowground Inputs.............................................136 Interpretation .................................................................................................................138
Conclusion ............................................................................................................................140
LIST OF REFERENCES.............................................................................................................147
BIOGRAPHICAL SKETCH .......................................................................................................165
8
LIST OF TABLES
Table page 2-1 Published mineral nutrition studies reporting total nutrient concentrations used to
investigate the chemical diversity (CDQ) and species richness relationship. ....................32
3-1 Initial foliar total nutrient concentrations of 10 tropical tree species used in the incubation study. ................................................................................................................64
3-2 Species composition and abbreviation codes of the 21 leaf mixtures observed in the incubation study. ................................................................................................................65
3-3 Chemical diversity indices of leaf mixtures according to the mode of chemical characterization of material: total nutrient concentration (Total), mid-infrared (MIR) and near-infrared spectroscopy (NIR). ..............................................................................66
3-4 Kendall’s Tau correlation coefficient of total nutrient concentrations with decomposition responses of leaf-mixtures.........................................................................67
3-5 Summary of repeated measures ANOVA for effects of mixture composition on net respired CO2 of 21 leaf mixtures at Day 80.......................................................................68
3-6 Summary of factorial ANOVA for effects of mixture composition on the carbon mineralization rate constant, a, of 21 leaf mixtures...........................................................68
3-7 Expected and observed carbon mineralization rates and non-additive effect of paired litter treatments in incubation study...................................................................................69
3-8 Summary of factorial ANOVA for effects of mixture composition on interactions of net CO2, I net CO2, and the C mineralization rate constant, I a, of 21 leaf mixtures............70
3-9 Summary of selected successful models† relating chemical identity (1/λ) to net evolved CO2 and C mineralization rate of 21 leaf mixtures. .............................................71
3-10 Summary of successful models† relating chemical identity (1/λ) and chemical diversity (CDQ) on C mineralization rate of 21 leaf mixtures. ..........................................71
3-11 Summary of successful models† of the relationship of chemical identity and chemical diversity on the decomposition responses of INGA leaf mixtures (n=7). ........................72
3-12 Summary of successful models† of the relationship of chemical identity and chemical diversity on the decomposition responses of EMBA leaf mixtures (n=7).........................73
3-13 Summary of successful models† of the relationship of chemical identity and chemical diversity on the decomposition responses of PAUM leaf mixtures (n=7).........................73
9
3-14 Discriminant analysis of non-additive effects observed in the C mineralization rates of 21 leaf mixtures, performed on the first four PCA factor scores of the spectra of key and companion species................................................................................................74
3-15 Spectral traits of key species (MIR) contributing to successful models of leaf-mixture decomposition processes and their potential physiochemical interpretation. ....................75
3-16 Kendall’s Tau correlation between microbial functional diversity (Gini coefficient) and mean substrate use of 31 EcoPlate substrates and the decomposition responses of mixed leaf treatments.. .......................................................................................................76
4-1 Soil chemical properties and particle size distribution of cacao and secondary forest study sites in the Assentamento Frei Vantuy, Ilhéus, Bahia, Brazil. ...............................109
4-2 Summary of the floristic diversity of four cabruca sites (0.25 ha) located in the Frei Vantuy Settlement, Ilhéus, Bahia, Brazil.........................................................................110
4-3 Summary of the floristic diversity of four secondary forest sites (0.25 ha) located in the Frei Vantuy Settlement, Ilhéus, Brazil.......................................................................111
4-4 Phytosociological parameters (excluding cacao) of cabrucas in southern Bahia observed in other studies..................................................................................................112
4-6 Tree species (≥ 5 cm DBH) identified in the selected study areas of cabruca and secondary forest in the Assentamento Frei Vantuy, Ilhéus, Bahia, Brazil. .....................114
4-7 Mean total nutrient concentrations of surface soil (0-10cm) and litter and total annual input of nutrients through litterfall in traditional cacao systems and adjacent secondary forest sites in the Assentamento Frei Vantuy, Ilhéus, Bahia, Brazil. .............118
10
LIST OF FIGURES
Figure page 2-1 Chemical diversity (CDQ) as a function of species richness of eight litter species
native to the Atlantic Forest region of southern Bahia. .....................................................33
2-2 The effect of the removal of species from a mixture on chemical diversity under assumption of equal species abundance, and DRIFTS characterization of the assemblage in Fig. 2-1. ......................................................................................................34
2-3 The relationship of foliar CDQ diversity with species richness of two different temperate forest communities. . ........................................................................................35
2-4 The idiosyncrasy of additive studies..................................................................................36
3-1 Diffuse reflectance infrared spectra of 10 litter species.....................................................77
3-2 Chemical diversity (normalized to the maximum value) of all possible mixtures of 10 selected tropical species as a function of species richness, as characterized by total nutrient concentrations. ....................................................................................................78
3-3 Chemical diversity (normalized to the maximum value) of all possible mixtures of 10 selected tropical species as a function of species richness, as characterized by mid-infrared spectroscopy (MIR). ............................................................................................79
3-4 Chemical diversity (normalized to the maximum value) of all possible mixtures of 10 selected tropical species as a function of species richness, as characterized by near-infrared spectroscopy (NIR). ............................................................................................80
3-5 Carbon mineralization of ten individual tropical species and the control treatment. . ......81
3-6 The effect of key species on decomposition responses after 80 days................................82
3-7 Dynamic nature of non-additive effects on cumulative evolved CO2 between litter pairs over the course of incubation. ...................................................................................83
3-8 Interaction effects of nitrogen-fixing and non nitrogen-fixing leaf mixtures.. ..................84
3-9 Predicted versus observed interactions occurring in C mineralization rate constant and net evolved CO2 (Day 80) of the 21 mixed-litter treatments. ...................................85
3-10a Chart of spectral traits, one each from key and companion species, contributing to successful models of the form response ~ 1/λ key + 1/λ comp where response is net evolved CO2 (Day 30).. ......................................................................................................86
11
3-10b Chart of spectral traits, one each from key and companion species, contributing to successful models of the form response ~ 1/λ key + 1/λ comp where response is net evolved CO2 (Day 80).. ......................................................................................................87
3-11 Chart of spectral traits, one each from key and companion species, contributing to successful models of the form response ~ 1/λ key + 1/λ comp where response is the carbon mineralization rate constant, a. ..............................................................................88
3-12 Principal components analysis of the single-substrate utilization profile of the 31 microbial communities litter treatments extracted from the final time step the incubation study. ................................................................................................................89
3-13 Mean substrate use (AWCD) and functional diversity (GINI) of microbial communities extracted from 21 litter-pair treatments grouped by key species. ................90
4-1 Location of study area among land use types in the municipality of Ilhéus, Bahia, Brazil................................................................................................................................119
4-2 Approximate locations of observed traditional cacao and secondary forest study sites in the Assentamento Frei Vantuy, Ilhéus, Bahia, Brazil..................................................120
4-3 Example of forest structure of secondary forest fragments.. ...........................................121
4-4 Example of forest structure of cabruca systems. .............................................................122
4-5 Diameter class distribution of stems in two secondary forest fragments.........................123
4-6 Monthly litterfall in cacao and secondary forest collected from the period September 2005 to August 2006........................................................................................................124
4-7 Standing biomass on the forest floor of forested plots in May and November, 2005.. ...125
4-8 Total organic matter content and organic matter distribution among particle size fractions of cacao and secondary forest sites.. .................................................................126
A-1 Sum of squared differences from of organic inputs plotted against each other at the 0-5 cm and 5-10 cm depths. . ..........................................................................................143
A-2 Isotopic carbon signatures of 18 single-species plots. .....................................................144
A-3 Sum of squared differences between forest floor litter and root litter in a multi-species forested system atop a clayey soil, based on non-combusted and combusted (cleaned) spectra. .............................................................................................................145
A-4 Sum of squared differences between forest floor litter and root litter in a multi-species forested system atop a sandy soil, based on non-combusted and combusted (cleaned) spectra. .............................................................................................................146
12
Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy
LINKING SPECIES RICHNESS, LITTER CHEMICAL DIVERSITY,
AND SOIL CARBON DYNAMICS IN THE ATLANTIC FOREST, BAHIA, BRAZIL
By
Kimberly Y. Epps
May 2009
Chair: Nicholas B. Comerford Major: Soil and Water Science
The high productivity and high species diversity of tropical forests amidst the supposed
nutrient limitation of tropical soils prompt the question, “Does the botanical diversity of tropical
forests assist in the maintenance of nutrient bioavailability?” The primary objective of this
research was to establish the relationship, if any, between plant-litter diversity and litter
decomposition employing tree species of the biodiversity hotspot of the Atlantic Rain Forest of
southern Bahia, Brazil. The central hypothesis was that the chemical diversity of leaf mixtures is
a predictor of microbially mediated processes such as leaf litter decomposition. Two aspects
central to this work were (1) the development of a chemical diversity index of leaf mixtures and
(2) the use of infrared spectroscopy to chemically characterize each species.
Field and laboratory studies targeted the impact of plant litter diversity on soil properties
and processes. The objective of the field study was to evaluate and compare cocoa production
systems and adjacent areas of secondary forest for tree species diversity, inflows of leaf litter
mass and nutrients, and stocks of nutrients in surface soils. Results indicated that despite slightly
lowered tree density and tree diversity compared to secondary forest, traditional cocoa
production systems constituted an agroforestry system that closely resembles secondary forest in
terms of biomass production and carbon and nutrient inputs to surface soil.
13
14
A laboratory incubation study was conducted to test the ability of the chemical diversity
index and chemical identity—the comprehensive chemical fingerprint of individual species by
infrared spectroscopy—to predict rates of carbon mineralization of species mixtures. Infrared
spectral regions were identified that explained as much as 86% of the variation in carbon
mineralization of leaf mixtures. Chemical diversity as a solo parameter showed no correlation
with mixture decomposition, but in conjunction with chemical traits, predicted rates of carbon
mineralization when mixtures were grouped by the presence of a “key” species. Results suggest
that chemical diversity may be most effective in predicting the decomposition of mixtures
dominated in quantity or behavior by a single species.
CHAPTER 1 INTRODUCTION
The transformation of plant litter into soil organic matter and the attendant liberation of
nutrients form the backbone of the terrestrial biogeochemical cycle. Organic matter
decomposition is a complex series of interchanges between plant material, mineral substrate and
the decomposer community that is the precursor to an array of ecosystem functions such as the
recycling of nutrients from dead to living matter that drives ecosystem productivity and carbon
sequestration. Confident as we have become in predicting the degradation of litter derived from
single species using litter quality parameters such as initial nitrogen (N) content, or lignin:N
ratios, our predictions fail when we apply the same relationships to the decay of mixed organic
substrates. The departure of mixtures from the mean response of their individual constituents,
known as “non-additive” effects, signals a gap in our understanding of the mechanisms operating
in the breakdown of multiple and single substrates. Given the prevalence of mixed litter
decomposition, either in the context of multi-species assemblages or differently aged litterfall of
the same species, soil organic carbon models that do not incorporate the phenomenon of non-
additive effects of diverse litter mixtures are inadequate to predict soil C dynamics and potential
modifications brought by climate change.
The aim of this work was to predict the non-additive response of leaf mixtures and to
identify the initial chemical traits correlated with them. The effort to define leaf mixture
diversity in terms of leaf chemistry led to supporting investigations in the application of diffuse
reflectance infrared Fourier transform spectroscopy (DRIFTS) on the relationship between the
chemistry of organic matter inputs to soil and soil properties. Lastly, live systems of contrasting
tree species diversity of similar climate, soil and annual rates of organic biomass to soil through
15
16
litterfall provided the basis of a field-level comparison of the effect of litter chemical diversity on
soil properties.
The role of the definition of diversity on the ability to establish firm connections between
litter mixture diversity and decomposition was addressed in Chapter 2. A new index of chemical
diversity based was introduced based on the chemical composition of constituents within the
plant assemblage.
Chapter 3 relates an incubation study directed at (i) modeling the decomposition of
mixtures constructed from the leaves of 10 tropical tree species, using either the chemical traits
of mixture constituents, the chemical diversity of the mixture, or both and (ii) determining the
relationship between the non-additive effects of leaf mixtures and the functional diversity of their
associated microbial communities.
In Chapter 4, the tree species diversity and soil nutrient status of traditional cocoa
production systems, representing a “low diversity” land use was compared to that of proximal
secondary forests within the “biodiversity hotspot” of the Atlantic Rain Forest in Bahia, Brazil.
Lastly, Chapter 5 highlights the key findings of the work, notes the inherent limitations of
the approaches taken, recommends ways of overcoming them and suggests investigatory
pathways to further elucidate the role of plant chemical diversity in terrestrial biogeochemical
cycling.
CHAPTER 2 CHEMICAL DIVERSITY – HIGHLIGHTING A SPECIES RICHNESS AND ECOSYSTEM
FUNCTION DISCONNECT1
Introduction
The spatial distribution of tree species, fungal hyphal networks, intersecting root systems
and successions of litterfall ensure that litter interactions are the rule rather than the exception in
forest ecosystems. It is well established that the chemical nature of individual leaf litters affects
rates of decomposition, the primary engine of nutrient recycling in terrestrial ecosystems (Swift
et al. 1979). However, the role of plant litter diversity on litter decomposition and nutrient
mineralization continues to be the subject of inquiry in natural and planned ecosystems of
tropical and temperate climate zones (Finzi & Canham, 1998; Rothe &Binkley, 2001; Zimmer
2002; Gama-Rodrigues et al., 2003). To date, studies of diversity effects on varied aspects of
nutrient turnover have recognized the “non-additive effect” of mixed litter interactions, but have
failed to generalize either the magnitude or direction of the interactions in relation to either the
number of species present or to functional group richness (Gartner & Cardon, 2004;
Hättenschwiler et al., 2005). In fact, the recurrent outcome of these works suggests that species
identity surpasses species diversity as the main driver of decomposition-related processes
(Wardle et al., 1997; Gastine et al., 2003; de Deyn et al., 2004). Furthermore, recent works have
shown mixture interactions among even intra-species litters, suggesting that genetic variations
within species leading to litter quality differences may be of equal or greater significance than
species designations (Madritch & Hunter, 2005; Schweitzer et al., 2005).
The hypothesis that litter interactions during decomposition stem, in part, from the effects
of resource heterogeneity on fungal, bacterial and arthropod composition and activity is not new
1 This chapter was published in Oikos Volume 116, No. 11, pp. 1831-1840, and is cited as Epps et al., 2007 in the remainder of the text.
17
(Seastedt, 1984; Chapman et al., 1988; Blair et al., 1990). However, the principal approach
employed thus far to test this hypothesis—generating mixtures by varying species or functional
group richness—may be partly responsible for the current failure to validate the hypothesis.
Two assumptions encumber these experiments. First, species richness is made equivalent to
“difference” in the trait or traits pertinent to the observed activity and, second, this difference is
presumed to increase systematically with richness. Containing little explicit information of litter
quality, diversity values based on taxonomy or growth habit provide little basis for the prediction
of biochemically mediated ecosystem functions such as decomposition and mineralization.
In the few experiments that define litter mixtures by initial nutrient concentrations
(Hoorens et al., 2003), the problems persist. The selection of the grouping characteristic
assumes that the characteristic is a driving factor of the ecosystem function under investigation.
Furthermore, the criteria used to assign mixtures into categories based on the chosen nutrient
(e.g. high, medium and low concentrations or low-contrast versus high-contrast mixtures) are
arbitrary. While bearing more relevant information than taxonomic categorizations, simple
classification schemes are nonetheless problematic because they are limited to characterization
via a single nutrient. The articulation of a clear relationship between litter diversity and nutrient
cycling processes—if it exists—requires a continuous variable representative of compositional
heterogeneity that is founded on multiple attributes of litter quality.
With the future goal of linking plant biodiversity with ecosystem processes related to
nutrient cycling, I explicitly define the functional diversity of a litter (or foliar) mixture as the
degree of compositional heterogeneity between its constituents. In this paper, I present an index
of chemical diversity (CD) to quantify the compositional heterogeneity of litter mixtures as
constructed from three main components: 1) the traits used to describe the composition of the
18
members of the set, 2) a coefficient of dissimilarity based on these traits, and 3) a summary
measure of the trait dissimilarity between members.
The use of foliar chemical and physical attributes as functional traits is an appropriate and
rational choice when the function to be described is decomposability. Yet, the selection of traits
is not trivial. Within a given climate regime, univariate characteristics such as total nitrogen (N),
phosphorus (P), or lignin:N ratios have been correlated to the decomposition of single-species
litters (Melillo et al., 1982; Nwoke et al., 2004; Palm & Sanchez ,1991, respectively). However,
no single litter quality variable can predict even monospecific decomposition behavior under all
conditions (Smith et al., 1998). Similarly, the traits that give rise to litter interactions may vary
according to environmental conditions, making a priori trait selection to compute a relevant
chemical diversity index a gamble. Further complicating the issue, the time and cost of the
discrete analyses commonly employed in plant tissue characterization discourage the undertaking
of exhaustive chemical inventories for a large number of species.
As a solution, I explore diffuse reflectance infrared Fourier transform spectroscopy
(DRIFTS), as a rapid and cost-effective alternative to a battery of wet chemistry assays. Infrared
spectra contain qualitative and quantitative information of inorganic and organic compounds and
improved calibration techniques enable its increasingly effective use to determine nutrient
contents and carbon profiles of feeds, manures and soils (Reeves, 1998), and even to predict
decomposition patterns of single species and organic residues (Shepherd et al., 2005).
Furthermore, DRIFTS is a non-destructive technique that can preserve potentially key
information that may be lost in traditional digests. Coupled with multivariate techniques, such as
principal components analysis and advanced linear or non-linear modeling approaches, DRIFTS
19
may facilitate both the post hoc weighting of suites of traits that contribute most to chemical
diversity, and the identification of those that exert more control over litter interactions.
The choice is similarly challenging in deriving a single value to express the diversity of the
selected traits. Recently proposed continuous measures of functional diversity able to
accommodate multiple traits are functional attribute diversity, FAD2, (Walker et al., 1999);
average functional attribute diversity (Heemsbergen et al., 2004); FD (Petchey & Gaston, 2002)
and; quadratic entropy (Rao, 1982; Botta-Dukát, 2005). The advantages of quadratic entropy
that recommend it as the basis of a chemical diversity index are its ability to incorporate relative
abundance and pairwise trait differences, its flexibility to integrate multiple traits, and its
insensitivity to the number of species.
Rao’s Quadratic Entropy is defined as the average dissimilarity dij between two randomly
selected species i and j from a pool of R species with replacement. The chemical diversity index,
CDQ, is interpretable, therefore, as the mean chemical heterogeneity of a mixture, and is
calculated as
(2-1) ijj
R
i
R
jiQ dCD ρρ∑∑
= =
=1 1
where ρi and ρj are the relative abundances by biomass of species i and j, respectively, and
dij, the dissimilarity coefficient representing the compositional difference between them. I adopt
a standardized Euclidean distance dij, as a dissimilarity coefficient,
( )∑=
−=n
kjkikij tt
nd
1
21 (2-2)
20
where tik and tjk are the values of chemical trait k of species i and j, respectively in
chemical trait space of dimension n. The square root, sum of squared trait differences, is divided
by the number of traits, n, such that the index is not directly affected by the number of traits
(Botta-Dukát 2005).
At once, it is necessary to consider the following contested properties of quadratic entropy
that warrant close attention to its application as a measure of functional diversity:
i) The maximum value of diversity is not necessarily reached at uniform frequency
distribution of species (evenness);
ii) The elimination of species can result in an increase in the diversity value.
The mathematical properties of quadratic entropy that delineate its behavior are well
treated in recent works by Shimatani (2001), Izsák and Szeidel (2002), Pavoine et al. (2005) and
Ricotta (2005). In summary, the metric used to calculate dissimilarity, dij, dictates the
conformity or nonconformity of quadratic entropy to accepted diversity index axioms. The ideal
metric for a well-behaved index founded on Rao’s quadratic entropy is one that is ultrametric
(see Pavoine et al., 2005). Commonly used dissimilarity measures in ecology based on rooted
trees, such as taxonomic distinctness (Clarke & Warwick, 1998), or on gene frequencies, as in
Nei’s genetic distance (Nei, 1972), are ultrametric. However, the options of dissimilarity metrics
conducive to continuous, quantitative variables such as nutrient concentrations— Euclidean
distance, Bray-Curtis coefficient (Bray & Curtis, 1957), and the Gower distance (Gower, 1971),
for example — are not. The unavoidable adoption of a dissimilarity metric that leads to
unconventional patterns of functional diversity with species richness need not discount its
validity. Chemical diversity, expressed as a weighted mean of compositional dissimilarity, lends
it great interpretive power of the description of litter mixtures per unit biomass.
21
The aim of the current study was to observe how chemical diversity changes with species
richness and how this relationship varies under differing environments or assemblages. In brief,
the study shows that the relationship between chemical diversity and species richness suggests
that detritivore resource heterogeneity, as represented by chemical diversity, varies within levels
of species richness, revealing the weakness of number of species as a predictor of litter-quality-
dependent processes. Moreover, it is shown that species identity can be confounded with
chemical diversity such that the presence of a species with unique composition can alter the
heterogeneity of a mixture. Lastly, it is also confirmed that the relationship between species
diversity and chemical trait diversity for identical species assemblages changes in response to the
environmental context. The points discussed below illustrate why predictive patterns of
decomposition behavior as a function of richness have remained elusive, and how chemical
diversity may yield new insight into the nutrient-supplying function of plant communities.
Materials and Methods
Characterization of Leaf Tissue
The examples illustrated are drawn from published and recently gathered data from
tropical and temperate forest-systems (Table 2-1). For each dataset of R species an R x n trait
matrix was constructed and standardized by subtracting the mean trait value and dividing by the
mean absolute deviation of each trait before the calculation of the dissimilarity matrix. Leaf
traits were designated as the total tissue concentrations of nitrogen (N), phosphorus (P),
potassium (K), calcium (Ca), and magnesium (Mg) expressed as percent of total dry weight.
These analyses were chosen for an initial investigation of the behavior of chemical diversity
because they are the most widely published in forest mineral nutrition studies, of which very few
provide information at a species level.
22
Two methods of chemical characterization, total nutrient analysis and diffuse reflectance
infrared Fourier transform spectroscopy, were performed on recently gathered material from the
Atlantic Forest region of southern Bahia, to assess their relative practicality. Freshly fallen leaf
litter from 35 year-old, replicated plots of eight native species were collected from the arboretum
of the Ecological Station Pau Brasil in Porto Seguro, Brazil (16º 23′ S 39º 11′ W). The site is
situated on Ultisols derived from Tertiary sediments and experiences mean annual temperature of
23ºC and mean annual precipitation of 1696 mm. Leaves were dried at 60ºC for 48 hours and
ball milled in a SPEX 8000M Mixer/Mill (SPEX SamplePrep LLC, Metuchen, NJ). Total N was
analyzed by dry combustion and total tissue concentrations of P, K, Ca and Mg were determined
by HCl digestion after combustion in a muffle furnace for 5 hours at 500º C (Miller, 1998).
Total P was determined by colorimetry using the molybdate blue method, and total cation
concentrations were measured by atomic absorption spectroscopy. For characterization via
DRIFTS, spectra were obtained from non-diluted ground samples using a Digilab FTS-7000
spectrophotometer (Varian, Inc., Walnut Creek, CA) equipped with a KBr beam-splitter, DTGS
detector, and an AutoDIFF autosampler (Pike Technologies, Madison, WI). Spectral grade KBr
was used as a background. Spectra were collected as interferograms at a resolution of 4 cm-1 in
the mid-infrared range (4000 to 400 cm-1). Sixty-four scans were averaged to form each spectral
signature and expressed in units of pseudo-absorbance (log reflectance-1).
Mathematical Treatment of Spectral Data
Before statistical analyses, raw spectral data were subjected to baseline and multiplicative
scatter correction followed by first derivative transformation to minimize variations arising from
optical effects due to particle size (Martens & Næs, 1989). Multiplicative scatter correction
compensates for additive and/or multiplicative effects in spectral data. Normalization (sample-
wise scaling) of data was not used in order to preserve information on the absolute and relative
23
amounts of chemical constituents. Each baseline-corrected spectral point was treated as a “trait”
and the Euclidean distance between the average spectra of each species was calculated as
described above. Clustering was performed by the unweighted pair-group clustering method
using arithmetic averages (UPGMA). All calculations and graphical output were performed
using the R environment for statistical computing (R Development Core Team, 2006) and code
written by the author.
Results and Discussion
The CDQ of all possible litter combinations generated from eight species native to the
Atlantic Forest characterized by total nutrient concentrations (a) and DRIFTS (b) are represented
in Fig. 2-1. Each point represents a unique combination of species drawn from the total pool, R,
at each level of species richness, S, with the number of combinations appearing at each richness
level determined by the combinatorial R!/S! (R-S)!. In all mixtures, species are present at
uniform distribution. Gray points represent mixtures containing the most chemically distinct
species, as determined by the last clustered species. To permit comparison between the two
techniques, the CDQ values were scaled to a 0-to-1 range through division by the maximum
value.
Juxtaposition of the two forms of characterization demonstrates the effect of trait selection
on CDQ. Imbiruçu (Bombax macrophyllum L.), a waxy, broadleaf species, is the most distinct
species with respect to total elemental concentrations. In contrast, using DRIFTS distinguishes
vinhático (Plathymenia foliolosa L.), a legume. DRIFTS is particularly attractive for chemical
characterization for several reasons. First, its adoption reduces the risk of neglecting critical
traits that govern litter interactions. Total nutrient concentration profiles do not contain explicit
information on carbon, whose forms, including cellulose, lipids and secondary compounds,
affect leaf decomposability. Furthermore, in a single “snapshot”, an infrared spectrum can
24
supply a time-dependent biochemical definition of substrate availability to decomposers that
takes into account the shifting importance of litter quality parameters over the course of
decomposition (Joffre et al., 2001). DRIFTS also facilitates the use of post hoc statistical
regression techniques to identify and assign weights to spectral regions (chemical traits) most
accountable for response variables, permitting an optimization process that can determine the
maximum contribution of chemical diversity to the observed response. Advantageously, this can
be accomplished without prior knowledge of the driving traits or the covariance between them.
Lastly, the inclusion of a greater number of traits, inherent in the employment of DRIFTS, also
has the effect of stabilizing computed dissimilarities between species (Ehrlich, 1964) and,
consequently, the geometry of the relationship.
One of the anticipated features of the chemical diversity versus species richness curve is
the occurrence of maximum diversity before maximum species richness is reached. This implies
that the addition of a species to a collection may reduce the overall chemical diversity of the
mixture and, conversely, that species removal may increase chemical diversity (Fig. 2-2). In its
relevance to the CDQ, the loss or gain in species is equivalent to a change in relative abundance
because functional diversity is computed across units of biomass and not species, per se.
Analogously, the attainment of maximum heterogeneity can occur at abundances outside of
evenness as a result of the dissimilarity matrix. This contentious behavior of quadratic entropy
reflects a desirable property of the CDQ in terms of exhibiting “trait dilution” of the mixture.
The occurrence of overlapping composition of foliar material means that even the inclusion of a
chemically distinct species will contribute to the homogenization of the assemblage. While not a
startling observation (the possibilities of leaf and litter chemistries are constrained by their
function as terrestrial plant tissue), the CDQ maximum suggests an upper limit to diversity effects
25
caused by compositional differences. The strength of this feature is that it enables the valid
comparison of mixtures across values of species richness and across species assemblages
because CDQ is a summary of the chemical heterogeneity per unit biomass.
The relative abundance of species combined with species characteristics determines the
overall substrate heterogeneity as viewed by decomposers (Dangles & Malmqvist, 2004). The
addition of an identical species results in a reduction of CDQ because of the increase in that
species’ relative abundance. Predictably, mixtures that approximate monocultures earn low CDQ
values. It should be noted, however, that low values of CDQ occur between similar, high-quality
species mixes (or high-quality monocultures) as it does between similar low-quality species
mixes (or low-quality monocultures). Thus, the index has the capacity to isolate the effect of
mixture heterogeneity from that of mean mixture quality on observed processes, especially at
low values of chemical diversity.
Another recognizable aspect of the CDQ-richness curve is that of Huston’s “variance
reduction effect” (Huston, 1997)—the decrease in the range of diversity values with increasing
richness as a consequence of overlapping species pools. Because the rate of this reduction is not
the same for all mixtures, it may be interpreted as a characteristic of the assemblage in question.
Figure 2-3 contrasts the envelopes of two temperate forest communities as described by total
nutrient concentrations. A slow reduction of the variation in CDQ suggests strong clustering
among the member-species such that within-group CDQ is very low but across-group
heterogeneity is high. Compact curves indicate a more gradual differentiation of species (less
clustering). Therefore, the curve is a restatement of the topology of the cluster dendrogram.
Areas worthy of further exploration include the investigation of the behavior of CDQ under
other dissimilarity metrics, as well as the comparison of CDQ with measures of molecular
26
diversity or complexity of litter mixtures. The establishment and monitoring of species
assemblages of varying CDQ in litter-mix experiments is the first call to action to test its
correlation with quality-dependent ecosystem processes.
Experimental Ramifications of Chemical Diversity
The variation of functional trait diversity with species richness is fundamental in
establishing the species diversity-ecosystem function relationship (Petchey & Gaston, 2006).
The relationship between species richness and chemical diversity consistently shows that species
diversity cannot act as a proxy for functional diversity (Naeem & Wright, 2003). Therefore,
attempts to predict ecosystem functions such as decomposition and mineralization using the
number of species are inherently doomed.
The range in chemical diversity within levels of species richness reveals the weakness of
species number to describe litter diversity in the context of detritivore-mediated processes. This
range within species richness mimics the wide error bars associated with the responses observed
in many litter-mix experiments. Because logistical constraints discourage the inclusion of all
possible combinations of species mixtures, an idiosyncratic response to increased richness is
highly probable when mixtures are chosen at random. Even studies based on additive designs do
not offer a failsafe solution because the successive addition of species is not guaranteed to
correspond with proportional increases in substrate heterogeneity. As a result, investigations of
mixtures relying on species richness are encumbered by both the chemical diversity within
richness levels and the non-uniform directionality of chemical diversity with increasing richness.
Negative, positive and idiosyncratic relationships between chemical diversity and species
richness are all possible from additive series of assemblages generated from the same pool (Fig.
2-4a). Assuming that resource heterogeneity elicits a positive (or negative), linear response of
litter decomposition, the results seen to date in litter-mix studies may be the direct consequence
27
of the choice of a richness gradient as the independent variable. The use of a monotonic
independent variable such as chemical diversity has the potential to yield a consistent
unidirectional response of decomposition with litter functional diversity.
The categorical nature of species richness also favors the conclusion that species identity,
rather than species diversity, dictates mixture behavior. As illustrated, combinations containing
the most chemically distinct species tend to exhibit greater chemical diversity, showing that
species identity helps to define mixture diversity. While unique traits may not be synonymous
with unique behavior, the fact that the addition or removal of a key species may simultaneously
result in altered mixture heterogeneity underscores the importance of appropriate experimental
design to gauge the relative influence of species diversity and species identity on nutrient release
processes.
Monitoring Potential of the Chemical Diversity Index
Foliar and litter chemical composition are routinely used as predictors of single-species
behavior under specific environmental conditions. As a summary of plant community chemical
makeup, the chemical diversity index is, by nature, the product of multiple abiotic and biotic
interactions, including but not limited to, inter-species genetic differences, inter-specific
competition, inherent soil nutrient bioavailability, climate, and, even, herbivory. Beyond the
realm of litter decomposition studies, an immediate and compelling application of the chemical
diversity index is that of the monitoring of the chemical diversity of vegetation communities in
conjunction with standard biodiversity inventories. The addition of chemical diversity in
diversity assessments permits the comparison of the functional potential of assemblages over
time and over environmental and climatic gradients, which species lists cannot provide. Figure
2-5 illustrates the patterns in chemical diversity versus richness of identical species assemblages
28
occurring over a soil chronosequence (Fig. 2-5a) and of different but proximate forest types (Fig.
5b). Figure 2-5a demonstrates that chemical traits and by extension, functional traits, are not
conserved across environments. Therefore, identical communities, as defined by taxonomical
character, are not likely to exert the same control on ecological processes in all environments
because their relative functions are defined by the very environments in which they are
encountered.
When mapped on the landscape, such information may help reveal broad scale patterns of
resource heterogeneity that may impact related ecosystem properties such as detritivore and
herbivore communities and activity (Armbrecht et al., 2004; Dehlin et al., 2006; Swan & Palmer,
2006). While mean foliar or litter nutrient concentrations may appear equal across systems, how
these nutrients are packaged and distributed among its member species has the potential to
explain trophic responses inasmuch as they are driven by resource heterogeneity. The value of
the chemical diversity index may lie in its ability to monitor the biochemical variability of plant
tissue within communities that simultaneously determines both consumer preference and
decomposition (Pastor & Cohen, 1997) (See Fig. 5b).
The mechanisms that bring about the co-existence of species may also determine how
species express themselves in the environment (Cardinale et al., 2000; Mouquet et al., 2002).
Aboveground species diversity and productivity already show firm positive relationships under
conditions of niche differentiation (Tilman, 1999). The effect of the diversity-productivity
relationship on foliar and litter chemical composition and relative species abundance may play
an important role in linking aboveground species diversity to belowground diversity and
processes. Because litter and foliar properties (species, abundance and chemistry) of plant
communities vary ontogenetically and seasonally, such aboveground-belowground linkages are
29
necessarily contextually dependent (Blomqvist et al., 2000). The chemical diversity index is a
dynamic variable that awaits broad-scale testing and application in varied systems under multiple
conditions.
The current limitation of the index is the lack of sufficient data with which to determine the
statistical significance of diversity differences between mixtures. Due to the variability of
composition within species and the propagation of error in the calculation of the index, it is
difficult to discriminate levels of chemical diversity between assemblages with confidence. This
is especially true with increased species richness as the pair-wise error in trait differences is
compounded during the computation of the index (Fig. 2-4b). However, for low species richness
it is possible to confirm significant differences between mixtures. Yet, the variation of
litter/foliar composition within species and even within individuals need not compromise the
viability of the chemical diversity index. Rather, it calls for careful and systematic sampling
techniques in order to first make credible claims to species differences and later to predict
responses that may occur as a result.
Conclusion
The poor predictability of litter species diversity on decomposition processes may be a
consequence of our current failure to appropriately measure litter functional diversity. A
measure of functional diversity based on species chemical composition, the chemical diversity
index, CDQ, is shown to vary within species richness levels, which may account, in part, for the
idiosyncratic and species-dependent responses observed in early litter-mix studies. The potential
for chemical diversity to correlate with species interactions as they occur in litter-mix studies is
tenable because it is derived from variables known to predict the behavior of individual species.
Because of its ability to distill community patterns of foliar nutrient distribution under
different environmental conditions, the chemical diversity index may also prove a valuable
30
31
monitoring tool, supplementing existing indices based on taxonomy. Whether chemical diversity
can be used to predict how plant diversity influences environmental processes on a larger scale
remains to be established.
Table 2-1. Published mineral nutrition studies reporting total nutrient concentrations used to investigate the chemical diversity (CDQ) and species richness relationship.
Region Leaf form/ context foliage (f); litter (l)
Reference
Brazilian Atlantic Forest l, arboretum, tropical lowland rain forest
(Montagnini et al., 1995)
New Hampshire* f, temperate forest (NERC, 2006) Estonia f, meadow (Niinemets & Kull, 2003) Jamaica f, tropical montane rain forest (Tanner et al., 1977) Hawai’i f, tropical montane rain forest (Vitousek et al., 1995) * Data from the “MAPBG Project” was selected where all five nutrient concentrations (N, P, K, Ca and Mg) were reported and more than three samples were listed. Nutrient concentration values were averaged within each species.
32
33
A B
Figure 2-1. Chemical diversity (CDQ) as a function of species richness of eight litter species
native to the Atlantic Forest region of southern Bahia as characterized by A) total nutrient concentrations, N, P, K, Ca and Mg, and by B) diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) under equal species distribution. Gray points are assemblages containing the most chemically distinct species as illustrated in their respective dendrograms. The most chemically distinct species shifts from imbiruçu (Bombax macrophyllum L.) to vinhático (Plathymenia foliolosa L.) employing DRIFTS. Assemblages are shown under assumption of equal species abundance (This study, 2007).
-20% -15% -10% -5% 0% 5%
Amescla
Arapaçu
Arapatí
Gindiba
Imbiruçu
Jatobá
Sapucaia
Vinhático
Spec
ies
Rem
oved
Relative Change From CDQFinal
Figure 2-2. The effect of the removal of species from a mixture on chemical diversity under
assumption of equal species abundance, and DRIFTS characterization of the assemblage in Fig. 2-1.
34
35
Figure 2-3. The relationship of foliar CDQ diversity with species richness of two different
temperate forest communities. A) Estonia, wooded meadow (Niinemets & Kull, 2003); B) New Hamphire, conifer forest (NERC, 2006).
A A B
36
Figure 2-4. The idiosyncrasy of additive studies. A) Negative, positive and idiosyncratic
relationships between chemical diversity and species richness are possible from three series of additive mixtures derived from the same species pool. B) The same series depicted with error bars (Montagnini et al., 1995).
B
CD
Q
0
0.1
0.2
0.3
0.4
0.5
0.6
0 4 8 12 16 20 24
Species Richness
NegativePositiveIdiosycratic
-0.2
0
0.2
0.4
0.6
0.8
1
0 4 8 12 16 20 24
Species Richness
CD
Q
A
37
A
B
Figure 2-5. Potential of CDQ as a monitoring tool across communities. A) Identical species assemblages across a soil chronosequence (Vitousek et al., 1995). B) Different species assemblages in different environments (Tanner, 1977). Assemblages are shown under assumptions of equal species abundance.
CHAPTER 3 THE EFFECT OF CHEMICAL IDENTITY AND CHEMICAL DIVERSITY ON THE
DECOMPOSITION OF TROPICAL LEAF MIXTURES
Introduction
Decomposition of plant litter releases nutrients from dead organic matter and makes them
potentially available again for uptake by plants and is a primary conduit of carbon into the soil
(Berg & McClaugherty, 2008). In the latter half of the 20th century, concern over accelerated
rates of loss of plant diversity and shift in species composition in vegetation communities fueled
research into the role of individual species and plant diversity on ecosystem processes including
soil carbon dynamics (Hobbie, 1992; Silver et al., 1996; Hooper & Vitousek, 1998; Díaz et al.,
2004; Rodríguez-Loinaz et al., 2008) Gleixner et al., 1995; Bunker, 2005; Sauer et al., 2007;
Gnankambang et al., 2008; Schulp et al., 2008). Although the basic factors governing litter
decomposition—litter quality, environment, and the decomposer community—have been
recognized for nearly a century (Tenney &Waksman 1929), the decay of mixtures comprising
multiple species, which typify most ecosystems, has proven more resistant to generalization
(Gartner & Cardon, 2004 ). Understanding the processes involved in the breakdown of mixed-
species litter is important in understanding the role of species diversity in natural ecosystem
functioning and in designing agronomic ecosystems that cycle nutrients and sequester carbon
effectively. The usual approach in attempting to understand mixed-species decay involves
observation of mixtures of species, defined by species or functional group richness. This study
took another approach that defined individual and mixtures of foliar material by chemical
identity and chemical diversity based on infrared spectroscopy. Subjugating taxonomic identity
for chemical composition, I sought to predict the respiration rates of leaf mixtures.
The enigma of mixed-substrate decay persists, in part, because of our incomplete
understanding of single-substrate decomposition. The principal controls of litter decomposition
38
are climate, soil decomposer community, soil fertility and litter physical and chemical properties.
Under constant environmental conditions of climate and soil mineralogy, litter biochemistry is
the determinant of litter decomposition (Swift et al., 1979; Coûteaux et al., 1995). Controlling
for physical properties such as particle size that contribute to “quality” (Rovira & Vallejo, 2002),
litter chemistry has been shown to dictate litter mass loss (Preston & Trofymow, 2000; Pérez-
Harguindeguy et al., 2001; Raich et al., 2007), nutrient mineralization rates (Nwoke et al., 2004)
and the composition and activity of the decomposer community (Salamon et al., 2006).
Chemical traits shown to predict single-litter decomposition rates include C:N, lignin:N,
polyphenolic content, and N, P, Ca, and, recently Mn concentrations (Aber et al., 1990; Palm &
Sanchez, 1991; Hättenschwiler & Vitousek, 2000; Trofymow et al., 1995; Nwoke et al., 2004;
Berg et al., 2007). However, the relative importance of chemical traits and their relationship
with litter decomposition vary with environmental context and, in particular, with site fertility
(Vanlauwe et al., 1997; Seneviratne, 1999). Hence, generalizations of the relationship between
litter chemistry and decomposition rates of single-species litter remain understandably coarse.
The phenomenon of “non-additive effects” of mixed-litter decomposition is the
observation that when two species decay together, the rate of decomposition of the resultant
mixture is not the mean of the rates as weighted by their appearance in the mixture. At times
positive and at others negative, the commonness of non-additive effects warns that patterns of
individual species are insufficient to predict the decay of litter mixtures (Gartner & Cardon,
2004; Hättenschwiler et al., 2005). Contrasting litter chemistries has been proposed to control
the interactions observed in mixed litter decomposition (Seastedt, 1986; Chapman et al., 1988)
and the roster of initial litter characteristics purported to drive mixed-litter decomposition—
lignin:N, Ca, P and N concentrations—mirror those that predict single-litter decay (Hu et al.,
39
2006; Liu et al., 2007; Amatangelo, 2008; Gnankambary et al., 2008; Pérez-Harguinedeguy et
al., 2008).
Characterization of litter mixtures by contrasting chemical qualities (high versus low C:N
or initial N concentrations), as in the studies by Smith and Bradford (2003) and Hoorens et al.
(2003), offer representations of quality heterogeneity of litter mixtures. However, the
shortcomings of this approach are the a priori selection of quality characteristics that may not be
relevant to the response variable under the conditions of the study and the arbitrary classification
of “high” and “low” nutrient levels.
A potential solution to the advance selection of chemical traits is diffuse reflectance
infrared Fourier transform spectroscopy (DRIFTS). The infrared absorption spectrum of a
sample provides a summary of its mineral and organic composition, as molecular bonds absorb
energies in proportion to their concentration (Stuart, 1997). Extensive applications of infrared
spectroscopy in the analysis of organic compounds has led to the identification of spectral
regions and their associated chemical moieties, including compounds relevant to leaf chemistry
and decomposition, such as lignin and cellulose (Crawford & Crawford, 1980; Card et al., 1988;
Reeves, 1993). Infrared spectroscopy as the basis for the characterization of leaf species and a
means for the determination of the compositional diversity of leaf mixtures may reduce the risk
of omitting litter quality parameters critical to the analysis. Furthermore, the use of infrared
spectroscopy in conjunction with the chemical diversity index, a summary of the compositional
heterogeneity of species assemblages (Epps et al., 2007), enables a continuous, quantitative
metric of the diversity of leaf mixtures.
The purpose of this study was to predict rates and non-additive effects of microbial
respiration on leaf mixtures using chemical identity and chemical diversity and to identify traits
40
or suites of traits that consistently correlated with mineralization rates. In seeking to understand
the underlying microbial interactions that give rise to litter interactions, I tested the hypothesis
that positive non-additive effects (elevated microbial respiration) were correlated with greater
microbial functional diversity as defined by community-level physiological profiling.
Materials and Methods
Leaf Material Collection and Characterization
Fresh leaves were collected from mature trees of 15 species common to the Atlantic Forest of
southern Bahia and the “Zona da Mata” of Minas Gerais in the arboretum of the Federal
University of Viçosa in the state of Minas Gerais, Brazil, (20° 46' S, 42° 52' W; 651 m mean
altitude, 21°C MAT, 1450-1800 mm MAP) between February 2007 and May 2007. Samples
were gathered from one to five individuals, typically from the lower canopy, and composited, in
order to gain sufficient material for subsequent analyses. Leaves were discarded if they were
very young, very old or showed signs of excessive herbivory or other damage. To remove dust,
insects, and foreign matter, leaves were immersed quickly in a mild detergent solution and the
surface residues were loosened by gentle rubbing. Afterward, leaves were quickly submerged in
deionized water to rinse away residues; blotted dry; loosely packed in large brown paper bags;
and dried in a convection oven at 65°C for 72 hours, then ground to 10 mesh (< 1mm). Petioles
were removed before grinding, but the rachis and petiolules of compound leaves were included.
Ground leaf material was stored in paper envelopes or small brown paper bags within sealed
plastic bags and refrigerated at 4°C until analyzed.
Total nutrient concentrations of foliar material were determined using combustion
followed by an HCl-digest (Miller et al., 1998). Total P was determined by the ascorbic acid
reduction molybdate-blue procedure and read for blue color development at 880nm in a 1-cm
cell. Total K was determined using atomic absorption spectroscopy (Perkin Elmer AA200,
41
Perkin Elmer, Inc., Watham, MA) with an oxygen-acetylene flame. Total Ca and Mg
concentrations were also determined by atomic absorption spectroscopy using a nitrous oxide-
acetylene flame after the addition of 0.25% LaO3 at 1:10 (w:v). Total carbon and nitrogen were
determined by gas combustion using a Shimadzu TOC-VCPH Analyzer (Shimadzu Scientific,
Columbia, MD). Analyses were performed in quintuplicate with the inclusion of NIST standard
reference material (SRM) 1545 (peach leaves) and SRM 1515 (apple leaves).
Foliar material was also characterized using diffuse reflectance infrared Fourier transform
spectroscopy in the near-infrared (NIR: 1000 – 2500 nm) and the mid-infrared (MIR: 2500 –
25000 nm) regions. Spectra were obtained on a DigiLab FTS-7000 FTIR instrument (Varian,
Inc., Walnut Creek, CA) equipped with a Pike AutoDiff 60-cup autosampler (PIKE
Technologies, Madison, WI), and either a KBr beam-splitter and a deuterated trigylcine sulfate
(DTGS) detector (MIR) or a quartz beam-splitter and indium antinomide (InSb) detector (NIR).
Spectra were collected as interferograms normalized against reference spectra of KBr and sulfur
in the mid- and near-IR, respectively, in order to remove the effects of the detector and the
environment (moisture, CO2). Each spectrum resulted from the average of 64 scans taken at a
resolution of 4 cm-1 and were recorded in units (1/log reflectance). Spectra were baseline
corrected using standard normal variate correction (Martens & Næs, 1989). In addition to the ten
species, 21 leaf mixtures, prepared by homogenizing the ground foliage of two species in equal
proportions of dry mass (see Incubation Study), were also scanned for a “composite” spectrum.
Calculation of Chemical Diversity
Using three separate approaches of chemical characterization of leaf material (total nutrient
concentrations, MIR and NIR spectroscopy), the chemical diversity index, CDQ, a measure of
chemical heterogeneity of a mixture (Epps et al., 2007) was calculated for 21 leaf mixtures (see
42
Incubation Study). The chemical diversity of a mixture is defined as the mean dissimilarity in
chemical traits of two randomly selected species from a mixture of R species
ijj
R
i
R
jiQ dCD ρρ∑∑
= =
=1 1
(3-1)
where ρi is the relative abundance by dry weight of species i and the dissimilarity, dij, is the
Euclidean distance between all n traits describing species i and j (3-2). For this experiment R = 2.
( )∑=
−=n
kjkikij ttd
1
2 (3-2)
Incubation Study
In February 2008, a decomposition experiment was conducted to measure the rate of carbon
mineralization of leaf mixtures. Ten tropical tree species common to the secondary forests of the
Atlantic Rain Forest region in southern Bahia (Table 3-1) were selected for this study from a
preliminary study of 15 species. In the preliminary study, 0.5 g of each foliage type was
incubated alone in 25 g of soil at 60% field capacity in the dark for 30 days at 28°C. Treatments
were replicated four times. Evolved CO2 was assayed at regular intervals using the base trap
method to gauge the relative microbial respiration rates associated with each species. Inga
affinis (= INGA), and Balfourodendron reidelianum (= PAUM) exhibited lowest and highest
mineralization rates o the 15 species and were. Cecropia sp. (EMBA), displayed a middle
response. The remaining seven species were selected in order to maximize the representation of
species among vegetation families and to span the broadest possible range of litter chemical
composition as determined by the chemical diversity index of pairs based on MIR
characterization. Only one of the species chosen was an exotic species. Artocarpus heterophylla
(= JAQU) was included because of its dominance in many secondary forest systems in southern
Bahia (Mori et al., 1983; Sambuichi et al., 2008). Based on these preliminary findings (not
43
reported), 32 treatments were generated: 10 single-species treatments, 21 litter-pairs and a
control with no leaf additions (Table 3-2). The paired treatments were formed on the basis of the
three “key” species, each one of which was paired with seven “companion” species.
Ground leaves were mixed with acid-washed and rinsed sand (100 mesh) in sterilized,
sealed plastic containers in the ratio of 0.500 (± 0.001) g litter to 25.000 (± 0.005) g sand,
equivalent to 2.0% OM by dry weight. Quartz sand was used in order to minimize the effects of
sorption of nutrients (especially phosphorus). Leaf mixtures consisted of equal dry mass of each
species. To each container, 2.0 ml of inoculant was added (8% gravimetric water content or
approximately 60% field capacity). The inoculant was prepared from a suspension made by
incubating 10 g of topsoil with 1 L of sterilized, deionized water for three days in the dark at
28°C. The suspension was vacuum-filtered through a 0.45 μm filter, and the resultant filtrate
was diluted 1:10 with sterilized deionized water. A control treatment (sand plus inoculant, no
foliar material), and a blank treatment (no sand, tissue or inoculant) were included. All
treatments were replicated four times. A completely randomized design was used for the
placement of microcosms in the incubator. Microcosms were incubated in the dark at 28°C, in a
forced-air convection incubator (Isotemp, Thermo Fisher Scientific, Waltham, MA).
Microbially respired CO2 was measured by the alkaline trap method at 11 intervals over 80 days
(2, 4, 10, 14, 22, 30, 40, 50, 60, 70, and 80 days). A scintillation vial containing 20.0 ml of 0.5
M NaOH was placed in each container. At each time interval, the exposed base traps were
removed and capped and the vials placed in sealed containers until titration. Cumulative evolved
CO2 was expressed per gram of added carbon, calculated as the total dry weight of added
substrate multiplied by the loss on ignition.
44
Specific carbon mineralization rates were obtained by fitting the model (3-3) to the
change in cumulative CO2 evolved over time, t, measured in days.
tatCOCum ln)(2 = (3-3)
The time-dependent rate, k, is equal to the rate constant a divided by the time t. Because the
quantity, ln t, is a constant at a specified time interval and the same for all treatments, the value
of the rate constant, a, was used interchangeably with the carbon mineralization rate.
Microbial Community-Level Physiological Profiling
Potential catabolic activity of the microbial community associated with each of the leaf
treatments investigated in the above incubation study was assessed using Biolog EcoPlatesTM.
Each plate contains 96 wells consisting of three replicates of 31 single carbon substrates and a
water blank. The wells also contain tetrazolium dye which produces formazan, a blue
compound, in proportion to microbial catabolic activity upon the substrate. The catabolic
profiles of the incubated treatments were determined as follows:
At the end of the 80-day incubation period, a bulked tissue/ sand sample was formed by
combining equal dry weight of material from all replicates. Under aseptic conditions, 1.0 g of
equivalent dry weight of composited litter/sand material into a 50 ml centrifuge tube to which
was added 10.0 mL of sterilized 50 μmol phosphate buffer solution adjusted to pH 6.5 (the mean
pH of foliar extracts used in the study). Each tube was subjected to two vortex/ sonication
cycles. This suspension was immediately passed through a sterilized Whatman GF/C filter to
remove large particulate matter and dissolved organic carbon and again through a Whatman 0.22
μm filter to concentrate the bacteria onto the filter. The filter was transferred into a sterilized
centrifuge tube and eluted with 15 ml of the same sterilized, pH-adjusted buffer solution. Using
this extraction approach, the introduction of carbon sources with the inoculant was greatly
45
reduced. Consequently, the absorbance values read from each well could be attributed more
faithfully to the microbial use of the intended substrate rather than the use of plant-derived,
added substrate.
Wells were inoculated with 250μl of microbial extracts. Color development in the wells
was quantified using a Victor3 Multilabel Plate Reader (Perkin Elmer, Inc., Waltham, MA) to
record absorbance at 590nm. Plates were incubated in the dark at 28°C over seven days and read
at time intervals of 0, 16, 40, 64, 88, and 160 hours. Key species, control, and blank treatments
were duplicated.
Differences in the microbial biomass supported by a food resource are a valid effect of
resource quality. Optical density of the control (water) well was used as an estimate of inoculum
density. To standardize treatment responses by microbial biomass, substrate absorbance values
were normalized of by that of the water control at each time interval (Compton et al., 2004).
Individual carbon-source utilization was determined at each time interval as the mean of
replicated normalized absorbance values. Because the rate of substrate utilization is also a
critical indicator of the microbial community, the time course of color development for each
substrate was incorporated by calculating total substrate use as the trapezoidal area under the
absorbance versus time curve for each carbon source (Guckert et al., 1996).
Two measures, average well color development (AWCD) and the Gini coefficient (GINI)
were used as summaries of microbial activity that incorporated all substrates. Average well
color development for each treatment was calculated for all time intervals as
wii
aAWCD a931 93
1−= ∑
=
(3-4)
where a is the absorbance at 590 nm for the ith well, āw is the mean absorbance of the water blank
and 93 is the total number of wells (Garland & Mills, 1991). Microbial functional diversity was
46
quantified using the Gini coefficient, a measure of the unequal use of carbon sources. Following
Harch et al. (1997), the coefficient was calculated as
∑∑= =
−=N
i
N
jjia
aNGINI
1 12 a
21 (3-5)
where ai and aj are the mean absorbance values of wells i and j, respectively, and ā is the
AWCD. A Gini value of zero represents and even use of substrates or perfect equality, whereas
a Gini value of 1 is perfect inequality (only one substrate consumed). Greater inequality in
substrate use is synonymous with decreased functional diversity or preferential resource
utilization. Thus, a smaller Gini coefficient indicates greater functional diversity.
Statistics
Leaf chemical traits were either discrete analyte concentrations of the selected nutrients or,
in the case of spectral characterization, absorbance values at individual wavelengths. Spectral
functional traits, or simply “spectral traits”, refer to absorbance at a particular energy level
designated by wavenumber units or cm-1. Spectral traits are henceforth abbreviated in the text as
1/λ.
The effects of species identity and mixture composition on initial nutrient parameters;
cumulative evolved CO2; C mineralization rates; inoculum density; AWCD; and single-carbon
substrate utilization were tested with one way ANOVA followed by pair-wise comparisons of
means using Tukey’s HSD or, where appropriate, Tukey’s HSD test for unequal n. Correlations
between leaf chemical traits and decomposition responses (cumulative evolved CO2, C
mineralization constant, net CO2 interaction and mineralization rate interaction) were evaluated
using Kendall’s Tau.
Interactions, or non-additive effects, of the two response variables net evolved CO2 and C
mineralization rate, were calculated as the relative difference between the expected and the
47
observed responses (Wardle et al., 1997). To estimate the variance of expected values, the
average response of all possible combinations of the four replicates of both mixture constituents
was calculated. Two-tailed t-tests were used to evaluate the difference between the expected and
observed values from zero. Interactions occurring in substrate utilization profiles were
calculated and tested in the same manner.
A generalized linear model was used to determine the significance of chemical
composition and chemical diversity on the decomposition of single and mixed treatments.
Continuous response variables were final net evolved CO2 after 80 days (net CO2), the carbon
mineralization rate coefficient, a, and their respective non-additive effects, I net CO2 and Ia.
Mixture composition was defined either categorically by the identities of the key and companion
species, or by continuous variables vis-à-vis one of the three chemical characterization
approaches. All possible one- and two-parameter combinations of chemical identity and
chemical diversity were generated and incorporated into linear regression models of the form
( )21 , traittraitfy = (3-6)
or ( )QCDtraitfy ,1= . (3-7)
When chemical identity and chemical diversity were included in the same model, both
parameters employed the same characterization method (i.e. only the NIR-derived diversity
index was included in a model with NIR spectral traits). To reduce computational time,
regressions were performed on spectra that had been smoothed by averaging every fourth data
point. In this manner, NIR spectra were reduced from 3113 points to 778 points and MIR from
1868 points to 467 points. “Successful” models were defined as those in which all model
parameters were significant at p<0.05 in univariate analysis and the whole model adjusted R2
exceeded 0.65.
48
Principal components analysis was performed separately on the covariance matrix of the
spectral profiles of single-species litters in the near- and mid-IR regions. Discriminant analysis
using the PCA factor scores of the spectra of key and companion species comprising the 21 litter
pairs was performed to determine whether spectral information was adequate to classify mixture
non-additive effects as antagonistic, additive, or synergistic. Principal components analysis was
also conducted on the carbon-source utilization profiles of all 31 litter treatments in order to
observe the relationships between their respective microbial communities.
Discriminant analysis was performed on the substrate-use profiles to determine whether
potential microbial catabolic activity could differentiate the litter interactions observed in the
incubation study.
Where necessary, response variables were natural log-transformed in order to meet the
assumptions of statistical tests. Normality was determined using the Kolmogorov-Smirnov test
for normality and homoskedasticity was tested with the Brown and Forsythe homogeneity of
variance. When transformation did not succeed in meeting the assumptions, the Kruksal-Wallis
non-parametric test for the comparison of multiple means was used.
All statistical analyses were performed at significance level α=0.05 and were conducted in
R (R Foundation for Statistical Computing, 2005) and Statistica 7 (StatSoft, Inc., 2004).
Results
Species Identity on Carbon Mineralization of Mixtures
The ten tropical species selected for this study represented seven families and included
leguminous and non-leguminous species and expressed a wide range of initial nutrient
composition (Table 3-1, Fig 3-1) and carbon mineralization rates (Fig 3-5). Each mode of
characterization captured a different portrait of chemical quality and, by extension, chemical
diversity (Table 3-3, Figs. 3-2 through 3-4). Leaf mixtures with the most contrasting
49
compositions differed by characterization—Balfourodendron reidelianum and Plathymenia
fololiosa (PVIN) by total nutrient composition; Cecropia sp. and Cabralea canjerana (ECANJ)
by MIR; and B. reidelianum and Chorisia speciosa (PAPAI) by NIR (Table 3-3).
Identity of the “key” species was a strong predictor of leaf-mixture respiration rates (Table
3-6) and the C mineralization of mixtures mirrored the trend observed in the mineralization of
individual key species (PAUM>EMBA>INGA) (Fig. 3-6). Interactions between litters were
common. Of the 21 litter-pairs, 11 experienced significant interactions in C mineralization rates,
eight of them positive (Table 3-7). Non-additive effects in net evolved CO2 changed in both
direction and magnitude over time, declining along the course of decomposition (Fig. 3-7).
As with whole-mixture mineralization rates, interactions were also correlated with the
presence of key species. Mixtures containing B. reidelianum showed greater synergistic
interactions than mixtures based on Cecropia sp. or I. affinis (Fig. 3-6).
Residues of nitrogen-fixing species have been demonstrated to promote the decomposition
of litters of “poorer” quality (Fyles & Fyles, 1993; Rothe & Binkley 2001; Xiang & Bauhus,
2007). Two of the ten base species were nitrogen-fixers, I. affinis and P. fololiosa (10 pairs).
Despite the fact that the nitrogen fixers generated the least net microbial respiration and the
slowest C mineralization rates, the ten pairs containing N-fixers exhibited stronger positive non-
additive effects than those without (Fig. 3-8).
Chemical Identity on Carbon Mineralization of Mixtures
A summary of successful models of the 21 litter pairs appear in Tables 3-9 and 3-10.
Individual spectral traits from key species and companion species explained as much as 87% of
the variation in both net evolved CO2 and C mineralization rate of the 21 leaf mixtures (Table 3-
9). Models that included interaction terms between the key and companion spectral traits were
not successful according to the criteria. No models based solely on the spectral traits of one
50
mixture component were successful. By contrast, lignin:N of leaf mixtures, based on the ratio of
associated spectral peaks from 1515-1500cm-1 (lignin) and 1680-1630cm-1 (proteins) (Reeves,
1993), explained only 71% of mixture carbon mineralization (Data not shown).
Chemical diversity alone supplied no significant relationships to mineralization rate or net
respired CO2. Models incorporating chemical identity and mixture chemical diversity as
dependent variables showed clear relationships for net CO2 and C mineralization rates of all 21
leaf pairs (Table 3-10). Under characterization by total nutrient concentrations, a factorial
regression model incorporating chemical heterogeneity and the total phosphorus concentration of
the key species accounted for 88% of the variability in net CO2 evolved from leaf mixtures.
Only the MIR region provided significant spectral traits in conjunction with chemical diversity to
explain 67% of variation in C mineralization rate of mixtures (Table 3-10).
The same model-building approach was applied to subsets of the 21 leaf mixtures, grouped
by key species (n=7). When the presence of key species was incorporated into the model, the
chemical identity of the companion species and mixture chemical diversity explained as much as
98% of the variation in all four decomposition response variables, including the strength and
magnitude of interactions (Tables 3-11 to 3-13). Figure 3-9 shows a summary of the predicted
versus observed decomposition response patterns of each pair-group in which the regression
results of all successful spectral traits–chemical diversity pairings are featured in one plot for a
given response variable, for succinctness.
Discriminant analysis on the PCA factor scores of the spectra of the key and companion
species comprising the mixture successfully classified the 21 litter pairs into three distinct groups
by non-additive effect, with a confidence interval of 95% (Table 3-14). Both spectral regions
51
provided sufficient information for discrimination by non-additive effect, whereas mineral
concentration profiles were inadequate.
Identification of Functional Chemical Traits
The search for spectral traits that yielded successful models highlighted distinct regions of
the mid- and near-infrared spectra that correlated with leaf mixture decomposition. Considerable
overlap of the spectral traits correlated with single and mixed litter decomposition occurred, but
traits unique to each litter context (single vs. mixed), and each process, were apparent and, in the
case of the dynamic variable, cumulative CO2 released, the spectral traits contributing to
successful models and the regression coefficient associated with the model changed with time
(Data not shown). Table 3-15 lists the complete inventory of significant spectral traits
contributing to the model y = f(1/λ key, 1/λ comp) and their associated molecular functional groups.
As defined earlier, a spectral trait refers to an absorbance value at a particular location in the
infrared spectrum that represents the vibrational frequency of molecular bonds. However, it
should be noted that outside of pure compounds, there is no unique correlation between a
spectral location and a particular functional group due to overlapping; thus, the functional groups
listed here represent the most probable contributors to the absorbance feature.
However, when the mixtures were grouped by key species, models formed from
companion spectral traits and mixture chemical diversity consistently explained 93%-99% of the
variation in net respired CO2 and C mineralization rate (Tables 3-11 through Table 3-13). The
spectral traits contributing to successful models varied for each decomposition response, but
most importantly, for each key species. For example, using CDQ as the index of mixture
diversity, net respired CO2 was explained with spectral traits 1/λcomp =1258.5 cm-1, 1683 cm-1
and 1143 cm-1 in mixtures of INGA, EMBA and PAUM, respectively.
52
Whether using condensed spectral information as represented by latent vectors in
discriminant analysis or individual spectral traits, this study provides first evidence that non-
additive effects can be predicted using litter chemical traits.
Microbial Functional Diversity and Non-additive Effects
On a per substrate basis, fewer than half of the 31 substrates exhibited significant
correlations with net evolved CO2 at 80 days or the C mineralization rate constant. Interestingly,
significant non-additive responses in the utilization of individual substrates were also observed.
That is, for a mixture comprised of litter A and litter B, the resultant microbial community of the
mixture did not behave as the mean response of the original microbial communities, A and B,
plus the inoculum. Principal component analysis of the substrate utilization profiles of litter
pairs showed strong clustering of mixed-litter communities around that of their “key” species
(Figure 3-12). Discriminant analysis of single substrate use was unable to categorize treatments
by the non-additive effects.
Mean carbon utilization (AWCD) of microbial communities of litter treatments was not
significantly correlated with C mineralization rate and net evolved CO2 of the incubation study.
The Gini coefficient of microbial communities associated with each litter treatment was
negatively correlated with mean substrate use (Table 3-16) suggesting that greater functional
diversity of the microbial community corresponded with low carbon substrate utilization. The
Gini coefficient also showed significant, positive correlation with C mineralization rate and net
evolved CO2 of the leaf-mixtures incubation study, as well as their non-additive effects (Table 3-
16). Thus, decreasing microbial functional diversity was associated with greater mineralization
response and positive non-additive effects. In contrast to litter incubation responses, mean
carbon utilization was greater in litter pairs associated with the slowly respiring I. affinis than
with either Cecropia sp. or B. reidelianum for nearly all time intervals. Pairs containing I. affinis
53
also exhibited significantly greater substrate-use diversity than Cecropia sp. or B. riedelianum at
64 and 88 hours (Figure 3-13).
Discussion
Decomposition of mixed species is the rule in agricultural and natural environments. The
non-additive behavior of heterogeneous substrates has been documented for well over a century
(Heal et al., 1995) and although it is accepted that the simple weighted average of the
decomposition response of component species are inadequate to describe the behavior of
mixtures, the solution to the mixed-species effect is forthcoming. This experiment sought to
model patterns of microbial respiration – the outcome of substrate preference and substrate-use
efficiency – in response to mixed substrates.
Given that the richness level of the mixtures observed in this study was constant at R=2,
the effect of species richness on decomposition response could not be investigated. However,
the wide variation in microbial respiration in response to substrates that was observed within this
richness level (over 200% from the slowest and fastest rates) underscores the weakness in using
richness to differentiate mixture behavior when it is evident that composition will play a pivotal
role.
Importance of Species Identity and Chemical Identity on Carbon Mineralization
Mixed litter behavior was well determined by the decomposition patterns of their key
species (Fig. 3-6), with key species accounting for 7% to 73% of the total sum of squares (Tables
3-5 and 3-6). The experiment was designed in anticipation of this result, based on the
assumption that under additive effects the mean response of the three groups would be separated
by the difference in response generated by the three base-species. However, as stated earlier in
the description of the experimental design, the individual performance of Cecropia sp. and B.
riedelianum in this experiment did not differ significantly, contrary to the results of a trial run
54
conducted in another soil medium in which C mineralization rate and net evolved CO2 of B.
riedelianum was greater. Yet, the behavior of leaf mixtures founded on B. riedelianum differed
significantly from those based on Cecropia sp. Both the mineralization rate and net evolved CO2
of PAUM-mixtures was significantly greater than those of EMBA. Thus, while additive effects
of species may be similar due to their express decomposition trajectories, their non-additive
effects can significantly diverge.
The results corroborate those of Ball et al. (2008) which state that species identity dictates
both additive and non-additive effects. From this, it can be inferred that species associated with
non-additive effects should possess quality characteristics responsible for such effects. Thus it is
likely that the underlying cause of species identity or mixture composition on the behavior of
litter mixtures is the chemical nature of the species.
While very few published works have used a quantitative measure of the chemical
similarity of litter mixtures, a couple of studies have notably generated mixtures with the
awareness of litter quality in mind. Chapman et al., 2007, in creating mixtures of four temperate
tree species comprising conifers and broadleaf species, found the strongest synergisms to occur
between most similar pairings (conifers). Using diverse organic substrates ranging from wood
chips to feces, Dehlin et al. (2006) found no significant differences between expected and
observed values for basal microbial respiration of organic matter pairs of obvious contrasting
quality.
In this experiment, chemical heterogeneity, by itself, showed no significant correlation
with mixture decomposition response. The litter pair most similar in composition as assessed by
CDQ in all modes of chemical characterization contained the species Inga affinis and
Plathymenia foliolosa (IVIN), which, while showing significant, synergisms in both net CO2 and
55
C mineralization rate, did not display the most dramatic interactions. The next most similar
pairing was that of EPAI (neither of these leguminous species) which showed only additive
effects. Thus, the finding of Chapman et al. (2007) that positive non-additive effects are more
common among more homogenous mixtures is not generalizable. The mismatch between mass
loss as was followed in the Chapman study, and microbial respiration, as followed here may
prevent the direct comparison of results. Microbial respiration is indicative of the rate with
which the decomposing soil community utilizes the substrate. Because the amount of CO2
respired per unit of carbon incorporated is a matter of microbial use efficiency, without direct
measure of microbial growth and energy expenditure through production of exudates and
enzymes the two cannot be directly correlated. However, in studies in which both mass loss and
microbial respiration were monitored during the decomposition of mixed litters, where mass -
loss yielded non-additive results, the same was true of microbial respiration, i.e. synergies in
mass loss corresponded to synergies in microbial respiration (Briones & Ineson, 1996; Bardgett
& Shine, 1999).
In the present study, chemical diversity and chemical identity explained 67% of the
variation in net respired CO2. However, two parameters of chemical identity proved better
predictors than one spectral trait and chemical diversity (R2=0.86). This is a similar but stronger
relationship than that uncovered by to Pérez-Harguinedeguy et al. (2008) who also constructed
significant models of mixed-litter decomposition using chemical diversity in conjunction with a
litter quality trait. In that study, chemical heterogeneity was defined as the coefficient of
variation in initial nitrogen and fiber contents, lignin, cellulose, and hemicellulose, between
mixture components. The linear model incorporating nitrogen content and chemical
heterogeneity accounted for 29% of the decomposition response of 26 litter mixtures containing
56
two to five species under field conditions (litterbags, unground). Their success in constructing a
significant model using un-ground foliar material under field conditions provides further
evidence that mixture chemical diversity and chemical identity are critical functional traits that
contribute to determining mixture behavior.
While not providing successful models for mixed-litter respiration, total nutrients K, Ca
and Mg significantly correlated with net CO2, and C mineralization rate and their interactions
(Table 3-4). In their investigation of mixed temperate species Briones and Ineson (1996) found
evidence of transfer of Ca and Mg during the decomposition of eucalyptus when paired in
mixtures with oak or ash, both pairs exhibiting enhanced respiration. In this experiment, K
correlated significantly with mineralization rate interaction which suggests that elevated transfer
of K between species may correspond with increased respiration rate.
Which Traits?
Over 250 spectral traits were found to be predictors of C mineralization rates,
corresponding to polysaccharides, proteins and even water content. The generalization of the
predictive capacity of pairs of spectral traits that formed successful models in this experiment to
mixtures of other species is undetermined. While the pair of spectral traits 1644 cm-1 and 1428
cm-1 of leaf mixtures provided greater explanation of net respired CO2 than lignin:N in this
experiment, further studies are required to determine whether these two traits, or others, can be
used reliably in other contexts. Context dependence of trait significance is restricted not only to
conditions of the physical environment (Jonsson & Wardle, 2008) but also to the resource
environment. This is to say that while most of the traits are present in some quantity, whether in
single or mixed litter situations, for some spectral traits, their relative importance is enhanced by
the presence of other traits. That the sets of spectral traits relevant to net evolved CO2 and
mineralization rate were not identical, may distinguish quality characteristics linked to
57
palatability (recalcitrance) from quality characteristics associated with the bioavailability of
supporting nutrients. Recalcitrance may cause a lag in substrate use as enzymes are produced,
consequently suppressing mineralization rates, whereas nutrients availability may allow the
greater substrate utilization, which in turn can stimulate microbial growth and, ultimately, CO2
production.
The same is certainly true of spectral traits that are predictors in conjunction with the
chemical diversity index. In the case of leaf mixtures grouped by key species, the suite of
critical, mixture-defining traits of the companion species varied with each key species,
suggesting species-specific diversity effects. It is interesting to note that for a given spectral
trait, the model yielded a positive or negative coefficient. The full implications of this are not
yet clear. Yet, it can be speculated that depending on the chemical moiety represented by the
spectral trait, the mechanism underlying the decomposition response may be positively affected
by chemical heterogeneity. For example, given a spectral trait which by itself is a strong
predictor, associated with an inhibitor, (a negative coefficient), a model showing the significant
contribution of heterogeneity may mean that the effects of inhibitory compounds are mitigated
(or intensified) by increased (decreased) complexity. This might be considered a dampening or
dilution effect. For B. reidelianum, the relationships between the companion spectral trait and
mixture diversity and net CO2 were negative suggesting that any deviation of the companion
species away from the independently rapidly mineralizing, B. reidelianum was likely to result in
lowered microbial activity.
In contrast, where interactions arise from the transfer of nutrients or the use of available
carbon sources to aid in the degradation of more recalcitrant carbon forms or nutrient poor litters
(priming or nurse effect), a positive coefficient may appear. Knowledge of the physiochemical
58
quality associated with the spectral trait as well as the sign of the interaction term between the
spectral trait and diversity index as portrayed in the model may provide clues as to the nature of
the mechanism. Because more than one spectral trait can create a successful model, it can be
inferred that more than one mechanism may be underway (Hättenschwiler, 1995).
Within the scope of this study, more cannot be concluded about specific litter quality traits
or specific mechanisms triggered from the microbial interactions resulting from these qualities.
However, these results hold promise for the ability to demystify the underpinnings of mixed
substrate behavior in the immediate future.
Microbial Underpinnings of Non-Additive Effects
In this experiment, the use of ground leaves and its setting in a controlled laboratory
environment focused on the activity of the microbial community as the sole agents of litter
decomposition. “Litter” interactions are, in fact, the result of interactions occurring in the
microbial community. The response of a community to food resource quality, of which diversity
may play a part, can include changes in microbial species composition, relative abundance and
trophic interactions (Little et al., 2008).
Plant species identity can also play a significant role in the composition of the microbial
community via their influence on the quantity and quality of carbon resource (Carney & Matson,
2006). As observed in the behavior of net evolved CO2 and C mineralization rate of the 31 litter
treatments, the overall patterns of single-C-source utilization were strongly dependent on the key
species. Microbial responses to food resource diversity were non-additive. Based on the
assumption that microbial substrate activity is a direct consequence of microbial species
demographics, the response of microbial community composition to litter mixtures was also non-
additive. The initial microbial communities of a mixture comprised of species A and B were the
community residing on the foliage of species A, plus that of species B, as well as the inoculum
59
derived from the forest soil. In natural environments, this should be expected, each litter
bringing with it its associated phyllosphere community and the surface soil community
introducing its own members. In this study, microbial composition, as proxied by activity,
tended more toward that of the key species (Fig. 3-12). These two observations that plant
species dictate microbial composition and that substrate diversity affects functional diversity of
the microbial community are examples of top down effects of resource on the decomposer food
web.
In all, these and other results indicate a greater complexity to the microbial response to
substrate heterogeneity which may be partitioned into the separate but related effects of litter
colonization and resource use. Results of microbial community analysis should be interpreted
with extreme caution. Conclusions were drawn from a single Biolog plate for most treatments
and where treatments were duplicated, as in the case of key species and the blank differences in
the response of some wells. The high variation of response observed within the same plate also
advises replication of treatments at the plate-level. Also, the microbial communities drawn from
litter mixtures was extracted at the final stage of the litter incubation study and therefore may be
identical nor perform analogously to the community present at time zero. Lastly, Biolog results
are the consequence of culturable bacteria and therefore may not be representative of the
communities responsible for actual observed activities on litters.
Conclusion
The influence of forest species diversity on soil-centered ecosystem processes has taken on
increased importance because of the rate at which species composition and distribution of
vegetation communities are rapidly changing. Until now, our ability to predict these changes has
limited by the fact that the functional traits driving litter interactions have not been identified
even while these interactions are commonly observed.
60
Number of species or taxonomic measures of diversity may be valid for assessing resource
use but not resource supply. Interspecies differences and phenotypic differences emerge such
that a given species assume different litter traits and characteristics under differing
environmental conditions, including aboveground plant diversity. Moreover, should species
richness prove a good predictor in a particular setting, it may not be applicable to other groups of
species. Functional traits to describe individual species and species mixtures relevant to the
process being observed are desirable.
I hypothesized that a strictly quantitative approach of characterizing species identity
(chemical identity), as well as mixture diversity (chemical diversity), would enable the
elucidation of clear, and potentially transferable relationships between litter makeup and
decomposition behavior. The exploration of simple models constructed solely on chemical
identity – the infrared spectral fingerprint – and chemical diversity yielded robust relationships
with all four process variables. The incorporation of species identity, chemical identity and
chemical diversity of mixtures enabled the prediction the strength and magnitude of interactions.
The main questions of the study and their conclusions are:
• Can interactions and net mixture decomposition patterns be predicted by comprehensive initial leaf chemistry profiles or mixture chemical diversity?
Microbial respiration in response to leaf mixtures was successfully modeled by leaf chemistry of
component species and by the foliar chemistry of one species in conjunction with the chemical
diversity of the mixture. The use of leaf chemistries (chemical identity) of both mixture
constituents provided a model with greater explanatory power than the model constructed with
only one species and the mixture diversity. However, when the identity of the “key” species of
the pair was taken into account, C mineralization rate, net evolved CO2 and, most importantly,
the deviations of these values from the expected were also modeled with the chemical identity of
61
only one member of the mixture plus mixture chemical diversity. This is the first time that the
direction and magnitude of mixed-litter interactions have been strongly correlated with leaf
mixture chemistry.
• Which diversity measure and which traits can best model mixture trajectories?
All three analytical methods employed to chemically characterize individual leaf species and
mixture diversity (total nutrient concentrations and near- and mid-infrared spectroscopy)
provided successful models of CO2 evolution rates and interactions. However, the flexibility and
capacity for calibration of both carbon forms and mineral nutrients of mid-infrared diffuse
reflectance spectroscopy recommends it over the other modes.
• Do microbial interactions mirror litter interactions over the course of decomposition?
Single-substrate use patterns of microbial communities extracted from leaf-mixture treatments
did not provide ready interpretation of litter interactions. In general, catabolic profiles of
microbial communities derived from litter mixtures clustered around those of the single key
species as was observed in the respiration of the litter mixtures, themselves. I hypothesized that
strong positive interactions would be significantly correlated with high functional diversity.
Where correlations were significant, the reverse was true: lower catabolic diversity signifying
specialization in substrate use, correlated with greater net CO2 evolution.
The implication of this work is that the potential exists for chemical diversity to be used in
conjunction with vegetation chemistry to model carbon dynamics in real systems. More explicit
representation of litter “quality” in complex litter systems is possible. This can prove a valuable
asset in investigating tropical forest systems with high plant species diversity and chemical
composition (Hättenschwiler et al., 2008; Townsend et al., 2008) and even more possibilities
arise in landscape level analysis with the advent of new technologies to capture snapshots of
forest canopy chemical composition as in satellite-infrared analysis (Huber et al., 2008). The
62
63
representation of litter quality using infrared spectra in to model litter decomposition and N
mineralization has already been established (Gillon et al., 1999; Bruun et al., 2005; Shepherd et
al., 2005). The ability to efficiently analyze litter chemistry with DRIFTS along the course of
decomposition with minimal sample preparation and alteration recommends this approach to the
investigations of the dynamic aspects of microbial response to shifting litter composition. As
with other chemometric analyses, the calibration of non-additive effects is plausible, in which the
decomposition trajectory of a set of litter mixtures can be monitored under the required
environmental conditions and used as a “training” set for unknown, experimental mixtures.
Lastly, the interactions of biochemical constituents that drive the interactions in rates
which in turn may lead to greater understanding of the microbial impetus and general food web.
The effect of extrinsic factors the perception of litter quality should follow the approach of
Rodriguez-Loinaz et al. (2008) in which soil enzyme activity is investigated with vegetation
diversity since enzymes control rates of nutrient cycling processes.
Table 3-1. Initial foliar total nutrient concentrations of 10 tropical tree species used in the incubation study.
Scientific name Family Abbrev. P (%)
K (%)
Ca (%)
Mg (%)
Ash (%)
Alchornea sp. Euphorbeaceae DOCE 0.18f 0.98ab 1.21cd 0.04b 6.17b Artocarpus heterophylla Moraceae JAQU 0.13c 1.68c 1.41d 0.06c 12.21f Balfourodendron reidelianum Rutaceae PAUM 0.18f 3.05f 0.96bc 0.21f 10.74d Cabralea canjerana (Vell.) Mart. Meliaceae CANJ 0.11a 1.08b 2.04f 0.09c 8.56c Cassia ferruginea Caesalpinaceae CANA 0.14cd 0.97a 2.23g 0.05b 9.20c Cecropia sp. Cecropiaceae EMBA 0.16e 2.59e 1.39d 0.11d 11.32de Chorisia speciosa Bombaceae PAIN 0.14cd 1.99d 1.98e 0.11d 10.12d Inga affinis Mimosaceae IGA 0.13c 0.96a 0.78b 0.02a 5.40b Lecythis pisonis Lecythidaceae SAPU 0.14cd 0.96a 1.42d 0.19e 9.01c Plathymenia fololiosa Mimosaceae VINHA 0.12b 0.88a 0.28a 0.03a 3.17a
Different lower-case letters indicate significant differences in composition between species (p=0.05, Tukey’s HSD).
64
Table 3-2. Species composition and abbreviation codes of the 21 leaf mixtures observed in the incubation study.
Companion “Key” species species Inga affinis Cecropia sp. Balfourodenron
reidelianum
Alchornea sp. IDOC EDOC PDOC Artocarpus heterophylla IJAQ EJAQ PJAQ Cabralea canjerana ICANJ ECANJ PCANJ Cassia ferruginea ICANA ECANA PCANA Chorisia speciosa IPAI EPAI PAPAI Lecythis pisonis ISAP ESAP PSAP Plathymenia fololiosa IVIN EVIN PVIN
65
Table 3-3. Chemical diversity indices of leaf mixtures according to the mode of chemical characterization of material: total nutrient concentration (Total), mid-infrared (MIR) and near-infrared spectroscopy (NIR).
Treatment CDQ Total MIR NIR
ECANA 0.464 0.462 0.375 ECANJ 0.547 0.846 0.441 EDOC 0.472 0.539 0.697 EJAQ 0.305 0.433 0.644 EPAI 0.262 0.295 0.317 ESAP 0.414 0.675 0.335 EVIN 0.727 0.537 0.576 ICANA 0.463 0.548 0.272 ICANJ 0.472 0.526 0.463 IDOC 0.333 0.382 0.215 IJAQ 0.455 0.473 0.320 IPAI 0.521 0.570 0.738 ISAP 0.499 0.458 0.572 IVIN 0.205 0.252 0.157 PAPAI 0.512 0.538 0.858 PCANA 0.729 0.618 0.465 PCANJ 0.779 0.529 0.468 PDOC 0.634 0.524 0.347 PJAQ 0.583 0.625 0.440 PSAP 0.507 0.368 0.671 PVIN 0.864 0.439 0.412
66
Table 3-4. Kendall’s Tau correlation coefficient of total nutrient concentrations with decomposition responses of leaf-mixtures. Bold values are significant at p<0.05.
Response P (%) K (%) Ca (%) Mg (%) Ash
Net CO2 0.200 0.454 0.273 0.303 0.428 a 0.222 0.424 0.303 0.402 0.432 I Net CO2 -0.068 0.231 0.167 0.145 0.249 I a -0.054 0.262 0.235 0.176 0.335 Net CO2 = net respired CO2 at 80 days; a = C mineralization rate constant; I Net CO2 = interaction of Net CO2; Ia = interaction of C mineralization rate where interaction is calculated as the percent difference from expected value.
67
Table 3-5. Summary of repeated measures ANOVA for effects of mixture composition on net respired CO2 of 21 leaf mixtures at Day 80.
Net CO2 df SS MS F p
Intercept 1 2258969 2258969 7452.203 0.000000Key 2 86277 43139 142.311 0.000000Companion 6 16497 2749 9.070 0.000000Key x Companion 12 6900 575 1.897 0.0517Residuals 63 19097 303 Whole model R2
adj = 0.92 (F = 18.09; df = 20; p = 0.00).
Table 3-6. Summary of factorial ANOVA for effects of mixture composition on the carbon mineralization rate constant, a, of 21 leaf mixtures.
a df SS MS F p
Intercept 1 145304.0 145304.0 10695.78 0.000000Key 2 6988.8 3494.4 257.22 0.000000Companion 6 918.5 153.1 11.27 0.000000Key x Companion 12 849.3 70.8 5.21 0.000006Residuals 63 828.7 13.6
Whole model R2adj = 0.89 (F = 32.5; df = 20; p = 0.00).
68
Table 3-7. Expected and observed carbon mineralization rates and non-additive effect of paired litter treatments in incubation study. Values presented are means ± standard error, in parentheses.
Treatment apair expected apair observed
Additive = No interaction
ISAP 22.5 (3.3) 24.8 (3.6) ICANA 24.5 (2.3) 29.4 (3.8) IDOC 26.3 (2.5) 27.2 (0.7) ICANJ 30.6 (2.9) 30.3 (2.0) IJAQ 31.7 (8.6) 31.7 (0.4) EDOC 41.9 (2.2) 38.4 (6.0) ECANJ 46.2 (2.6) 42.9 (6.3) PCANJ 47.5 (3.6) 47.4 (1.7) EPAI 53.8 (2.5) 52.9 (4.8)
Antagonistic = Negative interaction
IPAI 38.2 (2.8) 34.8 (0.8) PAPAI 53.5 (6.2) 48.8 (2.8)
Synergistic = Positive interaction IVIN 19.6 (2.9) 25.2 (3.6) EVIN 35.2 (2.6) 39.9 (2.3) PVIN 36.1 (4.2) 44.1 (2.0) ESAP 38.1 (3.0) 52.4 (5.6) PSAP 38.8 (4.8) 49.7 (2.9) ECANA 40.1 (1.9) 52.1 (4.3) PCANA 41.3 (3.4) 51.1 (4.0) PDOC 42.7 (4.1) 55.2 (1.8) PJAQ 44.7 (15.6) 54.6 (3.6) EJAQ 45.9 (11.2) 54.4 (5.0) Non-additive designations were determined by comparison of the observed and expected values by two-tailed t-tests at p < 0.05.
69
Table 3-8. Summary of factorial ANOVA for effects of mixture composition on interactions of net CO2, I net CO2, and the C mineralization rate constant, I a of 21 leaf mixtures.
I net CO2 df SS MS F p
Intercept 1 23731.5 23731.52 66.90 0.000000Key 2 23572.3 11786.15 33.23 0.000000Companion 6 12100.6 2016.76 5.69 0.000017Key x Companion 12 8161.6 680.14 1.92 0.033843Residuals 209 74133.3 354.70
I a df SS MS F p
Intercept 1 2993.6 2993.6 26.82 0.000230Key 2 300.7 150.4 1.35 0.296687Companion 6 3004.9 500.8 4.49 0.013059Residuals 12 1339.6 111.6
Whole model (I net CO2) R2adj = 0.31 (F = 6.2; df = 20; p = 0.000).
Whole model (I a) R2adj = 0.52 (F = 3.7; df = 8; p = 0.021).
70
Table 3-9. Summary of selected successful models† relating chemical identity (1/λ) to net evolved CO2 and C mineralization rate of 21 leaf mixtures.
Estimate Std. error t-value p Whole model
Net CO2
Intercept -410.3 68.9 -5.95 1.2 e-05 F = 27.1 1/λ key (1644 cm-1) 292.1 28.5 10.26 6.0 e-09 df = 2,18 1/λ comp (1428 cm-1) 157.1 52.7 2.98 0.008 p =1.6 e-08 R2 = 0.86 R2
adj = 0.85
a
Intercept -960.69 92.99 -10.33 5.4 e-09 F = 60.2 1/λ key (2315 cm-1) -886.60 84.17 -10.53 4.0 e-09 df = 2,18 1/λ comp (611 cm-1) 28.61 9.31 3.07 0.0066 p =1.1 e-08 R2 = 0.87 Radj
2 = 0.86 † Model: response = trait1 + trait 2. Table 3-10. Summary of successful models† relating chemical identity (1/λ) and chemical
diversity (CDQ) on C mineralization rate of 21 leaf mixtures. Estimate Std. error t-value p Whole model
apair Intercept -170.1 62.6 -2.72 0.0146 F = 11.3 1/λ key (2585 cm-1) -338.3 99.1 3.42 0.0033 df = 3,17 CDQ (MIR) 270.0 126.9 2.13 0.0482 p = 0.0026 1/λ key * CDQ 421.9 198.3 2.13 0.0483 R2 = 0.67 Radj
2 = 0.61 † Model: response = trait1 * CDQ.
71
Table 3-11. Summary of successful models† of the relationship of chemical identity and chemical diversity on the decomposition responses of INGA leaf mixtures (n=7).
Estimate Std. error t-value p Whole model
Net CO2
Intercept -2126 262 -8.10 0.0039 F = 105 1/λ key (1258.5 cm-1) 1865 223 8.36 0.0036 df = 3, 3 CDQ (MIR) 4675 553 8.46 0.0035 p = 0.0016 1/λ key * CDQ -3880 470 -8.26 0.0037 R2 = 0.99 R2
adj = 0.98
apair Intercept -355.5 65.2 -5.46 0.012 F = 31.9 1/λ key (1174 cm-1) 312.0 54.5 5.73 0.011 df = 3, 3 CDQ (MIR) 769.1 136.2 5.65 0.011 p = 0.0089 1/λ key * CDQ -623.3 114.1 -5.46 0,012 R2 = 0.97 R2
adj = 0.94
I net CO2 Intercept 3508 438 8.00 0.0041 F = 36.3 1/λ key (1258.5 cm-1) -2948 372 -7.92 0.0042 df = 3, 3 CDQ (MIR) -7199 923 -7.80 0.0044 p = 0.0074 1/λ key * CDQ 6061 784 7.73 0.0045 R2 = 0.97 R2
adj = 0.95
I a Intercept 371.5 94.8 3.92 0.030 F = 76.8 1/λ key (2531cm-1) 444.9 138.6 3.21 0.049 df = 3, 3 CDQ (MIR) -1110.5 192.1 -5.78 0.010 p = 0.0025 1/λ key * CDQ -1391.2 277.9 -5.01 0.015 R2 = 0.99 R2
adj= 0.97 † Model: response = trait1 * CDQ.
72
Table 3-12. Summary of successful models† of the relationship of chemical identity and chemical diversity on the decomposition responses of EMBA leaf mixtures (n=7).
Estimate Std. error t-value p Whole model
Net CO2
Intercept -269.4 46.9 -5.74 0.01050 F = 116 1/λ key (1683 cm-1) 501.6 49.3 10.17 0.00203 df = 3, 3 CDQ (MIR) 1307.2 103.1 12.68 0.00106 p = 0.0013 1/λ key * CDQ -1450.4 109.8 -13.21 0.00094 R2 = 0.99 R2
adj = 0.98
I net CO2 Intercept -54.8 40.8 -1.34 0.272 F = 32.7 1/λ key (3271.75 cm-1) 194.8 49.3 3.95 0.029 df = 3, 3 CDQ (MIR) -555.3 115.2 -4.82 0.017 p = 0.0086 1/λ key * CDQ 433.7 120.2 3.61 0.037 R2 = 0.97 R2
adj =0.94 † Model: response = trait1 * CDQ. Table 3-13. Summary of successful models† of the relationship of chemical identity and
chemical diversity on the decomposition responses of PAUM leaf mixtures (n=7). Estimate Std. error t-value p Whole model
Net CO2
Intercept 1752 210 8.34 0.0036 F = 26.4 1/λ key (1143 cm-1) -1308 171 -7.66 0.0046 df = 3, 3 CDQ (MIR) -2607 371 -7.03 0.0059 p = 0.0177 1/λ key * CDQ 2192 298 7.35 0.0052 R2 = 0.96 R2
adj = 0.93
apair Intercept 520.7 54.8 9.50 0.0025 F = 36.2 1/λ key (1135 cm-1) -394.4 44.6 -8.84 0.0031 df = 3, 3 CDQ (MIR) -772.9 93.0 -8.31 0.0037 p =0.0074 1/λ key * CDQ 649.2 75.2 8.63 0.0033 R2 = 0.97 R2
ad j = 0.95 † Model: response = trait1 * CDQ.
73
74
Table 3-14. Discriminant analysis of non-additive effects observed in the C mineralization rates of 21 leaf mixtues, performed on the first four PCA factor scores of the spectra of key and companion species in the mid- (MIR) and near-infrared (NIR).
Region = MIR ADD SYN
SYN D = 7.015 F(6,13) = 3.999 p = 0.017
ANT D = 15.945 F(6,13) = 3.141 p = 0.039
D = 26.292 F(6,13) = 5.747 p = 0.006
Region = NIR ADD SYN
SYN D = 5.711 F(6,13) = 3.256 p = 0.035
ANT D = 17.920 F(6,13) = 3.529 p = 0.027
D = 26.278 F(6,13) = 5.272 p = 0.006
Distances are the squared Mahalanobis distance. ADD= additive; SYN=synergistic; ANT=antagonistic.
Table 3-15. Spectral traits of key species (MIR) contributing to successful models of litter-pair decomposition processes and their potential physiochemical interpretation.
Trait (cm-1)
Functional group
Litter quality
Trait (cm-1)
Functional group
Litter quality
Trait (cm-1)
Functional group
Litter quality
Trait (cm-1)
Functional group
Litter quality
3649.5 2462 1706 1413 δCH2, δCH3
(1450-1375) alkanes
3642 2454 1698 1405 3634 2446 1690.25
νC=O (1705-1685)
aliphatics, unsaturated
1397 3588 2438.75 1683 1390 3580 2431 1675 1382 3572.25 2423 1667 1374 3565 2415.25 1659.75 1366.5 3557 2408 1652 1359 3549 2400 1644 1351 3541.75 2392 1636.25 1297 3534 2385 1629 1289 3526 2377 1621
νC=O (1680-1630) δN-H (1650-1640) δN-H (1640-1550) δN-H, (1640-1550)
amides
1282 3518.25 2369 1613 1066 3511 2361.5 1582.5 1019.25 3503 2354 1575 1012
νC-O-C (1300-1050) νC-O (1200-1020)
esters, alcohols
3495
νC-O, νNH (3600-3200)
alcohols, phenols, amides, amines
2346 1567 1004 2546.75 2338 1521 973 2531 2331 1513 965.25 2523.5 2323 1505
(1509) lignin 958
2516 2315 1498 950 2508 2307.5 1490 765
2500 2300 1482
757 2492.75 2292
νC≡C, νC≡N, νS-H, (2600-2100)
alkynes, nitriles, thiols
1459 749.5 2485 1775 1451 742 2477 1767.75 1428 734 2469.25
νOH (3400-2500) νC≡C, νC≡N, νS-H, (2600-2100)
carboxylic acid
1760 1420.5
δCH2, δCH3 (1450-1375) alkanes
726
δCH, γCH (1000-650) γCH (900-650)
alkenes, aromatics
Values in plain text are spectral traits shared by C mineralization rate and Net CO2. Bold values are significant contributors to models of Net CO2 only. Bold, italicized values indicate spectral traits significant for C mineralization rate only. Adapted from Barbosa (2007) Table 3-1.
75
Table 3-16. Kendall’s Tau correlation between microbial functional diversity (Gini coefficient) and mean substrate use of 31 EcoPlate substrates and the decomposition responses of mixed leaf treatments. Bolded values denote significance at p < 0.05.
AWCD Rate Constant, a
Net CO2 I a I Net CO2
Gini -0.588 0.432 0.505 0.242 0.500
76
B
A
Figure 3-1. Diffuse reflectance infrared spectra of 10 litter species. A) Mid-infrared spectra; B)
Near-infrared spectra.
77
B A
Figure 3-2. Chemical diversity (normalized to the maximum value) of all possible mixtures of 10 selected tropical species as a function of species richness, as characterized by total nutrient concentrations. Each point represents a potential litter mixture. Points in blue are mixtures that contain the most chemically distinct species. Nested dendrograms show the clustering of species by their compositional similarity as characterized by the approach.
78
B A
Figure 3-3. Chemical diversity (normalized to the maximum value) of all possible mixtures of 10 selected tropical species as a function of species richness, as characterized by mid-infrared spectroscopy (MIR). Each point represents a potential litter mixture. Points in blue are mixtures that contain the most chemically distinct species. Nested dendrograms show the clustering of species by their compositional similarity as characterized by the approach.
79
80
Figure 3-4. Chemical diversity (normalized to the maximum value) of all possible mixtures of 10 selected tropical species as a function of species richness, as characterized by near-infrared spectroscopy (NIR). Each point represents a potential litter mixture. Points in blue are mixtures that contain the most chemically distinct species. Nested dendrograms show the clustering of species by their compositional similarity as characterized by the approach
B A
Figure 3-5. Carbon mineralization of ten individual tropical species and the control treatment
(no litter addition). Error shown is ± SE.
81
B
D C
A
Figure 3-6. The effect of key species on decomposition responses after 80 days. A) net evolved CO2; B) C mineralization constant, a; C) percent deviation from expected evolved net CO2 (net CO2 interaction) and; D) percent deviation from expected C mineralization rate constant (rate interaction).
82
Figure 3-7. Dynamic nature of non-additive effects on cumulative evolved CO2 between litter pairs over the course of incubation A) after 30 days; B) after 80 days. Asterisks (*) denote significant interactions at confidence limit = 0.95.
83
B A
Figure 3-8. Interaction effects of nitrogen-fixing and non nitrogen-fixing leaf mixtures. A) Net CO2 interactions; B) C mineralization rate interactions.
84
Figure 3-9. Predicted versus observed interactions occurring in C mineralization rate constant and net evolved CO2 (Day 80) of the 21 mixed-litter treatments. Predicted values are the result of linear regression analysis of the form (response ~ 1/λ comp *CD) when separated by key species. The predicted values represent all successful models for the respective decomposition response variable. A) C mineralization rate constant interaction; B) Net CO2 (Day 80) interaction.
85
C
A
B
Figure 3-10a. Chart of spectral traits, one each from key and companion species, contributing to successful models of the form response ~ 1/λ key + 1/λ comp where response is net evolved CO2 (Day 30). A) The intersection of key-species and companion-species spectral traits; B) Spectral traits of the key species and the coefficient of determination (R2) of the associated model; C) Spectral traits of the companion species and the coefficient of determination (R2) of the associated model.
86
Figure 3-10b. Chart of spectral traits, one each from key and companion species, contributing to
successful models of the form response ~ 1/λ key + 1/λ comp where response is net evolved CO2 (Day 80). A) The intersection of key-species and companion-species spectral traits; B) Spectral traits of the key species and the coefficient of determination (R2) of the associated model; C) Spectral traits of the companion species and the coefficient of determination (R2) of the associated model.
C
B
A
87
B
C
A
Figure 3-11. Chart of spectral traits, one each from key and companion species, contributing to successful models of the form response ~ 1/λ key + 1/λ comp where response is the carbon mineralization rate constant, a. A) The intersection of key-species and companion-species spectral traits; B) Spectral traits of the key species and the coefficient of determination (R2) of the associated model; C) Spectral traits of the companion species and the coefficient of determination (R2) of the associated model.
88
Figure 3-12. Principal components analysis of the single-substrate utilization profile of the 31 microbial communities litter treatments extracted from the final time step the incubation study.
89
90
Figure 3-13. Mean substrate use (AWCD) and functional diversity (GINI) of microbial
communities extracted from 21 litter-pair treatments grouped by key species, as determined by Biolog Ecoplate community-level physiological profiles analysis. A) 64 hours; B) 88 hours; C) 160 hours.
B
C
A
CHAPTER 4 TREE SPECIES DIVERSITY AND NUTRIENT CYCLING POTENTIALS OF SMALL-
SCALE CACAO PRODUCTION SYSTEMS AND SECONDARY FOREST FRAGMENTS IN SOUTHERN BAHIA, BRAZIL
Introduction
Cacao production is limited to three principal growing regions in the world – Africa, Latin
America and Southeast Asia – and occurs nearly exclusively in the world’s zones of high
floristic and faunal diversity, the so-called biodiversity hotspots (Myers et al., 2000). The area
of south Bahia, imbedded in the Atlantic Rain Forest, may be the most environmentally
significant region of cacao production in the world due to its high plant and animal endemism
and its rapidly shrinking area (Donald, 2004; Thomas et al., 1998). According to some accounts,
less than eight percent of the original cover of the Atlantic Forest remains and these persist as
isolated fragments in an agricultural and increasingly urbanized landscape (Morellato & Haddad,
2000). While forest fragmentation can adversely affect tree species and structure through
geometry, isolation and disturbance (Hill & Curran, 2003), the floristic diversity of Atlantic
Forest remnants remains high (Tabanez & Viana, 2000; Thomas, 2003). Despite the
encroachment of urban and agricultural development, 300 new species and five new genera were
reported between 1978 and 1980 (Dean, 1995). Within the heart of Brazil’s “Cacao Coast” a
“hot zone” of floristic diversity was recently recognized as supporting the second greatest tree
species density in the world (Martini et al., 2007) and discoveries of new plant species remain
unabated (Barneby & Grimes, 1994; Thomas, 1997; Amorim, 2002; de Oliveira et al., 2004;
Goldenberg & Amorim, 2006).
Traditional cacao production, locally known as cabruca, an agroforestry system
characterized by selective thinning of the forest understory prior to the establishment of cacao
seedlings (Rice & Greenberg, 2000) comprises roughly 70% of cacao production in southern
91
Bahia (Araújo et al., 1998). Since the economic blows dealt to the region by declining cocoa
prices in the world market and the introduction of Moniliophthora pernciosa, the fungal
pathogen causing witch’s broom disease, the grand plantation system of cacao that had shaped
the culture and economy of southeastern Bahia gave way. Large fazendas were abandoned,
converted to pasture or sold to the government as part of a plan of land reform (da Fonseca et al.,
2003). The result is that today, small-scale producers with average land-holdings of < 10 ha
significantly outnumber larger plantations (Landau et al., 2003). Cabruca’s role as the destroyer
or preserver of forest cover has been the focus of many discussions centered on regional forest
conservation strategies (Vinha & Silva 1982; Alger 1998; Greenberg, 1998; da Fonseca et al.,
2003). Yet, sufficient scientific evidence to support the claims of sustainability and biodiversity
of cabrucas, especially in the context of increasingly balkanized land-holding is still lacking.
At the same time, secondary forests are increasingly the most prevalent forest type around
the globe (Brown & Lugo, 1990). The FAO estimates that degraded secondary and heavily
logged forests constitute 60% of the world’s forest cover (FAO, 2005). Since the arrival of the
Portuguese to Porto Seguro in southern Bahia, the Atlantic Rain Forest of Brazil has been
reduced to 2-8% of its original cover (CIT). In 1974, primary forest accounted for 7.03% of total
land use whereas secondary forest and cabruca comprised 12.62% and 3.84%, respectively, of
land use in southern Bahia (CEPLAC, 1976). Recent geographical assessments of the district of
Ilhéus, the center of Brazil’s cacao producing region, attribute 68.3 thousand ha (40%) of its land
cover to cacao, and 26.75% to areas of conservation or preservation (Santana et al., 2003).
Excluding environmental protection zones, natural forest fragments in the district account for a
meager 5% of land cover, as estimated by optical methods (Saatchi et al., 2001). Even as the
concept of pristine native forests is being challenged, the diminishing area of “native forests”
92
necessitates the better understanding of the interactions occurring in the broad array of
ecosystems termed “secondary” forests.
Efforts to conserve endangered species and to restore vanishing forest cover in the
Brazilian Atlantic Forest highlight the role that economically significant activities, such as
traditional shade-cacao production, can play in restoring and maintaining ecosystem services
formerly supplied by native forest (da Fonseca et al., 2003). While several studies have been
recently undertaken on the structure and diversity of shade-cacao systems in southeastern Bahia
(Schulz et al., 1994; Sambuichi, 2002; Sambuichi & Haridasan 2007) few studies have focused
on small-holder cacao systems which comprise the majority of cacao production (see Machado et
al., 2005) and even fewer on the status of the secondary forests in their proximity (see Lobão,
2007). Given the dominance of secondary forests in the landscape over primary forest in the
region, the monitoring of secondary forests is essential in order to provide a reference for more
managed agroforest systems and to question the validity of the use of secondary forests as such a
reference. Studies of species diversity and nutrient cycling of both systems at the level of local
management are integral to inform conservation strategies of both secondary forests and
cabrucas to deliver ecosystem services such as soil conservation, carbon and nutrient storage and
habitat afforded by floristic biodiversity, as well as economic services generated by cacao and
non-timber forest products.
The objectives of this study were: 1) to assess differences in floristic composition between
cabrucas and secondary forest 2) to determine the overall inputs to and stocks of nutrients and to
the soil and lastly, 3) to compare the tree diversity and nutrient cycling status of forests
fragments with primary forests as depicted in the literature.
93
Materials and Methods
Site Selection and Characterization
The study sites were located within the Projeto Assentamento Frei Vantuy, an agricultural
settlement occupied by a former landless community, in the district of Ilhéus, Bahia, (14º 48′ 09″
S, 39º 07′ 54″ W) situated alongside the principal highway connecting Ilhéus and Itabuna (Fig. 4-
1). The native vegetation of the area has been categorized as lowland Southern Bahian Wet
Forest and displaying the classic rain forest structure of emergent, canopy, understory and
herbaceous layers (Thomas, 2003). A mean annual temperature of 23.3°C and rainfall between
1800-2000 mm (CEPLAC, 2006), without a distinct dry season classify the region as Af under
the Köppen climate system. The settlement resides on a patchwork soils and contain the Itabuna
and CEPEC soil units classified in the Brazilian and U.S. taxonomic systems as Alissolos
Crômicos (Typic and Orthoxic Hapludults) and Luvissolos Crômicos (Typic Hapludalfs),
respectively, and in the low lying areas Neossolos Flúvicos (Typic Tropofluvents) (Santana et
al., 2003). With the exception of the Fluvents, these soils are unusually high in inherent fertility
relative to other soil types in the district and as such are prime areas for cacao production.
The Frei Vantuy Settlement was established in 2003 by 33 families as a part of the national
program of agrarian reform. The settlement, totaling 479 ha comprises 159 non-contiguous
hectares of forest reserve, of which, approximately 80 ha is forest in various stages of
regeneration. The land is partitioned among the families. As decreed by law, each landholder
must set aside a percentage of his or her holdings to forest cover. The primary appeal of this site
was the nature of the community – small-scale, family-owned cacao systems, in contrast to
larger, plantation systems. Furthermore, the proximity of the community to the research and
extension activities of the campus of the State University of Santa Cruz (UESC) as well as that
of the Executive Commission on the Cacao Management Plan (CEPLAC) suggested that the
94
growers have been exposed to and sensitized to programs on the environmental education, the
importance of the maintenance of forest cover and diversity. As such, the secondary forests in
this study were assumed to be representative of “best-case management scenarios”.
Eight paired plots were chosen to encompass a range of soil properties (Table 4-1) and
topographies occurring in the site with the stipulation that forest and cacao plots were as close as
possible and reasonably mature (> 50 years old) (Fig. 4-2). Plot history and management were
determined from interviews with the landholders and corroborated with long-term residents and
workers from the area. Individual study plots of 50 m x 50 m were demarcated within the forest
and cabruca sites (Figs. 4-3, 4-4) to investigate tree community composition, nutrient delivery
through litterfall and soil nutrient stocks. Plots were subdivided into 25-10 m x 10 m sub-plots
to facilitate the inventory and sampling processes.
Species Inventory and Tree Diversity
The common name and occurrence of trees and palms with stems ≥ 5 cm DBH (height =
1.3m) within each site were first identified with the aid of a knowledgeable local farmer raised
and residing on the grounds of the settlement. Tree cuttings were collected and used to identify
trees to the family, genus and, where possible, to the species level, with the consultation of the
herbarium of CEPLAC. Voucher specimens were not catalogued. Trees classified as “unknown”
were used only in stem density counts and not in the enumeration of species richness. For two of
the forest fragment plots, JR and VR, the family and stem size of all individuals was recorded to
determine tree horizontal structure. The complete area (2500 m2) of VR was surveyed for
diameter class distribution but, due to time constraints, only 6 of the 25 subplots in JR (600 m2)
were measured.
95
Forest Floor
Standing organic matter was harvested in April/ May 2005 and again in November 2005.
Forest floor was sampled by a random toss method. For each plot, ten subplots were chosen
beforehand by a random number generator. A 50 cm x 50 cm metal frame was thrown within
each selected subplot and all matter within the frame was collected with the aid of a knife.
Material was dried at 60ºC for 48 hours, weighed and averaged to calculate the dry weight of
standing biomass per hectare for each plot. The removal and burning of floor material in May is
common practice in cacao production systems in preparation for cacao harvest. Forest floor
measures were taken prior to clearing in all plots except one (VIC).
Litterfall Production
Wooden collection boxes of internal area 1.0 m2 were constructed with nylon mesh
bottoms with openings of 1 mm2, resting 20 cm above the ground. Ten boxes were evenly
distributed in a regular grid formation in each 0.25 ha plot and rotated every two weeks to
maximize area coverage. Material was collected bi-weekly, dried at 60ºC for 48 hours and
weighed.
Soil Nutrient Status
Soil was sampled within each of the 25- 10 m x 1 0m subplots at each site. Ten soil
samples were taken with a soil corer from the 0-10 cm depth and bulked to create a single sample
for each subplot. Composite soil samples for each site were generated by combining aliquots of
equal weight from the 25 subplot samples after air-drying and sieving to < 2 mm.
For composite-plot samples, particle size distribution was determined by the hydrometer
method (Gee & Bader, 1986); pH was determined in H2O (1:2 soil to solution ratio) and 1.0M
KCl (1:2 ratio) using a glass electrode pH meter; and exchangeable cations by ammonium-
acetate extraction (CSIRO 1985). Gravimetric water content was determined by drying the soil
96
at 105°C for 24 hours. Total nutrient concentrations (P, K, Ca, Mg, Zn) were determined at the
sub-plot and whole plot levels using the sulfuric acid - peroxide digest method (Gasparatos &
Haidouti, 2001). Phosphorus in the extracts was determined colorimetrically by the vanadium
blue method, and the absorbance of the blue color was read at 725nm (Olsen & Sommers, 1982).
Calcium and magnesium were determined by atomic absorption (Perkin Elmer AAnalyst 200,
Perkin Elmer, Inc., Waltham MA), using an acetylene-nitrous oxide flame after addition of
0.25% w:v La2O3. Potassium concentrations were determined using atomic absorption
spectrometry with an oxygen-acetylene flame. For each study plot, a 1-kg composited soil
sample was formed from all 25 subplots. In triplicate, a 250g subsample was sieved into four size
classes (<53μm, 53-150μm, 150-250μm and 250μm-2 mm) by dry-sieving for 5 minutes using a
mechanical shaker. Each size class was weighed for size-class distribution and further analyzed
for total P and loss on ignition (LOI). LOI was determined by combustion at 400ºC for 16 hours
in order to remove organic matter. The low temperature was set to minimize the
dehydroxylation of clay minerals (Nelson & Sommers, 1996) during the combustion of organic
matter. Total C and N of whole, composited soils were determined by dry combustion on a
Shimadzu TOC-V total carbon and nitrogen analyzer equipped with an SSM-5000A solid sample
combustion unit (Shimadzu Scientific Instruments, Columbia, MD).
Statistical Analysis
Within plot species diversity as described by Margalef’s diversity (DMG), and the Shannon
information index (H′) were employed in order to assist the comparison of species diversity
across studies. While imperfect in compensating for differing sampling intensities and areas
(Magurran, 2003), both measures incorporate richness and number of individuals and provide a
better alternative to the direct comparison of richness values. Jaccard’s similarity coefficient was
used to compare the similarity of species composition within cabruca systems and secondary
97
forest plots. Differences between the eight study plots in soil and litter nutrients, and litterfall
and standing biomass were determined by one-way analysis of variance (ANOVA) followed by
Tukey’s Honestly Significant Differences test (HSD). Comparisons of measured parameters
between the two land uses were performed using the two-tailed, paired Student’s t-test. Where
necessary, data were log-transformed to conform to the assumptions of normality and
homogeneity of variance; if transformed data failed to meet the required assumptions, the
Kruksal-Wallis ANOVA and median test were used to distinguish differences between plots and
the Wilcoxon signed rank test was used to separate differences between land uses. Significant
differences in monthly litterfall rates and seasonal standing biomass within plots and within land
uses were also determined using dependent t-tests. All statistical analyses were performed at a
95% confidence interval using STATISTICA 7 (StatSoft, Inc., 2004).
Results and Discussion
Tree Density and Species Inventory
Stem density, species richness and diversity of all cabruca and forest plots are summarized
in Tables 4-2 and 4-3. In all, 733 individuals representing 41 species in 23 families were
identified within the four cacao systems observed. Cabrucas were dominated by Moraceae,
Araliaceae, Anacardaceae, Fabaceae and Urticaceae. Within the secondary forest sites, 1108
individuals were surveyed, comprising 96 species from 35 families. Excluding palms, Moraceae,
Araliaceae, Lauraceae, Lecythidaceae and Fabaceae dominated secondary forest assemblages,
with the species Arthocarpus sp., Didymopanax morototonii and Eschweilera ovata constituting
37% of the individuals encountered. Stem density in the cabruca averaged 183±13 and 44±10
individuals ≥ 5cm DBH per 0.25 ha with and without cacao, respectively, and 277±33
individuals in secondary forest. Given the same sampling efforts in both land uses, cabrucas
presented one-third the number of species found in adjacent forest fragments. These findings are
98
consistent with those of Rolim and Chiarello (2004) which showed depressed species richness in
areas of cabruca relative to lesser disturbed forest. Of the 38 native species observed within the
cabruca systems, 30 coincide with species observed in the combined plots of secondary forest, or
approximately 30% of native species encountered in secondary forest sites. This result has
implications for cabruca systems acting as a potential seed source for adjacent secondary forests.
Compared to published reports of other cacao production systems in the region (Table 4-
6), the cabrucas observed here are typical. On a per hectare basis, the tree density encountered in
these systems was over twice that as in cabrucas reported by Hummel (1995) and Sambuichi
(2002) with the same size limit. However, in a survey of cabrucas located approximately 40 km
from the study site, 35 to 133 ind. ha-1 ≥ 10 cm DBH excluding cacao were observed (Alves,
1990 as cited in Sambuichi, 2002), suggesting that the stem density of cabrucas in the present
study were sparse in comparison. With regard to tree species diversity, the cabrucas observed in
this study were relatively impoverished, as measured with and without the presence of cacao
H′avg = 2.64 ± 0.09 and H′avg = 1.17± 0.08, respectively. According to recent literature, the least
floristically diverse cabruca system in the region is an abandoned cabruca site (O1) with a
Shannon diversity value, H′= 3.31 (Sambuichi & Haridasan, 2007). Cabrucas established in
another former landless community settlement, potentially similar in parcel size as those
experienced in this study still demonstrated higher diversity values with the inclusion of cacao,
where Shannon diversity H′ = 1.86 (Machado et al., 2005). Among themselves, the four
cabrucas were dissimilar in floristic composition. High variability in tree density and species
diversity and composition are trademarks of agroforestry systems and reflect not only
environmental controls such as geography and climate, but also land-owner preference and land
management history (Schroth & Harvey, 2007).
99
The distribution of plant families in cacao systems differ markedly from that of typical
Atlantic Forest. Myrtaceae, Sapotaceae, Caesalpinaceae, Lauraceae and Chrysobalanaceae
comprise the five most important species in native south Bahian moist forest in terms of stem
density and basal area (Mori et al., 1983). In an assessment of three local cabruca systems,
Fabaceae, Moraceae and Lecythidaceae emerged as the most important families by stem density
and basal area (Lobão, 2007). Dominance by Moraceae due to Ficus spp. as observed in the sites
of this study is common to disturbed areas (Sambuichi & Haridasan, 2007). However the
reestablishment of Myrtaceae, Sapotaceae and Chrysobalanaceae and their attendant shade-
tolerant species in abandoned cabrucas suggest that cabrucas have the capacity to regenerate and
revert to a state closer to that of “primary” forest with time (Sambuichi & Haridasan, 2007).
Several measures indicate that the secondary forest sites observed in this study were under
or recovering from disturbance (Table 4-3). Total stem density (1108 ind. ha-1) was lower than
primary forest in the region but within the range of other secondary forests and fragmented
systems of lowland humid Atlantic forest. Native stands, an example of which is a site in the
Serra do Conduru State Park, held as many as 2450 ind. ha-1 (Martini et al., 2007) in the same
diameter class limit. Based on the size distribution of the two forest fragments JR and VR (Fig.
4-5), the frequencies of stems exceeding 10 cm DBH were 138 and 148 ind. 0.25 ha-1, (or 550
and 598 ind. ha-1) respectively, compared to 891.26 ha in the nearby Una Biological Reserve
(Mori et al., 1983), and 493-842 ind. ha-1 in montane tropical forests in upper Amazonia (Gentry,
1988). Other secondary forests of the Atlantic rain forest exhibit densities in the range of 800
ind. ha-1 (Elias Jr., 1998). Declines in stem density from interior forest to fragmented forest to
forest edge to secondary forest correlate with increased forest disturbance (Santos et al., Oliveira
100
et al., 2008). The relative sparseness of the observed secondary sites relative to interior primary
forests in the region alludes to a history of disturbance in the secondary sites.
In the currency of species richness, the observed secondary forest sites (44-57 species ≥ 5
cm DBH per 0.25 ha; H′avg = 2.74 ± 0.18, DMGavg = 8.45 ± 1.22) were impoverished relative to
other patches of secondary forest. In the same diameter class limit, Elias Jr. (1998) catalogued
106 individuals in 800 species (H′ =3.96); at ≥ 15 cm DBH fragments in middle and advanced
stage exhibited tree diversity indices 3.16 and 3.77, respectively. In contrast, a site of primary
forest within the Serra Grande National Park, less than 50 km away from the study sites, held
2450 stems ≥ 5cm DBH and 458 species in one hectare (Thomas et al., 2008) and in the Una
Reserve, 178 species in 600 sampled individuals were catalogued in one hectare (Mori et al.,
1983). These two native sites render Margalef diversity indices of 58.6 and 27.7, dwarfing the
most floristically diverse of the secondary forest sites observed in this study (DMG=10.26). The
crowning glory of the region is the Serra do Conduru State Park where 144 tree species ≥ 5 cm
DBH were inventoried in 0.1 hectare. (Martini et al., 2007).
Considerable overlap in the composition of forest species encountered in this work and in
the published species rosters of secondary and primary forests in the region cited above was
evident. Despite their proximity, secondary forest sites exhibited low floristic similarity (Table 4-
3) a phenomenon regularly observed in native stands. Spatial heterogeneity of vegetation species
is the outcome of multiple factors including history and geography which determine the regional
species pool, as well as the interactions between plant and animal species (Ricklefs, 1987).
Human intervention also plays a part. The presence of exotic plantation trees would be expected
in cabrucas, due to their historical economic importance in the region; however the impact of
introduced species on un-managed forested systems was striking. Exotic species appearing in all
101
secondary forest areas were dendê palm (Elaeis guineensis N. J. Jacquin), and jackfruit
(Artocarpus heterophyllus Linn.). Jackfruit accounted for the majority of all stems in three of
the four secondary forest plots, accounting for 20-26 % of the individuals counted
Jussara palm (Euterpe edulis Mart.) was the most dominant species, appearing in all four
forest sites, and comprising between 6% and 23% of stems. Didymopanax morototoni, was the
second most frequently occurring species and alongside other pioneers, Tapirira guianensis and
Pterocarpus violaceaus, are indicative of early succession stands (Santos et al., 2008). Of the
four secondary sites, JR situated in the largest forest fragment appeared to be of least disturbance
as evidenced by the greater presence of shade tolerant species such as Sloanea obtusifolia, Virola
gardneri, Mabea occidentalis (Santos et al., 2008), the high density of jussara palm, an
economically lucrative and ecologically threatened species (Matos & Bovi, 2004), as well as the
relatively low frequency of introduced species. In contrast, the site VR appeared to be the most
disturbed site with 51% of its basal area dominated by jackfruit.
Litter Production, Standing Biomass
Mean monthly litterfall produced from September 2005 to August 2006, which included
leaves, branches, flowers, fruits and seeds, was 0.81 ± 0.32 Mg ha-1 in areas of cacao production,
and 0.80 ± 0.32 Mg ha-1 in secondary forest. The peak period of litterfall in cabruca systems
corresponded to the leaf-drop cycle of cacao trees in October. Forest systems also experienced
peak litterfall from October through December, in accordance with the observations of native
forests by Mori et al., (1983). On an annual basis, the amount of litterfall was statistically
insignificant between cabruca and secondary forest, but rates of monthly deposition differed
between the land uses (Fig. 4-6). Annual litterfall rates for the two systems were 9.45 ± 0.25 Mg
ha-1 and 9.42 ± 0.25 Mg ha-1 in cabruca and forest, respectively. Equal rates of organic inputs
within the two systems are interesting in light of the fact that the stem density of cabrucas was
102
significantly lower than forests. These rates are within the range observed for lowland tropical
rain forests, annual 5.5 to 14.4 Mg ha-1 (Greenland et al., 1992) and cacao production systems in
Brazil, 9.0 and 14.0 Mg ha-1 when cacao was shaded with Erythrina or Ficus (de Oliveira Leite
& Valle, 1990).
The amount of standing litter material in areas of cabruca and secondary forest areas was
5.6 ± 0.3 Mg ha-1 and 5.9 ± 0.8 Mg ha-1, respectively for the month of May 2005 and 5.4 ± 1.1
Mg ha-1 and 4.9 ± 1.7 Mg ha-1 in November (Fig. 4-7). These two periods represent six months
and one month after the peak leaf-drop period. With the exception of one plot (JC), no
significant seasonal differences in standing biomass appeared within individual plots and no
significant differences between land-uses were found.
Nutrient Inflows through Litterfall
Litterfall is one of the principal mechanisms for the transfer of nutrients into the soil. While
nutrients are extracted from cabruca systems through seed harvest, much is returned through
litter. Litter nutrient concentrations of both cacao and secondary forest sites reflect the relatively
high nutrient status of the soils on which they are situated (Table 4-7). Total P and K
concentrations of secondary forest foliar litter were two to three times that of the mean total P
and K concentrations of native species assemblages on coastal tabuleiro soils (Montagnini et al.,
1995; Silva 1990). Cabruca litterfall was nutrient rich in comparison to secondary forest sites for
elements P, K, and Mg. However, in conjunction with litter biomass, only annual rates of P
addition through litterfall showed significant differences between the two systems. Again,
cabrucas exceeded total P inputs relative to secondary forests at a rate of 6.2± 0.3 kg ha-1 yr-1
versus 4.4 ± 0.3 kg ha-1 yr-1. Because litter nutrients were assessed only once with material
gathered over the course of a year, it was not possible to observe the probable seasonal
103
differences in inputs which may occur as a function of both litter biomass and litterfall
concentrations.
Soil Nutrient Status
Total nutrient concentrations in surface soil from the 0 to 10cm depth differed
significantly between cabruca and secondary forest systems with respect to P, Ca and Mg. Bulk
densities of paired plots were assumed equivalent and concentrations directly comparable. This
could be attributed to the addition of fertilizers to boost cacao production. The most commonly
used fertilizers are superphosphate or triple superphosphate which would augment both P and Ca
levels. Liming is also recommended at the rates 2,000 kg ha-1 yr-1 for clayey soils and 1,000 ha-1
yr-1 for sandy soils (Chepote et al., 2003). However, according to the land proprietors, no
additions of fertilizer or lime were applied after the establishment of cacao seedlings and each
plot is at least 50 years old. Calcium and magnesium which did not exhibit significant
differences in annual inputs through litter between the two land uses, nonetheless exhibited
greater accumulation in cabruca soils, suggesting a history of the use of the calcareous
amendments.
Despite the possibility of fertilization at low levels within the cabruca systems, it is also
likely that the “extra” P supplied to cabrucas derives from the litter of cacao itself, as supported
by the greater quantities entering into the soil as litter. Phosphorus sufficiency is commonly
observed in cacao agroforestry systems (Ofori-Frimpong et al., 1999). The difference in Mg
levels observed between the land uses may also arise from the surface deposition of cacao leaves.
Annan-Afful et al. (2004) found significantly greater concentrations of Mg in cacao leaves and
of exchangeable Mg in cacao-plantation topsoils than in other land uses along a toposequence
which included primary forest, fallow, mixed cropping, and traditional rice farming.
104
Significantly greater organic matter concentrations were also demonstrated cabrucas
relative to secondary forest. Despite the observation that total inputs of OM through litterfall do
not differ on an annual basis between the two systems, the concentration of soil organic matter in
the whole soil and in all soil particle size fractions of cabruca systems surpassed that of
secondary forest. Figure 4-8 illustrates the distribution of organic material among soil
aggregates of varying size. For both land uses, the greatest quantity of organic material resided
in the largest size fraction (250μm to 2.0 mm), while the highest concentrations consistently
appeared in the smaller fractions (not shown), a common trend. Higher soil Ca content may
facilitate OM sequestration and the formation of Ca-humic compounds in cabruca soils.
Proper design and management of agroforestry systems, including cabruca, can make them
effective carbon sinks (Montagnini & Nair, 2004). Based upon the rate of carbon inputs through
litterfall, standing biomass and the total quantity of carbon in the soil, cabruca as managed in this
region is as effective at C sequestration as secondary forest. In cacao systems of southern
Cameroon, soil carbon levels exceeded that of secondary forest (Duguma et al., 2001). Similar
trends were documented in West African cacao plantations that after 25 years approached the
litterfall production rates and soil carbon stores of primary forest (Annan-Afful et al., 2005; Isaac
et al., 2005).
However, as important as the quantity of carbon stored is, so is its long-term stability. Soil
organic matter is afforded physical or chemical protection from microbial degradation via three
principle mechanisms—its inherent biochemical recalcitrance, associations with surface
minerals, and occlusion within soil aggregates (Sollins et al., 1996; Six et al., 1999). Assuming
for a moment, similar litterfall chemistries between the two land uses, the greater partitioning of
SOM to smaller soil size fractions in secondary forests may mean greater stores of more stable
105
carbon. Even in forested sandy soils where the formation of aggregates is weak compared to
soils richer in clay, showed that with decreasing particle size, the greater the energy required to
access the associated organic matter and the greater protection afforded (Sarkhot et al., 2007).
More work in determining the relative stability of the carbon present in cabrucas and secondary
forests is necessary.
Conclusion
The predominance of secondary forests warrants deeper understanding of their nutrient
dynamics and species composition to ensure wise use and management for future generations
(Ewel, 1979). In the specific case of the Atlantic Forest which also supports the cacao producing
region of southern Bahia, the regeneration of primary forest may be beyond reach, but a
functioning and profitable “mosaic of cabruca and secondary forest” (Rolim & Chiarello, 2004)
is presently attainable. Realizing this goal requires the acknowledgment of and reconciliation of
potentially competing conservation goals.
The results of this study show that despite slightly lowered tree density and tree diversity
compared to secondary forest, traditional cacao production presents an agroforestry system
which closely resembles secondary forest in terms of biomass production, carbon inputs and
nutrient delivery. Because these results show that lowered floristic diversity of cabrucas can
maintain or increase stocks of soil carbon as secondary forest, it is crucial to distinguish the
importance between the potentially conflicting ecosystem services of enhancing floristic
diversity (and by extension wildlife diversity) and carbon sequestration—the two additional
commodities often promoted in environmentally-conscious cacao products.
Future research efforts should include landscape-level models on the effect of different
degrees of forest fragmentation interspersed with cabrucas and agricultural activities to establish
critical limits beyond which various ecosystem services such as wildlife habitat, carbon
106
sequestration, water are (Cassano et al., 2008). Understanding the specific thresholds to each
particular ecosystem service will enable a more informed discussion of the trade-offs between
increasing productivity and economic ambitions and biodiversity goals (Franzen & Mulder,
2007).
Threats to secondary forests and fragments of varying stages of regeneration include
selective extraction, fragmentation, and biological invasion (Phillips, 1997). Evidence of all
three activities was apparent in the fragments investigated in this study. Connectivity with
healthy forest is essential to maintain secondary forests. Proximity to native forest stands is
crucial as decay, degradation, and in reverse, recovery, is dependent on dispersal factors
(Cardoso da Silva & Tabarelli, 2000). However, recovery, as measured by forest structure and
tree species diversity can take on the order of 300 years (Liebsch et al., 2008). Addressing the
problems instigated by edge effects and forest separation is crucial to landscape regeneration and
the maintenance of forest landscape (Laurance et al., 2002; Santos et al., 2008). The severe
fragmentation observed in this study with many forest fragments < 0.5 ha may be a common
occurrence in the cacao growing region of southern Bahia, and may accelerate with the
settlement of hundreds of families in the movement of land reform. It is therefore critical to
consider the management practices or partitioning of land within land settlements as part of a
broader landscape management plan.
It is important to note that three significant reserves lie within 50 km of the study sites –
the Una Biological Reserve, Serra do Conduru State Park and the Serra Grande National Park. If
at their best cabrucas reflect the historical diversity of the region and are recipients or
benefactors of local diversity sources, the disappearance or fragmentation of these refuges will
precipitate the decline in the diversity of cabruca systems as well.
107
108
Programs that promote environmental certification in a similar vein as organic certification
have been proposed to encourage the adoption of practices such as reduced weeding in cabrucas
to allow the recruitment of native forest species (Bedê et al., 2007). While such measures could
prove beneficial, the local farming communities and environmental policy makers should first
articulate priorities of ecosystem services for solutions to forest degradation and economic
stagnation to be effective.
Table 4-1. Soil chemical properties and particle size distribution of cacao and secondary forest study sites in the Assentamento Frei Vantuy, Ilhéus, Bahia, Brazil.
Site pH Coarse sand
Fine sand
Silt Clay Base saturation
1:2 (H2O)
1:2 (KCl) % Dry weight %
Cacao JC 5.1 4.3 21.2 15.1 25.3 38.4 39.9 EVC 5.1 4.5 13.5 11.6 23.1 51.8 47.3 AC 5.3 4.9 22.2 12.9 31.5 33.4 70.2 VIC 5.1 4.5 48.7 19.2 16.9 15.2 57.1 Secondary forest JR 4.7 4.0 22.0 19.6 25.2 33.2 16.4 BR 4.9 4.6 32.4 16.9 23.0 27.7 62.6 BBR 5.0 4.8 27.9 16.3 24.4 31.4 67.3 VR 4.4 3.9 64.5 20.7 6.1 8.7 22.9 Site names were formed by the name of the proprietor (e.g. J= Josadabes; EV= João Evangelista; A= Aldienor; B= Brito; V=VI= Virgínia) and land-use type (C= Cacao/ Cabruca; R= Forest Reserve).
109
Table 4-2. Summary of the floristic diversity of four cabruca sites (0.25 ha) located in the Frei Vantuy Settlement, Ilhéus, Bahia, Brazil. Values in parentheses represent diversity parameters excluding cacao.
JC EVC AC VIC No. of individuals 172 (41) 182 (37) 201 (59) 178 (40) No. of species 18 (17) 21 (20) 24 (23) 16(15) Margalef’s Diversity Index 3.31 (4.58) 3.85 (5.54) 4.34 (5.40) 2.89 (3.80) Shannon Information Index (H′) 1.14 (2.59) 1.05 (2.92) 1.42 (2.77) 1.05 (2.29) Pielou’s Evenness Index (J′) 0.39 (0.90) 0.35 (0.96) 0.45 (0.88) 0.38 (0.85) Jaccard’s Similarity Coefficient
JC EVC
AC VIC
--
0.26 0.19 0.21
0.26
-- 0.19 0.21
0.19 0.19
-- 0.22
0.21 0.21 0.22
-- Site names were formed by the name of the proprietor (e.g. J = Josadabes; EVC = João Evangelista; A = Aldienor; VI=Virgínia) and land-use type (C = Cacao/Cabruca). Stem densities include unidentified species. Diversity parameters do not include unidentified species.
110
111
Table 4-3. Summary of the floristic diversity of four secondary forest sites (0.25 ha) located in the Frei Vantuy Settlement, Ilhéus, Brazil.
JR BR BBR VR No. of individuals 308 301 260 239 No. of species 46 44 46 57 Margalef’s Diversity Index 7.88 7.55 8.13 10.26 Shannon Information Index (H′) 2.85 2.49 2.87 2.76 Pielou’s Evenness Index (J′) 0.79 0.66 0.75 0.67 Jaccard’s Similarity Coefficient
JR BR
BBR VR
--
0.22 0.18 0.17
0.22
-- 0.38 0.20
0.18 0.38
-- 0.23
0.17 0.20 0.23
-- Site names were formed by the name of the proprietor (e.g. J = Josadabes; B =BB= Brito; V = Virgínia) and land-use type (R = Forest Reserve). Stem densities include unidentified species. Diversity parameters do not include unidentified species.
Table 4-4. Phytosociological parameters (excluding cacao) of cabrucas in southern Bahia observed in other studies.
Author Location System Area (ha) ≥ DBH (cm) †Stems Species H′
Sambuichi, 2002
Fazenda Novo Horizonte, Ilhéus-BA, 14° 37′ S, 39° 16′ W
Cabruca 2.6 5 138 41 3.35
Hummel, 1995
Fazenda Serra Grande, Ilhéus-BA, 14° 45′ S, 39° 14′ W
Cabruca 2.6 5 145 40 NA
Sambuichi & Haridasan 2007
Ilhéus-BA, 14° 41′-14° 44′ S, 39° 09′-39° 12′ W O1 O2 O3 N1 N2
Cabruca
“Old” “Old” “Old” “New” “New”
3 10
47 158 175 102 355
46 85 113 82 180
3.31 3.34 3.99 3.54 4.22
Lobão, 2007 Fazenda Pau Brasil, Ibirapitanga-BA, 14° 10′ S, 39° 24′ W
Fazenda Bom Retiro, Piraí do Norte-BA, 13° 48′ S, 39° 24′ W
Fazendo Vapor, Ubatã, BA 14° 13′S, 39° 3′ W
Cabruca 1‡ 15 152
120
180
43
36
52
3.29
3.24
3.97
NS= Not specified †Stem densities in cabruca systems do not include cacao ‡Results reported on a per hectare basis; point-centered quarter method
112
113
Table 4-5. Phytosociological parameters of forest fragments in southern Bahia, observed in other studies.
Author Location System Area (ha)
≥ DBH (cm) †Stems Species H′
Lobão, 2007 Fazenda Nova Esperança, Itapé-BA, 14° 10′ S, 39° 24′ W
Fazenda Marinêda, Jussari-BA, 15° 11′ S, 39° 30′ W
2º Forest “Medium”
“Advanced”
1‡ 15 80
249
35
72
3.16
3.77
Elias Jr., 1998 Veracruz Florestal Ltd., Eunapolis-BA, 16° 22′ S, 39° 34′ W
2º Forest
1 ‡ 5 800 106 3.96
Martini et al., 2007
Serra do Conduru State Park, Bahia, Brazil Old growth forest (OGF)
Old logged forest (OLF)
Recently logged forest (RLF)
1º Forest
0.1 4.8
254
263
253
144
137
134
NS
NS= Not specified; NA= Not available †Stem densities in cabruca systems do not include cacao ‡Results reported on a per hectare basis; point-centered quarter sampling method
Table 4-6. Tree species (≥ 5 cm DBH) identified in the selected study areas of cabruca and secondary forest in the Assentamento Frei Vantuy, Ilhéus, Bahia, Brazil.
Occurrence Family/ Scientific name Common name 2º Forest Cabruca
ANACARDACEAE Anacardium occidentale L. Cajueiro vermelho 1 Tapirira sp. Pau-pombo 14 Astronium sp. Aroeira 1 Spondias sp.1 Cajarana 2 Spondias sp. 2 Cajazeira 6 15
ANNONACEAE Duguetia magnolioidea Maas. Bacumuxá 6 2 Guatteria sp. Pindaiba preta 7 Xylopia sp. 1 Pindaiba 5 Xylopia sp. 2 Pindaiba vermelho 2
APOCYNACEAE Aspidosperma sp. Pitiá 2 Himatanthus sp. Janauba 5 Macoubea guianensis Aublet Pau de leite 1
ARALIACEAE Didymopanax morototoni Decne & Planch.
Matatauba 59 17
ARECACEAE Bactris sp. Tucum mirim 8 Elaeis guineensis N.J. Jacquin Dendê 98 Euterpe edulis (Mart.) Jussara 163 Polyandrococos caudescens (Mart.) Burí 1 Syagrus oleracea Patí 11
BOMBACEAE Eriotheca macrophyllum Imbiruçu 5 1 Hydrogaster sp. Bomba de água 1
BORAGINACEAE Cordia sp. 1 Baba de boi 1 Cordia sp. 2 Claraiba 6 2
BURSERACEAE Protium sp. Amescla 12
CARYOCARACEAE Caryocar sp. Pequi doce 1
CHRYSOBALANACEAE Couepia sp. Oití 1
114
Table 4-6. Continued. Occurrence Family/ Scientific name Common name 2º Forest Cabruca
CLUSIACEAE Garcinia sp. Olandí 2 Symphonia sp. Landirana 1 Tovomita sp. Mangue 15
COMBRETACEAE Terminalia brasilensis Araçá da água 2
ELAEOCARPACEAE Sloanea obtusifolia Gindiba 5 1
ERYTHROXILACEAE Erythroxylum sp. Peroba-rosa 10
EUPHORBIACEAE Cnidoscolus marcgravii Pohl. Penão 1 Hyeronima alchorneoides Cajueiro 2 1 Mabea sp. 1 Cambreaçu 7 Mabea sp. 2 Cambreaçu mirim 2 Tetrorchidium rubruvenium Coarana brava 3 1
FABACEAE (CAESALPINOIDEAE) Apuleia sp. Jitaí-peba 1 Arapatiella sp. Arapatí Caesalpinia ferrea leiostachya Pau ferro 1 1 Cassia multijuga Cobí 6 16 Copaifera sp. Pau óleo 1 Tachigalia multijuga Ingauçu 6 1 Zollernia sp. Mocaítaba 1
FABACEAE (MIMOSOIDEAE)
Inga affinis Ingá cipo 2 2 Inga sp. 1 Ingá 4 Inga sp. 2 Ingá vermelho 1 Pithecolobium polycephalum Monzê 1
FABACEAE (PAPILIONOIDEAE)
Andira sp. Angelim 1 Erythrina sp. Eritrina 2 Lonchocarpus sp. Cabelouro 1 Machaerium angustfolium Vog. Sete cascas 15 2 Pterocarpus sp. Pau sangue 3 2
FLACOURTIACEAE Banara sp. 1 Carpotroche brasiliensis Sapucainha 1
115
Table 4-6. Continued. Occurrence Family/ Scientific name Common name 2º Forest Cabruca
LAURACEAE Cryptocarya sp. Louro cravo 33 5 Nectandra sp. 1 Louro canela 1 Nectandra sp. 2 Louro graveto 3 Nectandra sp. 3 Louro sabão 11 7 Nectandra sp. 4 Louro verdadeiro 6 Nectandra sp. 5 Louro bosta 1
LECYTHIDACEAE Cariniana sp. Jequitibá 1 Eschweilera ovata (Cambess.) Miers Biriba 47 Lecythis lurida (Miers.) Mori Inhaiba 2 Lecythis pisonis Cambess. Sapucaia 1 1
MALVACEAE Sterculia chicha (St. Hill) Samuma 35 1 Theobroma cacao Cacaueira 6 556
MELASTOMATACEAE Miconia sp. 1 Mundururu 13 1 Miconia sp. 2 Mundururu branco 3 Miconia sp. 3 Mundururu ferro 3 3 Miconia sp. 4 Mundururu folha de lixa 5 Miconia sp. 5 Mundururu tresquinha 2 Tibouchina sp. Mundururu vermelho 2
MELIACEAE Cedrela sp. Cedro-rosa 4 Guarea sp. 1 Carrapeta 5 Guarea sp. 2 Cedro cabacinha 2 1
MONIMIACEAE Mollinedia sp. Café preto 28
MORACEAE Artocarpus sp. Jaqueira 195 30 Brosimum sp. Conduru 13 Ficus sp. Gameleira 1 Ficus insipida Willd. Gameleira-branca 2 2 Ficus gomelleira Kunthe et Bouché Gameleira-preta 5 3 Helicostylus sp. Amora 2 Sorocea sp. Amora branca 1
MYRISTICACEAE Virola gardnerii Bicuíba 1
116
Table 4-6. Continued. Occurrence Family/ Scientific name Common name 2º Forest Cabruca
MYRTACEAE Eugenia sp. 1 Batinga 4 Eugenia sp. 2 Murta 5 1 Eugenia sp. 3 Murta branca 11 5 Eugenia sp. 4 Murta preta 4 1 Psidium sp. Araçá branco 3 Syzygium aromaticum Cravinho 1
NYCTAGINACEAE Guapira sp. Farinha-seca 15
OLACACEAE Olacaceae sp. 1
POLYGONACEAE Coccoloba sp. Pê de bolo 1 2
RUBIACEAE Genipa americana Jenipapo 5 2 Rubiaceae sp. Pau cravo 2
RUTACEAE Citrus sp. Tangerina 1 Zanthoxylum sp. Paparaiba de espinho 3 8
SAPINDACEAE Cupania sp. Camboatã 9
SAPOTACEAE Manilkara sp. Parajuzinho 1 Manilkara salzmanii Maçaranduba 1 Pouteria sp. 1 Abiu da mata 5 Pouteria sp. 2 Bapeba 3 1
SIMAROUBACEAE Simarouba sp. Paparaiba 30 10
ULMACEAE Trema micrantha Curindiba 1
URTICACEAE Cecropia sp. Embauba 1 14 Pouroma guianensis Aubl. Tararanga 1
UNIDENTIFIED Unidentified sp. 1-3 6 Unidentified sp. 4-9 48 Total 1108 733
117
Table 4-7. Mean total nutrient concentrations of surface soil (0-10cm) and litter and total annual input of nutrients through litterfall in traditional cacao systems and adjacent secondary forest sites in the Assentamento Frei Vantuy, Ilhéus, Bahia, Brazil.
N P K Ca Mg
Cabruca Soil (ug g-1) 336.6*
(5.1) 102.0 (5.1)
621.8* (27.5)
521.8* (25.4)
Litter (% dry weight) 0.064* (0.002)
0.43* (0.05)
1.46 (0.09)
0.37* (0.03)
Litter (kg ha-1 yr-1) 6.24* (0.45)
42.04 (5.74)
141.33 (9.16)
35.77 (2.00)
Secondary Forest Soil (ug g-1) 248.2*
(11.2) 92.9 (4.3)
439. * (26.7)
284.6* (12.1)
Litter (% dry weight) 0.046* (0.003)
0.31* (0.03)
1.46 (0.19)
0.28* (0.05)
Litter (kg ha-1 yr-1) 4.42* (0.20)
30.02 (3.17)
141.74 (22.40)
26.23 (4.06)
Error reported in parentheses is ± SE. Values marked with an asterisk represent significant differences between land uses at P< 0.05.
118
Maps courtesy of Faria Filho & Araujo, 2003
Figure 4-1. Location of study area among land use types in the municipality of Ilhéus, Bahia, Brazil.
119
1 km
6 5
4
3 1 2
1 Josadabes’ Cabruca (JC); Josadabes’ Forest Reserve (JR) 2 Brito’s Forest Reserve (BR); Brito’s Second Forest Reserve (BBR) 3 Aldienor’s Cabruca (AC) 4 Joao Evangelista’s Cabruca (EVC) 5 Virginia’s Cabruca (VIC); Virginia’s Forest Reserve (VR) 6 Low inundated clearing
LEGEND
N
Figure 4-2. Approximate locations of observed traditional cacao and secondary forest study sites in the Assentamento Frei Vantuy, Ilhéus, Bahia, Brazil.
120
C
B A
Figure 4-3. Example of forest structure of secondary forest fragments. A) Josadabes forest reserve (JR); B) Looking up at the canopy of Brito’s forest reserve II (BBR); C) Understory of Josadabes reserve (JR).
121
C
B A
Figure 4-4. Example of forest structure of cabruca systems. A) The shade and cacao layers of Virginia’s cabruca (VIC); B) Josadabes cabruca (JC); C) Floor of Virginia’s cabruca (VIC).
122
0
25
50
75
100
125
150
175
200
0-10
10-2
0
20-3
0
30-4
0
40-5
0
50-6
0
60-7
0
70-8
0
80-9
0
90-1
00
> 10
0
DBH (cm)
No.
Indi
vidu
als
0.1 ha, 199 ind.
A
0
25
50
75
100
125
150
175
200
0-10
10-2
0
20-3
0
30-4
0
40-5
0
50-6
0
60-7
0
70-8
0
80-9
0
90-1
00
> 10
0
DBH (cm)
No.
Indi
vidu
als
0.25 ha, 289 ind.
B
Figure 4-5. Diameter class distribution of stems in two secondary forest fragments. A) Josadabes’ Reserve (JR); B) Virginia’s Reserve (VR).
123
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug
Mg/
ha
CabrucaForest
Figure 4-6. Monthly litterfall in cacao and secondary forest collected from the period September 2005 to August 2006. Error bars represent ± SE.
124
Figure 4-7. Standing biomass on the forest floor of forested plots in May and November, 2005. A) All eight forest systems. B) Cabruca versus secondary forest systems. Error bars represent ± SE.
CACAO FOREST
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
10.0
JC EVC AC VIC JR BR BBR VR
Mg/
haMayNovember
A
0.0
2.0
4.0
6.0
8.0
10.0
MAY NOVEMBER
Mg/
ha
CacaoForest
B
125
126
0.0
20.0
40.0
60.0
80.0
100.0
120.0
140.0
JC EVC AC VIC JR BR BBR VR
Plot
mg
OM
/ g s
oil
< 53um150um - 53um250um - 150um2 mm - 250um
A CACAO FOREST
0%
20%
40%
60%
80%
100%
JC EVC AC VIC JR BR BBR VR
Plot
% T
otal
OM
< 53um150um - 53um250um - 150um2 mm - 250um
B CACAO FOREST
Figure 4-8. Total organic matter content and organic matter distribution among particle size fractions of cacao and secondary forest sites. A) Total organic matter content per gram of oven dry soil; B) Organic matter distribution by soil particle size class. Error bars represent ± SE.
CHAPTER 5 OVERVIEW AND SYNTHESIS
The role of species diversity in maintaining critical ecosystem functions ranks among the
most contentious scientific debates of our time. How we respond to the rapid changes in the
redistribution of species across the planet in this age of accelerated globalization rests as much
upon our understanding of how species behave and interact within different environments as
upon our priorities between biodiversity preservation and resource management.
Today, carbon is arguably the most talked about chemical element, thanks to the
mainstreaming of discussions surrounding global warming. With popularity come fads and
scams. Throughout the developing world, “carbon sequestration”, along with “biodiversity” and
“organic” are terms pitched to small farming communities as an environmental cash cow.
Because of the high expectations of the carbon market to offset past and future CO2 production,
scrutiny of the conditions under which long-term carbon storage is possible is necessary as are
multidisciplinary efforts to quantify realistic carbon sequestration rates under well-defined
environmental and human-use conditions. Real scientific data are necessary to form the basis of
the carbon market for both producers and consumers alike.
The importance of vegetation diversity on ecosystem functions such as carbon dynamics,
including sequestration is yet undetermined. The inability to predict the long-term effect of plant
diversity or even of individual species on the terrestrial carbon cycle is a liability in a growing
carbon credit market. Because of its intersection with the global economy, the question of
diversity and ecosystem function is not strictly an academic exercise. The models used to
estimate soil carbon storage potentials should be the best available and suited to the particular
environment in which a carbon management plan is to be executed.
127
The aim of this work was to establish a firm relationship between plant litter diversity, and
the soil ecosystem function of decomposition, whose principal objectives were i) to compare the
capacity of small-scale traditional agroforestry system (low tree species diversity) and secondary
forest (high tree species diversity) to provide selected ecosystem services and ii) to predict the
decomposition behavior of litter mixtures under controlled conditions. Chapter 2 developed
analytical and conceptual tools used in the litter-mix prediction efforts described in Chapter 3. In
Chapter 4, two land-uses provided tropical forest systems of differing tree diversity in which to
contrast the direct and indirect effects of species composition and diversity on soil functions.
The setting was staged in an area dependent on dwindling forest resources for income and
where the term “biodiversity” was perhaps of greater economic interest than ecological for those
concerned. In such an environment, the conflation of science and philosophy can be detrimental
to both the producer who may not receive the crop (or monetary) yields promised or to the
consumer who may pay to support a false product. Selling biodiversity as a scientifically
grounded management practice to growers, especially vulnerable small-scale growers demands
universally validated and repeatable results.
In the Atlantic Rain Forest region of southern Bahia, a land formerly dominated by cacao
production, reforestation projects flourish in the effort to reclaim lands degraded by the
proliferation of pasture following the devastation of the cacao production due to witch’s broom
in the 1990s. In particular, the resurgence of cabruca, the traditional, method of cacao
production has become favored by institutions formerly predisposed toward high-input, full-sun,
large-scale plantation production.
This work has shown that selected functions related to nutrient delivery such as litterfall
production, carbon storage, and soil nutrient status were supplied equally well by managed,
128
production-oriented forest systems as by “natural” stands. However, the functional parity is
misleading. This is due in part to the great variation in what is designated secondary forest and
also to the selected functions observed. Hence the choice of ecosystem function will determine
the level of importance of diversity in maintaining it. If the function desired can be
accomplished by a few productive species as by multiple species, priorities between potentially
conflicting goals of conservation and service must be drawn. For this region, such results
underscore the importance of dialog between the conservation and cacao-producing communities
on the priorities of cabruca systems and the need to investigate the effect of size of land-holding
on species diversity, forest structure and soil properties.
The prediction of soil ecosystem services as a function of land cover and land use is the
cornerstone of soil organic carbon models from Century to NuCM. Yet these models ignore the
predominance of mixed substrates and use simplified estimates based on homogenous soil inputs
which has been shown by many studies to be inadequate to describe rates of carbon
mineralization and nutrient release. In tropical forest systems, the shortcomings of such models,
even if conservatively estimated at only a few percent in prediction accuracy, when integrated
over the area covered by tropical forests, could represent a miscalculation on the order of
gigatons of carbon either sequestered in the soil or released to the atmosphere.
Few studies have sought to predict litter mixture behavior, particularly the magnitude and
direction of litter interactions that occur during mixed litter decomposition. Fewer have
attempted to devise an alternative measure of litter mixture diversity based on multiple chemical
traits. The principal contribution of this work was the successful modeling of the carbon
mineralization of leaf mixtures on the basis of chemical identity and a novel construct, and
mixture chemical diversity. These two concepts supersede the former taxonomic descriptions of
129
mixtures by species identity and species richness. By adopting litter chemistry as relevant
functional traits to decomposition processes I have demonstrated the possibility to predict not
only aggregate mixture behavior but also the strength and magnitude of litter interactions—a feat
not yet noted in the published literature. While limited in scope because of its controlled
laboratory setting, the use of ground litter tissue and the limitation to two species mixtures, the
results presented here, in conjunction with continuous evolution models of decomposition
incorporating IR, provide a mechanistic footing on which to advance SOM models.
As a summary of chemical heterogeneity the chemical diversity index (CDQ) offers a
snapshot of resource heterogeneity that can also be extended to live-vegetation systems. An
unanticipated finding was uniqueness of the relationship between foliar chemical diversity and
species richness for a vegetation community. Because foliar chemistry is the outcome of
interspecies and environmental interactions, an index that summarizes chemistries of many
species can serve as a monitoring tool in conjunction with other measures. Furthermore, the
index is dynamic because even while the species composition of a vegetation community may
remain constant, its chemical profile will not.
Future Work
The importance of vegetation diversity on ecosystem functions such as wildlife
preservation and aesthetics are well established. Its relevance to soil-based functions, in
particular carbon dynamics, including sequestration is yet undecided. The inability to predict the
effect of plant diversity or even of individual species on the terrestrial carbon cycle over the long
term is a liability in a growing carbon credit market. Because of its intersection with global
economy, the question of diversity and ecosystem function is not strictly an academic exercise.
The models used to estimate soil carbon storage potentials should be the best available and suited
to the particular environment in which a carbon management plan is to be executed.
130
131
Further development of this work should include the following:
• Documentation of the spatial and temporal variability in the chemical heterogeneity of litter mixtures under field conditions
• Expansion to mixtures comprised of more than two species;
• Robustness to changes in relative abundance (unequal species distribution);
• Field-level studies which include the effect of litter particle size and the importance of litter chemistry on macro and mesofauna and;
• Investigation of the environmental mediation (mineralogy, moisture, soil nutrients, pH, CO2 enrichment) of microbial community and activity on the utilization of mixed substrates.
• Field testing of reforestation using chemical heterogeneity as a tool in defining site biodiversity, while limiting the number of species required for reforestation.
Changes in climate, such as shifts in precipitation patterns, elevated CO2 and increased
temperature, stand to influence biogeochemical cycles via their affect on the world’s vegetation.
This may entail alterations in the species composition of plant communities, enhanced above-
and belowground productivity or modifications in plant chemistry and morphology. In a CO2-
rich world, the impact of relatively swift changes in vegetation chemistry and seasonality on the
millennia of give and take between insects and plants carries implications for pest incidence and
pollination. The use of litter chemistry as the starting point of decomposition pathways enhances
our ability to foresee the changes that climate change can impart on soil ecosystem functions.
APPENDIX A USING INFRARED SPECTROSCOPY TO DETERMINE THE RELATIVE
CONTRIBUTION OF ORGANIC INPUTS TO SURFACE SOIL ORGANIC MATTER
Introduction
Soil boasts the largest reservoir of organic carbon in terrestrial ecosystems, (Post et
al. 1982, Schimel 1995) whose stores are principally subsidized by the death and
transformation of plant biomass – litterfall, root turnover and root exudates. The relative
contribution of aboveground and belowground inputs to soil carbon (quantity and quality)
is an important element in determining management strategies for the maintenance and
sequestration of soil carbon in forested systems. Past studies using carbon isotope
methods and/or litter- and root-exclusion experiments in temperate and tropical forest
systems have reported root respiration and root decomposition as greater contributors to
soil CO2 efflux than that deriving from the decomposition of aboveground litter
deposition (Bowden et al., 1993; Li et al., 2004, Sulzman et al., 2005). Root biomass has
been also shown to be the dominant source of new and stable carbon in surface soil
horizons (Balesdent & Balabane, 1996; Gale & Cambardella, 2000; Hobbie et al., 2004;
Rasse et al., 2005). The majority of studies monitoring the dynamics of soil carbon
employ field designs based on root and litter exclusion experiments followed by carbon
isotopic labeling or nuclear magnetic resonance (NMR) spectroscopy, both expensive and
time-consuming methods. Diffuse reflectance infrared spectroscopy (DRIFTS) is a
technique swiftly gaining popularity in the environmental science community because of
the facility, economy and versatility with which it can determine an array of properties of
organic and inorganic materials.
This study explored the use of DRIFTS to assess the relative contribution of leaf-
and root-derived inputs to soil carbon. Using replicated suites of foliage, litter, soil and
132
roots of mono-specific plots of 10 tropical forest species, I asked: How does the spectral
fingerprint of soil organic matter under a single species compare to those of foliage, litter
and fine roots of the same species and how does this relationship change with depth? The
inquiry was extended to natural and managed multi-species forested systems to see
whether the trends observed under neat, mono-specific conditions were maintained in
polycultural conditions.
Materials and Methods
Study Site and Sampling
Single-species stands
Samples were drawn from a 35-year-old arboretum established in the Pau Brasil
Ecological Station, in Porto Seguro, Brazil in the state of Bahia (16°23' S, 39° 11'W)
whose mean annual temperature and precipitation classify the climate as Af in the
Koppen system. Soils were classified as Oxisols (Montagnini et al., 1995) with sandy
surface textures. Live foliage, litter and soil were sampled in June 2004, from 19 plots
representing ten tropical tree species native to the Brazilian Atlantic Rain Forest of
southern Bahia (Table A-1). Plots measuring approximately 2 m x 2 m contained three to
five mature trees of one species. Understory growth was supressed due to frequent
cutting. Species were selected based on the facility to collect sufficient foliage and litter
material for subsequent analysis, with preference for those with replicate plots. Freshly
senesced litter was collected from the soil surface by hand and selected on the basis of
color and apparent lack of fungal colonization. Foliage was collected from the lower
canopy using a tree pruner. Leaves showing signs of extreme insect damage were
discarded. Soil was collected from the 0-5 cm and 5-10 cm depth with a shovel from
three random locations within the plot after the removal of the litter layer. From this soil
133
material, coarse and fine roots of undetermined viability were extracted by sieving and by
hand. Fine roots (< 2mm) were retained for later analysis.
Mixed-species stands
Material from mixed-species stands were gathered from forest fragments of the
Atlantic Rain Forest in southern Bahia, located in the heart of the cacao-producing region
between Itabuna and Ilhéus (14°39' S, 39° 10'W) within the same climate zone as the
research station. Sixteen 10 m x 10 m plots in high clay soils (54%) under secondary
forest and nine 10 m x 10 m plots on sandy soils (70%) from secondary forest and
cabruca, a traditional cacao production system were sampled (Chapter 4). Traditional
cacao production is characterized by the thinning of the forest cover and planting cacao in
the understory, thereby preserving some of the inherent tree species diversity (CIT). By
the accounts of local inhabitants, the areas under cabruca had been under production for
at least 50 years. The presence of the palm jussara, Euterpe edulis, and the absence of
common plantation trees, such as jackfruit (Artocarpus heterophylla) marked the forest
stand on clay soil as mature. The fragment of forest on sandy soils had been uncut for at
least 50 years but showed signs of disturbance on its perimeter. Mean stem density of the
secondary forest and cacao plots was 12 trees per 10 x 10 m plot, with 2-5 species per
100 m2 subplot in the instance of the cacao-production area, and 4-13 species per 100 m2
in secondary forest (Chapter 4). In a departure from the mono-specific stands, in lieu of
freshly senesced litter, forest floor was collected from a 50 cm x 50 cm frame randomly
tossed three times within each subplot and pooled to form a composite sample. Soil was
collected from the 0-10 cm depth using a shovel, again from three random locations
within each subplot. Roots were extracted from a 20 cm x 20 cm x 10 cm block carved
from the soil near the center of each subplot.
134
Sample Preparation
Soils were air dried, sieved to < 2mm and ball-milled after the removal of large bits
of organic matter. Roots, foliage and forest floor were dried at 60°C for 48 hours and
ball-milled prior to spectral analysis. Sub-samples of ground soil and tissue materials
were combusted at 400°C for 16h (Nelson & Sommers, 1996) and 500°C for five hours
(CIT), respectively, to generate “ashed” samples.
Spectral Analysis and Isotopic Carbon Analysis
Combusted and non-combusted soil and tissue samples were chemically
characterized using diffuse reflectance infrared Fourier transform spectroscopy
(DRIFTS). Spectra were scanned on a Digilab7000 Series infrared spectrometer without
KBr dilution in the mid-infrared region (4000 cm-1 to 400 cm-1) at a resolution of 4 cm-1.
A total of 64 scans, normalized against an IR-grade KBr standard, were averaged to
generate a final spectrum for each sample. Spectra were transformed using the standard
normal variate method to correct for baseline drift (Barnes et al., 1989).
Prepared samples of soil (0-5cm), litter and roots derived from the single-species
plots were analyzed for natural 13C isotopic ratios using an inductively-coupled plasma
mass-spectrometer (Elan 9000, Perkin Elmer, Waltham, MA).
Statistical Analysis
The sum of squared differences (SSD) between the soil spectra of 18 single species
plots and the spectra of their respective inputs was used as a simple graphical comparison
to gauge the similarity between soil organic matter and organic inputs. The SSD was
computed by summing the difference between the absorbance values of spectra of
organic inputs (foliage, litter and roots) from soil spectra over all wavepoints, W or
135
( 2
1iORGiSOM
W
iabsabsSSD −= ∑
=
) . (3-1)
Ash-subtraction is a mathematical technique that amplifies the signal of the organic
component by reducing the spectral contribution of the inorganic fraction (Westerhaus et
al., 2004). The SSD between soil and organic inputs was determined for ash-subtracted
and non-ash subtracted spectra. Values of the SSD (both whole and ash-subtracted)
calculated from roots (y) were plotted against those of litter and foliage (x). The
clustering of points on the y=x line indicates equal contribution of the two inputs, while a
deviation from the line indicates dominance of one of the inputs. Points clustered below
the 1:1 line exhibit a smaller SSD of the input represented along the vertical axis relative
to the input represented along the horizontal axis. The regression line best representing
the points was used to test the alternative hypothesis “HA≠1”.
Significant differences in isotopic ratios between soil and litter and soil and roots
were determined by the two-tailed Student t-test for dependent samples, performed at
95% confidence level. Analysis was conducted in R version 2.5.1 (R Development Core
Team, 2006).
Results and Discussion
Relative Importance of Above- and Belowground Inputs
In the monoculture systems, comparison of the single-species SSD relative to soil
organic matter of three inputs–foliage, litter, and roots–showed greater similarity between
the soil and root spectra than between soil and foliage, and soil and litter spectra (Fig. A-
1). As means of comparison, the relationship between SSDlitter and SSDfoliage was plotted
and found not to differ significantly from unity (Fig. A-1A). The equivalence of litter
136
and foliage relative to soil organic matter (SOM) was conserved with soil depth,
confirming the interpretability of the graphical method.
The slopes of SSDroot versus SSDfoliage and SSDroot versus SSDlitter differed
significantly from 0 and 1 within 95% confidence interval (Fig. A-1 B,C). This
relationship was exaggerated at the 5-10 cm depth; the spectra of roots were closer to
those of the soil than were foliar or litter spectra, as suggested by the decrease in slope.
However, while a trend of increasing similarity of SOM to root matter with depth was
observed, the effect was not significant.
The use of ash-subtracted spectra enhanced the organic matter signal, particularly
in soil spectra. While the slopes were greater in the analysis of ash-subtracted soil, litter
and root matter, the associated confidence intervals were much narrower and significant
difference from unity was maintained (Fig. A-1D).
Interestingly, the computed SSD of root, litter and foliage were not similar between
replicates of single-species plots. Such discrepancies could arise from site differences that
can lead to subtle changes in the quantity and quality of carbon supplied to the soil by
roots and foliage, including but not limited to, light availability, soil moisture content and
surrounding tree density. The most distal points in Fig. A-1 were replicates of species
clustered below the1:1 line but occurred on poorly drained soils in a low-lying section of
the arboretum.
Delta 13C data of soil, root and litter inputs of the single-species plots also indicated
greater similarity between soil and roots (Fig. A-2). The isotopic signature of 13C
differed significantly for all three substrates (P<0.001), and the difference between soil
137
and root signatures were significantly lower than that of soil and litter signatures
(P<0.001).
In the multi-species systems, evidence to suggest the dominance of root-derived
carbon in the SOM was ambivalent. Following the same analysis, the spectral signature
of the top 10 cm of the clayey soil under secondary forest bore greater similarity to the
non-combusted spectral profile of roots than to that of surface-deposited litter (Fig. A-3).
The relationship was strengthened when employing ash-subtracted spectra. In contrast,
neither roots nor litter dominated in the secondary forest and cabruca systems situated on
the sandy soil. A possible explanation for this is the relative instability of these systems.
Because of the high activity expressed through the extraction of species in the cabruca
system and the influence of new, introduced species in the forest fragment, sufficient
time may not have passed for an equilibrium between aboveground vegetation and
belowground soil carbon to occur, thereby blurring the parentage soil carbon signature.
Interpretation
This study centered on the application of infrared spectroscopy to discern the
relative similarity in chemical composition of surface and subsurface inputs to that of soil
organic matter. Soil organic matter spectra were found to resemble root matter more
closely than litter in mono-specific plots and in one of the multi-species plots. The direct
comparison of initial inputs and SOM signatures assumes that the decomposition
trajectories taken by root and litter are sufficiently unique that their end-products can be
directly associated with them. This includes the possibility in which recalcitrant
compounds of inputs are conserved despite transformations brought on by the action of
the decomposer community and, therefore, able to serve as biomarkers. In the case of
138
litter-derived material, such identifiers may be secondary metabolites rich in phenolics,
and, in the case of roots, polymers such as suberin (Rasse et al., 2005).
Natural carbon isotope abundance has been used to track above- and belowground
carbon allocation in grassland and forested systems (Peterson & Fry, 1987; Staddon,
2004) and to monitor soil carbon dynamics (Balesdent et al., 1987) arising from shifts in
vegetation communities such as cropping systems (Vitorello et al., 1989), C3 versus C4
plants (Dzurec et al., 1985) or legumes and non-leguminous species (Rao, 1994).
Because the isotopic signature of organic matter is strongly influenced by its consumers,
δ13C is also a helpful tool in charting food webs (Rothe & Gleixner, 2000; Ruess et al.,
2005). Preferential incorporation of lighter 12C by soil fauna leads to the enrichment of
residual materials in the heavier 13C isotope.
The enrichment in natural 13C from litter to roots (Fig. A-2) arises from multiple
processes by which photosynthetic C is fractionated and partitioned amongst plant organs
(Badeck et al., 2005). The observed proximity of the δ13C soil and roots has several
viable interpretations. The first is that, assuming the same degree of decomposition and
decomposition trajectory undergone by root and litter, root carbon is a significantly
greater contributor to SOM than plant carbon. However, if litter were preferentially
consumed by decomposers over root matter, litter-derived end products could constitute
the majority of SOM through the greater exposure of litter to microbial activity, even
while its original isotopic signature is more distant to SOM. In the absence of information
about decomposition pathways, interpretation of δ13C values is inconclusive.
Consumers of litter detritus exhibit preference for carbon forms and leaf litter has
been commonly to be the preferred substrate over root matter (Kramer & Gleixner,
139
2008). However, a recent study conducted in a temperate forest system demonstrated a
stark contrast in the degrading communities of root- and leaf-derived materials, with the
majority of soil invertebrates obtaining their carbon from roots (Pollierer et al., 2007).
Under such a scenario, the proximity of the carbon isotopic signature of roots and SOM
supports the interpretation that root-derived OM constitutes a greater proportion of SOM
than litter-derived OM and that the spectral analysis is corroborated.
Already used successfully to calibrate and predict NMR-determined alkyl C and O-
alkyl C contents (Leifeld, 2006), which are key functional groups used for studying the
structure and turnover of SOM, DRIFTS presents a potential tool in the quantifying the
relative contribution of various inputs to soil carbon. Its utility in calibration and
prediction can permit greater sample throughput which can facilitate the expansion of
spatial coverage for landscape-level assessments as well as finer scaled sampling over
soil particle size fractions. An advantage over isotopic analysis is that no assumptions of
decomposition trajectory or decomposition community have to be made. If successful,
this approach can help to identify the dominant source of soil organic matter and its final
destination may be useful in evaluating soil carbon management strategies in forested
systems.
Conclusion
Using DRIFTS as the basis for chemical characterization, greater similarity
between soil organic matter and fine-root inputs than between SOM and leaf litter was
observed. Many studies have used root- and litter-exclusion experiments or isotopic
ratios for determining the relative contribution of organic inputs to soil carbon. This
method potentially offers a faster, cost-effective technique for a qualitative assessment
with no field-level experimental manipulation and minimal sample preparation. Once
140
141
validated by traditional methods, DRIFTS analysis presents a valuable monitoring tool
for land management practices in which carbon sequestration is a primary objective.
Table A-1. Ten tropical forest species used to analyze foliage, litter, root inputs to soil organic matter.
Common name Scientific name Duplicated?
Amescla mirim Protium sp. Y Angico branco Anadenathera peregrina N Arapaçu Sclerolobium chrysophyllum Y Arapati Arapatiella psilophylla Y Gindiba Sloanea obtusifolia Y Imbiruçu Bombax macrophyllum Y Jatobá Hymenaea sp. Y Pau marfim Balfourodendron riedelianum Y Sapucaia Lecythis pisonis Y Vinhático Plathymenia foliolosa Y
142
D C
B A
Fig. A-1. Sum of squared differences from of organic inputs plotted against each other at the0-5
cm and 5-10 cm depths. A) Foliage vs. Litter. B) Roots vs. Foliage. C) Roots vs. Litter. D) Roots vs. Litter for ash-subtracted soil.
143
Fig. A-2. Isotopic carbon signatures of 18 single-species plots.
144
B
A
Fig. A-3. Sum of squared differences between forest floor litter and root litter in a multi-species
forested system atop a clayey soil, based on non-combusted and combusted (cleaned) spectra A) Non-combusted; B) Ash-subtracted spectra.
145
146
Fig. A-4. Sum of squared differences between forest floor litter and root litter in a multi-species
forested system atop a sandy soil, based on non-combusted and combusted (cleaned) spectra A) Non-combusted; B) Ash-subtracted spectra.
B
A
LIST OF REFERENCES
Aber JD, Melillo JM, McClaugherty CA. 1990. Predicting long-term patterns of mass-loss, nitrogen dynamics, and soil organic-matter formation from initial fine litter chemistry in temperate forest ecosystems. Canadian Journal of Botany 68: 2201-2208.
Alger K. 1998. The reproduction of the cocoa industry and biodiversity in southern Bahia, Brazil. Proceedings of the Smithsonian Institution, Washington, DC. http://nationalzoosiedu/ConservationAndScience/ MigratoryBirds/ Research/Cacao/paperscfu (accessed April 2006).
Amatangelo KL. 2008. Biomass and biochemistry: litter impacts on community and ecosystem processes. Ph.D thesis, Stanford University, USA.
Annan-Afful E, Iwashima N, Otoo E, Owusu-Sekyere E, Asubonteng KO, Kamidohzono A, Masunaga T, Wakatsuki T. 2005. Land-use dynamics and nutrient characteristics of soils and plants along topo-sequences in inland valley watersheds of Ashanti Region, Ghana. Soil Science and Plant Nutrition 50: 633-647.
Araújo M, Alger K, Rocha R, Mesquita CAB. 1998. A Mata Atlântica do sul da Bahia: situação atual, ações e perspectivas. Reserva da Biosfera da Mata Atlântica-MAB-UNESCO. Caderno 8:1-36.
Armbrecht I, Perfecto I, Vandermeer J. 2004. Enigmatic biodiversity correlations: Ant diversity responds to diverse resources. Science 304: 284-286.
Badeck FW, Tcherkez G, Nogues S, Piel C, Ghashghaie J. 2005. Post-photosynthetic fractionation of stble carbon isotopes between plant organs—a widespread phenomenon. Rapid Communications in Mass Spectrometry 19: 1381-1391.
Balesdent J, Balabane M. 1996. Major contribution of roots to soil carbon storage inferred from maize cultivated soils. Soil Biology and Biochemistry 28: 1261-1263.
Balesdent J, Mariotti A, Guillet B. 1987. Natural C-13 abundance as a tracer for studies of soil organic-matter dynamics. Soil Biology and Biochemistry 19: 25-30.
Ball BA, Hunter MD, Kominoski JS, Swan CM, Bradford MA. 2008. Consequences of non-random species loss for decomposition dynamics: experimental evidence of additive and non-additive effects. Journal of Ecology 96: 303-313.
Barbosa LCA. 2007. Espectroscopia no infravermelho na caracterização de compostos orgânicos. Viçosa, Brazil: Editora UFV.
Bardgett RD, Shine A. 1999. Linkages between plant litter diversity, soil microbial biomass and ecosystem function in temperate grasslands. Soil Biology and Biochemistry 31: 317-321.
Barnes RJ, Dhanoa MS, Lister SJ. 1989. Standard normal variate transformation and de-trending of near-infrared diffuse reflectance spectra. Applied Spectroscopy 43: 772-777.
147
Bedê L, Araújo M, Faria D, Schroth G. 2007. Biodiversity friendly certification of cocoa landscapes in Bahia, Brazil Accessed on the web 01/16/09 at http://web.catie.ac.cr/cd_multiestrata/presentaciones_orales/session3/Biodiversity.pdf
Beer J. 1988. Litter production and nutrient cycling in a coffee (Coffea arabica) or cacao (Theobroma cacao) plantations with shade trees. Agroforestry Systems 7: 103-114.
Ben-Dor E, Banin A. 1995. Near-infrared analysis as a rapid method to simultaneously evaluate several soil properties. Soil Science Society of America Journal 59: 364-372.
Berg B, McClaugherty C. 2008. Plant litter: Decomposition, humus formation, carbon sequestration, 2nd Edition. Berlin, Germany: Springer-Verlag.
Berg B, Steffen KT, McClaugherty C. 2007. Litter decomposition is dependent on litter Mn concentrations. Biogeochemistry 82: 29-39.
Blair JM, Parmelee RW, Beare MH. 1990. Decay rates, nitrogen fluxes and decomposer communities of single- and mixed-species foliar litter. Ecology 7: 1976-1985.
Blomqvist MM, Olff H, Blaauw MB, Bongers, T, van der Putten WH. 2000. Interactions between above- and below-ground biota: importance for small-scale vegetation mosaics in a grassland ecosystem. Oikos 90: 582-298.
Bobo KS, Waltert M, Sainge NM, Njokagor J, Fermon H, Mühlenberg M. 2006. From forest to farmland: Species richness patterns of trees and understorey plants along a gradient of forest conversion in Southwestern Cameroon. Biodiversity and Conservation 15: 4097-4117.
Botta-Dukát Z. 2005. Rao’s quadratic entropy as a measure of functional diversity based on multiple traits. Journal of Vegetation Science 16: 533-540.
Bowden RD, Nadelhoffer KJ, Boone RD, Melillo JM, Garrison JB. 1993. Contributions of aboveground litter, belowground litter, and root respiration to total soil respiration in a temperate mixed hardwood forest. Canadian Journal of Forest Research 23: 1402-1407.
Bray JR, Curtis JT. 1957. An ordination of the upland forest communities of Southern Wisconsin. Ecological Monographs 27: 325-349.
Briones MJI, Ineson P. 1996. Decomposition of eucalyptus leaves in litter mixtures. Soil Biology and Biochemistry 28: 1381-1388.
Brown S, Lugo AE. 1982. The storage and production of organic matter in tropical forests and their role in the global carbon cycle. Biotropica 14: 161-187.
Brown S, Lugo AE. 1990. Tropical secondary forests. Journal of Tropical Ecology 6: 1-32.
148
Bruun S, Stenberg B, Breland TA, Gudmundsson J, Henriksen TM, Jensen LS, Korsæth A, Luxhøi J, Pálmason F, Pedersen A, Salo T. 2005. Empirical predictions of plant material C and N mineralization patterns from near infrared spectroscopy, stepwise chemical digestion and C/N ratios. Soil Biology and Biochemistry 37: 2283-2296.
Bunker DE, DeClerck F, Bradford JC,Colwell RK, Perfecto I, Phillips OL, Sankaran M, Naeem S. 2005. Species loss and aboveground carbon storage in a tropical forest. Science 310: 1029-1031.
Card DH, Peterson DL, Matson PA. 1988. Prediction of leaf chemistry by the use of visible and near infrared spectroscopy. Remote Sensing of Environment 26:123-157.
Cardinale BJ, Nelson K, Palmer MA. 2000. Linking species diversity to the functioning of ecosystems: On the importance of environmental context. Oikos 91: 175-183.
Cardoso da Silva JM, Tabarelli M. 2000. Tree species impoverishment and the future flora of the Atlantic forest of northeast Brazil. Nature 404:72-74.
Carney KM, Matson PA. 2006. The influence of tropical plant diversity and composition on soil microbial communities. Microbial Ecology 52: 226-238.
Cassano CR, Schroth G, Faria D, Delabie JHC, Bedê L. 2008. Landscape and farm scale management to enhance biodiversity conservation in the cocoa producing region of southern Bahia, Brazil. Biodiversity and Conservation DOI 10.1007/s10531-008-9526-x.
CEPLAC Comissão Executiva Do Plano Da Lavoura Cacaueira. 1976. Dinâmica do uso da terra: Diagnóstico sócio-econômico da região cacaueira, Volume 3. de Oliveira Leite J, ed. Ilhéus, Brazil: IICA/CEPLAC.
CEPLAC Comissão Executiva Do Plano Da Lavoura Cacaueira. 2006. Departamento de Meteorologia, CEPLAC-CEPEC.
Chapman K, Whittaker JB, Heal OW. 1988. Metabolic and faunal activity in litters of tree mixtures compared with pure stands. Agriculture, Ecosystems and the Environment 24: 33-40.
Chapman SK, Koch GW. 2007. What type of diversity yields synergy during mixed litter decomposition in a natural forest system? Plant and Soil 299:153-162.
Chepote RE, Sodré GA, Reis EL, Pacheco RG, Marrocos PCL, Serôdio MH, Valle RR. 2005. Recomendações de corretivos e fertilizantes na cultura do cacaueiro no Sul da Bahia - 2ª aproximação. Ilhéus, Brazil: CEPLAC/CEPEC.
Clarke KR, Warwick RM. 1998. A taxonomic distinctness index and its statistical properties. Journal of Applied Ecology 35: 523-531.
149
Cohen MJ, Prenger JP, DeBusk WF. 2005. Visible-near infrared reflectance spectroscopy for rapid, nondestructive assessment of wetland soil quality. Journal of Environmental Quality 34: 1422-1434.
Compton JE, Watrud LS, Porteus LA, DeGrood S. 2004. Response of soil microbial biomass and community composition to chronic nitrogen additions at Harvard forest. Forest Ecology and Management 196: 143-158.
Conn C, Dighton J. 2000. Litter quality influences on decomposition, ectomycorrhizal community structure and mycorrhizal root surface acid phosphatase activity. Soil Biology and Biochemistry 32: 489-496.
Coûteaux MM, Bottner P, Berg B. 1995. Litter decomposition, climate and litter quality. Trends in Ecology and Evolution 10: 63-66.
Crawford DL, Crawford RL. 1980. Microbial degradation of lignin. Enzyme and Microbial Technology 2:11-22.
da Fonseca GAB. 1985. The vanishing Brazilian Atlantic forest. Biological Conservation 34: 17-34.
da Fonseca GAB, Alger KN, Pinto LP, Araújo M, Cavalcanti R. 2003. Corredores de biodiversidade: O corredor central da Mata Atlântica. In: Prado PI, Landau EC, Moura RT, Pinto LPS, Fonseca GAB, Alger K, orgs. Corredor de Biodiversidade da Mata Atlântica do Sul da Bahia. Publication on CD-ROM, Ilhéus, Bahia: IESB/ CI/ CABS/ UFMG/ UNICAMP.
Dalal RC, Henry RJ. 1986. Simultaneous determination of moisture, organic carbon, and total nitrogen by near IR reflectance spectrophotometry. Soil Science Society of America Journal 50: 120-123.
Dangles O, Malmqvist P. 2004. Species richness-decomposition relationships depend on species dominance. Ecology Letters 7: 395-402.
de Deyn GB, Raaijmakers CE, van Ruijven J, Berendse F, van der Putten WH. 2004. Plant species identity and diversity on different trophic levels of nematodes in the soil food web. Oikos 106: 576-586.
de Oliveira Leite J, Valle RR. 1990. Nutrient cycling in the cacao ecosystem: Rain and throughfall as nutrient sources for the soil and cacao tree. Agriculture, Ecosystems and the Environment 32: 143-154.
Dean W. 1995. With broadax and firebrand—the destruction of the Brazilian Atlantic Forest. Berkeley, USA: University of California Press.
Dehlin H, Nilsson MC, Wardle DA. 2006. Aboveground and belowground responses to quality and heterogeneity of organic inputs to the boreal forest. Oecologia 150: 108-118.
150
Díaz S, Hodgson JG, Thompson K, Cabido M, Cornelissen JHC, Jalili A, Montserrat-Martí G, Grime JP, Zarrinkamar F, Asri Y, Band SR, Basconcelo S, Castro-Díez P, Funes G, Hamzehee B, Khoshnevi M, Pérez-Harguindeguy N, Pérez-Rontomé MC, Shirvany FA, Vendramini F, Yazdani S, Abbas-Azimi R, Bogaard A, Boustani S, Charles M, Dehghan M, de Torres-Espuny L, Falczuk V, Guerrero-Campo J, Hynd A, Jones G, Kowsary E, Kazemi-Saeed F, Maestro-Martínez M, Romo-Díez A, Shaw S, Siavash B, Villar-Salvador P, Zak MR. 2004. The plant traits that drive ecosystems: Evidence from three continents. Journal of Vegetation Science 15: 295-304.
Donald PF. 2004. Biodiversity impacts of some agricultural commodity production systems. Conservation Biology 18: 17-37.
Duguma B, Gochowski J, Bakala J. 2001. Smallholder cacao (Theobroma cacao Linn) cultivation in agroforestry systems of West and Central Africa: Challenges and opportunities. Agroforestry Systems 51: 177-188.
Dzurec RS, Boutton TW, Caldwell MM, Smith BN. 1985. Carbon isotope ratios of soil organic matter and their use in assessing community composition changes in Curlew Valley, Utah. Oecologia 66: 17-24.
Ehrlich PR. 1964. Some axioms of taxonomy. Systematic Zoology 13: 109-123.
Elias Jr. E. 1998. Florística e estrutura fitossociológica de fragmentos de Floresta Atlântica do município de Eunápolis-Bahia. Master’s thesis, Federal University of Viçosa, Brazil.
Epps KY, Comerford NB, Reeves JB, Cropper WP, Araujo QR. 2007. Chemical diversity-highlighting a species richness-ecosystem function disconnect. Oikos 116: 1831-1840.
Ewel JJ. 1979. Natural systems as models for the design of sustainable systems of land use. Agroforestry Systems 45: 1-21.
Faria D, Paciencia MLB, Dixo M, Laps RR, Baumgarten J. 2007. Ferns, frogs, lizards, birds and bats in forest fragments and shade cacao plantations in two contrasting landscapes in the Atlantic forest, Brazil. Biodiversity and Conservation 16: 2335-2357.
Faria Filho AF, Araújo QR. 2003. Zoneamento do meio fisico do município de Ilhéus, Bahia, Brasil, utilizando a técnica de geoprocessamento. CEPLAC/CEPEC. Boletim Técnico No. 187. 20 pp.
Finzi AC, Canham CD. 1998. Non-additive effects of litter mixtures on net N mineralization in a southern New England forest. Forest Ecology and Management 105: 129-136.
Franzen M, Mulder MB. 2007. Ecological, economic and social perspectives on cocoa production worldwide. Biodiversity and Conservation 16: 3835-3849.
Fyles JW, Fyles IH. 1993. Interaction of Douglas-fir with red alder and salal foliage litter during decomposition. Canadian Journal of Forest Research 23: 358-361.
151
Gale WJ, Cambardella CA. 2000. Carbon dynamics of surface residue- and root-derived organic matter under simulated no-till. Soil Science Society of America Journal 64: 190-195.
Gama-Rodrigues AC, Barros NF, Santos ML. 2003. Decomposição e liberação de nutrientes do folhedo de espécies florestais nativas em plantios puros e mistos no sudeste da Bahia. Revista Brasileira de Ciência de Solo 27: 1021-1031.
Garland JL, Mills AL. 1991. Classification and characterization of heterotrophic microbial communities on the basis of patterns of community-level sole-carbon-source utilization. Applied and Environmental Microbiology 57: 2351-2359.
Gartner TB, Cardon ZG. 2004. Decomposition dynamics in mixed-species leaf litter. Oikos 104: 230-246.
Gasparatos D, Haidouti C. 2001. A comparison of wet oxidation methods for determination of total phosphorus in soils. Journal of Plant Nutrition and Soil Science 164: 435-439.
Gastine A, Scherer-Lorenzen M, Leadley PW. 2003. No consistent effects of plant diversity on root biomass, soil biota and soil abiotic conditions in temperate grassland communities. Applied Soil Ecology 24: 101-111.
Gee GW, Bader JW. 1986. Particle-size analysis-Hydrometer method. In: Klute A, ed. Methods of soil analysis: Part 1 Physical and mineralogical methods, 2nd Edition. Madison, USA: ASA-SSSA, 383-411.
Gentry AH. 1988. Tree species richness of upper Amazonian forest. Proceedings of the National Academy of Sciences 85: 156-159.
Gillon D, Houssard C, Joffre R. 1999. Using near-infrared reflectance spectroscopy to predict carbon, nitrogen and phosphorus content in heterogeneous plant material. Oecologia 118: 173-182.
Gleixner G, Kramer C, Hahn V, Sasche D. 1995. The effect of biodiversity on carbon storage in soils. In: Scherer-Lorenzen M, Körner C, Schulze ED eds. Forest diversity and function. Berlin, Germany: Springer, 165-183.
Gnankambary Z, Bayala J, Malmer A, Nyberg G, Hien V. 2008. Decomposition and nutrient release from mixed plant litters of contrasting quality in an agroforestry parkland in the south-Sudanese zone of West Africa. Nutrient Cycling in Agroecosystems 82: 1-13.
Gower JC. 1971. A general coefficient of similarity and some of its properties. Biometrics 27: 857-871.
Greenberg R. 1998. Biodiversity in the cacao agroecosystem: Shade management and landscape considerations. Proceedings of the First International Workshop on Sustainable Cocoa Growing, Panama, 3/30-4/02, 1998, Smithsonian Migratory Bird Center, Panama. http://www.nationalzoosiedu/smbc/Research/Cacao/ cacaoasp.
152
Greenland PJ, Wild A, Adams D. 1992. Organic Matter Dynamics in Soils of the Tropics-From Myth to Complex Reality. In: Lal R, Sanchez PA, eds. Myth and science of soils of the tropics. Madison, USA: SSSA/ ASA, 17-23.
Harch BD, Correll RL, Meech W, Kirkby CA, Pankhurst CE. 1997. Using the Gini coefficient with Biolog substrate utilization data to provide an alternative quantitative measure for comparing bacterial soil communities. Journal of Microbiological Methods 30: 91-101.
Hartemink AE. 2005. Nutrient stocks, nutrient cycling, and soil changes in cocoa ecosystems: A review. Advances in Agronomy 86: 227-253.
Hättenschwiler S. 1995. Effect of tree species diversity on litter quality and decomposition. In: Scherer-Lorenzen M, Körner C, Schulze ED eds. Forest diversity and function. Berlin, Germany: Springer, 149-164.
Hättenschwiler S, Aeschlimann B, Couteaux M-M, Roy J, Bonal D. 2008. High variation in foliage and leaf litter chemistry among 45 tree species of a neotropical rainforest community. New Phytologist 179: 165-175.
Hättenschwiler S, Tiunov AV, Scheu S. 2005. Biodiversity and litter decomposition in terrestrial ecosystems. Annual Review of Ecology, Evolution and Systematics 36: 191-218.
Hättenschwiler S, Vitousek PM. 2000. Role of polyphenols in terrestrial ecosystem nutrient cycling. Trends in Evolution and Ecology 6: 238-243.
Heal OW, Anderson JM, Swift MJ. 1995. Plant litter quality and decomposition: An historical overview. In: Cadisch G, Giller KE (eds) Driven by nature: Plant litter quality and decomposition, pp 3-32.
Hector A, Joshi J, Scherer-Lorenzen M, Schmid B, Spehn EM, Wacker L, Weilenmann M, Bazeley-White E, Beierkuhnlein C, Caldeira MC, Dimitrakopolous PG, Finn JA, Huss-Danell K, Jumpponen A, Leadley PW, Loreau M, Mulder CPH, Neβhöver C, Palmborg C, Read DJ, Siamantziouras ASD, Terry AC, Troumbis AY. 2007. Biodiversity and ecosystem functioning: Reconciling the results of experimental and observational studies. Functional Ecology 21: 998-1002.
Heemsbergen DA, Berg MP, Loureau M, van Hal JR, Faber JH, Verhoef HA. 2004. Biodiversity effects on soil processes explained by interspecific functional dissimilarity. Science 306: 1019-1020.
Heffernan B. 1985. Exchangeable cations and cation exchange capacity of acid soils In: B Heffernan, ed. A handbook of methods of inorganic chemical analysis for forest soils, foliage and water. Canberra, Australia: CSIRO, 113-120.
Hendricks JJ, Nadelhoffer KJ, Aber JD. 1993. Assessing the role of fine roots in carbon and nutrient cycling. Trends in Ecology and Evolution 8: 174-178.
153
Hervé BD, Vidal S. 2008. Plant biodiversity and vegetation structure in traditional cocoa forest gardens in southern Cameroon under different management. Biodiversity and Conservation 17: 1821-1835.
Hill JL, Curran PJ. 2003. Area, shape and isolation of tropical tree forest fragments: Effects on tree species diversity and implications for conservation. Journal of Biogeography 30: 1391-1403.
Hobbie EA, Johnson MG, Rygiewicz PT, Tingey DT, Olszyk DM. 2004. Isotopic estimates of new carbon inputs into litter and soils in a four-year climate change experiment with Douglas-fir. Plant and Soil 259: 331-343.
Hobbie SE. 1992. Effects of plant-species on nutrient cycling. Trends in Ecology and Evolution 7: 336-339.
Hooper DU, Vitousek PM. 1998. Effects of plant composition and diversity on nutrient cycling. Ecological Monographs 68: 121-149.
Hoorens B, Aerts R, Stroetenga M. 2003. Does initial litter chemistry explain litter mixture effects on decomposition? Oecologia 137: 578-586.
Hu YL, Wang SL, Zeng DH. 2006. Effects of single Chinese fir and mixed leaf litters on soil chemical, microbial properties and soil enzyme activities. Plant and Soil 282: 379-386.
Huber S, Kneubühler M, Psomas A, Itten K, Zimmerman NE. 2008. Estimating foliar biochemistry from hyperspectral data in mixed forest canopy. Forest Ecology and Management 256: 491-501.
Huston MA. 1997. Hidden treatments in ecological experiments: Re-evaluating the ecosystem function of biodiversity. Oecologia 110: 449-460.
Isaac ME, Gordon AM, Thevathasan N, Oppong SK, Quashie-Sam J. 2005. Temporal changes in soil carbon and nitrogen in west African multistrata agroforestry systems: A chronosequence of pools and fluxes. Agroforestry Systems 65: 23-31.
Izsák J, Szeidl L. 2002. Quadratic diversity: Its maximization can reduce the richness of species. Environmental and Ecological Statistics 9: 423-430.
Joffre R, Ågren G, Gillon D, Bosatta E. 2001. Organic matter quality in ecological studies: Theory meets experiment. Oikos 93: 451-458.
Jonsson M, Wardle DA. 2008. Context dependency of litter-mixing effects on decomposition and nutrient release across a long-term chronosequence. Oikos 117: 1674-1682.
Kanmegne J, Smaling EMA, Brussaard L, Gansop-Kouomegne A, Boukong A. 2006. Nutrient flows in smallholder production systems in the humid forest zone of southern Cameroon. Nutrient Cycling in Agroecosystems 76: 233-248.
154
King JS, Pregitzer KS, Zak DR, Holmes WE, Schmidt K. 2005. Fine root chemistry and decomposition in model communities of north-temperate tree species show little response to elevated atmospheric CO2 and varying soil resource availability. Oecologia 146: 318-328.
Kramer C, Gleixner G. 2008. Soil organic matter in soil depth profiles: Distinct carbon preferences of microbial groups during carbon transformation. Soil Biology and Biochemistry 40: 425-433.
Krebs CJ. 1989. Ecological methodology. New York, USA: Harper & Row.
Laurance WF, Lovejoy TE, Vasconcelos HL, Bruna EM, Didham, RK, Stouffer PC, Gascon C, Bierregaard RO, Laurance SG, Sampaio E. 2002. Ecosystem decay of Amazonian forest fragments: A 22-year investigation. Conservation Biology 16: 605-618.
Leifeld J. 2006. Application of diffuse reflectance FT-IR spectroscopy and partial least squares regression to predict NMR properties of soil organic matter. European Journal of Soil Science 57: 846-857.
Li Y, Xu M, Sun OJ, Cui W. 2004. Effects of root and litter exclusion on soil CO2 efflux and microbial biomass in wet tropical forests. Soil Biology and Biochemistry 36: 2111-2114.
Liebsch D, Marques MCM, Goldenberg R. 2008. How long does the Atlantic Rain Forest take to recover after disturbance? Changes in species composition and ecological features during secondary succession. Biological Conservation 141: 1717-1725.
Little AEF, Robinson CJ, Peterson SB, Raffa KF, Handelsman J. 2008. Rules of engagement: Interspecies interactions that regulate microbial communities. Annual Review of Microbiology 62: 375-401.
Liu P, Sun OJ, Huang J, Li L. 2007. Nonadditive effects of litter mixtures on dcomposition and correlation with initial litter N and P concentrations in grassland plant species of northern China. Biology and Fertility of Soils 44: 211-216.
Lobão DEVP. 2007. Agroecossistema cacaueiro da Bahia: Cacau-cabruca e fragmentos florestais na conservação de espécies arbóreas. Ph.D thesis, Universidade Estadual Paulista, Brazil.
Machado ELM, Higashikawa EM, Macedo RLG, Venturin N, Naves ML, Gomes JE. 2005. Análise da diversidade entre sistemas agroflorestais em asssentamentos rurais no sul da Bahia. Revista Científica Eletrônica de Engenharia Florestal 5:1-14 http://www.revista.inf.br/florestal/.
Madritch MD, Hunter MD. 2005. Phenotypic variation in oak litter influences short- and long-term nutrient cycling through litter chemistry. Soil Biology and Biochemistry 37: 319-327.
Magurran AE. 2003. Measuring biological diversity. Oxford, UK: Blackwell.
155
Martens H, Næs T. 1989. Multivariate calibration. Chichester, UK: Wiley.
Martini AMZ, Fiaschi P, Amorim AM, da Paixão JL. 2007. A hot-point within a hot-spot: A high diversity site in Brazil’s Atlantic Forest. Biodiversity and Conservation 16: 3111-3128.
Mathers NJ, Jalota RK, Dalal RC, Boyd SE. 2007. 13C-NMR analysis of decomposing litter and fine roots in the semi-arid Mulga Lands of southern Queensland. Soil Biology and Biochemistry 39: 993-1006.
McCarty GW, Reeves III JB, Reeves VB, Follett RF, Kimble JM. 2002. Mid-infrared and near-infrared diffuse reflectance spectroscopy for soil carbon measurement. Soil Science Society of America Journal 66: 640-646.
McTiernan KB, Ineson P, Coward PA. 1997. Respiration and nutrient release from tree leaf litter mixtures. Oikos 78: 527-538.
Melillo JM, Aber JD, Muratore JF. 1982. Nitrogen and lignin control of hardwood leaf litter decomposition dynamics. Ecology 63: 621-626.
Miller RO. 1998. High temperature oxidation: Dry ashing. In: Kalra YP, ed. Handbook of reference methods for plant analysis. Boca Raton, USA: CRC Press, 53-56.
Miltner A, Zech W. 1998. Beech leaf litter degradation and transformation as influenced by mineral phases. Organic Geochemistry 28: 457-463.
Montagnini F, Fanzeres A, Guimarães da Vinha S. 1995. The potentials of 20 indigenous tree species for soil rehabilitation in the Atlantic Forest region of Bahia, Brazil. Journal of Applied Ecology 32: 841-856.
Montagnini F, Nair PKR. 2004. Carbon sequestration: An underexploited environmental benefit of agroforestry systems. Agroforestry Systems 61-2: 281-295.
Morellato LPC, Haddad CFB. 2000. Introduction: The Brazilian Atlantic Forest. Biotropica 32: 786-792.
Mori SA, Boom BM, Carvalho AM, Santos TS. 1983. Southern Bahian moist forest. The Botanical Review 49: 155-232.
Mouquet N, Moore JL, Loreau M. 2002. Plant species richness and community productivity: Why the mechanism that promotes coexistence matters. Ecology Letters 5: 56-65.
Murphy J, Riley JP. 1962. A modified single solution method for the determination of phosphate in natural waters. Analytica Chimica Acta 27: 31-36.
Myers N, Mittermeier RA, Mittermeier CG, da Fonseca GAB, Kent J. 2000. Biodiversity hotspots for conservation priorities. Nature 403: 853-858.
156
Naeem S, Wright JP. 2003. Disentangling biodiversity effects on ecosystem functioning: Deriving solutions to a seemingly insurmountable problem. Ecology Letters 6: 567-579.
Nair PKR. 1989. Agroforestry systems in the tropics. Dordrecht, the Netherlands: Kluwer Academic/ICRAF.
Nei M. 1972. Genetic distances between populations. American Naturalist 106: 283-292.
Nelson DW, Sommers LE. 1996. Total Carbon, Organic Carbon and Organic Matter. In: Bartels JM, ed. Methods of soil analysis, Part 3: Chemical methods. Madison, USA: SSSA-CSA-ACS, 961-1010.
NERC Northeastern Ecosystem Research Cooperative. 2006. Foliar chemistry database. USDA Forest Service Northeastern Research Station and University of New Hampshire Complex Systems Research Center. www.folchem.sr.unh.edu.
Niinemets U, Kull K. 2003. Leaf structure vs nutrient relationships vary with soil conditions in temperate shrubs and trees. Acta Oecologica 24: 209-219.
Nwoke OC, Vanlauwe B, Diels J, Sanginga N, Osonubi O. 2004. Impact of residue characteristics on phosphorus availability in West African moist savanna soils. Biology and Fertility of Soils 39: 422-428.
Nye PH. 1961. Organic and nutrient cycles under a moist tropical forest. Plant and Soil 13: 333-346.
Oliveira MA, Santos AMM, Tabarelli M. 2008. Profound impoverished of the large-tree stand in a hyper-fragmented landscape of the Atlantic forest. Forest Ecology and Management 256: 1910-1917.
Orwin KH, Wardle DA, Greenfield LG. 2006. Ecological consequences of carbon substrate identity and diversity in a laboratory study. Ecology 87: 580-593.
Palm CA, Sanchez PA. 1991. Nitrogen release from the leaves of some tropical legumes as affected by their lignin and polyphenolic contents. Soil Biology and Biogeochemistry 23: 83-88.
Pasquini C. 2003. Near infrared spectroscopy: Fundamentals, practical aspects and analytical applications. Journal of the Brazilian Chemical Society 14: 198-219.
Pastor J, Cohen Y. 1997. Herbivores, the functional diversity of plants species, and the cycling of nutrients in ecosystems. Theoretical Population Biology 51: 165-179.
Pavoine S, Ollier S, Pontier D. 2005. Measuring, diversity from dissimilarities with Rao's quadratic entropy: Are any dissimilarities suitable? Theoretical Population Biology 67: 231-239.
157
Pérez-Harguindeguy N, Diaz S, Cornelissen JHC, Vendramini F, Cabido M, Castellanos A. 2000. Chemistry and toughness predict leaf litter decomposition rates over a wide spectrum of functional types and taxa in central Argentina. Plant and Soil 218: 21-30.
Pérez-Harguindeguy N, Blundo CM, Gurvich DE, Díaz S, Cuevas E. 2008. More than the sum of its parts? Assessing litter heterogeneity effects on the decomposition of litter mixtures through leaf chemistry. Plant and Soil 303: 151-159.
Petchey OL, Gaston KJ. 2002. Functional diversity (FD), species richness and community composition. Ecology Letters 5: 402-411.
Petchey OL, Gaston KJ. 2006. Functional diversity: Back to basics and looking forward. Ecology Letters 9: 741-758.
Peterson BJ, Fry B. 1987. Stable isotopes in ecosystem studies. Annual Review of Ecology and Systematics 18: 293-320.
Phillips OL. 1997. The changing ecology of tropical forests. Biodiversity and Conservation 6: 291-311.
Pollierer MM, Langel R, Körner C, Maraun M, Scheu S. 2007. The underestimated importance of belowground carbon input for forest soil animal food webs. Ecology Letters 10: 729 - 736.
Post WM, Emanuel WR, Zinke PJ, Stangenberger AG. 1982. Soil carbon pools and world life zones. Nature 298: 156-159.
Preston CM, Trofymow JA. 2000. Variability in litter quality and its relationship to litter decay in Canadian forests. Canadian Journal of Botany-Revue Canadienne de Botanique 78: 1269-1287.
Qu XH,Wang JG. 2008. Effect of amendments with different phenolic acids microbial biomass, activity, and community diversity. Applied Soil Ecology 39: 172-179.
R Development Core Team. 2006. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0-URL http://www.R-project.org.
Raich JW, Russell AE, Bedoya-Arrieta R. 2007. Lignin and enhanced litter turnover in tree plantations of lowland Costa Rica. Forest Ecology and Management 239: 128-135.
Rao CR. 1982. Diversity and dissimilarity coefficients: A unified approach. Theoretical Population Biology 21: 24-43.
Rao IM, Ayarza MA, Thomas, RJ. 1994. The use of carbon-isotope ratios to evaluate legume contribution to soil enhancement in tropical pastures. Plant and Soil 162: 177-182.
158
Rasse DP, Rumpel C, Dignac M-F. 2005. Is soil carbon mostly root carbon? Mechanisms for a specific stabilization. Plant and Soil 269: 341-356.
Rassmussen C, Southard RJ, Horwath WR. 2006. Mineral control of organic matter mineralization in a range of temperate conifer forest soils. Global Change Biology 12: 834-847.
Reeves III JB. 1993. Infrared spectroscopic studies on forage and by-product fibre fractions and lignin determination residues. Vibrational Spectroscopy 5: 303-310.
Reeves III JB. 1998. Improvement in Fourier near- and mid-infrared diffuse reflectance spectroscopic calibrations through the use of a sample transport device. Applied Spectroscopy 50: 965-969.
Rice RA, Greenberg R. 2000. Cacao cultivation and conservation of biological diversity. Ambio 29: 167-173.
Ricotta C. 2005. A note on functional diversity measures. Basic Applied Ecology 6: 479-486.
Rodríguez-Loinaz G, Onaindia M, Amezaga I, Mijangos I, Garbisu C. 2008. Relationship between vegetation diversity and soil functional diversity in native mixed-oak forests. Soil Biology and Biochemistry 40: 49-60.
Rolim SG, Chiarello AG. 2004. Slow death of Atlantic Forest trees in cocoa agroforestry in southeastern Brazil. Biodiversity and Conservation 13: 2679-2694.
Rothe A, Binkley D. 2001. Nutritional interactions in mixed species forests: A synthesis. Canadian Journal of Forest Research 31: 1855-1870.
Rothe J, Gleixner G. 2000. Do stable isotopes reflect the food web development in regenerating ecosystems? Isotopes in Environmental and Health Studies 36: 285-301.
Rovira P, Vallejo VR. 2002. Mineralization of carbon and nitrogen from plant debris, as affected by debris size and depth of burial. Soil Biology and Biochemistry 34: 327-339.
Ruess L, Tiunov AI, Haubert D, Richnow HH, Häggblom, Scheu S. 2005. Carbon stable isotope fractionation and trophic transfer of fatty acids in fungal based soil food chains. Soil Biology and Biochemistry 37: 945-953.
Saatchi S, Agosti D, Alger K, Delabie J, Musinsky J. 2001. Examining fragmentation and loss of primary forest in the southern Bahian Atlantic Forest of Brazil with radar imagery. Conservation Biology 15: 867-875.
Salamanca EF, Kaneko N, Katagiri S. 1998. Effects of leaf litter mixtures on the decomposition of Quercus serrata and Pinus densiflora using field and laboratory microcosm methods. Ecological Engineering 10: 53-73.
159
Salamon JA, Alphei J, Ruf A, Schaefer M, Scheu S, Schneider K, Suhrig A, Maraun M. 2006. Transitory dynamic effects in the soil invertebrate community in a temperate deciduous forest: Effects of resource quality. Soil Biology and Biochemistry 38: 209-221.
Sambuichi RHR. 2002. Fitossociologia e diversidade de espécies arbóreas em cabraca na região sul da Bahia, Brasil. Acta Botanica Brasileira 16: 89-101.
Sambuichi RHR, Haridasan M. 2003. Nutrição mineral das espécies arboreas em cabrucas na região sul da Bahia. VI Congresso de Ecologia do Brasil, 9-14, November, CE. Anais de trabalhos Completos 3: 351-352.
Sambuichi RHR, Haridasan M. 2007. Recovery of species richness and conservation of native Atlantic forest trees in the cacao plantations of southern Bahia in Brazil. Biodiversity and Conservation 16: 3681-3701.
Santana SO, Ramos JV, Ruiz MAM, Araújo QR, Almeida HA, Faria Filho, AF, Mendonça JR, Santos LFC. 2003. Zoneamento agroecológico do municipio de Ilhéus, Bahia, Brasil. CEPLAC/CEPEC. Boletim Técnico No. 186. 44pp.
Santos BA, Peres CA, Oliveira MA, Grillo A, Alves-Costa CP, Tabarelli M. 2008. Drastic erosion in functional attributes of tree assemblages in Atlantic forest fragments of northeastern Brazil. Biological Conservation 141: 249-260.
Sarkhot DV, Comerford NB, Jokela EJ, Reeves III JB, Harris WG. 2007. Aggregation and aggregate carbon in a forested southeastern coastal plain Spodosol. Soil Science Society of America Journal 71: 1779-1787.
Sato S, Comerford NB. 2006. Organic anions and phosphorus desorption and bioavailability in a humid Brazilian Ultisol. Soil Science 171: 695-705.
Sauer TJ, Cambardella CA, Brandle JR. 2007. Soil carbon and tree litter dynamics in a red cedar-scotch pine shelterbelt. Agroforestry Systems 71: 163-174.
Savitzky A, Golay MJE. 1964. Smoothing and differentiation of data by simplified least squares procedures. Analytical Chemistry 36: 1627-1639.
Scheu S. 1995. Linkages between tree diversity, soil fauna and ecosystem processes. In: Scherer-Lorenzen M, Körner C, Schulze ED eds. Forest diversity and function. Berlin, Germany: Springer, 211-233.
Schimel DS. 1995. Terrestrial ecosystems and the carbon-cycle. Global Change Biology 1: 77-91.
Schimel JP, Hättenschwiler S. 2007. Nitrogen transfer between decomposing leaves of different N status. Soil Biology and Biochemistry 39: 1428-1436.
160
Schulp CJE, Nabulars GJ, Verburg PH, de Waal RW. 2008. The effect of tree species on carbon stocks in forest floor and mineral soil and implications for soil carbon inventories. Forest Ecology and Management 256: 482-490.
Schulz B, Becker B, Götsch E. 1994. Indigenous knowledge in a ‘modern’ sustainable agroforestry system-a case study from eastern Brazil. Agroforestry Systems 25: 59-69.
Schulze ED. 2006. Biological controls of the terrestrial carbon sink. Biogeosciences 3: 147-166.
Schweitzer JA, Bailey JK, Hart SC, Whitham TG. 2005. Non-additive effects of mixing cottonwood genotypes on litter decomposition and nutrient dynamics. Ecology 86: 2834-2840.
Schwendener CM, Lehmann J, de Camargo PB. 2005. Nitrogen transfer between high- and low-quality leaves on a nutrient-poor Oxisol determined by 15N enrichment. Soil Biology and Biochemistry 37: 787-794.
Schroth G, Harvey CA. 2007. Biodiversity conservation in cocoa production landscapes: An overview. Biodiversity and Conservation 16: 2237-2244.
Seastedt TR. 1984. The role of microarthropods in decomposition and mineralization processes. Annual Review of Entomology 29: 25-46.
Seneviratne G. 2000. Litter quality and nitrogen release in tropical agriculture: A synthesis. Biology and Fertility of Soils 31: 60-64.
Shepherd KD, Vanlauwe B, Gachengo CN, Palm CA. 2005. Decomposition and mineralization of organic residues predicted using near infrared spectroscopy. Plant and Soil 277: 315-333.
Shepherd KD, Walsh MG. 2002. Development of spectral libraries for characterization of soil properties. Soil Science Society of America Journal 66: 988-998.
Shimatani K. 2001. On the measurement of species diversity incorporating species differences. Oikos 93: 135-147.
Silver WL, Brown S, Lugo AE. 1996. Effects of changes in biodiversity on ecosystem function in tropical forests. Conservation Biology 10: 17-24.
Six J, Guggenberger K, Paustian L, Haumaier, Elliott ET, Zech W. 2001. Sources and composition of soil organic matter fractions between and within soil aggregates. European Journal of Soil Science 52: 607-618.
Smith CK, Gholz HL, De Assis Oliveira F. 1998. Fine litter chemistry, early-stage decay and nitrogen dynamics under plantations and primary forest in lowland Amazonia. Soil Biology and Biochemistry 30: 2159-2169.
161
Smith VC, Bradford MA. 2003. Do non-additive effects on decomposition in litter-mix experiments result from differences in resource quality between litters? Oikos 102: 235-242.
Sonwa DJ, Nkongmeneck BA, Weise SF, Tchatat M, Adesina AA, Janssens MJJ. 2007. Diversity of plants in cocoa agroforests in the humid forest zone of Southern Cameroon. Biodiversity and Conservation 16: 2385-2400.
Staddon PL. 2004. Carbon isotopes in functional soil ecology. Trends in Ecology and Evolution 19: 148-154.
Stahl PD, Christensen M. 1992. In vitro mycelial interactions among members of a soil microfungal community. Soil Biology and Biochemistry 24: 309-316.
StatSoft, Inc. 2004. STATISTICA (data analysis software system), version 7. www.statsoft.com.
Stuart B. 1997. Biological applications of infrared spectroscopy. Chichester, UK: Wiley.
Sulzman EW, Brant JB, Bowden RD, Lajtha K. 2005. Contribution of aboveground litter, belowground litter, and rhizosphere respiration to soil CO2 efflux in an old growth coniferous forest. Biogeochemistry 73: 231-256.
Swan CM, Palmer MA. 2006. Composition of speciose leaf litter alters stream detritivore growth, feeding activity and leaf breakdown. Oecologia 147: 469-478.
Swift MJ, Heal OW, Anderson JM. 1979. Decomposition in terrestrial ecosystems. Oxford, UK: Blackwell.
Tabanez AAJ, Viana VM. 2000. Patch structure within Brazilian Atlantic Forest fragments and implications for conservation. Biotropica 34: 925-933.
Tanner EVJ. 1977. Four montane rain forests of Jamaica: A quantitative characterization of the floristics, the soils and the foliar mineral levels and a discussion of the interrelations. Journal of Ecology 65: 883-918.
Tenney FG, Waksman SA. 1929. Composition of natural organic materials and their decomposition in the soil: IV The nature and the rapidity of decomposition of the various organic complexes in different plant materials, under aerobic conditions. Soil Science 28: 55-84.
Thomas WW. 2003. Natural Vegetation Types in Southern Bahia. In: Prado PI, Landau EC, Moura RT, Pinto LPS, Fonseca GAB, Alger K, orgs. Corredor de biodiversidade da Mata Atlântica do sul da Bahia. Publicação em CD-ROM, Ilhéus, Brazil: IESB/ CI / CABS / UFMG / UNICAMP.
Thomas WW, de Carvalho AAMV, Amorim AMA, Garrison J, Arbelaez AL. 1998. Plant endemism in two forests in southern Bahia, Brazil. Biodiversity Conservation 7: 311-322.
162
Thompson J, Proctor J, Viana V, Milliken W, Rater JA, Scott DA.1992. Ecological studies on a lowland evergreen forest on Maraca Island, Roraima, Brazil. I. Physical environment, forest structure and leaf chemistry. Journal of Ecology 80: 689-703.
Tian G, Brussard L, Kang BT. 1995. Plant residue decomposition in the humid tropics-influence of chemical composition and soil fauna. In: Reddy VM ed. Soil organisms and litter decomposition in the tropics. Boulder, USA: Westview Press.
Tilman D. 1999. The ecological consequences of changes in biodiversity: A search for general principles. Ecology 80: 1455-1474.
Townsend AR, Asner GP, Cleveland CC. 2008. The biogeochemical heterogeneity of tropical forests. Trends in Ecology and Evolution 23: 424-431.
Trofymow JA, Moore TR, Titus B, Prescott C, Morrison I, Siltanen M, Smith S, Fyles J, Wein R, Camiré C, Duschene L, Kozak L, Kranabetter M, Visser S. 2002. Rates of litter decomposition over 6 years in Canadian forests: Influence of litter quality and climate. Canadian Journal of Forest Research 32: 789-804.
Trofymow JA, Preston CM, Prescott CE. 1995. Litter quality and its potential effect on decay rates of materials from Canadian forests. Water Air and Soil Pollution 82: 215-226.
Trumbore SE. 2006. Carbon respired by terrestrial ecosystems-recent progress and challenges. Global Change Biology 12: 141-153.
Vanlauwe B, Diels J, Sanginga N, Merckx R. 1997. Residue quality and decomposition: An unsteady relationship? In: Cadisch G, Giller KE (eds) Driven by nature: Plant litter quality and decomposition, pp157-166.
Vinha SG, da Silva LAM. 1982. Árvores aproveitadas como sombreadoras de cacaueiros no sul da Bahia e norte do Espírito Santo. Ilhéus, Brazil: CEPLAC-CEPEC.
Violle C, Navas ML, Vile D, Kazakou E, Fortunel C, Hummel I, Garnier E. 2007. Let the concept of trait be functional! Oikos 116: 882-892.
Viscarra-Rossel RA. 2008. ParLeS: Software for chemometric analysis of spectroscopic data. Chemometrics and Intelligent Laboratory Systems 90: 72-83.
Vitorello VA, Ceri CC, Andreux F, Feller C, Victoria RL. 1989. Organic-matter and natural C-13 distribution in forested and cultivated Oxisols. Soil Science Society of America Journal 53: 773-778.
Vitousek P, Gerrish G, Turner D, Walker L, Muellerdombois D. 1995. Litterfall and nutrient cycling in 4 Hawaiian montane rain forests. Journal of Tropical Ecology 11: 189-203.
Waksman SA, Tenney FG. 1928. Composition of natural organic materials and their decomposition in the soil. III: The influence of nature of plant upon the rapidity of its decomposition. Soil Science 26: 155-171.
163
164
Walker B, Kinzig A, Langridge J. 1999. Plant attribute diversity, resilience, and ecosystem function: The nature and significance of dominant and minor species. Ecosystems 2: 95-113.
Wardle DA, Bonner KI, Nicholson KS. 1997. Biodiversity and plant litter: experimental evidence which does not support the view that enhanced species richness improves ecosystem function. Oikos 79: 247-258.
Weber KP, Grove JA, Gehder M, Anderson WA, Legge RL. 2007. Data transformations in the analysis of community-level substrate utilization data from microplates. Journal of Microbiological Methods 69: 461-469.
Westerhaus M, Workman Jr. J, Reeves II JB, Mark H. 2004. Quantitative analysis. In: Roberts CA, Workman Jr. J, Reeves III JB, eds. Near Infrared Reflectance Spectroscopy, Agronomy Monograph no. 44. Madison, USA: ASA/ CSA/ SSSA, 133-174.
Williams BL. 1994. Interactions between tree species and their effects on nitrogen and phosphorus transformations in the forest floor. In: Boyle TJB, Boyle CEB, eds. Biodiversity, temperate ecosystems, and global change. Berlin, Germany: Springer-Verlag, 357-370.
Xiang W, Bauhus J. 2007. Does the addition of leaf litter from N-fixing Acacia mearnsii accelerate leaf decomposition of Eucalyptus globulus? Australian Journal of Botany 55: 576-583.
Zimmer M. 2002. Is decomposition of woodland leaf litter influenced by its species richness? Soil Biology and Biochemistry 34: 277-284.
BIOGRAPHICAL SKETCH
The author was born and raised in Jamaica, New York in an auspicious year for New
Yorkers (Subway Series, Woodstock, alleged men on Moon). At the age of five, after hearing
that stars were round, self-luminous, gaseous spheres instead of the silver things that adorned her
kindergarten drawings, she decided on a future in astrophysics. Epps was given a good start at
Eastwood Elementary (aka P.S. 95Q), the Immaculate Conception School, and the Bronx High
School of Science. She remained faithful to her original plan and entered Harvey Mudd College
in Claremont, California, from which she graduated with a degree in physics with an emphasis in
astrophysics. Immediately thereafter, the author strayed from the heavenly spheres and headed
south, literally, to New Orleans, Louisiana and attempted to establish a career in photography but
was instead waylaid by music after spotting a ‘cello for sale in the window of Werlein’s. Some
years of lessons and music production followed, to varying degrees of success, and her return to
the West Coast marked an exciting and eclectic period during which she served the community
as a waitress, fisherman, cook, store clerk, and SAT tutor. The inescapable necessity of reliable
income led the author to the San Francisco School where she began as a part-time algebra
teacher and left four years later as the eighth grade teacher. After graduating a second time from
the eighth grade, Ms. Epps entered the United States Peace Corps as a volunteer in the
Agroforestry Program in Cameroon, Central Africa, where she was introduced to soil in all of its
glory and, became fascinated with phosphorus.
Returning early from Peace Corps service, Ms. Epps entered graduate school at the
University of California Davis in the Program in International Agricultural Development with an
emphasis in soil science. She later applied to the Department of Soil Science from which she
earned her master’s degree. A visit to the University of Florida to resolve laboratory trouble
with a method of sequential phosphorus fractionation of wetland sediments led her to converse
with Dr. Nicholas B. Comerford to whom she is grateful for all that followed.