biological processes influencing nutrient...
TRANSCRIPT
1
BIOLOGICAL PROCESSES INFLUENCING NUTRIENT LIMITATION IN A LOWLAND TROPICAL WET FOREST IN COSTA RICA
By
SILVIA ALVAREZ CLARE
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
2012
2
© 2012 Silvia Alvarez Clare
3
To my parents, my rudder to Chuck, my anchor
and to Lucia Jane, my shining star in dark seas
4
ACKNOWLEDGMENTS
I would like to thank my advisor Michelle C. Mack. Her great ideas, unconditional
support of my research, and patience have made this dissertation possible. Her passion
for science and mechanistic approach to ecological questions, have greatly shaped my
views as a scientist. I would also like to thank my committee members Emilio Bruna,
Nick Comerford and Ted Schuur for their support and mentorship during this long
process. All my professors at UCR and at UF have influenced me and helped me to
achieve this goal, and I thank them.
I acknowledge the National Science Foundation, the Tropical Conservation and
Development Program at the University of Florida, the South Eastern Alliance for
Graduate Education and Professoriate (SEAGEP), Elizabeth Clare-Rhoades, and
Patricia Clare for financial support. The Department of Biology and the School of
Natural Resources and the Environment at the University of Florida, as well as the
Escuela de Agricultura de la Región del Trópico Húmedo (EARTH), provided valuable
institutional support.
I would like to extend my deepest appreciation to all the people that came to the
field with me during these six years and that endured mud, rain, mosquitoes, spines,
snakes, and endless hours of fieldwork: Agustin Alvarez, Adolfo Artavia, Adrián
Villalobos, Andy Retzler, Balbina García, Carlos (Pelón), Catherine Cardelús, Charles
Knapp, Danielle Pallow, Eduardo Chacón, Elida Madrid, Enrique Salicetti, Faeleen Tais,
Fernanda Arhernas, Hanna Lee, Ismael Herrera, Jenny Bermudez, Jonathan Artavia
(La Selva), José Antonio Dominguez, José Zuniga, Josué Beltetón, Karla Ayala, Katy
Evans, Laura Morales, Laura Schreeg, Ma Alexandra Chicaciza, Maga Gei, María de
los Angeles, Melania Fernandez, Michelle Mack, Ondřej Vybíral, Rady Ho, Rigoberto
5
Gonzalez, Rolo, Sara Kovachich, Shantelle Bartra, Tony Arévalo, Violeta Rodriguez,
Virginia Guillén, and Yariela Ugalde. I also thank the people who helped me during what
seemed like endless hours in the laboratory; especially Grace Crummer and Julia
Reiskind, who with their advice and friendship made my life better during these years.
All members of the Mack and Schuur labs were great colleagues and friends,
contributing to exciting science discussions and providing support during difficult times.
My office mates Jennie DeMarco, Jenny Shafer, and Caitlin Hicks were great friends,
who I will always remember.
During my field work at EARTH University, many people provided logistical and
technical support. Professors Bert Kolhlmann and Ricardo Russo served as my local
mentors. Warner Vargas, Checo, and all the staff at the “Unidad de Ingeniería Agrícola”
allowed me to use the workshop and the tractor to fertilize my plots. Melissa Arce and
Alejandra Carvajal helped with lodging. Carlos Sandí and the staff at the “Vivero
Forestal” provided help in the field and with processing seedlings. Herbert Arrieta and
the staff at the “Laboratorio de Suelos” provided invaluable support, which included
work and storage space, assistance with laboratory analysis, and their friendship.
My friends, both the “ticos” and “the gang” were critical to keep me sane during my
PhD. I would like to thank them for making these years some of the most special of my
life. I do not have enough words to thank my family (my parents, my brothers, my
abuelitas) for their unconditional love and support. My mother is my example to follow; I
would like to thank her and tell her: you are next! This dissertation should probably be
co-authored with Chuck Knapp. He has read and edited every page, provided feedback,
and supported me every step of the way. His love, patience, and encouragement have
6
been infinite. My baby Lucia Jane, who was conceived, born, and turned one during my
PhD, reminds me every day of what really matters in life. Finally, I thank God, in his
universal non-denominational form, for giving me health, strength and perseverance to
fulfill this goal.
7
TABLE OF CONTENTS page
ACKNOWLEDGMENTS .................................................................................................. 4
LIST OF TABLES .......................................................................................................... 10
LIST OF FIGURES ........................................................................................................ 12
LIST OF ABBREVIATIONS ........................................................................................... 14
ABSTRACT ................................................................................................................... 15
CHAPTER
1 BACKGROUND AND MOTIVATION ...................................................................... 17
2 INFLUENCE OF PRECIPITATION ON SOIL AND FOLIAR NUTRIENTS ACROSS NINE COSTA RICAN FORESTS ............................................................ 27
Introduction ............................................................................................................. 27 Methods .................................................................................................................. 31
Study Site ......................................................................................................... 31 Soil Sampling and Analysis .............................................................................. 31
Foliage Sampling and Analysis ........................................................................ 33 Data Analysis ................................................................................................... 34
Results .................................................................................................................... 34 Soil Characteristics ........................................................................................... 34 Foliar Measurements ........................................................................................ 35
Relationship Between Soil and Foliar Measurements ...................................... 36 Discussion .............................................................................................................. 36
Patterns of Soil Nutrient Availability .................................................................. 36 Pattern of Foliar Nutrients ................................................................................. 39 Conclusions ...................................................................................................... 41
3 DIRECT TEST OF NUTRIENT LIMITATION TO NET PRIMARY PRODUCTIVITY IN A LOWLAND TROPICAL WET FOREST ............................... 53
Introduction ............................................................................................................. 53
Methods .................................................................................................................. 58
Site Description ................................................................................................ 58 Experimental Design ........................................................................................ 61 Soil Measurements ........................................................................................... 61 Tree Diameter Measurements .......................................................................... 63 Seedling Measurements ................................................................................... 65 Litterfall Collection ............................................................................................ 66 Root Measurements ......................................................................................... 67
8
Statistical Analysis ............................................................................................ 68
Results .................................................................................................................... 70 Soil Measurements ........................................................................................... 70
Tree Growth ..................................................................................................... 71 Seedling Measurements ................................................................................... 73 Litterfall and Litter Pool Measurements ............................................................ 73 Root Biomass and Productivity ......................................................................... 74
Discussion .............................................................................................................. 75
Soil Measurements ........................................................................................... 75 Nutrient Limitation to NPP ................................................................................ 76 Nutrient Limitation and Tree Size ..................................................................... 79 Effect of Taxa on Nutrient Limitation ................................................................ 81 Conclusions ...................................................................................................... 82
4 EFFECT OF NUTRIENT ADDITIONS ON FOLIAR, LITTER AND ROOT CHEMISTRY ......................................................................................................... 109
Introduction ........................................................................................................... 109
Methods ................................................................................................................ 113 Experimental Design ...................................................................................... 113 Chemical Analysis .......................................................................................... 115
Statistical Analysis .......................................................................................... 116 Results .................................................................................................................. 117
Foliar Nutrients ............................................................................................... 117 Litterfall Nutrients ........................................................................................... 119 Root Nutrients ................................................................................................ 120
Discussion ............................................................................................................ 121 Effects of Fertilization on Foliar Nutrients ....................................................... 121
Influence of Tree Size on Foliar Nutrients ...................................................... 123 Influence of Taxa on Foliar nutrients .............................................................. 124
Effects of Fertilization on Litterfall and Root Nutrients .................................... 125 Total Soil P as a Driver of Tissue Nutrient Concentrations ............................. 126 Conclusions .................................................................................................... 127
5 CONCLUSIONS AND LESSONS LEARNED ....................................................... 146
Conclusions .......................................................................................................... 146 The Myth of P Limitation in the Tropics .......................................................... 147 Heterogeneous Nutrient Limitation ................................................................. 147
Environmental and Biological Processes Influence Nutrient Limitation and Carbon Cycling ............................................................................................ 148
Lessons Learned .................................................................................................. 149 APPENDIX
A SPATIAL AND TEMPORAL VARIATION OF LIGHT REACHING THE UNDERSTORY IN A WET TROPICAL FOREST IN COSTA RICA ...................... 156
9
B BIOTIC AND ABIOTIC FACTORS INFLUENCING TREE GROWTH IN A LOWLAND TROPICAL WET FOREST: A MIXED MODEL APPROACH ............. 160
C ADDITIONAL TABLES AND FIGURES ................................................................ 167
LIST OF REFERENCES ............................................................................................. 174
BIOGRAPHICAL SKETCH .......................................................................................... 190
10
LIST OF TABLES
Table page 1-1 Fertilization experiments conducted in tropical forests where components of net
primary production were measured ........................................................................ 24
2-1 Characteristics of study sites in Costa Rica ............................................................ 43
2-2 Species and number of samples collected per site in nine sites in Costa Rica ....... 44
2-3 Soil nutrients and isotopic signatures (mean + SE) for the nine study sites in Costa Rica .............................................................................................................. 46
2-4 Foliar carbon, phosphorus, nitrogen and stable isotopic signatures (mean + SE) for the nine study sites in Costa Rica ..................................................................... 48
3-1 Particle sizes for soils from the top 10cm of the study plots .................................... 84
3-2 Means (with standard errors) for various soil parameters ....................................... 84
3-3 Means (with standard errors) for various soil parameters. ...................................... 85
3-4 Taxa selected to study the effect of fertilization on different functional groups ....... 86
3-5 Results from repeated measures MANOVAs for several soil variables measured at three depths ........................................................................................................ 87
3-6 Results from two-way ANOVA analyses for the percent difference between 2 yrs and pre-fertilization values for several soil parameters ........................................... 88
3-7 Results from repeated measures MANOVAs for “total basal area increase” by tree size class ......................................................................................................... 89
3-8 Results from repeated measures MANOVAs for “proportion of tree growth” by tree size class ......................................................................................................... 90
3-9 Contingency table describing the proportion of trees that grew or did not grow between 2 and 2.7 yrs after initial fertilization ......................................................... 91
3-10 Results from Pearson chi-square tests for seedling variables measured 1yr and 2 yrs after fertilization ..................................................................................... 93
3-11 Results from repeated measures MANOVAs for foliar (leaves and sticks <2mm diameter), reproductive (flowers and fruits), and coarse litterfall ........................... 93
3-12 Results from two-way ANOVA analyses for litterpool fractions ............................. 94
11
3-13 Results from repeated measures MANOVAs for fine (<2mm diameter) and large (>2mm diameter) root biomass collected at 0-15 cm depth.......................... 94
4-1 Results from repeated measures MANOVAs for foliar chemistry by tree size class .................................................................................................................... 130
4-2 Results from repeated measures MANOVAs comparing foliar %N, P (mg g-1) and N:P ratios among the six most common taxa ............................................... 131
4-3 Results from repeated measures MANOVAs for foliar %N by species. ................ 132
4-4 Results from repeated measures MANOVAs for foliar P by species. .................... 133
4-5 Results from repeated measures MANOVAs for litterfall chemistry ...................... 134
4-6 Results from repeated measures MANOVAs for root chemistry ........................... 135
5-1 Summary of f responses ratios (RR) of treatments relative to the control in the fertilization experiment .......................................................................................... 154
B-1 Parameters used in the mixed models .................................................................. 163
B-2 Models used to test the hypotheses ..................................................................... 164
B-3 Models used to test the hypotheses organized by increasing AIC values ............ 164
B-4 Maximum likelihood estimates, their standard error, and T-value for parameters included in model M1 ............................................................................................ 165
C-1 Species of trees found in the study plots .............................................................. 167
C-2 Results from repeated measures MANOVAs for foliar N:P ratios by tree size class ..................................................................................................................... 170
C-3 Results from repeated measures MANOVAs for foliar N:P ratios by species ....... 171
12
LIST OF FIGURES
Figure page 1-1 Diagram showing the evolution of nitrogen (N) and phosphorus(P) availability
through soil development. ...................................................................................... 26
1-2 Plant-soil- microbial feedbacks after addition of the limiting nutrient ....................... 26
2-1 Map of sites where soil and foliar samples were collected in Costa Rica ................ 49
2-2 Relationship between mean annual precipitation (+ SE) and soil parameters ....... 50
2-3 Mean N:P ratios (+ SE) .......................................................................................... 51
2-4 Mean soil phosphorus (+ SE) for the top 10 cm mineral layer of nine study sites in Costa Rica ......................................................................................................... 52
3-1 Monthly average maximum and minimum air temperatures (dashed lines) and precipitation (solid lines) at the study site .............................................................. 95
3-2 Floristic description of the study site ...................................................................... 96
3-3 Distribution of plots and blocks within the EARTH forest reserve ........................... 97
3-4 Mean + SE soil parameters ..................................................................................... 98
3-5 Mean + SE percent change in various soil parameters two years after fertilization .............................................................................................................. 99
3-6 Mean (+ SE) total basal area increase per treatment ............................................ 100
3-7 Mean (+ SE) percentage of trees that grew per plot ............................................ 101
3-8 Box plots of relative growth rates (RGR) measured between 0.4 and 2.7 yrs after initial fertilization, in the four nutrient addition treatments ............................. 102
3-9 Box plots showing relative growth rates for six common tree species, measured between 0.4 and 2.7 yrs. after initial fertilization, in the four nutrient addition treatments............................................................................................................. 103
3-10 Mean (+ SE) percent of seedlings ....................................................................... 104
3-11 Mean (+ SE) foliar litterfall production (foliage + sticks <2 mm in diameter) ........ 105
3-12 Mean (+ SE) coarse litterfall production .............................................................. 106
3-13 Mean (+ SE) fine root biomass (roots <2mm diameter) ...................................... 107
13
3-14 Box plot (25th percentile, median and 75th percentile) for root production from ingrowth cores ..................................................................................................... 108
4-1 Mean (+ SE) foliar N and P for trees with DBH between 5-10 cm ........................ 136
4-2 Box Plot comparing variability in N:P ratios at EARTH forest before fertilization and 2yrs after fertilization .................................................................................... 137
4-3 Mean (+ SE) foliar %N for six common tree species for the four nutrient addition treatments ............................................................................................................ 138
4-4 Mean (+ SE) foliar P for six common tree species for the four nutrient addition treatments ............................................................................................................ 139
4-5 Mean (+ SE) foliar N:P ratios for six common tree species for the four nutrient addition treatments .............................................................................................. 140
4-6 Relationship between several soil variables and plot-averaged foliar %N ............ 141
4-7 Relationship between several soil variables and plot-averaged foliar P ............... 142
4-8 Mean (+ SE) litterfall chemistry ............................................................................. 143
4-9 Mean litterfall nutrient concentrations by taxa ....................................................... 144
4-10 Mean (+ SE) root chemistry. ............................................................................... 145
5-1 Diagram representing how environmental and biological factors could interact to influence nutrient cycling in a diverse tropical forest ............................................. 155
A-1 Mean (+ SE) transmitted diffuse light for the four nutrient addition treatments ..... 158
A-2 Hemispherical canopy photograph taken in the same position at three successional dates ............................................................................................... 159
B-1 Box plots showing stem diameter increase (calculated as relative growth rate, see methods) for the four fertilizer treatments ...................................................... 165
B-2 Box plots showing stem diameter increase (calculated as relative growth rate, see methods) for the most common canopy palm, Socratea exorrhiza ................ 166
B-3 Relationship between initial diameter at breast height (DBH) and stem diameter increase ................................................................................................................ 166
C-1 Relationship between several soil variables and plot-averaged foliar N:P ratios .. 172
C-2 Relationship between soil Total P and Melich P ................................................... 173
14
LIST OF ABBREVIATIONS
BA Basal area (m2)
DBH Diameter at breast height (1.3 m above ground)
DIN Dissolved Inorganic Nitrogen (ug N g-1)
IVI Importance value index. Measured from the relative frequency, the relative density and the relative basal area of a species in a plot
K Potassium
N Nitrogen
NPP Net primary productivity (Mg C ha-1yr-1)
P Phosphorus
RGR Relative growth rate (ln(mm yr-1)). Calculated as the slope from a line of all the diameter measurements for each stem
15
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
BIOLOGICAL PROCESSES INFLUENCING NUTRIENT LIMITATION IN A LOWLAND
TROPICAL WET FOREST IN COSTA RICA
By
Silvia Alvarez Clare
May 2012 Chair: Michelle Caitlin Mack Major: Interdisciplinary Ecology
Nutrient limitation by nitrogen (N) and phosphorus (P) is pervasive in most
ecosystems on Earth. Thus, increases of available nutrients due to agricultural practices
and pollution may alter key ecosystem processes, such as net primary productivity
(NPP), and carbon sequestration in soils. However, our understanding of how nutrients
influence C dynamics in tropical forests remains far from complete. During my
dissertation work, I conducted two studies (one observational and one experimental) to
explore how environmental and biological processes influence nutrient limitation in
lowland tropical forests.
In my first study, I explored patterns of soil and foliar nutrients (specifically N and
P) across nine, relatively wet, mature lowland forests in Costa Rica. My objective was to
investigate the relationship between rainfall and plant or soil nutrients to better
understand the potential long-term effects that alterations in MAP could have on the
nutrient dynamics of wet forest plant communities. Across the gradient, soil N was
relatively more abundant than P but was also more sensitive to changes in MAP. I
concluded that complex feedbacks and interactions among environmental and biological
factors make it difficult to predict, by conducting observations on current patterns, how
16
changes in climate (e.g., MAP) or plant resources (e.g., nutrient availability) will
influence nutrient limitation in these forests.
In my second study, I conducted a fertilization experiment to directly test if N or P
availability limit NPP in a lowland tropical forest in Costa Rica and to explore how
biological factors, such as tree species composition and size, responded to added
nutrients. There was no significant effect of either N or P fertilization on tree diameter
increase, litterfall production, or root biomass two years after initial fertilization.
However, there were interesting and contrasting responses among tree species and
size classes. Results stemming from this experiment suggest that although soils have
high N relative to P, NPP is not necessarily limited by P, or at least does not respond to
either N or P fertilization in the short term, and highlights the importance of considering
biological factors, such as species composition and life history traits, before making
generalizations regarding nutrient limitation in tropical forests.
.
17
CHAPTER 1 BACKGROUND AND MOTIVATION
Tropical forest ecosystems provide numerous vital services for people and society
such as, biodiversity preservation, biomedical exploration, water resources, food, and
fiber (Zarin et al. 2004). Additionally, tropical forests contribute 59% of global forest
vegetation, 27% of soil carbon (C) pools (Dixon et al. 1994). They also account for at
least one-third of the annual biosphere-atmosphere carbon dioxide (CO2) exchange
(Field et al. 1992, Grace et al. 2001, 2006), thus making these systems an important
component of the global C cycle.
Contemporary human activities worldwide, however, are altering tropical forests at
an alarming rate. Perturbations include not only changes in climatic variables, such as
temperature and precipitation, but also an increase in limiting resources used by plants
and microbes. Nutrient availability, for example, has been largely increased by
atmospheric nutrient deposition that results from changing global biogeochemical cycles
(Smil 2000, Tilman et al. 2001, Galloway and Cowling 2002, Galloway et al. 2004). In
recent years global mobilization of phosphorus (P) has roughly tripled compared to its
natural flows (Smil 2000). In addition, fire, land use change, and indiscriminate use of
fertilizers have significantly increased atmospheric nitrogen (N) and P inputs in the
tropics (Galloway et al. 2004, Okin et al. 2004, Mahowald et al. 2008).
Nutrient limitation, especially of N and P, is pervasive in many ecosystems on
earth (Vitousek and Howarth 1991, Elser et al. 2010, Vitousek et al. 2010), and has
been shown to be an important control on C storage and cycling in multiple ecosystems
(Elser et al. 2007). Therefore, human-caused increases in nutrient availability could
have profound effects on tropical forest ecosystems by altering how nutrients influence
18
C accumulation through net primary productivity (NPP). However, we have an
incomplete understanding of the mechanisms that link nutrient limitation and availability
with C storage and cycling in tropical forests.
Traditionally, it has been assumed that in tropical forests, NPP is limited by P
(Figure 1-1, Walker and Syers 1976, Vitousek 1984, Vitousek and Farrington 1997).
Tropical forests usually occur in old soils, where the original P-rich parent material has
been weathered and most of the remaining P is occluded on iron and aluminum oxides
(Sanchez 1976, Miller et al. 2001). Nitrogen, by contrast, accumulates over time through
biological fixation, and is therefore relatively more available than P in old soils.
However, multiple abiotic and biotic factors can influence nutrient availability, so that not
all tropical forests occurring in old soils are limited only by P. In fact, direct evidence
from the few existing fertilization experiments conducted in the tropics (Table 1-1) and
indirect evidence from other biochemical and biological parameters (e.g., Reich and
Oleksyn 2004, McGroddy et al. 2004, Cleveland et al. 2011) suggest that N, P, or other
nutrients (e.g., potassium; Wright et al. 2011) can limit NPP in tropical forests. In this
dissertation, I explored how environmental and biological processes influence nutrient
limitation in moist-to-wet lowland tropical forests in Costa Rica. I used two
complementary approaches, consisting of an observational gradient study and an
experimental nutrient manipulation.
In Chapter 2, I present results from a study concerning patterns of soil and foliar
nutrients (specifically N and P) across nine, relatively wet (mean annual precipitation
(MAP) > 3500mm), mature lowland forest sites in Costa Rica. My objective was to
investigate the relationship between rainfall, and plant or soil nutrient characteristics to
19
better understand the potential long-term effects that alterations in mean annual
precipitation (MAP) could have on the nutrient dynamics of wet forest plant
communities. A gradient approach was a good alternative for this study because it
provided an integrated view of the long term effects that precipitation regimes can have
on ecosystem processes. I focused on the high-end of the precipitation spectrum (MAP
from 3500 to 5500 mm) because limited information exists concerning the
biogeochemical effects of abundant precipitation in wet tropical forests (Wieder et al.
2009). In areas with abundant precipitation, high water inputs can cause removal of
mobile nutrients in the soil solution via leaching (Radulovich and Sollins 1991), reduced
mineralization of nitrogen (N) and phosphorus (P) in poorly drained soils or anaerobic
microsites (Schuur and Matson 2001), slower decomposition of organic matter (Schuur
2001), and more intense weathering over time (Walker and Syers 1976, Crews et al.
1995). As a result, areas with high precipitation, where nutrients have been either
leached or occluded, can have soils with low nutrient supply rates. In this study, I tested
if patterns of soil and foliar nutrient concentrations followed patterns of MAP regardless
of co-varying factors such as altitude and species composition.
To complement the gradient study, and to learn more about the mechanisms
driving the patterns of soil and foliar nutrients observed across the nine Costa Rican
forests, I conducted one of the few fertilization experiments existing in the lowland
tropics to date. This experiment was established at the EARTH (Escuela de Agricultura
de la Región del Trópico Húmedo) Forest Reserve, in the lowlands of the Caribbean
slope of Costa Rica. As part of this experiment, I tested if N or P availability limits NPP
in this forest and explored how biological factors, such as tree species composition and
20
size, responded to added nutrients. Although this study included a significantly smaller
area than the gradient study and explored the effect of nutrient additions in a short time
scale, it provided drastically more detailed information than the gradient study on the
effects that changes in nutrient availability can have on tropical forest processes.
As a theoretical background to formulate hypotheses in the fertilization
experiment, I utilized a mechanistic model proposed by Vitousek (2004), which stems
from long-term experiments in the Hawaiian archipelago (Figure 1-1). According to this
model, in a system limited by P (as is traditionally expected of tropical forests in old
soils), increased P availability results in a positive feedback that reinforces high nutrient
availability and causes the P-limited forest to resemble a naturally P-rich system (Figure
1-2A). Specifically, increased P availability after fertilization leads to decreased P
residence times in the forest canopy, decreased overall P-use efficiency, increased P
concentrations in plant tissues and in leaf litter, increased rates of litter decomposition,
and more rapid regeneration of P from decomposing litter. The Hawaii experiments
provide a remarkable starting point for understanding the mechanisms reinforcing
nutrient limitation in the tropics; however, these experiments occurred in montane wet
forests dominated by a single tree species, Metrosideros polymorpha. By contrast, more
diverse continental tropical forests, such as the one at EARTH, are likely to contain a
wider array of plant functional traits related to nutrient use. Consequently, plant-soil-
microbial feedbacks in diverse tropical forests may respond differently to an increase in
nutrient availability than the monodominant forests in Hawaii.
One key aspect in which monodominant and diverse forests may differ is that in
nutrient-limited, diverse continental forests, species composition and identity could
21
influence the ecosystem response to nutrient addition. For example, in a diverse P-
limited forest, the more successful and thus more abundant species are expected to
have low P metabolic demand or high P-use efficiency (Chapin, Vitousek and Van
Cleve 1986). At the same time, these successful (i.e., abundant) species are expected
to contribute to the majority of the litterfall production of the system. If, for example,
these species respond to increased P availability by increasing their leaf (and litterfall)
production, but not their litterfall nutrient concentrations, then an alternative pathway
than that observed in the Hawaii experiments, may occur (Figure 1-2B). Moreover, if the
dominant species due to their low P metabolic demand do not respond to P fertilization,
no positive feedback to nutrient addition would be observed. In consequence, even if
the rest of the species in the community are strongly P limited and respond to P
additions, their response to fertilization will have no effect on community dynamics in
the short term (i.e., no feedback could be observed) because their litterfall mass is not
enough to dominate the plant community average. In the long-term, however, if P soil
availability remains high there could be a shift in species composition stemming from
differential regeneration of species with higher P demand.
In synthesis, nutrient limitation in diverse tropical forests is framed by individual
species and their interactions. Therefore, it is critical that species composition and life
history traits are considered in studies investigating nutrient limitation in diverse forests.
One innovative aspect of this dissertation is the exploration of taxa-specific responses
of some common tree species, with the objective of elucidating the role that individual
species have on the “community wide” signal of nutrient limitation.
22
Another understudied aspect of nutrient limitation in tropical forests is the
relationship between different demographic groups and nutrient cycling. In
monodominant forests (e.g., Hawaii), population growth rates (λ) will only be determined
by the vital rates of a single tree species and thus changes in nutrient availability will not
have consequences for community composition and biodiversity. By contrast in diverse
tropical forests, alterations in vital rates of different species due to changes in nutrient
availability can lead to shifts in community composition (Ceccon et al. 2004), which in
turn affect plant-soil-microbial feedbacks. Additionally, there may be resource
partitioning in space and time among age groups and different age groups may be
limited by different nutrients. For example adult trees, which intercept the majority of
light reaching the canopy and possess greater root area, may assimilate more nutrients
than light-limited saplings and seedlings (Lambers et al. 1998). Seedlings in the
understory may access nutrients added in fertilizer, but may be unable to incorporate
them in their tissues due to light co-limitation (Burslem et al. 1995). In this dissertation I
tested the effect of fertilization on stem diameter increase for trees from different size
classes, including measurements on tree seedlings.
In Chapter 3, I report the effects of a two-year N and P fertilization experiment on
different components of NPP, including tree diameter increase, litterfall productivity, root
productivity, and seedling growth and survival. In Chapter 4, I report the effects of
fertilization on nutrient (N and P) concentrations of leaves from the most common tree
species, of litterfall, and of roots collected during the experiment. Finally, in Chapter 5 I
provide some overall conclusions and discuss lessons learned and potential further
directions.
23
Overall, my dissertation work encompasses a comprehensive investigation of
nutrient limitation in the humid-to-wet lowland tropical forests of Costa Rica. Soil and
foliar nutrient data from the sites included in the gradient study provide a useful baseline
for future investigations in these sites. In addition, by conducting one of the few factorial
fertilization experiments in a lowland tropical wet forest, I hope to advance the general
understanding of plant-soil-microbial feedbacks in these systems and the role that these
feedbacks play for carbon sequestration and biodiversity preservation.
24
Table 1-1. Fertilization experiments conducted in tropical forests where components of net primary production were measured. Trt. = treatment, n.s = non-significant results, TSP = Triple Super Phosphate, ky = 1,000 yrs.
Source Duration (yrs)
Plot size and replicates
Trt. Amount (Kg Ha
-1
yr-1
)
Component of primary production Location Comments
Wood Fine roots Fine litter Coarse debris
Tanner et al. 1992
4.5yr 12x12 m x 5 reps
+N +P +NP
225/150 (urea) 75/50 (TSP)
No data No data Increased
No data n.s. n.s. Increased
No data Montane forest, Venezuela
Changed fertilizer quantities after 2 yrs
Mirmanto et al. 1999
Litterfall =1 yr Girth = 5 yrs
50 x 50 m x 5 reps
+N +P +NP
225 (urea) 75 (TSP)
n.s. n.s. n.s.
No data Increased Increased Increased
No data Lowland primary forest, Kalimantan, Borneo
Dipterocarp dominated forest
Newbery et al. 2002
2 yrs +P n.s. No data No data Cameroon
Hawaii (Summarized in Vitousek 2004)
variable Tree-centered 10 x 10 m or 15 x 15 m reps 4-6
+N +P +T* +NT +NP +PT +NPT
100 (NH4NO3 + urea) 100 (TSP)
Data exist for all these fractions but responses varied among sites and experiments.
Montane forests, Hawaii
Parent material ranging in age from 0.3 ky to 4100 Ky
Davidson et al. 2004
3 yrs (pre- fert. + 2 post-fert.)
20 x 20 m x 3 reps
+N +P +NP
100 (urea) 50 (TSP)
Increased
n.s
Increased
No data No data No data 6 yr-old secondary forest, Amazon basin, Brazil
25
Table 1-1. Continued.
Source Duration (yrs)
Plot size and replicates
Trt. Amount (Kg Ha
-1
yr-1
)
Component of primary production
Location Comments Wood Fine roots Fine litter
Coarse debris
*T = Other elements essential to plant growth, except N and P. These are Ca, Mg, K, S, Fe, Mn, Zn, Cu, B, Mo. For quantities and forms of elements added, refer to Vitousek 2004, Table 5.1)
Wright et al. 2011
11 yrs 40 x 40 m
x 4 reps +N +P +K +NK +NP +PK +NPK
125 (urea) 50 (TSP) 50 (KCl)
+P: Less decrease (large trees) +NK: Less decrease (small trees)
+K: Decrease
+P: small Increase
No data >200 yr secondary seasonal forest, Panama
There was a general decrease in growth over time but the decrease was less in several treatments
26
Figure 1-1. Diagram showing the evolution of nitrogen (N) and phosphorus(P) availability through soil development. Modified from Walker and Syers (1976).
Addition
of limiting
Nutrient
+ Nutrient
Availability -MRT
- NUE
+ Foliar
nutrient
concentration
+ Decomposition
rate
+ Litter nutrient
concentration
+ Regeneration
of nutrients
A + Nutrient
Availability
= MRT
= NUE
= Foliar
nutrient
concentration
= Decomposition
rate
= Litter nutrient
concentration
+ Regeneration
of nutrients
Addition
of limiting
Nutrient
+ Litterfall pool
+ Foliage
production
B
Figure 1-2. Plant-soil- microbial feedbacks after addition of the limiting nutrient. (A)
Evidence from Hawaii (Vitousek 2004). (B) Potential alternative feedback for a diverse tropical forest (this study) if the dominant species increases productivity, but not foliar nutrient concentrations, after P addition of the limiting nutrient. “MRT” refers to mean residence time and “NUE” to nutrient use efficiency. A plus sign indicates an increase, a minus sign a decrease, and an equal sign no change in the specific process.
N
27
CHAPTER 2
INFLUENCE OF PRECIPITATION ON SOIL AND FOLIAR NUTRIENTS ACROSS
NINE COSTA RICAN FORESTS
Introduction
Within the tropics, precipitation is expected to be one of the most important
variables affected by global climate change (Enquist 2002, Douville et al. 2006, Neelin
et al. 2006, IPCC 2007). Indeed, alterations in frequency, intensity and seasonality of
rainfall have the potential to affect multiple aspects of tropical ecosystems. Alterations in
precipitation can affect some ecosystem processes, such as forest regeneration (Sack
and Grubb 2002, Bunker and Carson 2005), net primary productivity (NPP; Clark et al.
2001), and litter decomposition (Powers et al. 2009, Wieder et al. 2009) within a short
time scale. Other consequences of altered precipitation regimes, however, are likely to
take decades to millennia to become evident. For example, changes in tree species
composition (Engelbrecht et al. 2007), biodiversity (Bazzaz 1998), soil development
(Walker and Syers 1976) and multiple aspects of nutrient cycling (Miller et al. 2001,
Schuur and Matson 2001, Vitousek 2004) usually occur over long periods.
Two approaches to investigate how changes in precipitation affect ecosystem
processes include water-manipulation experiments and observations across natural
precipitation gradients. There are benefits and limitations to both approaches.
Manipulation experiments, such as water addition (Yavitt and Wright 2008) or
precipitation exclusion (Nepstad et al. 2002), have the advantage of changing only one
or few factors at a time (e.g., water and nutrients), and are appropriate to investigate
short-term effects of precipitation. These experiments, however, are expensive, usually
This chapter has been published in Biotropica, 43: 433–441.
28
performed at a small scale, and are not appropriate to investigate long-term effects of
changes in mean annual precipitation (MAP). Natural precipitation gradients present an
interesting alternative to explore the long term effects of MAP on ecosystems. Ideally, in
a gradient study all “state factors” (climate, parent material, topography, time and biota;
sensu Jenny 1941), are kept constant, except for the factor being studied. This is mostly
the case in the Hawaiian archipelago, where multiple studies using MAP gradients have
been conducted (Austin and Vitousek 1998, Schuur and Matson 2001, Idol et al. 2007,
Houlton et al. 2007). One limitation of gradient studies, however, is that in most non-
island systems, multiple state factors vary systematically, making it challenging to
separate direct effects of MAP, for example, from other environmental factors. Despite
these limitations, gradient studies are useful because they provide insight on the long
term effects that precipitation regimes can have on ecosystem processes and can help
predict the effects that changes in MAP will have on ecosystems.
In this study, I used a precipitation gradient in Costa Rica to investigate the
potential long term influence that alterations in precipitation could have on nutrient
cycling and nutritional status of diverse- forest communities. I attempted to keep other
environmental factors constant by selecting nine, predominantly lowland, mature- forest
sites where altitude ranged from 200 to 1200m elevation. I focused on the high-end of
the precipitation spectrum (MAP from 3500 to 5500 mm) because limited information
exists concerning the biogeochemical effects of the abundant precipitation characteristic
of wet tropical forests (Wieder et al. 2009). In areas with abundant precipitation, high
water inputs can cause removal of mobile nutrients in the soil solution via leaching
(Radulovich and Sollins 1991), reduced mineralization of nitrogen (N) and phosphorus
29
(P) in poorly drained soils or anaerobic microsites (Schuur and Matson 2001), slower
decomposition of organic matter (Schuur 2001), and more intense weathering over time
(Walker and Syers 1976, Crews et al. 1995). As a result, areas with high precipitation,
where nutrients have been either leached or occluded, can have soils with low nutrient
supply rates.
Foliar nutrient concentrations (especially N and P) can track soil fertility and
therefore can be good indicators of the nutritional status of an ecosystem. For example,
in a study across a precipitation gradient in Hawaii, Schuur and Matson (2001) reported
a decrease in soil and foliar N concentrations (but not P) with increased precipitation.
Increased precipitation along a gradient in Panama was associated with decreased
foliar P, calcium and magnesium (Santiago et al. 2004). Relative concentrations of foliar
nutrients can also provide insight regarding nutrient limitation in a system.
Stoichiometric analyses of foliar nutrient concentrations (especially N and P) have
been used as an index of nutritional status of plants (e.g., Aerts and Chapin 2000,
Vitousek 2004) and can provide insight into processes such as net primary productivity
(NPP), decomposition, nutrient mineralization, trace gas emissions, and leaching
losses. Several studies have shown that N:P ratios above 16 (mass based) are typical
of sites limited by P and values below 14 are typical of sites limited by N, although
there is a substantial amount of variation around these threshold values (Güsewell
2004). At a global scale, foliar P increases and N:P ratios decline with increasing
latitude, supporting the basic hypothesis that the tropics are relatively more P limited
than temperate zones (Reich and Oleksyn 2004). However multiple factors, in addition
to soil fertility, can influence foliar nutrient concentrations and their ratios. Two factors
30
that can strongly influence nutrient concentrations are life history strategy and
physiology. For example, potentially N-fixing leguminous species in several tropical
forests have higher foliar N concentrations than species from other families (Martinelli et
al. 2000, Townsend et al. 2007).
Nitrogen stable isotopic composition of leaves and soil is another measurement
used to indicate N sources, sinks, and losses in a system (reviewed by Handley and
Raven 1992, Högberg 1997). For example, if N is found in excess in a system, losses
as nitrate and trace gas fluxes can result in 15N enriched soils (Högberg 1997).
Because plants take up N from the soil and incorporate it into their tissues, foliage 15N
signatures can broadly trace soil signatures. Thus, forests with more open N cycles can
have soils and foliage with higher 15N signatures (Garten 1993, Martinelli et al. 1999).
At a global scale, as MAP increases 15N signatures for both soils and plants decrease
(Amundson et al. 2003). In the present study, I tested if patterns of soil and foliar
nutrient concentrations followed patterns of MAP regardless of co-varying factors such
as altitude and species composition. Specifically, I addressed the following questions:
(1) Do sites with higher precipitation have relatively lower soil nutrient concentrations?
(2) Do patterns of foliar nutrients track patterns of soil nutrients? (3) Do foliar nutrients
differ among genera or functional groups? and finally (4) Do other environmental factors
(e.g., altitude) interact with MAP to influence patterns of soil and foliar nutrients? I
predicted that sites with higher MAP would have lower soil nutrients (specifically N and
P). I expected greater differences in foliar nutrients (specifically N and P) among genera
within a site than in community-averaged foliar nutrients among sites. Within genera,
31
however, I expected that foliar nutrients would track measurements of soil nutrient
availability.
Methods
Study Site
In June and July 2006 I sampled nine forest sites in Costa Rica (Figure 2-1 and
Table 2-1) ranging from 3500 – 5500 mm MAP and from 100-1200 m in altitude.
Vegetation at all sites was mature tropical forest but physical and historical factors
varied, such as geology, topography, and degree of human intervention (Table 2-1). Our
study sites were: Universidad EARTH Forest Reserve (Earth), Instituto Tecnológico de
Costa Rica forest plots at Mogos (Mogos), forest fragment near La Palma (La Palma),
La Selva Biological Station (La Selva), Rancho Mastatal (Mastatal), private forest
reserve in Dos Brazos de Río Tigre de Osa (Dos Brazos), Alberto Jimenez Forest
Reserve in San Ramón (San Ramón), Parque Nacional Tapantí (Tapantí), and La
Gamba Reserve in Golfito (Golfito). In each site I walked trails until I found at least three
mature individuals from five common tree species that were likely to be found in all (or
most) study sites and that included at least one leguminous species. I chose replicate
trees from the same species that were separated from each other by at least 20 m. I
then collected foliage from these trees and soil samples from the surrounding area.
Soil Sampling and Analysis
At each site, I randomly collected five to eight 10 cm deep soil cores within areas
where foliage samples were collected. All soil cores were separated by at least 30 m.
When a visible organic horizon deeper than 5 cm was present, I analyzed it separately
and collected a 10cm -deep mineral core starting at the depth that the organic horizon
ended. I stored soils at 4°C for ≤ 48 hours before being extracted at the La Selva
32
Biological Station. Because of the remoteness of the Mogos and Golfito sites, these
soils were stored for five days at 4°C before we could conduct extractions, which could
potentially have decreased mineralization and nitrification values (Arnold et al. 2008).
However, from the two sites for which we had to store soils for more than 48hrs, Golfito
had not only low N mineralization and nitrification but also low foliar nutrients (which
were analyzed on dried foliage), suggesting that low N mineralization and nitrification
values are real and not an analytical artifact. Mogos had substantially higher
mineralization and nitrification values, and high foliar N. Once in the laboratory, I
homogenized the bulk soil samples by separating roots, rocks, and other coarse debris.
To measure NO3- and NH4
+ I extracted 10 g of fresh soil in 50 ml 2M KCl and the
solution was measured using an Astoria Pacific colorimetric autoanalyzer (Clackamas,
Oregon, USA). I calculated dissolved inorganic nitrogen (DIN) as the sum of NO3- and
NH4+
at this initial extraction. To obtain a measurement of potential mineralization, I
incubated 10 g of soil for eight days at field moisture and at room temperature
(approximately 25 °C); then I extracted and measured NO3- and NH4
+ as above. I
calculated net N mineralization from changes in the NO3- and NH4
+ concentrations
(Riley and Vitousek 1995, Robertson et al. 1999). I calculated net nitrification as the
change in NO3- concentration per gram dry soil mass divided by the time of incubation. I
measured soil pH and available P on air-dried soils that were forced with a rubber
stopper through a 2 mm sieve. Soil pH was measured using 2:1 water: soil ratio on an
electronic pH meter (Thermo Orion 250A+, Orion Research, Inc., Boston,
Massachusetts, USA).I extracted Soil P with a Melich I solution (Kuo 1996). The
resulting P in solution was read by colorimetric determination of ortho-phosphate
33
(Murphy and Riley’s (1962) method with modified volumes for analysis in a microplate
reader) using a spectrophotometer microplate reader (PowerWave XS Microplate
Reader, Bio-Tek Instruments, Inc., Winooski, Vermont, USA). I measured total percent
N, percent C, 15N and 13C on ground soils dried at 60 °C with an elemental analyzer
(ECS 4010, Costech Analytical, Valencia, California, USA) coupled with an isotope ratio
mass spectrometer (Delta Plus XL, ThermoFinnigan, Bremen, Germany).
Foliage Sampling and Analysis
I collected fully-expanded, sun-leaf samples, using a crossbow with a bolt affixed
with monofilament line, from five common tree genera occurring at most of our sites. I
collected samples from at least nine trees from the five most common evergreen
species at each site. I tried to maximize species– or at least genus– overlap among
sites, and included at least one abundant legume in each location (Table 2-2). I dried
the collected foliage at 60°C and ground resulting samples using a Wiley Mill (Thomas
Scientific, Swedesboro, NJ, USA) passed through a #40 screen for chemistry
measurements of foliage from each tree (samples were not bulked). I measured total
percent N, percent C, 15N and 13C of foliage with an elemental analyzer (ECS 4010,
Costech Analytical, Valencia, California, USA) coupled with an isotope ratio mass
spectrometer (Delta Plus XL, ThermoFinnigan, Bremen, Germany). I measured P in
foliage samples using an ash digestion (Jones and Case 1996) followed by colorimetric
determination of ortho-phosphate using a spectrophotometer microplate reader
(PowerWave XS Microplate Reader, Bio-Tek Instruments, Inc., Winooski, Vermont,
USA).
34
Data Analysis
I used linear regressions to describe relationships among variables and one-way
analysis of variance (ANOVA) to compare variables among sites. When ANOVAs were
significant I conducted Tukey tests to compare pairs of means. I tested distributions
obtained from the nine site averages using Shapiro-Wilk tests. In most cases
distributions were normal but when this was not the case, I ln-transformed data to
improve fit. If data still deviated from normality, I used non-parametric statistics. In all
figures, site means with respective standard errors are shown. For soil variables, each
site mean results from averaging 5-8 soil cores; for foliar variables, each site mean
results from averaging foliage from 9-23 trees (Table 2-2). I calculated regressions
using means from each site (Sample size = 9). Analyses were performed using JMP IN
5.0 (SAS Institute Inc., Cary, NC, U.S.A).
Results
Soil Characteristics
I observed differences among site means for several soil parameters. For some
parameters, however, within site variability was higher than among site variability (Table
2-3). Mean N concentrations for mineral soils, for example, ranged from 0.32 percent in
Dos Brazos to 0.87 percent in Tapantí but individual core measurements ranged from
0.26 percent to 1.42 percent in Mastatal alone. Average percent C in the mineral soil
ranged from 4.13 percent at La Palma to 11.8 percent at San Ramón. However, soils in
Tapantí (the highest altitude site), presented a distinct organic (O) horizon containing up
to 33.2 percent C (Table 2-3). This organic layer had higher nutrient levels than the
other mineral soils, and was thus not included in among-site comparisons. From the N
measurements, only DIN differed significantly among sites. Overall, net nitrification rates
35
were 100 percent or more of net mineralization rates, suggesting that most mineralized
N was converted rapidly to nitrate. All sites had very low (negative) net ammonification
rates and all sites had at least one soil core where net immobilization of N occurred
after incubation (data not shown). There was no relationship between mean soil N
concentrations and net mineralization or nitrification (r2 = 0.18, P = 0.26 for
mineralization and r2 = 0.01, P = 0.99 for nitrification), suggesting that differences in
these processes were not driven by the initial N pool size. In addition, there was a
positive relationship between mean soil C and N (r2 = 0.95, P < 0.01) but not between
soil C and P (r2 = 0.15, P = 0.30) or N and P (r2 = 0.06, P < 0.52).
Precipitation was a better predictor of soil N than soil P. There was a tendency
for net N mineralization to decrease with increasing MAP but only net nitrification
decreased significantly with increased MAP (Figure 2-2). In higher altitude sites, soils
had higher C (r2 = 0.49, P = 0.04) and N (r2 = 0.59, P = 0.02) concentrations but not
higher P (r2 = 0.01, P = 0.86) concentration.
Foliar Measurements
All foliar parameters measured varied significantly among sites (Table 2-4) but
were not correlated with MAP or altitude (data not shown). There was only a marginal
correlation between mean foliar percent C and N across sites (r2 = 0.42, P = 0.058).
When observing foliar nutrients by plant functional type, legumes had higher N:P ratios
than non-legumes because legumes had higher N concentrations than non-legumes (T
= 10.48, d.f. = 137, P < 0.001; Figure 2-3); legumes and non-legumes did not differ in
percent P concentrations (T = 1.08, d.f. = 135, P = 0.28). Foliar N:P ratios varied greatly
among genera. For the four most common genera present in the majority of the study
sites (Hyeronima, Inga, Protium, and Virola), there were significant N:P differences
36
among genera and across sites (two-way ANOVA for genera and site: Model F11, 75 =
11.99, P < 0.001; effect tests: genera F3 = 34.01, P < 0.001, site F8 = 3.04, P = 0.006;
Figure 2-3B), with Inga usually presenting the highest N:P ratios and Protium the
lowest.
Relationship Between Soil and Foliar Measurements
Soil Melich-extractable P was a better predictor of foliar N and P than soil N. Soil
Melich-extractable P was positively correlated with foliar P concentrations (Figure 2-4A)
and marginally correlated with foliar N concentrations; r2 = 0.34, P= 0.059), as well as
negatively correlated with foliar N:P (Figure 2-4B). Soil total N and DIN concentrations
were not correlated with foliar N concentrations (r2 = 0.14, P= 0.321 and r2 = 0.06, P =
0.543). Nitrogen fluxes (e.g., net mineralization and nitrification) were also not good
predictors of foliar N concentrations. Isotope δ15 N values, however, were well
correlated between soils and plants (r2 = 0.84, P < 0.001). Interestingly, both study sites
located in the Caribbean slope of the country (Earth and La Selva) had soils and plants
significantly more enriched in δ15N than the rest of the sites (Wilcoxon rank test: X2 =
4.2, d.f. = 1, P = 0.040; Tables 2-3 and 2-4). These two sites also had the smallest
relative offset between soil and foliar signatures (Earth = -0.57 and La Selva = -0.29).
Soil C concentration was not correlated with any foliar measurements (data not shown).
Discussion
Patterns of Soil Nutrient Availability
Across nine forest sites in Costa Rica, there was a decrease in net N
mineralization and nitrification rates with increased MAP, although the pattern was
much stronger for nitrification than for mineralization. Reductions in net nitrification,
specifically, could be due to the lysis of aerobic microorganisms under anaerobic
37
conditions (high rainfall) or by the activity of facultative anaerobes adapting to
fluctuating oxic and anoxic conditions (Davelaar 1993, Wright et al. 2001, Rinklebe and
Langer 2006). Decreased net N mineralization with high MAP has been observed in
other tropical forests (Chandler 1985, Schuur and Matson 2001) and has also been
attributed to anaerobic conditions. However, Pandey et al. (2009) in three land-use
systems in India found higher net N mineralization (but not net nitrification) during dry
spells of the wet season compared to net N mineralization during high precipitation
periods, highlighting different controls on these two biogeochemical processes.
Although I found a decrease in both net N mineralization and nitrification with increased
MAP, the higher variability in the net N mineralization patterns suggests that net N
mineralization rates could also be influenced by other factors such as litter quality
(Austin and Vitousek 2000, Schuur 2001, Santiago et al. 2005). In addition, soil total N
and DIN concentrations did not decrease with increasing MAP, suggesting that in my
study sites, excess water from high MAP may result in a decrease in the rate of N
cycling but not in N pool sizes.
In contrast to N availability, measurements of soil P did not decrease with
increasing MAP. Potentially in the wettest sites, anoxic conditions resulting from high
precipitation can cause low redox potentials in soils, which can release mineral-sorbed
P (Miller et al. 2001, Schuur and Matson 2001, Schuur et al. 2001). Phosphorus
released by this mechanism could offset slow organic matter decomposition and
mineralization, resulting in reduced available N but not P at the wettest sites.
Alternatively, we may have failed to detect a correlation between P and MAP in this
study because we did not measure the P fractions that are most susceptible to changes
38
in MAP. For example, in a study in a precipitation gradient in Hawaii only the most
soluble P fraction, “resin-extractable P”, decreased with increasing MAP (Idol et al.
2007). In another Hawaii study, however, Miller et al. (2001) reported an increase of
labile P (defined as resin extractable P + bicarbonate extractable P) and a decline of
recalcitrant inorganic P with increasing MAP. These examples demonstrate that
different P forms may be influenced differently by MAP, making it challenging to detect
patterns. In my study, although I did not measure multiple P fractions, the strong
positive correlation between Melich-extractable P and foliar P suggests that we
captured a significant part of the “plant available” P with my analysis. Thus, I conclude
that at least this P fraction is not correlated with MAP in my gradient, although the
mechanisms for this result are unclear.
Percent C in the soil was also not correlated with MAP either. Most likely, the
lack of correlation is caused by the interaction of multiple environmental factors, such as
variation in altitude (Tanner et al. 1998), parent material (Jenny 1941), and topography
(Porder et al. 2006). For example, there was an interaction between elevation and
precipitation that caused a dramatic effect on soils in Tapantí. In this site, located at
1200 m elevation and with a MAP of 5000 mm, soils exhibited a definite organic horizon
with significantly higher C (and N and P) than the rest of the sites (Table 2-3). Thus,
Tapantí had the highest soil percent carbon and nutrient concentrations due to this
organic horizon. The distinct organic horizon in Tapantí soils is probably the result of
slow decomposition of organic matter caused by a combination of lower temperatures,
higher elevation, and lower litter quality (Tanner et al. 1998, Schuur 2001, McGroddy
and Silver 2004, Cusack et al. 2009). Surprisingly the other two high elevation sites –
39
San Ramón (elevation 1000 m; MAP 4500mm) and Mastatal (elevation 900 m; MAP
4000mm) – did not show a distinct organic horizon and thus had significantly lower soil
nutrients and C than Tapantí. From the data presented here, it is difficult to determine
the cause of this difference. I suspect, however, that a combination of higher
temperature (due to slightly lower elevation), lower MAP, and potentially higher litter
quality due to different species composition could impede the formation of an organic
horizon in San Ramón and Mastatal.
Pattern of Foliar Nutrients
Consistent with other studies, foliar P was a good predictor of soil P (e.g., Schuur
and Matson 2001, Santiago et al. 2005, Wu et al. 2007). This implies that plants are
incorporating in their tissues as much P as is available at each site, and supports the
idea that foliar concentrations are good predictors of soil nutrient availability (Vitousek
and Farrington 1997, Aerts and Chapin 2000, Vitousek 2004); moreover, it suggests
that across the study sites P may be more limiting to plant growth than N. Tapantí,
however, did not have higher foliar nutrients than the rest of the sites, even though it
had the highest concentrations of soil extractable P and N, when considering the
organic horizon. There are several possible explanations for this result. One possibility
is that high soil nutrient concentrations in Tapantí do not result in higher nutrient pools
than the rest of the sites due to relatively lower bulk density. However, we did not
measure bulk density and therefore we cannot directly test this hypothesis.
Alternatively, Tapantí may in fact have the largest soil nutrient pools but other factors,
such as floristic composition in Tapantí, may be characterized by species with inherent
low foliar nutrient concentrations (area based) as is typical in other pre-montane and
40
montane forests (Tanner et al. 1998; Table 2-2), thus not reflecting soil nutrient
availability.
In contrast to foliar P, foliar N was not a good predictor of soil N (or soil P),
suggesting that factors other than soil N availability are influencing foliar N
concentrations across the plant community. Indeed, I found higher foliar N (but not P) in
leguminous vs. non-leguminous species, with legume abundance varying among sites
(personal observation). Thus in a given ecosystem, a higher proportion of legumes can
result in a higher community-wide foliar N (regardless of MAP), which highlights the
importance that species (or functional group) composition can have for a site’s
biogeochemistry (Hooper and Vitousek 1998, Townsend et al. 2007). High foliar
concentrations of N in leguminous species have been observed in other studies (e.g.,
Martinelli et al. 2000, Townsend et al. 2007) and are typical for nodulating and non-
nodulating legumes (McKey 1994). In contrast to foliar N, foliar δ 15N was well
correlated with soil δ 15N (and with foliar %N). The two Caribbean sites –La Selva and
Earth– had soils and foliage most enriched in δ 15N. One possible explanation for this
pattern is that these two sites could have a more open N cycle, with more N losses and
thus a more positive δ 15N signature (Martinelli et al. 1999). In addition, more enriched
δ 15N signatures could reflect differences in parent material or N deposition rates
(Eklund et al. 1997). Further investigations are required to obtain a conclusive
mechanism for the observed pattern.
Across sites, foliar N:P ratios above 16 (Figure 2-3) support the general view that
P limits plant productivity in tropical forests (Vitousek 1984, Hedin et al. 2003, Reich and
Oleksyn 2004). However, when observing individual genera within sites, sizable
41
differences in N:P ratios suggest that the community averaged, foliar N:P-ratio
thresholds used in temperate systems to infer nutrient limitation may not be applicable
to tropical forests. In my study, variability among genera and functional type (legumes
vs. non-legumes) reflects the high diversity of life history traits and physiological
strategies in tropical forests. Moreover, it stresses that even though individual-taxa foliar
nutrient concentrations can provide useful information regarding multiple ecosystem
processes, community averaged N:P ratios do not always reflect nutrient limitation. In
fact, physiological properties and life history strategies may be more important than
environmental limitations in determining foliar nutrient concentrations (Ågren 2004,
Niklas et al. 2005). To confidently draw conclusions concerning the nutritional status of
a plant community based on foliar nutrient ratios, these values should be calibrated by
conducting a direct test of nutrient limitation, such as a nutrient addition experiment
(Chapin et al. 1986, Koerselman and Meuleman 1996).
Conclusions
The results from this study suggest that soil N dynamics may be more sensitive to
increasing precipitation than soil P dynamics in these tropical wet forests in Costa Rica.
However, the lack of relationship between soil N dynamics and foliar N makes the
consequences of this sensitivity uncertain. On the contrary, although factors other than
MAP may control soil P in these forests, the positive relationship between soil and foliar
P suggests that plant foliar nutrients are more responsive to changes in soil P than in N.
Large differences in foliar nutrient concentrations and N:P ratios within sites stress the
key role that plant functional group composition plays in tropical ecosystem
biogeochemistry. For example within a site, a decrease in N availability could be
irrelevant (or even beneficial) for N-fixing legumes, but detrimental for non-legume
42
species with high N metabolic requirements, such as Virola sp. that has high alkaloid
concentrations (Dominy et al. 2003). Thus, my study highlights the importance of
considering species composition and identity before making community-wide
generalizations regarding nutrient limitation in tropical forests. Furthermore, interactions
among multiple environmental and biological factors, such as altitude, temperature,
redox potential, and decomposability of soil organic matter, can influence the sensitivity
of N and P dynamics to changes in MAP in each site, making predictions extremely
difficult and inferences regarding nutrient limitation uncertain.
With climate change, several models predict a decrease in MAP for certain areas
in Costa Rica (Neelin et al. 2006). My results suggest that this decrease could lead to
an increase in available N from mineralization in the wetter sites of my gradient. More
investigation is required to reach conclusive results regarding the effect of MAP on soil
nutrient availability and the relationship between soil and foliar nutrients in these forests.
Key measurements conducted at the specific sites should include (1) meteorological
measurements, including temperature, (2) detailed soil characterization including bulk
density, cation exchange capacity, and texture, (3) investigation of parent material, and
finally (4) direct tests of nutrient limitation. Nevertheless, my study provides important
baseline biogeochemical data that can be used to improve understanding of nutrient
cycling and limitation in tropical forests, and to encourage future studies that identify the
possible consequences that alterations in MAP resulting from climate change will have
in tropical systems.
43
Table 2-1. Characteristics of study sites in Costa Rica. Altitude = meters above sea level, MAP = mean annual precipitation, Dry months = average number of months with less than 100 mm precipitation. Data from Proyecto Atlas Digital de Costa Rica 2008.
a Soil order from Proyecto Atlas Digital de Costa Rica 2008 and Sub-group from Pérez et al. (1978)
b Recently reclassified as Oxisols (Kleber et al. 2007)
Site Location Altitude
(m) MAP ( mm)
Dry months
Parent material Soil order and
Sub-groupa
Earth 10° 10’ 21.1” N 83° 36’ 9.8” W
100 3500 1 Alluvial and colluvial deposits and distal facles of modern volcanic rocks
Inceptisols and Ultisols (Oxic palehumult)
Mogos 8° 45’ 8.6” N 83° 23’ 7.1”W
100 3500 1 Sea floor basalts Ultisols (Typic tropohumult)
La Palma 8° 39’ 5.7” N 83° 23’ 36.9” W
100 4000 1 Alluvial and colluvial deposits Entisols or Mollisols (Fluvaquentic hapludoll)
La Selva 10° 26’ 24.8” N 84° 0’ 56.0” W
100 4000 1 Alluvial and colluvial deposits and distal facles of modern volcanic rocks
Inceptisols and Ultisols b
(Oxic dystropept)
Mastatal 9° 40’ 52.1” N 84° 22’ 34.0” W
900 4000 3 Deep water sedimentary rocks Inceptisols and Ultisols (Typic Dystropept)
Dos Brazos
8° 31’ 59.3” N 83° 23’ 30.3” W
200 4500 1 Deep water sedimentary rocks Ultisols (Typic tropohumult)
San Ramón
10° 9’ 4.8” N 84° 29’ 15.2” W
1000 4500 3 Volcanic intrusive rocks from the Tertiary Inceptisols (Typic dystrandept)
Tapantí 9° 45’ 59.1” N 83° 48’ 9.9” W
1200 5000 2 Volcanic intrusive rocks from the Tertiary and alluvial and colluvial deposits
Inceptisols (Andic humitropept)
Golfito 8° 38’ 39.8” N 83° 9’ 54.8” W
200 5500 1 Sea floor basalts and deep water sedimentary rocks
Ultisols (Typic tropohumult)
44
Table 2-2. Species and number of samples collected per site in nine sites in Costa Rica. See methods section for site description. DB=Dos Brazos, EAR=Earth, GOL= Golfito, LP=La Palma, LS=La Selva, MOG=Mogos, MAS= Mastatal, SR= San Ramón, TAP=Tapantí
Species DB EAR GOL LP LS MOG MAS SR TAP Total
Alchornea latifolia
2 3 5
Ardisia sp.
1
1
Billia hippocastanum
1 1
Billia rosea
1
1
Brunellia standleyana
3 3 6
Citharexylum caudatum
3 3
Dendropanax arboreus 3 3
2
3
11
Elaegia auriculata
3
3
Hasseltia quinquenervia
1
1
Hyeronima alchorneoides 3 2
3
3
11
Hyeronima oblonga
1
2
3
Inga acrocephala 1
1
Inga barbourii
3
3
Inga nobilis
1
1
Inga oerstediana 2
1 3
Inga pezizifera
1
1
Inga polita
1
1
Inga sp.
1 1
Inga sp. Leonis
1
1
Inga sp. tonduzii
1 1
Inga thibaudiana
3
1
4
Inga venusta
2
2
Inga vera
3
3
Laetia procera
4
4
45
Table 2-1. Continued
Species DB EAR GOL LP LS MOG MAS SR TAP Total
Marila pluricostata
2
2
Micropholis meloniana
2
2
Pentaclethra macroloba
1
1
2
Protium aracouchini
3
3
Protium confusum
1
1
Protium glabrum 3
1
1
5
Protium panamense
4
4
Protium pittieri
3
1
4
Protium ravenii
2
2
Rollinia pittieri
1
1
Saurauria montana
2 2
Saurauria rubiformis
1 1
Virola guatemalensis 3
3
2
8
Virola koschnyi
3 1 1 1
4
10
Virola multiflora
1
1
Virola sebifera
1
1
2
Virola surimamensis
1
1
Vochysia allenii
2
2
Vochysia ferruginea
2 2
4
8
Warzewiczia coccinea
1
3
4
Total number of samples 15 16 12 16 9 11 23 19 16 137
46
Table 2-3. Soil nutrients and isotopic signatures (mean + SE) for the nine study sites in Costa Rica shown in order of increasing mean annual precipitation (Table 2-1). Soils were sampled 10 cm deep in the mineral layer, except for Tapantí, where a distinct organic layer was present and thus organic (O) and mineral (M) horizons were separated. Soil pH was measured in water; DIN = dissolved inorganic nitrogen (nitrate + ammonium), Net min = net mineralization, and Net nit = net nitrification. Shown are also F tests from one-way ANOVA comparing sites (N = 9) with significant values bolded (* = P < 0.05, ** = P< 0.001).For significant ANOVAs, means that do not share a common letter superscript are significantly different. Organic soils were excluded from the analysis.
Site pH %C
δ13C Melich P (µg g-1)
%N
EARTH 4.77 + 0.20bc 5.17 + 0.98ab -28.36 + 0.56 1.24 + 0.34a 0.48 + 0.08
Mogos 4.86 + 0.08bc 4.92 + 0.42ab -28.01. + 0.21 0.28 + 0.04a 0.36 + 0.04
La Palma 5.32 + 0.26b 4.13 + 0.66b -27.56 + 0.13 0.42 + 0.14a 0.38 + 0.07
La Selva 4.16 + 0.04c 5.42 + 0.63ab -27.72 + 0.28 0.89 + 0.19a 0.46 + 0.04
Mastatal 5.12 + 0.17b 5.57 + 1.45ab -28.11 + 0.72 0.80 + 0.21a 0.52 + 0.13
Dos Brazos 6.25 + 0.06a 3.42 + 0.48b -28.28 + 0.29 1.51+ 0.80a 0.32 + 0.05
San Ramón 5.15 + 0.06b 11.88 + 1.74a -28.15 + 0.08 0.19 + 0.03a 0.87 + 0.14
Tapantí M 4.70 + 0.08bc 7.96 + 2.34ab -28.64 + 0.81 0.90 + 0.32a 0.63 + 0.17
Tapantí O 3.93 + 0.05 33.23 + 4.94 -29.32 + 0.42 25.67 + 11.17 2.21 + 0.34
Golfito 5.06 + 0.11b 4.65 + 0.49ab -28.44 + 0.19 0.28 + 0.03a 0.37 + 0.03
F values 12.33** 2.48* 0.29 2.44* 1.58
47
Table 2-3. Continued.
Site δ15N DIN (µg g-1)
Net N min (µg N g-1 d-1)
Net N nit (µg N g-1 d-1)
EARTH 5.01 + 0.48ab 11.92 + 0.57 3.25 + 1.06 3.77 + 1.09
Mogos 2.77 + 0.49c 14.15 + 0.69 3.74 + 1.13 4.48 + 1.68
La Palma 2.99 + 0.23bc 11.49 + 0.33 0.57 + 0.41 0.78 + 0.58
La Selva 5.53 + 0.32a 12.68 + 0.47 2.50 + 0.81 2.72 + .77
Mastatal 2.23. + 0.38c 11.42 + 0.70 3.15 + 1.90 3.12 + 1.39
Dos Brazos 2.27 + 0.56c 11.94 + 1.30 0.87 + 0.3 1.94 + 0.62
San Ramón 3.19 + 0.30abc 14.42 + 0.70 1.53 + 0.65 3.64 + 1.16
Tapantí M 2.73 + 0.64bc 13.03 + 0.64 3.00 + 1.86 1.18 + 1.67
Tapantí O 0.88 + 0.45 40.77 + 11.51 15.50 + 7.24 10.03 + 1.93
Golfito 2.12 + 0.17c 12.92 + 0.40 0.73 + 0.55 1.05 + 0.16
F values 7.72** 2.09 0.71 0.92
48
Table 2-4. Foliar carbon, phosphorus, nitrogen and stable isotopic signatures (mean + SE) for the nine study sites in Costa Rica. For complete list of collected specimens refer to Table 2-2. Shown are also F tests from one-way ANOVA comparing sites (N = 9); significant values are bolded (* = P < 0.05, ** = P < 0.001).Within a column, means that do not share a common letter superscript are significantly different
Site %C δ13C %P %N δ15N
Earth 48.65 + 0.49a -31.92 + 0.25bc 0.13 + 0.01a 2.27 + 0.12ab 2.16 + 0.19b
Mogos 45.37 + 0.79bc -32.72 + 0.33c 0.09 + 0.01bc 1.88 + 0.14ab -0.75 + 0.26cd
La Palma 46.72 + 0.59abc -31.17 + 0.4bc 0.09 + 0.01bc 2.08 + 0.13ab -1.51 + 0.34de
La Selva 49.44 + 0.32a -30.41 + 0.39ab 0.12 + 0.01ab 2.5 + 0.15a 3.91 + 0.37a
Mastatal 46.81 + 0.59abc -31.66 + 0.45bc 0.11 + 0.01ab 2.27 + 0.09ab -1.42 + 0.29de
Dos Brazos 45.75 + 0.58bc -31.81 + 0.34bc 0.14 + 0.01a 2.41 + 0.13a 0.16 + 0.27c
San Ramon 45.39 + 0.4bc -31.36 + 0.33bc 0.12 + 0.01ab 2.3 + 0.18ab -0.82 + 0.38cd
Tapantí 47.63 + 0.46ab -29.57 + 0.24a 0.14 + 0.01a 2.32 + 0.16ab -0.46 + 0.40cd
Golfito 44.83 + 0.90c -31.60 + 0.32bc 0.07 + 0.01c 1.67 + 0.11b -2.60 + 0.17e
F values 6.01** 6.05** 9.78** 2.85* 28.34**
49
Figure 2-1. Map of sites where soil and foliar samples were collected in Costa Rica.
Mean annual precipitation is shown in shades of gray. Precipitation data obtained from Proyecto Atlas Digital Costa Rica (2008).
50
Ne
t N
Min
era
liza
tio
n (
g N
/g/d
)
0
1
2
3
4
5
6
MAP (m)
3.0 3.5 4.0 4.5 5.0 5.5 6.0
So
il P
(m
g/k
g)
0.0
0.5
1.0
1.5
2.0
2.5
3.0Earth
Mogos
La Palma
La Selva
Mastatal
Dos Brazos
San Ramon
Tapanti
Golfito
R2 = 0.07 P = 0.452
A
B
R2 = 0.31 P = 0.095
C
R2 = 0.63 P = 0.010
+ P
recip
ita
tio
n
Net
Nitrification (g N
/g/d
)
0
2
4
6
8
B R2 = 0.63 P = 0.010
Figure 2-2. Relationship between mean annual precipitation (+ SE) and soil
parameters, including (A) average net nitrogen mineralization rates, (B) net nitrification rates, and (C) soil Melich I phosphorus for the top 10 cm mineral layer of nine study sites in Costa Rica. Circles represent lowland sites (< 900 m) and triangles are mid-elevation sites (900-1200 m).
51
N:P
0
10
20
30
40
50
60
Legumes
Non-legumes
F1,135 = 84.9, P < 0.001 A
Earth
Mog
os
La P
alm
a
La S
elva
Mas
tata
l
Dos
Bra
zos
San R
amon
Tapan
ti
Golfito
N:P
0
10
20
30
40
50
60
Hyeronima (2 sp.)
Inga (13 sp.)
Protium (6 sp.)
Virola (5 sp.)
B Genera: F3 = 34.01, P < 0.001
Site: F8 = 3.04, P= 0.006
- Precipitation +
Figure 2-3. Mean N:P ratios (+ SE) for (A) legumes and non-legumes and (B) for the
most common genera (including the leguminous genus Inga) in nine study sites in Costa Rica. The number of species collected for each genus is shown in the legend. Dashed line represents the suggested threshold for N vs. P limitation suggested in the literature (Aerts and Chapin 2000, Güsewell 2004, Reich and Oleskyn 2004). For details of which species were collected at each site refer to Table 2-4. Sites are arranged by increasing MAP.
52
Folia
r %
P
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0.20
Soil P (ug/g)
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Folia
r N
:P
14
16
18
20
22
24
26
28
A
B
R2 = 0.58 P = 0.017
R2 = 0.44 P = 0.051
Pre
cip
itatio
n +
Earth
Mogos
La Palma
La Selva
Mastatal
Dos Brazos
San Ramon
Tapanti
Golfito
Figure 2-4. Mean soil phosphorus (+ SE) for the top 10 cm mineral layer of nine study sites in Costa Rica plotted against (A) mean (+ SE) foliar P and (B) mean (+ SE) foliar N:P ratios. Circles represent lowland sites (< 900 m) and triangles are mid-elevation sites (900-1200 m).
53
CHAPTER 3 DIRECT TEST OF NUTRIENT LIMITATION TO NET PRIMARY PRODUCTIVITY IN A
LOWLAND TROPICAL WET FOREST
Introduction
Nutrient availability controls key processes in all ecosystems on earth. Net
primary productivity (NPP), nutrient use efficiency (NUE) by plants, and nutrient
turnover through decomposition, are all processes affected by nutrient availability.
Nitrogen (N) and phosphorus (P), either individually or in combination, limit primary
productivity in most terrestrial ecosystems (Vitousek and Howarth 1991, Elser et al.
2007, Vitousek et al. 2010). In turn, plant adaptations to limitation by these nutrients
feed back strongly to control ecosystem rates of nutrient cycling (Chapin 1980, Hobbie
1992, Vitousek 2004). Most tropical forests occur on old soils, where P-rich parent
material has been weathered and most of the remaining P is occluded on iron and
aluminum oxides (Sanchez 1976, Miller et al. 2001). Nitrogen, by contrast, accumulates
over time through biological fixation, and is therefore expected to be relatively more
available than P in old, weathered soils. Thus, it is generally believed that NPP is limited
by P in these systems (Walker and Syers 1976, Vitousek 1984, Vitousek and Farrington
1997).
However, multiple abiotic and biotic factors can influence nutrient availability, so
that not all tropical forests occurring in old soils are limited by P. In montane tropical
forests, for example, erosion rates caused by topographic variation are sufficient to
provide renewal of solum P from weathering parent material. Thus, NPP in these forests
is generally limited by N, instead of P (reviewed by Tanner et al. 1998, Porder et al.
2006). Other factors, such as precipitation (Schuur 2003, Santiago et al. 2005, Alvarez-
Clare and Mack 2011), disturbance (Davidson et al. 2004), and life-form diversity
54
(Hiremath and Ewel 2001) can also influence nutrient availability in tropical forests,
resulting in limitation by nutrients other than P. In fact, data from the few fertilization
experiments conducted in the tropics, revealed that N limitation is pervasive among
tropical forests (Le Bauer and Treseder 2008), although this sample was dominated by
tropical montane forests.
Historically, our conceptual understanding of nutrient limitation was one derived
from Liebig’s law of the minimum, where the single scarcest nutrient in relation to plant
demand, usually N or P, was the most limiting (Liebig 1842). However, recent
investigations showing synergistic interactions between limited supplies of N and P are
widespread across aquatic and terrestrial systems (Elser et al. 2007, Harpole et al.
2011), and evidence of limitation by other nutrients, such as potassium, is emerging
(e.g., Kaspari et al. 2008, Wright et al. 2011). Consistent with these findings, nutrient
limitation in tropical forests is probably not an N versus P question, but instead one that
includes complex interactions among nutrient cycles, and their linkages with biological
processes. For example, N inputs have been shown to accelerate phosphorus cycling
rates through enhancement of soil and root phosphatase activity (Marklein and Houlton
2011), and tree species composition is known to influence rates of nutrient turnover
through decomposition (Cornwell et al. 2008, Wieder et al. 2009).
The influence of biological processes on nutrient cycling is expected to be of
particular importance in lowland tropical forests, where there is a high diversity of flora
(Losos et al. 2004). Here, a wide variety of tree functional traits related to nutrient
acquisition and use is likely to influence biogeochemical processes, creating complex
linkages with nutrient cycles (Townsend et al. 2008). For example, large interspecific
55
differences in foliar nutrient concentrations and resulting litter quality can influence
plant-soil microbial feedbacks related to nutrient use (Vitousek 2004, Wood et al. 2011).
In a lowland tropical forest in the south of Costa Rica, tree species variation in foliar P
and carbon (C) chemistry were associated with tree-specific differences in both free-
living N fixation and soil respiration (Reed et al. 2007, 2008, Wieder et al. 2008) In
addition, differences among species in the functional use of resources (e.g., nitrogen
fixers vs. non-fixers, shade tolerant vs. light demanding, species with superficial roots
vs. species with deep roots) can enable them to mitigate nutrient limitation by accessing
different nutrient pools (Vance 2003).Therefore, to understand nutrient limitation in
lowland tropical forests it is imperative to consider the influence of species-specific
effects on nutrient cycling.
Another understudied aspect of nutrient limitation in tropical forests is the
relationship between different demographic groups and nutrient cycling. In
monodominant forests (e.g., Hawaii, summarized by Vitousek 2004)), population growth
rates (λ) will only be determined by the vital rates of a single tree species and thus
changes in nutrient availability will not have consequences for community composition
and biodiversity. By contrast in diverse tropical forests, alterations in vital rates of
different species due to changes in nutrient availability can lead to shifts in community
composition (Ceccon et al. 2004), which in turn affect plant-soil-microbial feedbacks.
Additionally, there may be resource partitioning in space and time among age groups
and different age groups may be limited by different nutrients. For example adult trees,
which intercept the majority of light reaching the canopy and possess greater root area,
may assimilate more nutrients than light-limited saplings and seedlings (Lambers et al.
56
1998). Seedlings in the understory may access nutrients added in fertilizer, but may be
unable to incorporate them in their tissues due to light co-limitation (Burslem et al.
1995).
To date, only two fertilization experiments have been conducted in lowland tropical
forests to directly test nutrient limitation of NPP. One found evidence for N and P co-
limitation (Mirmanto et al. 1999), and one found evidence of limitation by N, P, and K
(Wright et al. 2011). These results differ from the traditional view that lowland tropical
forests are P limited. Moreover, the study by Wright et al. (2011) reported that in the
lowland tropical forest in Panama where their study was conducted, different tree age
groups and different fractions of NPP were limited by different nutrients. None of these
studies explored the effects of tree species composition and identity on the observed
responses to nutrient additions. The scarcity of fertilization experiments in lowland
diverse tropical forests, the conflicting results obtained from these few existing studies,
and the critical role that these forests have on the global C cycle motivated my
research. I conducted a factorial fertilization experiment with N and P additions, to
directly test nutrient limitation in a lowland tropical wet forest in Costa Rica.
My first objective was to test if N and P limited various components of NPP.
Because theory (Walker and Syers 1976) and indirect evidence from stoichiometry
(Hedin 2004, McGroddy et al. 2004, Reich and Oleksyn 2004) and soil microbial
dynamics (Cleveland and Townsend 2006) suggest that lowland tropical forests in
nutrient poor, clayey soils are P-limited, I hypothesized that P fertilization would
increase productivity by enhancing stem diameter growth and fine litterfall production.
Considering that fine litterfall has responded faster to fertilization than wood increment
57
in other studies (Mirmanto et al. 1999, Wood et al. 2009 but see Tanner et al. 1992),
and that litterfall constitutes the major component of productivity in tropical lowland
forests (Clark et al. 2001a), I expected NPP to increase mainly due to litterfall
production.
My second objective was to investigate the effect of fertilization on stem diameter
increase for trees from different size classes. Consistent with Wright et al. 2011, I
expected different age groups to respond differently to nutrient additions. Large trees (>
10 cm diameter at breast height (DBH)) usually have slow growth rates (Lieberman and
Lieberman 1994) or may utilize extra nutrients for reproduction instead of growth, and
small trees (5-10 cm DBH) are usually light limited (Lambers et al. 1998). Thus, I
predicted that intermediate sized trees (10-30cm DBH)) would have the largest
response to fertilization (mainly with P) because these trees have moderate access to
light but must reach the canopy to obtain full sun. These trees, therefore, would benefit
from investing extra nutrients to stem growth. Small trees and seedlings in the
understory may access nutrients added in fertilizer, but may be unable to incorporate
them in their tissues as biomass due to light co-limitation (Burslem et al. 1995). I
hypothesized no difference in stem diameter increase for small trees and no difference
in stem length or number of leaves in seedlings with increased nutrient availability.
My third objective was to explore the effect of fertilization on stem diameter
increase for trees from different taxa. In my experiment, I used four species and three
genera that were present in all treatments and that encompassed a variety of life history
traits to explore how fertilization affected diameter increase in different taxa (Table 3-1).
Because fast-growing, light-demanding species acclimate faster to increases in light or
58
nutrient regimes (Lambers et al. 1998), I predicted that fast-growing canopy species
(e.g., Pentaclethra macroloba and Goethalsia meiantha) would demonstrate greater
diameter increases after fertilization (especially with P) than slow-growing, shade-
tolerant trees (e.g., trees from the genus Virola). In addition, I predicted that subcanopy
trees with no access to appreciable light (e.g. Dendropanax arboreus or trees from the
genus Protium) would not demonstrate differences in diameter increase.
Methods
Site Description
The study was conducted at the Forest Reserve of the EARTH University (Escuela
de Agricultura de la Región del Trópico Húmedo), in Guácimo, Limón, Costa Rica (10°
11’ N and 84° 40’ W). This private reserve is located approximately 30 m above sea
level and consists of 900 ha of mature and regenerating wet forest and wetlands. Mean
annual temperature is 25.1 °C and mean annual precipitation (MAP) is 3,464 mm,
distributed in a bimodal pattern with peaks traditionally occurring in July-August and
November-December; March is the driest month with a MAP of 124 mm. Relative
humidity is above 80% during the year. An early ecological map (Tosi 1969) classifies
the EARTH forest within two life zones (sensu Holdridge 1971): tropical wet forest and
tropical moist forest with transition to wet. Throughout the duration of my study, MAP
averaged 3,844 mm ranging from 3,580 mm in 2007 to 4,097 in 2009. In 2009,
precipitation was unusually high during February and March (Figure 3-1). Mean
temperature between 2007 and 2009 was 24.8 °C, ranging from 23.2 °C in 2009 to 25.1
°C in 2008.
According to Sancho et al. (1990), the study area is located in the distal section
of a coalescence of alluvial fans. Parent material is from volcanic origin, mainly lava
59
flows and mud flows (lahars), boulders and pyroclastic materials. The undulating
topography is characterized with few eroded slopes and depressions where the eroded
material has been deposited. Soils in the area are mainly described in two units: 1)
Typic Dystrandepts, which are deep soils from volcanic origin, with low base saturation
and that remain humid throughout the year, or 2) Oxic Palehumult and Aeric
Tropaquept, described as deep, reddish soils that occur on old alluvial terraces and
suffer poor drainage. Detailed soil information exists for the rest of the EARTH property
but not for the forest reserve (Sancho et al. 1990).
Soil texture was analyzed in two ways. Texture analyzed in the traditional way with
a dispersant agent (Bouyoucous 1950) revealed that soils at the ERATH Forest
Reserve are clayey, with approximately 50% clay, 20% silt and 30% sand (Alvarez-
Clare, unpublished data). Results obtained with a Masterizer Particle Size Analyzer
revealed that most of that “clay” is actually very fine or fine silt with particle sizes
between 2-15.6 µm (Table 3-1). Overall, both total nitrogen (N) and phosphorus (P)
measurements are relatively high (Table 3-2), as well as most total micronutrients
(Table 3-3), although I did not conduct analyses of extractable ions for elements other
than P, which are probably low (Sancho et al. 1990).
The EARTH forest reserve is comprised of mostly secondary forest and disturbed
primary forest, which has not been altered in the past 25 yrs, since EARTH bought the
property and it became a reserve. Mean tree density is 390 trees with >10 cm diameter
at breast height (1.37 m) per hectare, with a basal area range of 15-25 m2 ha-1 (R.
Russo, unpublished data). In my study plots, which did not include swampy or riparian
areas, I found a higher density of trees (542 trees ha-1) and basal area (34.43 m2 ha-1).
60
Tree density is similar to that found at La Selva Biological Station, a well-studied wet
forest in Costa Rica (446 trees ha-1; Lieberman and Lieberman 1994). Although there
was variation in basal area among plots in my study site (Figure 3-2A), there were no
differences among blocks (F5, 23 = 0.64, P = 0.67) or among assigned treatments (F3, 23 =
1.72, P = 0.20), when the experiment was established.
Limited information exists regarding the floristic composition of the EARTH forest
reserve but within my study plots (total of 0.96 ha) I identified 104 tree species
comprising 82 genera and 46 families. Similar than other forests in the area (Hartshorn
and Hammel 1994), the forest at EARTH is dominated by the legume tree Pentaclethra
macroloba. This species contributed almost 30% of the total basal area within my study
plots (Figure 3-2B). High palm density is also characteristic of forests in the Caribbean
lowlands of Costa Rica (Hartshorn 1983, Hartshorn and Hammel 1994). At EARTH
forest, the second most important species was Socratea exohrriza, a canopy or
subcanopy palm with large leaves and distinctive stilt roots. In contrast with
Pentaclethra, however, this palm was important because of the high frequency and
density in which it occurs at EARTH, and not because of a large basal area (Figure 3-
2B).
Another inherent trait of these forests is their high dynamism. Canopy gaps are a
major source of environmental heterogeneity both in the canopy and in the forest floor
(Denslow and Hartshorn 1994). I characterized the variation in canopy openness among
plots using hemispherical photographs (Appendix A).
61
Experimental Design
The study plots were located in two flat, relatively high, non-flooding, mature-
forest areas (termed “Rancho” and “Rio”) within the forest reserve and were separated
by less than 1 km (Figure 3-3).In May 2007, I established 24 30 x 30 m plots and
assigned them randomly to three fertilizer treatments or a control in a complete block
design (n = 6). Three blocks were located at the Rancho site and three at the Rio site
(Figure 3-3). Though fertilization treatments are ongoing, data included here were
collected from August 2007 to March 2010. Besides the control plots, the three
treatments included +P (47 kg ha-1yr-1 of P as super triple phosphate), +N (100 kg ha-
1yr-1 of N applied as ammonium nitrate and urea), and +NP (N and P added together in
quantities as in +N and +P plots). Fertilizer was broadcast by hand twice a year on the
surface of the 900 m2 plots. All measurements were restricted only to the central 400 m2
of each plot (20 x 20 m) to reduce edge effects. The amount of fertilizer added is
consistent with similar studies conducted in montane (Harrington et al. 2001, Tanner
and Kapos 1992), secondary (Davidson et al. 2004), and lowland seasonal (Wright et al.
2011) tropical forests. In my experiment, all plots were separated by at least 50 m and
were selected to avoid leaf-cutter ant (Atta) nests, and extremely large trees that would
dominate an entire plot. To minimize the variability due to high tree species diversity,
each plot included at least one Pentaclethra macroloba tree and one Socratea exhorriza
palm, two of the most common species in this forest (Figure 3-2).
Soil Measurements
I collected soils before applying treatments and 1 and 2 years after the onset of
fertilization. At each collection time, soils were cored at depths of 0-10, 10-30 and 30-50
cm at randomly preselected, varying sites in each of four quadrants in a plot and then
62
combined into one composite sample per depth for each plot. I collected soil samples
using an open soil corer and kept them in a cooler for less than 6 hrs until they were
brought to the lab and stored in a 4°C cold room. All samples were homogenized by
separating roots, rocks, and other coarse debris, within 72 hrs of collection.
Homogenized soils were sub-sampled and extracted for chemical analysis.
Nitrate (NO3-), ammonium (NH4
+), dissolved inorganic nitrogen (DIN), net
mineralization, and net nitrification were measured on fresh soils by extracting 10 g of
soil in 50 ml 2M KCl at EARTH, freezing the extracts , and then conducting colorimetric
readings in an Astoria Pacific colorimetric autoanalyzer (Clackamas, Oregon, USA) at
the University of Florida. I calculated dissolved inorganic nitrogen (DIN) as the sum of
NO3- and NH4
+ at this initial extraction. To obtain a measurement of potential
mineralization, I incubated 10 g of soil for eight days at field moisture and at room
temperature (approximately 25 °C), then extracted and measured NO3- and NH4
+ as
above. I calculated net N mineralization from changes in the NO3- and NH4
+
concentrations (Riley and Vitousek 1995, Robertson et al. 1999) and net nitrification as
the change in NO3- concentration per gram dry soil mass divided by the time of
incubation.
Soil pH and extractable P were measured on air-dried soils that were forced with
a rubber stopper through a 2 mm sieve. Soil pH was measured using a 2:1 water: soil
ratio on a electronic pH meter (Thermo Orion 250A+, Orion Research, Inc., Boston,
Massachusetts, USA). I extracted soil P with a Melich I solution (Kuo 1996). The
resulting phosphate (PO4-3) in solution was read by colorimetric determination of ortho-
phosphate with modified volumes as described by Murphy and Riley (1962), using a
63
spectrophotometer microplate reader (PowerWave XS Microplate Reader, Bio-Tek
Instruments, Inc., Winooski, Vermont, USA). Total percent C and N were measured on
ground soils dried at 60 °C using an elemental analyzer (ECS 4010, Costech Analytical,
Valencia, California, USA). For soils collected in 2007 (pre-fertilization) and in 2009 (2
yrs after fertilization) at 0-10cm depth, a detailed “48-element 4-acid ICP-MS analysis”
(code ME-MS61) was conducted at ALS Chemex laboratories. A summary of Pre-
fertilization values is presented in Table 3-1.
Tree Diameter Measurements
In May 2007, all trees larger than 9 cm diameter at breast height (DBH) in each
20 x 20 m plot were identified to species or genus and labeled with a metal tag nailed to
the bole. A complete list of species of trees is included in Table C-1. Lianas with stems
larger than 10cm with roots in the plot were considered trees. I also included felled trees
with live leaves. For larger trees (approximately 10 per plot), I installed dendrometer
bands constructed with aluminum packing tape and springs as described by Keeland
and Joy Young (2004). I removed lianas and loose bark at point of attachment and
installed dendrometers ≥10 cm above identification tags or higher when buttresses were
present. Dendrometers were allowed to settle at least one month prior to initial marking.
In addition in each plot, I labeled 10 trees of common species with DBH of 4 to 9 cm
with plastic tags attached to a plastic string. I measured diameter increase with either
calipers (for trees with dendrometers) or with a DBH tape (for remaining trees) every six
months until approximately 2.5 yrs after initial fertilization, for a total of six censuses.
To test the community response to fertilization, I used the plot as the
experimental unit (n = 24). I calculated the basal area (BA) occupied by the trees with
the following formula: BA (m2) = (DBH/2) 2 * π. Total basal area increase for a plot (m2
64
ha-1 yr-1) is extrapolated from the total increase in tree diameter for a plot during a time
interval for each size class. This metric was calculated as the plot sum of all the
differences in BA between two consecutive measurements divided by the time interval,
and was dependent on the number and size of the trees in a plot. This metric only
considers trees that increased in diameter. Trees that died, broke, shrunk or did not
change in BA were excluded from the calculations. Overall, when a tree “shrunk”, this
was caused by (1) real shrinkage of the bole, probably due to a change in water
availability or time of day (Sheil 2003) (2) physical damage (one of the trunks was
broken or trunk split), (3) a measurement discrepancy caused by a new dendrometer. In
some cases, the original dendrometer was damaged and a second had to be installed.
Because it was hard to discern among causes of shrinkage, I eliminated these trees
from analyses. I also evaluated tree growth by calculating proportion of tree growth (%
trees that grew yr-1plot-1), which refers to the number of trees that increased in diameter
relative to the total number of trees in a plot for a given time interval. In this case I
considered all trees in a plot. I calculated a total of seven metrics of tree growth,
including relative growth rate (RGR), average of DBH per plot and maximum diameter
increase per plot. However, most metrics provided similar statistical results and
therefore I only present results from total BA increase and proportion of tree growth.
Because of high species diversity in tropical forests, it was challenging to find
trees from the same species in all plots or all treatments. Thus, I selected four common
species (Dendropanax arboreus, Goethalsia meiantha, Pentaclethra macroloba, and
Socratea exohrriza) and three common genera (Inga, Protium and Virola) of trees
ranging from 4 to 89 cm DBH to test how different taxa responded to fertilization
65
additions (Table 3-4). To analyze species responses to fertilization I used individual
trees as the experimental unit and RGR as the metric of tree growth. In this case I
considered all trees for analysis, including those that did not grow because the results of
the analysis do not change if trees that did not grow were included.
Seedling Measurements
In July 2007, I selected, marked, and photographed 20 seedlings or saplings
from the most common species in each plot that had expanded leaves and were less
than 1m high. I documented evidence of herbivory, and recorded stem length, number
of leaves, and canopy cover above each seedling using a spherical denisometer
(Forestry Supplies, Lincoln, NE, USA). In August 2008 and 2009, I conducted a census
of the marked seedlings and recorded survival, growth, number of leaves, and level of
herbivory. A seedling was considered dead if the stem was broken or dry and there was
no resprouting. Seedlings that could not be found were considered missing and were
not included in the analysis. Growth was defined as increase in stem length measured
with a ruler from the point where the root starts to the longest branch. However, some
seedlings showed no change, or a decrease in stem length, caused by a broken or
damaged stem. Other seedlings had multiple stems that were longer than the stem
measured in the previous census. This physical damage and resprouting of multiple
stems made it difficult to quantify plot-level growth. Therefore, I recorded the percent of
seedlings that grew, which I calculated as the proportion of seedlings with an increase
in stem length or number of stems relative to those with no change, or a decrease in
stem length relative to the previous census (1 yr – baseline; 2 yr – 1 yr). The percent of
seedlings that increased the number of leaves also refers to change between censuses
66
for live seedlings (1 yr – baseline; 2 yr – 1 yr). Finally, I recorded presence/absence of
herbivory at each census.
Litterfall Collection
I collected litterfall every two weeks for three years from the onset of the
experiment, from two traps installed in each plot. Traps measured 50 x 50 cm and were
constructed of mesh and pvc mounted on 1m-high metal rods. Samples from both traps
were combined, dried at 60 °C, separated into foliar, reproductive and woody
subsamples, and weighed. The foliar fraction included all leaves, petioles and fronds;
the reproductive fraction included flowers, peduncles, fruits and seeds; the woody
fraction included all sticks smaller than 2 mm in diameter. I discarded sticks larger than
2 mm in diameter, mosses, insects, and all unidentified material.
Not all litterfall-collection intervals were regular because at times inclement
weather prevented collection or traps were stolen. I therefore calculated a daily litterfall
rate (g biomass/day) for each collection by dividing dry litterfall mass by number of days
for that specific interval. If there was more than one collection per month I averaged the
rates for that month. For months where there were no collections, I used the average
from the previous and following month for that year. After obtaining all monthly rates, I
multiplied them by the number of days in that month to obtain a monthly production
value. The sum of all monthly productions 12 months post initial fertilization, was
labeled as “year one” (which corresponded to August 2007-July 2008). Litterfall
production for “year two” refers to August 2008-July 2009.
To measure coarse litter production, I marked two 1 x 1 m quadrants on the
forest floor in each plot (Clark and Clark 2001b). Here, I collected palm fronds and wood
pieces >2 mm in diameter, approximately every three months between February 2008
67
and July 2009, resulting in four collection periods. Samples from both quadrants were
combined, dried at 60 °C, and weighed. When samples were too large to transport to
the laboratory, fresh weight was obtained with a Pesola scale in the field and a
subsample collected. With the fresh and dry weight from the subsample, the %water
content of the large sample was estimated and the dry weight of the whole sample
calculated. In August 2009, I collected all standing litter in two 50 x 50 cm quadrats
placed randomly in the buffer zone of each plot in order to measure the pool of standing
litter on the forest floor. Samples were dried at 60 °C, separated into leaves,
reproductive parts (flowers and fruits), sticks <2 mm diameter, sticks >2 mm diameter or
wood pieces, and roots, and then weighed. Leaves and sticks <2 mm diameter were
combined as “foliar litter fraction” and the sum of all fractions is the total litterpool.
Surface roots in the litter were only found in seven plots, thus the root fraction was not
analyzed separately.
Root Measurements
To estimate the effect of fertilization on root biomass, I collected 0-15 cm deep
root cores using a pounding corer with a volume of 4.75 x15 cm. Prior to treatments,
followed by 1-yr and 2-yr post treatments, I collected root cores at four points per plot
adjacent to where soil samples were collected, and combined them into one composite
sample for each plot. Intact root cores were refrigerated at 4 °C for up to three months
and then separated into large (>2 mm diameter), small (<2 mm diameter), live and dead
roots. For samples collected in 2007 and 2008, I separated roots using variable-speed
electric drills as described by Espeleta and Clark (2007). For samples collected in 2009,
I dissolved the samples in two-gallon (7.6 L) pails filled with water but stirred them
manually and hand-sorted the roots using soil sieves. Each year, immediately after
68
cleaning and separating the roots, I dried them at 60 °C for at least one week and
obtained dry weight.
To estimate the effect of fertilization on fine root production I used ingrowth cores
(Cuevas and Medina 1988). In August 2007, I randomly chose two locations within each
plot and installed two cylindrical, closed-bottom root ingrowth cores (2 mm mesh, 10 cm
deep and 7.7cm diameter).To install the ingrowth cores, I collected soil cores, removed
all visible roots by hand (using latex gloves), placed the mesh ingrowth cores in the hole
where the core was removed, and then filled the cores with the root-free soil. In July
2009, I removed the cores and brought them to the laboratory, where I separated the
roots that had grown inside by washing the cores in a bucket with water and hand
sorting the roots using soil sieves. I separated roots by size (>2 mm or < 2 mm
diameter) and status (live/dead). I dried roots at 60 °C to a constant mass and recorded
dry weight.
Statistical Analysis
To test the effect of fertilization treatments on total basal area increase,
proportion of tree growth, litterfall and coarse debris production, and root biomass, I
used plot-averaged values (n=24) in repeated measures MANOVAs with measurements
at different times as dependent variables and treatment and block as independent
variables. I selected this approach over univariate repeated measures ANOVA because
in some cases the sphericity assumption was not met. I also calculated relative growth
rates (RGR) from individual trees and conducted one-way ANOVAs with RGR (mm yr-1)
as the dependent variable and treatment as the independent variable. I calculated RGR
for each tree as the slope from a line from the log-transformed DBH at each census (n =
6). I chose this approach to test fertilizer effects in addition to plot-averaged analyses
69
(MANOVAS) because this approach incorporates differences in tree size and can still
be calculated if one time measurement is missing. Because RGR is calculated as a
slope, it incorporates differences between measurements and provides an integrated
index of growth throughout the timescale of the experiment. I used Dunnett’s method to
test if treatment means were different from the control. I also used these one-way
ANOVAs with RGR as the dependent variable as a complementary analysis when there
was a treatment *time interaction in the repeated measures MANOVA and it was not
evident where the differences were. Finally, I used a contingency table to explore the
likelihood that a tree would grow in a control plot versus in a treatment plot. I calculated
odds ratios from this contingency table as (# trees that grew in treatment x/ # trees that
did not grow in treatment x)/ (# trees that grew in control/ # trees that did not grow in
control). I calculated these odds ratios for each tree size class and each treatment (x=
+N, +P, or +NP).
To explore which of the measured variables were important to predict tree RGR at
EARTH forest, I also utilized a mixed model approach. I fitted different models and
compared them using the AIC criteria. Results were similar than those obtained using
the repeated MANOVA approach and therefore were not included here. However, a
detailed description of the methodology and results from this analysis is included in
Appendix B.
To test for treatment and block effects in root production (results from ingrowth
cores) and litterpool samples I used a two-way ANOVA. Here, when necessary, data
were log-transformed to meet normality assumptions. To test for the effect of fertilization
on categorical variables deriving from seedling measurements, I used Pearson chi-
70
squared tests. For these analyses, I used individual seedlings (not plot-averaged
values) because the number of seedlings surviving in each plot varied significantly. For
tests of growth, number of leaves and herbivory, I only considered live seedlings.
In August 2008, just after my field season ended, a strong storm with severe
winds struck the EARTH Forest Reserve destroying plots 17 and 22 (Appendix A). Both
plots were on the Rio side and belonged to the +N treatment. This event resulted in
missing data for tree censuses four, five and six and for seedling census three (2 yrs
after fertilization), as well as for soil and root collections for 2009 (2 yrs after
fertilization). To avoid an unbalanced design in my statistical analyses, I replaced
missing data with averages from other plots for that treatment for that year. In the case
of litterfall collection, however, because of the significant amount of missing points I
eliminated these two plots from the analysis (n = 22). For the seedling analyses I
conducted two different tests for 1 yr and 2 yrs after fertilization and excluded seedlings
from plots 17 and 22 from the 2 yr analyses.
In all figures and tables as well as in the text, means (+ standard errors) are
shown. All figures were constructed using Sigmaplot 11.0 and analyses were conducted
in JMP 8.0(SAS Institute Inc., Cary, NC, USA).
Results
Soil Measurements
Repeated measures MANOVAs revealed that spatial and temporal variation in
soil parameters was larger than variation stemming from fertilization treatments, which
resulted in no treatment effect but significant block and time effects in most soil
parameters (Figure 3-4 and Table 3-5). Spatial variation was not only large among
blocks but also within blocks. For example at the 0-10 cm depth, DIN ranged between
71
3.2-26.6 µg g-1 within one block (a coefficient of variation (CV) of 40%) and Melich P
ranged between 2.1-6.7 µg g-1 (a CV of 58%).
This large variation in soil parameters within blocks made it difficult to detect
treatment effects. Therefore, to distinguish the effect of fertilization additions from
background heterogeneity, I compared the percent difference in soil parameters
between initial values (pre-fertilization) and those obtained two years after fertilization
(Table 3-6). Two years after the initiation of the experiment, there was no significant
percent change in pH, percent C, DIN, or percent N in any treatment (Figure 3-5A-D).
However, after two years of fertilization there was a significant increase in Melich P and
total P at the 0-10 cm depth in the +P and +NP treatments (Figure 3-5E, F). For Melich
P, there was a mean increase of 59 + 9.8 % and 87 + 25 % for the +P and +NP
treatments respectively. By contrast, there was a 9.3 + 23% and 20 + 5.6 % decrease in
Melich P in the control and +N treatments respectively. This increase in Melich P was
not observable in the 10-30 or 30-50cm depth increments (Figure 3-5E). Total P
increased 7.2 + 2.2 % and 8.6 + 2.5 % in the +P and +NP treatments respectively but
decreased by 1.8 + 1.0 % and 4.6 + 3.4 % in the control and +N treatments
respectively. The observed total P increase represents a recuperation of 97% and 93%
of the P added as fertilizer in the +N and +NP treatments, respectively. Melich P and
Total P were positively correlated before and 2yrs after initial fertilization (Figure C-2). I
did not analyze samples from deeper profiles for total P.
Tree Growth
At the community level there was no detectable difference in tree growth
measured as total increase in basal area (Figure 3-6 and Table 3-7). However, there
was a time*treatment interaction, as well as a significant block and time effect, in the
72
percentage of trees that grew per plot (Figure 3-7 and Table 3-8). This indicates that for
some blocks at different times there were significant treatment effects, although from
plotting the data and conducting ANOVAS at each time point, it was not possible to
discern where the differences occurred (non-significant ANOVAS not shown). To further
explore this result I compared the RGRs of all trees across treatments (experimental
unit here was individual tree as opposed to plot total). Tree RGRs were higher in the
+NP treatment relative to the control (F3,768 = 2.17, P = 0.08; Dunnett’s test = 2.35, P =
0.04 ). This difference was probably due to small trees (5-10cm DBH) growing more in
the +NP treatment (Figure 3-6a) and is consistent with the number of trees that grew in
each plot across time (Figure 3-7 and Table 3-8). This result was consistent if all trees
were included in the analysis or if only trees that grew were included in the analysis.
Overall, a tree in the +P treatment was 1.65 times more likely to grow than in a control
plot and a tree in the +NP treatment was 1.68 times more likely to grow than in a control
treatment, 2 yrs after initial fertilization (Table 3-9).
In addition to the community level response, I explored the effect of fertilizer
additions on trees of different size classes. For the 5-10 cm size class there was a time
effect and a time* treatment interaction in the number of trees that grew in each plot
across time (Figure 3-7 and Table 3-8). When considering tree RGRs among treatments
for this size class, there were significantly higher RGRs for trees in the +NP treatment
(F3,266 = 3.94, P = 0.01; Dunnett’s test = 2.36, P = 0.01; Figure 3-8). In addition for this
size class, trees in the +P and +NP treatment were 2.16 and 2.68 times more likely to
grow than trees in the control plots 2 yrs after initial fertilization (Table 2-9). For this size
class there was a positive correlation between percent change in total P in surface soils
73
(0-10 cm) after two years and the number of trees that grew between 2 and 2.5 yrs after
fertilization (r2 = 0.22, d.f. = 20, P = 0.03). For all the size classes combined there was
also a positive correlation between percent change in total P in surface soils (0-10 cm)
after two years and the number of trees that grew between 2 and 2.5 yrs after
fertilization (r2 = 0.33, d.f. = 20, P = 0.01).
I also studied the effect of fertilizer additions on trees from different taxa within
the community. From the four species and three genera where I analyzed RGR
individually (here again individual tree was the experimental unit), only Goethalsia
meiantha and Socratea exohrriza showed evidence of a response to fertilization (Figure
3-9). Goethalsia had the highest growth in the control treatment (Figure 3-9B), although
this pattern was highly influenced by a single tree and should be interpreted with
caution. Socratea exohrriza, a fast growing canopy palm, had higher growth with P
additions (Figure 3-9F). Protium had higher RGR when both N and P were added
together, although this difference was not statistically significant (P = 0.51).
Seedling Measurements
There was a significant treatment effect on the proportion of seedlings that
survived, grew, and showed an increase in leaf number two years after initial
fertilization. Presence of herbivory was common (more than 50% in every plot) and did
not differ among treatments (Figure 3-10 and Table 3-10).
Litterfall and Litter Pool Measurements
Mean foliar litterfall productivity was 5.0 + 0.27 Mg C ha-1 yr-1, mean reproductive
litterfall productivity was 0.72 + 0.19 Mg C ha-1 yr-1, and mean coarse litterfall
productivity was 1.98 + 0.34 Mg C ha-1 yr-1. There were no differences among
treatments in foliar, reproductive, or coarse litterfall productivities. However, foliar and
74
reproductive litterfall productivities decreased over time (Figures 3-11 and 3-12 and
Table 3-11) and the decrease was consistent across blocks (no time*block interaction).
The standing pool of litter measured in 2009 in 50 x 50 cm plots established in the
buffer zones of each plot averaged 2.77 + 0.18 Mg C ha-1 foliar litter, 0.56 + 0.17 Mg C
ha-1 reproductive litter, and 0.64 + 0.08 Mg C ha-1 sticks with diameters >2 mm. These
values summed up to a total litterpool of 6.76 + 0.36 Mg C ha-1. There was no difference
among treatments or blocks in foliar, reproductive or total standing litter or in the mass
of sticks with diameters >2 mm collected from these litterpool plots (Table 3-12).
Root Biomass and Productivity
Two years after initial nutrient additions there was no difference among
treatments in biomass of fine roots (< 2 mm diameter) or large roots (> 2 mm diameter)
in the top 15 cm of the soil profile. However, across treatments, there was a larger
biomass of fine roots two years after the onset of the experiment than in the two other
sampling times (Table 3-13 and Figure 3-13). Mean fine root biomass increased from
0.89 + 0.09 Mg C ha-1 pre-fertilization to 2.02 + 0.17 Mg C ha-1 two years after
fertilization. In addition, for both root sizes there was a significant block effect; overall,
plots in the Rancho site had more roots than plots in the Rio site (MANOVA by site: fine
roots F1, 19 = 19.4 P < 0.01 and large roots F1, 19 = 8.24 P < 0.01).
Root productivity measured with ingrowth cores in the 0-10cm soil profile, did not
differ among treatments or blocks (Ftreatment = 0.52, P = 0.67, Fblock = 1.57, P = 0.67 d.f. =
3,5 ; Figure 3-14). There was no relationship between root productivity and fine or large
root biomass before fertilization or after fertilization (data not shown).
75
Discussion
Soil Measurements
Baseline nutrient concentrations in soils at the EARTH Forest were high relative
to other tropical forests. Nitrogen measurements were similar to values obtained at La
Selva, another well-studied forest in Costa Rica with similar climatic and floristic
composition (Vitousek and Denslow 1986, Sollins et al. 1994). Superficial total P (0-10
cm) at EARTH Forest was 1601 µg g-1 (or 1136.71 kg ha-1), which was particularly high
compared to other forests where fertilization experiments have been conducted: a
secondary forest in Pará, Brazil (~225 µg g-1; Davidson et al. 2004), an evergreen forest
in Kalimantan, Borneo (<250 µg g-1; Mirmanto et al. 1999), a rain forest in Korup,
Cameroon (<300 µg g-1), a seasonal forest in Gigante Peninsula, Panama (772 µg g-1
;Yavitt et al. 2011), and a wet forest in Osa Peninsula (557 µg g-1; Townsend et al.
2002). However, total P values at EARTH were similar to the alluvial soils at La Selva
(1650 µg g-1 in alluvial soils and 579 µg g-1 in residual soils; Espeleta and Clark 2007,
Wood et al. 2009). Concentrations of immobile elements, such as niobium (Nb) and
zirconium (Zr), were similar to those reported for volcanic parent material in the area
(Table A-4; Meijer and Buurman 2003). This, in conjunction with high total P, suggests
that soils at EARTH were not significantly leached during development. Alternatively,
nearby volcanoes may have provided inputs of lahar or ash to the region recently, which
would have replenished soil fertility after leaching of original parent material (Meijer and
Buurman 2003, Porder et al. 2006). High concentrations of total P could imply that P
availability at EARTH forest is higher than in other tropical forests, where most rock
derived minerals have been leached or occluded, and P limits NPP (e.g., Kauai Island in
Hawaii; Herbert and Fownes 1995, and potentially Kalimantan, Borneo; Mirmanto et al.
76
1999). However, resin-extractable P (not shown but below 1µg g-1) and Melich-
extractable P (2.14 + 0.36 µg g-1) at EARTH Forest were low, which suggests that
although P pools are large, P bioavailability could be low. In these soils high clay
content (around 50%, Table A-2), high sorption capacity (Sanchez 1976), and microbes
(Olander and Vitousek 2004) can immobilize PO4-3 making it unavailable for plant
uptake (Vitousek and Denslow 1987, Vandecaar et al. 2009).
There was a large spatial and temporal variation in soil measurements, which
was evident as significant block effects, and could have masked increases in N or P
concentrations after fertilization. To distinguish the effect of fertilization additions from
background heterogeneity, I compared the percent difference in soil parameters
between baseline and two years after fertilization. After two years of fertilization, there
was a significant increase in soil Melich P and Total P relative to baseline levels but no
increase in any N parameters measured. These results are similar to soil responses to
fertilization in a secondary forest in the Brazilian Amazon (Davidson et al. 2004) and
could be attributed to the larger N than P background pools, or to higher N ion mobility.
Because Nitrate (NO3-) is more mobile than phosphate (PO4
3-) and soil samples were
collected almost 6 mo after fertilization, it is possible that all added N was leached
(Radulovitch and Sollins 1991) or immobilized by plants and/or microbes by the time
soils were sampled (Chapter 4).
Nutrient Limitation to NPP
After 2.5 yrs, there was no clear effect of fertilization on tree diameter increase
(or basal area (BA) increase) at the community level, litterfall productivity, or root
biomass or productivity. One possible explanation is that NPP at EARTH forest is
limited or co-limited by a nutrient other than N or P, as suggested by Kaspari et al.
77
(2008) for litterfall production and decomposition in Panama. At the EARTH forest there
are relatively high concentrations of N and P in soils (Table 3-2), foliage, and litterfall
(Chapter 4). In addition, litterfall production at EARTH Forest (11.44 Mg ha-1 yr-1) is on
the high spectrum compared to other tropical forests: 7.17 Mg ha-1 yr-1 in Kalimantan
(Mirmanto et al. 1999); 10.65 Mg ha-1 yr-1 in Gigante Peninsula (Kaspari et al. 2008); 12
Mg ha-1 yr-1 in Osa Peninsula (C. Cleveland unpublished data), and ~9 Mg ha-1 yr-1 in La
Selva (D. A. Clark unpublished data), which suggests that at EARTH trees are meeting
at least some of their nutrient demands to produce biomass. Therefore, it is possible
that as N and P requirements are met, the plant community becomes limited by another
nutrient, potentially potassium or a combination of micronutrients (Herbert and Fownes
1995, Kaspari et al. 2008, Townsend et al. 2011, Wright et al. 2011).
After 2.5 yrs there was no overall NPP response to fertilization but there was a
higher proportion of trees that showed some BA increase in the +P and +NP treatments,
which suggests that NPP at EARTH forest could be limited by P to some degree but
that this experiment failed to capture statistical differences among treatments.
Treatment effects could be obscured by the high spatial and temporal variability in
individual tree growth (Clark and Clark 1999, Clark and Clark 2011), litterfall (Wood et
al. 2009), and root biomass (Espeleta and Clark 2007), which are characteristic of
forests in this area. Potentially, microclimatic variables (e.g., light, water availability, or
soil temperature) could influence NPP more than nutrient availability on the short term
and at the plot scale, confounding any experimental treatments. If so, more intense
sampling and over a longer period would be required to observe treatment effects (Clark
and Clark 2001b, Clark and Clark 2011).
78
To compensate for spatial variation in litterfall, it could be argued that more traps
were needed to capture community wide responses to fertilization. In my experiment,
however, litterfall collection effort (area of traps/ plot area) was higher (6.26 m2 ha-1)
than in experiments at La Selva (4.5 m2 ha-1 in old growth and 2.56 m2 ha-1 in
secondary forest), where significant differences in litterfall production were observed
after a litter manipulation experiment equivalent to fertilization with 5-25 kg ha-1 of
organic P (Wood et al. 2009). In the case of diameter increase, all trees >10cm DBH
and 10 trees between 4-9 cm DBH were measured per plot, twice per year for 2.7 yrs.
These recordings are similar to other fertilization experiments where differences in tree
growth were observed (e.g., Tanner and Kapos 1992, Herbert and Fownes 1995,
Davidson et al. 2004). In contrast, superficial root biomass in these forests is largely
variable and influenced by the presence of palms (Espeleta and Clark 2007).Therefore,
it is likely that random location of palms relative to where samples were collected had
an influence in the observed results. In addition, the change in methodology used to
separate roots in the last sampling event (2 yrs after fertilization, see methods) may
have caused the apparent increase in root biomass over time. In conclusion, it is
unlikely that insufficient sampling effort was the cause for a lack of difference among
treatments in litterfall production or tree growth but it is possible that more intense root
sampling with consistent methodology over time would reveal differences in root
productivity or biomass.
Temporal variation in tree growth associated with climatic variables in this type of
forest is large and heterogeneous among size classes and species (Clark et al. 2003,
Clark and Clark 2011). Thus, it is possible that more time is needed to observe
79
treatment effects in diameter increase. In the case of litterfall production, treatment
differences have been observed in less than two years in other experiments (Herbert
and Fownes 1995, Mirmanto et al. 1999, Wood et al. 2009). However, in a montane
forest in Venezuela, there was an effect of both N and P additions on litterfall only after
4 yrs (Tanner and Kapos 1992). Because of high temporal variability in the system and
high background nutrient levels, it is possible that differences among treatments at
EARTH forest will only be observable after several more years of nutrient additions. As
recently highlighted by Clark and Clark (2011), temporal variability in NPP in lowland
tropical forests make long-term observations particularly important and necessary.
Nutrient Limitation and Tree Size
After 2.5 yrs, there was no difference among treatments in community-level BA
increase (Figure 3-6) but there was a higher proportion of trees that increased in BA in
+P plots (Figure 3-7). This suggests that the largest trees, which contribute the most to
total BA increase in each plot, are not limited by P (or did not respond in a 2.5 yr time
frame), but that smaller trees (5-10 cm DBH), which are more frequent, could be P
limited. This response of small trees was consistent with traditional expectations (i.e.,
Walker and Syers 1976) and was supported by a positive relationship between percent
difference in soil total P (0-10cm depth) and the proportion of trees showing some
diameter increase.
However, when observing RGR responses of all small trees (considering individual
trees as opposed to plot averages), it is evident, that although not statistically
significant, these small trees are limited not only by P but by N as well. RGRs of small
trees showed a sub-additive response to N and P (sensu Harpole et al. 2011),
increasing RGR by 15% with +N, 26% with +P and 38% with +NP relative to the control.
80
This form of co-limitation is incompatible with Liebig’s Law of the Minimum (Liebig 1842)
and supports recent evidence that proposes synergistic interactions between N and P
availability (Elser et al. 2007, Harpole et al. 2011, Cleveland et al. 2011).
Because of light-co-limitation of small trees, I had predicted that medium (10-30
cm DBH) trees would be more likely to respond to nutrient additions. However, small
trees were the only ones that differed among treatments, either because they are the
only ones limited by N and P or because small trees have a faster growth rate (Clark
and Clark 1992, Clark 1994). Although there was no difference among treatments in the
proportion of trees >30 cm DBH that grew, there was a high proportion of large trees
that grew overall (>40 % of trees for any given census for all treatments), which
suggests that EARTH Forest is a relatively young forest and that soils are relatively
nutrient-rich.
I hypothesized that seedlings would not respond to nutrient additions because
low light levels in the understory would prevent them from incorporating extra nutrients
as biomass (Denslow et al. 1990, Burslem et al. 1995). However, although this was
probably true in some cases, a higher proportion of P-fertilized seedlings survived,
grew, and increased leaf number after two years than seedling from other treatments.
This suggests that (1) the canopy at EARTH Forest is relatively open and enough light
reaches the understory to allow seedling growth (Lawrence 2003, Baroloto et al. 2006,
Appendix B), and (2) the seedling community at EARTH may be limited by P (Palow
and Oberbauer 2009). Interestingly, there seemed to be an effect on the +P but not the
+NP treatment. Perhaps, +N cancels the +P effect by increasing herbivory if N is
incorporated into tissues, increasing their palatability (Andersen et al. 2007 but see
81
Campo and Dirzo 2003). However, I did not collect foliage from seedlings and did not
observe a significant difference in the percent of seedlings exhibiting leaves with
herbivory. More experiments are required to test this hypothesis.
Effect of Taxa on Nutrient Limitation
I had predicted that fast growing, canopy species, such as Pentaclethra and
Goethalsia would respond to nutrient additions but this was not the case. The lack of
response of Pentaclethra is interesting because as the most important species in this
forest (up to 30% of the basal area and 14% stems, Figure 3-2); it has a strong
influence over the community-averaged response to fertilization. Pentaclethra is a
nitrogen-fixing legume and therefore I did not expect a response to N fertilization
(McKey 1994). However, the lack of response to the +P treatment suggests that either
this species is well adapted to the PO43- levels found at EARTH Forest or that it has
access to enough PO43- to fulfill metabolic needs. It has been suggested that N fixing
legumes, such as Inga and Pentaclethra, could have an advantage when available P is
low. These species can invest extra N in producing phosphatase enzymes, which
enable them to access organic P pools (Olander and Vitousek 2000, Treseder and
Vitousek 2001, Houlton et al. 2008). Measurements of phosphatase activities and
available P at the base of these trees could be used to test this hypothesis.
From the studied taxa, only Socratea exohrriza (the most abundant canopy palm;
Figure 3-2), significantly responded to nutrient additions. Individuals from this species
showed very fast growth in several plots where there was a combination of P fertilization
plus available light due to a tree or branch fall (personal observation). These palms,
however, usually reach a maximum stem diameter around 30 cm, so this response was
mainly due to small individuals. Interestingly, larger individuals responded by increasing
82
foliar P instead of stem diameter with P additions (Chapter 4). Finally, as predicted,
subcanopy trees (Dendropanax and Protium) did not grow more with fertilization
probably because of light co-limitation (Fetcher et al. 1994, Holste et al. 2011).
Conclusions
This study emphasizes the complexity of nutrient limitation in lowland diverse
tropical forests (Vitousek et al. 2010, Townsend et al. 2011, Cleveland et al. 2011) and
adds to the body of experimental evidence showing that these forests may not be
limited by P (Mirmanto et al. 1999, Newbery et al. 2002, Wright et al. 2011), or at least
do not respond to either N or P fertilization in the short term as is the case in montane
forests (Vitousek 2004, Tanner 1998). Most likely, as stated by Kaspari et al. (2008),
“these systems are non-Liebig worlds of multiple nutrient limitations”. More experiments
are needed to elucidate the mechanisms driving limitation by multiple nutrients, such as
proposed by Harpole et al. (2011).
Although the plant community as a whole at EARTH Forest does not appear to
be limited by N or P, my data suggest that seedlings and a palm species (Socratea
exohrriza) are limited by P, and that on average small trees (5-10 cm DBH) are co-
limited by N and P. The lack of response of Pentaclethra to fertilization implies that this
species, the dominant at EARTH Forest, is not limited by either N or P (Chapter 4). In
contrast, this species, and potentially others at EARTH Forest, may be limited by a
different nutrient or nutrients. Alternatively other factors, such as small scale variation in
light and water availability exert more important controls on tree growth than nutrient
availability.
Recent attention has been given to the complex nature of “community nutrient
limitation” (Vitousek et al. 2010, Harpole et al. 2011) and mechanisms for multiple
83
nutrient co-limitation have been proposed as an alternative to the historical Liebig’s Law
of the Minimum (Chapin et al. 1987, Gleason and Tilman 1992, Rastetter and Shaver
1992, Danger et al. 2008, John et al. 2007). In the case of this study, I would expand to
say that at EARTH forest there is “heterogeneous nutrient limitation”, not only driven by
variability in nutrient responses among tree species but also among size classes. This
heterogeneity highlights the importance of considering different aspects of the plant
community, such as forest structure and species composition, when making predictions
concerning nutrient limitation in lowland tropical forests. Furthermore, the differential
response of size classes and species suggests that changes in nutrient availability
could lead to changes in forest structure or even diversity in the long term and could
have important implications for plant-soil-microbial feedbacks concerning nutrient
limitation.
84
Table 3-1. Particle sizes for soils from the top 10cm of the study plots (Rancho and Rio
are two areas of the forest where plots were located). Data were obtained using a Mastersizer Particle Size Analyzer. Samples were dispersed in 1% solution of Calgon soap overnight.
Class clay very fine silt
fine silt medium silt
coarse silt
very fine sand
fine sand
medium sand and
greater
Size (um) < 2 2-7.8 7.8-15.6
15.6- 31
31-62.5
62.5-125
125-250
>250
Rancho 8.62 31.38 23.78 17.23 8.96 4.50 2.27 3.27
Rio 9.44 33.87 23.27 15.36 7.68 3.90 2.12 4.37
Average 8.99 32.51 23.54 16.38 8.38 4.23 2.20 3.77
Table 3-2. Means (with standard errors) for various soil parameters measured at three
depths on each plot at the beginning of the experiment. N = 24 plots. For details about analyses see methods section.
Measurement Depth (cm)
0-10cm SE 10-30cm SE 30-50cm SE
Bulk Density (mg cm-3) 0.71 0.03 0.78 0.02 0.83 0.04
pH H2O 4.04 0.04 4.20 0.03 4.32 0.02
Melich P(µg g-1) 2.14 0.36 1.40 0.25 1.02 0.25
Total %C 4.83 0.14 2.51 0.07 1.71 0.06
Total %N 0.49 0.01 0.28 0.01 0.18 0.01
DIN (µg N g-1) 19.67 1.47 10.86 1.23 6.89 0.56
Net min. (µg N g-1 d-1) 1.68 0.28 0.50 0.17 0.36 0.09
Net nit. (µg N g-1) 1.74 0.35 0.54 0.10 0.35 0.09
85
Table 3-3. Means (with standard errors) for various soil parameters measured at EARTH Forest at the beginning of the experiment. N = 24 plots, depth = 0-10cm. For details about analyses see methods section.
Element Mean SE
Al% 14.27 0.09
Ca % 0.04 0.00
Co (ppm) 27.40 1.49
Cu (ppm) 144.56 2.18
Fe % 8.78 0.10
K % 0.12 0.00
Mg % 0.17 0.00
Mn ppm 765.88 65.73
Mo (ppm) 2.67 0.07
Na % 0.04 0.00
Ni (ppm) 63.77 1.32
P (ppm) 1601.25 104.47
S % 0.10 0.00
Zn (ppm) 110.04 2.80
86
Table 3-4. Taxa selected to study the effect of fertilization on different functional groups. Shown is the percent of the plots in which a taxa were present (n = 24 plots) and the total number of trees measured (all trees were >4 cm DBH). All taxa had at least one individual in each treatment.
Species Family % plots
# trees
Functional properties at EARTH Forest
Dendropanax arboreus
Araliaceae 50 21 Subcanopy tree with soft wood, fast growth, and relatively high light demand.
Goethalsia meiantha
Malvaceae 50 49 Canopy or subcanopy tree with soft wood, fast growth, and relatively high light demand. Characteristic of disturbed areas.
Inga Mimosaceae 70 25 Seven species included1. Canopy or subcanopy trees with hard or semi hard wood, and mostly shade tolerant. Can fix Nitrogen.
Pentaclethra macroloba
Mimosaceae 100 91 Canopy tree with semi hard wood, medium growth, and shade tolerant. Most abundant tree at EARTH forest. Can fix Nitrogen.
Protium Burseraceae 75 41 Four species included2. Subcanopy trees (at EARTH forest) with relatively slow growth, and shade tolerant. Resinous compounds in leaves, stems, and fruits.
Socratea exohrriza
Arecaceae 100 154 Canopy/subcanopy palm with relatively fast growth and shade tolerant. Conspicuous stilt roots.
Virola Myristicaceae 70 24 Three species included3.Canopy trees with semi hard wood, relatively slower growth, and shade tolerant. High concentrations of alkaloids and other compounds in leaves, stems, and fruits. Dioecious.
Source McDade et al. (1994) and O. Vargas (personal communication). 1 I. alba, I. leocalycina, I. pezizifera, I. sapindioides, I. thiboudiana, I. umbilifera, I. venusta
2 P. confusum, P. panamense, P. pittieri, P. Ravenii
3 V. koschnyi, V. multiflora, V. sebifera
87
Table 3-5. Results from repeated measures MANOVAs for several soil variables measured at three depths. F-values for treatment, block and time were obtained from exact tests but time*treatment and time*block interactions are F-value approximations resulting from Wilk’s lambda multivariate tests. Significant effects are signaled with an asterisk
Treatment Block Time Treatment* Time
Block *Time
Variable F P F P F P F P F P
pH
0-10cm 0.12 0.95 1.21 0.37 1.93 0.01* 2.07 0.11 1.17 0.37 10-30cm 0.38 0.77 0.51 0.76 9.59 <0.01* 0.76 0.61 0.99 0.48 30-50cm 1.07 0.40 4.00 0.03* 41.19 <0.01* 1.26 0.32 2.08 0.07
%C
0-10cm 0.46 0.71 4.40 0.02* 9.00 <0.01* 1.33 0.29 2.57 0.03* 10-30cm 0.51 0.69 1.26 0.34 11.96 <0.01* 1.67 0.17 0.86 0.58 30-50cm 1.10 0.38 0.53 0.75 0.57 0.58 1.21 0.34 0.98 0.49
DIN (µg g-1)
0-10cm 0.93 0.46 1.44 0.27 15.65 <0.01* 1.26 0.31 3.84 <0.01* 10-30cm 1.27 0.33 1.02 0.45 2.90 0.09 0.33 0.91 1.11 0.39 30-50cm 3.70 0.04* 4.78 0.01* 8.32 0.01* 0.69 0.66 1.96 0.09
% N
0-10cm 0.09 0.97 3.66 0.03* 26.83 <0.01* 0.75 0.62 2.10 0.07 10-30cm 0.47 0.71 0.41 0.28 5.60 0.02* 0.90 0.51 0.60 0.80 30-50cm 1.23 0.34 0.41 0.83 0.07 0.93 1.43 0.24 1.02 0.46
Melich P (µg g-1)
0-10cm 0.11 0.95 3.23 0.04* 2.59 0.12 0.79 0.59 0.95 0.51 10-30cm 0.25 0.86 3.46 0.03* 30.84 <0.01* 0.44 0.85 2.46 0.03* 30-50cm 0.49 0.70 2.93 0.06 40.25 <0.01* 1.00 0.44 3.41 0.01*
Total P (µg g-1)
0-10cm 0.59 0.63 5.31 0.01* 3.6 0.08 7.42 <0.01* 1.17 0.38 Net mineralization (µg N g-1 d-1)
0-10cm 1.66 0.23 3.06 0.05 1.15 0.35 1.90 0.13 2.59 0.03* 10-30cm 0.68 0.58 2.10 0.14 1.99 0.18 1.08 0.40 1.68 0.15 30-50cm 0.40 0.76 2.41 0.09 1.92 0.19 0.30 0.94 0.17 0.01*
Net Nitrification (µg N g-1 d-1)
0-10cm 0.64 0.60 3.38 0.04* 0.74 0.49 1.69 0.17 2.03 0.07 10-30cm 0.68 0.58 0.51 0.76 8.44 <0.01* 0.69 0.66 1.98 0.09 30-50cm 2.49 0.11 3.14 0.04* 2.42 0.13 1.88 0.12 1.26 0.30
88
Table 3-6. Results from two-way ANOVA analyses for the percent difference between 2 yrs and pre-fertilization values for several soil parameters measured at three depths. Significant effects are signaled with an asterisk.
Parameter Depth df Treatment Block
F P F P
pH 0-10cm 3,4 2.61 0.12 1.55 0.27
10-30cm 3,5 1.10 0.38 0.47 0.70
30-50cm 3,5 2.06 0.16 1.13 0.39
% C 0-10cm 3,5 2.69 0.09 2.52 0.09
10-30cm 3,5 0.07 0.97 0.13 0.98
30-50cm 3,5 0.93 0.45 0.98 0.47
DIN (µg g-1) 0-10cm 3,5 2.05 0.16 6.52 <0.01*
10-30cm 3,5 1.43 0.28 4.40 0.01*
30-50cm 3,5 0.72 0.56 2.92 0.06
Total N (%) 0-10cm 3,5 1.76 0.21 4.75 0.01*
10-30cm 3,5 0.74 0.55 0.66 0.66
30-50cm 3,5 1.03 0.41 0.77 0.59
Melich P (µg g-1)
0-10cm 3,5 9.33 <0.01* 1.61 0.23
10-30cm 3,5 0.23 0.88 1.11 0.40
30-50cm 3,5 0.78 0.53 2.23 0.12
Total P (µg g-1) 0-10cm 3,5 11.27 <0.01* 2.30 0.11
89
Table 3-7. Results from repeated measures MANOVAs for “total basal area increase” by tree size class. F-values for treatment, block and time were obtained from exact tests but time*treatment and time*block interactions are F-value approximations resulting from Wilk’s lambda multivariate tests. In these cases degrees of freedom (df) are approximated as well. Significant effects are signaled with an asterisk.
Size class dfn,d F Probability
5-10 cm
Treatment 3,15 0.72 0.55
Block 5,15 1.15 0.38
Time 4,12 2.53 0.10
Time*treatment 12,32.04 1.15 0.20
Time*block 20,40.75 0.90 0.58
10-30 cm
Treatment 3,15 1.43 0.27
Block 5,15 3.87 0.02*
Time 4,12 6.80 0.04*
Time*treatment 12,32.04 1.58 0.15
Time*block 20,40.75 1.17 0.33
>30 cm
Treatment 3,15 1.16 0.36
Block 5,15 1.02 0.44
Time 4,12 4.23 0.02*
Time*treatment 12,32.04 0.67 0.76
Time*block 20,40.75 1.20 0.30
Total
Treatment 3,15 1.01 0.42
Block 5,15 2.50 0.08
Time 4,12 7.18 <0.01*
Time*treatment 12,32.04 1.02 0.45
Time*block 20,40.75 1.43 0.16
90
Table 3-8. Results from repeated measures MANOVAs for “proportion of tree growth” by tree size class. F-values for treatment, block and time were obtained from exact tests but time*treatment and time*block interactions are F-value approximations resulting from Wilk’s lambda multivariate tests. In these cases degrees of freedom (df) are approximated as well. Significant effects are signaled with an asterisk.
Size class dfn,d F Probability
5-10 cm Treatment 3,15 0.99 0.43 Block 5,15 1.58 0.23 Time 4,12 7.27 <0.01* Time*treatment 12,32.04 2.68 0.02* Time*block 20,40.75 1.17 0.33
10-30 cm Treatment 3,15 1.00 0.42 Block 5,15 7.01 <0.01* Time 4,12 18.13 <0.01* Time*treatment 12,32.04 1.55 0.16 Time*block 20,40.75 2.15 0.02*
>30 cm Treatment 3,15 3.12 0.06 Block 5,15 1.18 0.36 Time 4,12 2.07 0.15 Time*treatment 12,32.04 0.81 0.64 Time*block 20,40.75 0.85 0.65
Total Treatment 3,15 0.69 0.57 Block 5,15 2.96 0.05* Time 4,12 34.19 <0.01* Time*treatment 12,32.04 3.76 <0.01* Time*block 20,40.75 1.78 0.06
91
Table 3-9. Contingency table describing the proportion of trees that grew or did not grow between 2 and 2.7 yrs after initial fertilization (censuses 5 and 6), by tree size class, in the four nutrient addition treatments. The odds ratio refers to the likelihood of a tree growing in a given treatment relative to the control. In each cell total number, percent of total and percent for that category are shown.
# of trees Odds Ratio Size class no yes Total
5-10 cm control 26
11.82 40.63
38 17.23 59.38
64
+N 14 6.36 32.56
29 13.18 67.44
43 1.42
+P 13 5.91 24.07
41 18.64 75.93
54 2.16
+NP 12 5.45 20.34
47 21.36 79.66
59 2.68
10-30 cm control 34
10.30 34.69
64 19.39 65.31
98
+N 25 7.58 40.32
37 11.21 59.69
62 0.79
+P 23 6.97 26.44
64 19.39 73.56
83 1.44
+NP 28 8.44 33.77
55 16.67 66.27
87 1.04
>30 cm control 5
5.88 22.73
17 20.00 77.27
22
+N 1 1.18 7.14
13 15.29 92.86
14 3.82
+P 4 4.71 15.38
22 25.88 84.62
26 1.60
+NP 2 2.35 8.71
21 24.71 91.30
23 3.09
92
Table 3-9. Continued.
# of trees Odds Ratio Size class no yes Total
Total control 67
9.96 34.54
127 18.87 65.46
194
+N 43 6.39 33.33
86 12.78 66.67
129 1.05
+P 42 6.24 28.83
132 19.91 71.17
176 1.65
+NP 42 6.24 23.86
134 19.91 76.14
174 1.68
93
Table 3-10. Results from Pearson chi-square tests for seedling variables measured 1yr
and 2 yrs after fertilization. Survival= proportion of seedlings surviving, Growth = proportion of seedlings that grew, Number of leaves = proportion of seedlings that increased the number of leaves, and Herbivory = proportion of seedlings that showed evidence of leaf herbivory. Significant effects are signaled with an asterisk.
Parameter Time df n Treatment χ2 Probability
Survival 1 yr 3 476 2.81 0.42 2 yrs 3 375 9.86 0.02* Growth 1 yr 3 396 4.36 0.23 2 yrs 3 213 9.43 0.02* Number leaves 1 yr 3 387 2.23 0.53 2 yrs 3 217 13.86 <0.01* Herbivory 1 yr 3 406 4.26 0.23 2 yrs 3 219 2.11 0.55
Table 3-11. Results from repeated measures MANOVAs for foliar (leaves and sticks
<2mm diameter), reproductive (flowers and fruits), and coarse litterfall. F-values for treatment, block and time were obtained from exact tests but time*treatment and time*block interactions are F-value approximations resulting from Wilk’s lambda multivariate tests. In these cases degrees of freedom (df) are approximated as well. Significant effects are signaled with an asterisk. Plots 17 and 22 excluded from analysis (n= 22 plots).
Sample dfn,d F Probability
Foliar litterfall Treatment 3,13 0.14 0.93 Block 5,13 1.21 0.36 Time 2,12 16.02 <0.01* Time*treatment 6,24 0.59 0.34 Time*block 10,24 1.12 0.39
Reproductive litterfall Treatment 3,13 0.12 0.95 Block 5,13 1.21 0.36 Time 2,12 15.72 <0.01* Time*treatment 6,24 1.23 0.33 Time*block 10,24 0.90 0.54
Coarse litterfall Treatment 3,13 3.09 0.06 Block 5,13 1.18 0.37 Time 3,11 1.94 0.18 Time*treatment 9,27 0.58 0.80 Time*block 15,30 0.88 0.59
94
Table 3-12. Results from two-way ANOVA analyses for litterpool fractions. Foliar = leaves and sticks <2mm diameter, reproductive = flowers and fruits, large sticks= wood pieces and sticks > 2mm diameter, and total= the sum of all these plus surface roots.
Parameter df Treatment Block F Probability F Probability
Foliar 3,5 0.34 0.79 1.51 0.25 Reproductive 3,5 0.69 0.58 0.44 0.82 Large sticks 3,5 2.30 0.13 1.49 0.26 Total 3,5 0.31 0.82 1.77 0.19
Table 3-13. Results from repeated measures MANOVAs for fine (<2mm diameter) and
large (>2mm diameter) root biomass collected at 0-15 cm depth. F-values for treatment, block and time were obtained from exact tests but time*treatment and time*block interactions are F-value approximations resulting from Wilk’s lambda multivariate tests. In these cases degrees of freedom (df) are approximated as well. Significant effects are signaled with an asterisk.
Sample dfn,d F Probability
Fine root biomass Treatment 3,15 1.18 0.35 Block 5,15 3.41 0.03* Time 2,14 23.18 <0.01* Time*treatment 6,28 1.35 0.27 Time*block 10,28 1.40 0.23
Large root biomass Treatment 3,15 0.05 0.98 Block 5,15 4.68 <0.01* Time 2,14 2.43 0.12 Time*treatment 6,28 0.84 0.55 Time*block 10,28 1.55 0.17
95
Te
mp
era
ture
( C
)
15
20
25
30
35
Pre
cip
ita
tio
n (
mm
)
200
400
600
800
B 2008
Tem
pera
ture
( C
)
15
20
25
30
35
Pre
cip
itation (
mm
)
200
400
600
800
A 2007
(c) 2009
Month
Jan
Feb Mar
Apr
May Ju
nJu
lAug
Sep O
ctNov
Dec
Te
mp
era
ture
( C
)
10
15
20
25
30
35
Pre
cip
ita
tio
n (
mm
)
200
400
600
800
C 2009
Figure 3-1. Monthly average maximum and minimum air temperatures (dashed lines)
and precipitation (solid lines) at the study site for the three years encompassing the study. Arrows indicate fertilizer applications. Data collected by staff at the EARTH Meteorological station in the main campus.
96
Figure 3-2. Floristic description of the study site including (A) total basal area for each of
the study plots (and blocks) for trees >10 cm DBH at the beginning of the experiment, and (B)importance value indices (relative frequency + relative density + relative basal area) for the most important 20 species of trees (>10 cm DBH) found at EARTH forest. For full species list and complete species name refer to Table C-1.
A
B
97
Figure 3-3. Distribution of plots and blocks within the EARTH forest reserve. Each block
is composed of a control plot (circles), a plot where 100 kg ha-1yr-1 of nitrogen was added (crosses), a plot where approximately 50 kg ha-1yr-1 of phosphorus were added (triangles) and a plot where both nitrogen and phosphorus were added together (squares). Plot sizes are shown enlarged for illustrative purposes.
98
pH
3.8
3.9
4.0
4.1
4.2
4.3
4.4
4.5
(a)
Tota
l C
(%
)
3.0
3.5
4.0
4.5
5.0
5.5
Control
+N
+P
+NP
(b)
DIN
(ug g
-1)
6
8
10
12
14
16
18
20
22
24
26
28
(c)T
ota
l N
(%
)
0.40
0.42
0.44
0.46
0.48
0.50
0.52
0.54
Years after fertilization
Pre Fert. 1 yr 2 yr
Melic
h P
(ug g
-1)
0
1
2
3
4
5
6
7
Years after fertilization
Pre Fert. 1yr 2 yr
Tota
l P
(ug g
-1)
1000
1200
1400
1600
1800
2000
2200
(d)
(e) (f)
Figure 3-4. Mean + SE soil parameters, including (A) pH, (B) percent total carbon,
(C)dissolved inorganic nitrogen, (D) percent total nitrogen, (E) Melich extractable phosphorus, and (F) total phosphorus for soils collected at a depth of 0-10 cm in four different fertilization treatments before, 1 yr and 2 yrs after initial fertilization.
A B
C D
E F
99
% d
iffe
rence
in
pH
0
5
10
15
20
C
+N
+P
+NP
% d
iffe
rence
in
DIN
(u
g g
-1 )
-100
-50
0
50
100
% d
iffe
rence
in
to
tal C
(%
)
-40
-20
0
20
40
% d
iffe
rence
in
to
tal N
(%
)
-40
-20
0
20
40
Depth (cm)
0-10 10-30 30-50
% d
iffe
rence
in
Me
lich
P(u
g g
-1)
-40
-20
0
20
40
60
80
100
120
Depth (cm)
0-10 10-30 30-50
% d
iffe
rence
in
to
tal P
(ug g
-1)
-10
-5
0
5
10
15
20
(a) (b)
(c) (d)
(e) (f)
aa
b
b
b
aa
b
Figure 3-5. Mean + SE percent change in various soil parameters two years after
fertilization for samples collected at three different depths. Different letters represent significant differences among treatments.
A B
C D
E F
100
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9 B 10-30 cm dbh
B
asa
l are
a incre
ase
(m
2 h
a-1
yr--1
)
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35A 5-10 cm dbh
Years after fertilization
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Ba
sa
l are
a incre
ase
(m
2 h
a-1
yr-1
)
0.0
0.2
0.4
0.6
0.8
1.0
1.2 C >30 cm dbh
Years after fertilization
0.0 0.5 1.0 1.5 2.0 2.5 3.00.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
D All size classes
Control
+N+P+NP
Figure 3-6. Mean (+ SE) total basal area increase per treatment for trees within (A) 5-10
cm DBH, (B) 10-30 cm DBH, (C) >30 cm DBH and (D) all size classes combined, for the four nutrient addition treatments measured between 0.4 and 2.7 yrs after initial fertilization. Values for all trees per plot were summed and plots for each treatment averaged.
101
Control
+N+P+NP
B 10-30 cm dbh
Ave
rag
e p
rop
ort
ion o
f tr
ee
gro
wth
(% tre
es tha
t g
rew
yr-
1 p
lot-
1)
0
20
40
60
80
100
A 5-10 cm dbh
Years after fertilization
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Ave
rag
e p
rop
ort
ion o
f tr
ee
gro
wth
(% tre
es tha
t g
rew
yr-
1 p
lot-
1)
0
20
40
60
80
100
C >30 cm dbh
Years after fertilization
0.0 0.5 1.0 1.5 2.0 2.5 3.0
D All size classes
Figure 3-7. Mean (+ SE) percentage of trees that grew per plot for trees within (A) 5-10
cm DBH, (B) 10-30 cm DBH, (C) >30 cm DBH and (D) all size classes combined, for the four nutrient addition treatments measured between 0.4 and 2.7 yrs after initial fertilization.
102
Treatment
Control +N +P +NP
RG
R (
ln (
mm
yr
-1))
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2C > 30 cm dbh
RG
R (
ln (
mm
yr
-1))
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5A 5-10 cm dbh
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
B 10-30 cm dbh
Treatment
Control +N +P +NP-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5D All size classes
combined
Figure 3-8. Box plots of relative growth rates (RGR) measured between 0.4 and 2.7 yrs
after initial fertilization, in the four nutrient addition treatments. The three size classes are shown separately (A-C) and then combined (D).
103
RG
R (
ln(m
m y
r-1))
0.0
0.2
0.4
0.6
0.8
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
RG
R (
ln(m
m y
r-1))
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Treatment
Control +N +P +NP
RG
R (
ln(m
m y
r-1))
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
Treatment
Control +N +P +NP-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
A Dendropanax arboreus
F3,17 = 0.08, p = 0.97B Goethalsia meiantha
F3,46 = 3.48, p = 0.02
C Inga
F3,21 = 0.53, p = 0.66D Pentaclethra macroloba
F3,89 = 0.99, p = 0.40
E Protium
F3,39 = 0.79, p = 0.51F Socratea exohrriza
F3,152 = 3.71, p = 0.01
* *
* **
Figure 3-9. Box plots showing relative growth rates for six common tree species,
measured between 0.4 and 2.7 yrs. after initial fertilization, in the four nutrient addition treatments. Shown are also results from 1-way ANOVAs with treatment as dependent variable. An Asterisk indicates that the mean for the treatment differs from the control (Dunnett’s test).
104
% s
ee
dlin
gs a
live
0
20
40
60
80
100
1yr
2yr
% s
ee
dlin
gs t
ha
t gre
w
0
20
40
60
80
100
Treatment
Control +N +P +NP
% s
ee
dlin
gs in
cre
ase
d n
o.
lea
ve
s
0
20
40
60
80
100
Treatment
Control +N +P +NP
% s
ee
dlin
gs w
ith
he
rbiv
ory
0
20
40
60
80
100
(a) (b)
(c) (d)
Figure 3-10. Mean (+ SE) percent of seedlings (A) alive, (B) with increased stem length,
(C) with increased number of leaves, and (D) with evidence of herbivory for the four nutrient addition treatments one and two years after initial fertilization.
A B
C D
105
Fo
liar
litte
rfa
ll (M
g C
ha
-1 y
r-1)
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
Control
+N
+P
+NP
Years after fertilization
1 yr 2 yr 3 yr
Re
pro
du
ctive
litte
rfa
ll (M
g C
ha
-1 y
r-1)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
(a)
(b)
Figure 3-11. Mean (+ SE) foliar litterfall production (foliage + sticks <2 mm in diameter) (A) and reproductive litterfall production (flowers and fruits) (B) for the four nutrient addition treatments during the first, second and third year after initial fertilization.
A
B
106
Interval after fertilization (yrs)
0.4-0.8 0.9-1.2 1.3-1.6 1.7-2.0
Co
ars
e litte
rfa
ll (M
g C
ha
-1 y
r-1)
0
2
4
6
8
10
12
14
Control
+ N
+ P
+ NP
Figure 3-12. Mean (+ SE) coarse litterfall production (palm fronds and woody debris >2 mm diameter for the four nutrient addition treatments in four intervals between 0.4 and 2 yrs. after initial fertilization.
107
Fin
e r
oo
ts (
Mg C
ha
-1)
0.5
1.0
1.5
2.0
2.5
3.0
(a)
Years after fertilization
Pre Fert. 1 yr 2 yr
Larg
e r
oo
ts (
Mg C
ha
-1)
0
1
2
3
4
5
(b)
Control+N+P+NP
Figure 3-13. Mean (+ SE) fine root biomass (roots <2mm diameter) (A) and large root
biomass (roots >2mm diameter) (B) collected at 0-15cm depth for the four nutrient addition treatments before, 1yr, and 2yrs after initial fertilization.
A
B
108
Treatment
Control +N +P +NP
Root P
rod
uctivity (
Mg C
ha-1
yr-1
)
0.0
0.2
0.4
0.6
0.8
Figure 3-14. Box plot (25th percentile, median and 75th percentile) for root production
from ingrowth cores installed from 0-10cm depth and left in the field for two years for the four different treatments.
109
CHAPTER 4
EFFECT OF NUTRIENT ADDITIONS ON FOLIAR, LITTER AND ROOT CHEMISTRY
Introduction
Tissue nutrient concentrations have been used extensively as indicators of plant
nutritional status (e.g., Stone 1968, Bowen and Nambiar 1984, Drechsel and Zech
1991, Marschner 1995).Relative abundances of nitrogen (N) and phosphorus (P) in
leaves have been used to infer nutrient limitation of terrestrial net primary productivity
(NPP; Koerselmann and Meuleman 1996, Güsewell 2004, Reich and Oleskyn 2004,
McGroddy et al. 2004, Ågren 2008, Cleveland et al. 2011). To grow and accumulate
biomass through photosynthesis (i.e., NPP) plants need proteins, which are largely
constituted by N. To construct these proteins, however, plants need ribosomes and
ATP, which are largely constituted of P. Thus, recent studies on ecological
stoichiometry have proposed that plants require a defined ratio of N to P (usually 14-16
mass based) and that deviations from this ratio indicate a relative abundance (or
limitation) of one of the nutrients relative to the other (Sterner and Elser 2002, Ågren
2004). For example, across large spatial scales, foliar N:P ratios increase towards the
tropics, which suggests that at low latitudes N is relatively more abundant than P. Lower
N:P ratios in high latitudes, however, indicate that P is relatively more abundant than N.
These differences in N:P ratios have been used to infer that in the tropics, NPP is likely
limited by P and in the temperate systems by N (McGroddy et al. 2004, Reich and
Oleksyn 2004, Hedin 2004, Kerkhoff et al. 2005).
Other factors in addition to soil nutrient supply, however, can influence tissue
nutrient concentrations (and their ratio). For example, seasonality, life history traits,
species identity, and herbivory, have been shown to influence foliar N:P ratios (reviewed
110
by Ågren 2008). Thus, in some communities, N:P ratios co-vary predictably with
relative nutrient abundance or limitation; but in other cases, intrinsic physiology and
external environmental factors decouple plant N:P ratios from relative nutrient
availability (e.g., Townsend et al. 2007). Therefore, interpretation of foliar (and other
tissue) N:P ratios can be challenging because of partitioning of control between nutrient
availability and other environmental and physiological factors (Ågren 2008).
One approach to elucidate the relationship between NPP and tissue nutrient
concentrations in a given system is to “calibrate” N:P values by comparing them to NPP
responses to fertilization (Vitousek et al. 2010, Cleveland et al. 2011). However, only a
few studies have tested if in practice, N:P ratios correlate with responses to nutrient
additions in natural systems (e.g., Koerselmann and Meuleman 1996, Vitousek and
Farrington 1997, Ostertag 2001). Moreover, there are few studies that have explored
the variation in responses to nutrient additions among tree species or tree size classes
(Ostertag 2010). In this study, I conducted a fertilization experiment in a wet, lowland
tropical forest in Costa Rica to explore if trees respond to nutrient additions by changing
their tissue (foliar, litter and root) N and P concentrations, and if these responses relate
to relative growth rates (RGR) of trees.
Although most research has focused on foliar nutrients, N and P concentrations
in litterfall can provide information regarding the nutritional status of plant communities.
High N and P concentrations in green leaves are typically associated with higher
concentrations in litter (Kobe 2005, Hättenschwiler et al. 2008, Wood et al. 2009).
However, the few fertilization studies conducted in the tropics have demonstrated,
regardless of nutrient limitation to NPP, that litterfall P is more likely to increase with P
111
fertilization than litterfall N with either N or P fertilization (Tanner et al. 1992, Vitousek
1998, Vitousek 2004, Kaspari et al. 2008, Wright et al. 2011). This can be partly
attributed to a higher increase in green leaf P than in green leaf N after fertilization
(Ostertag 2010). Relative differences in nutrient resorption efficiency–the withdrawal of
nutrients from senescing plant tissue and the transport of those nutrients to other plant
tissues (Killingbeck 1996) – can also affect litterfall N and P concentrations. For
example, in low P soils, P resorption is generally higher than N resorption, making
N:Plitter larger than N:Pgreen (e.g., Kerkhoff et al. 2005, Richardson et al. 2008, Wood
et al. 2011). Thus, interpretation of N:P litter in fertilization experiments must be
conducted within the context of the responses observed for N:Pgreen and for individual
N and P concentrations.
Root N and P concentrations have not been widely used as indicators of nutrient
limitation but they are generally correlated positively with foliar nutrient concentrations in
natural systems (Kerkhoff et al. 2006, Ågren 2008, Elser et al. 2010). Consistent with
green leaves, there is a decline in root N:P ratio with latitude, although this relationship
is exponential rather than linear (Yuan et al. 2011). Root responses to fertilization, in
contrast, have not always been consistent with aboveground responses (Wright et al.
2011, Yavitt et al. 2011 but see Ostertag 2001). Fundamental physiological,
morphological and functional differences between leaves and roots may be the cause
for such discrepancy. For example, high N demand for RuBisCO enzyme (Ribulose-1,5-
bisphosphate carboxylase oxygenase) in leaves –but not in roots– can lead to a greater
N concentration per unit carbon lost in dark respiration for leaves than roots and can
also lead to green leaves requiring more N relative to P than fine roots (Reich et al.
112
2008). This in turn may result in differences in N to P stoichiometry between these two
types of tissues. However, in a compilation from various studies that included multiple
sites, Yuan et al. (2011) found no significant differences among mean C:N:P ratios of
green leaves (1,212:16:1), litter (1, 016:19: 1), or live roots (1,053:16:1).
I conducted a full factorial NP fertilization experiment at the EARTH (Escuela de
Agricultura de la Región del Trópico Húmedo) University Forest Reserve, Costa Rica.
There, I investigated whether foliar, litter, and root N and P concentrations increased
after fertilization, and whether these increases were related to tree growth indices. In
Chapter 3, I reported that 2.5yrs after fertilization, small trees (5-10 cm diameter) had a
higher growth in plots where both N and P were added simultaneously (+NP treatment)
but there were no differences among treatments in litterfall production, or root biomass
and production. Thus, I did not expect to observe community-wide responses in foliar,
litter, and root N and P concentrations when N or P were added individually (+N or +P
treatments). However, I expected a potential increase in nutrient concentrations in the
+NP treatment. Moreover, I expected to find mean foliar N:P ratios between 14-16,
indicating N and P co-limitation (Sterner and Elser 2002, Ågren 2004). I expected litter
and root nutrients to track foliar nutrient concentrations (Ågren 2008).
Lowland tropical forests are characterized by a high variation in foliar and litter
chemistry (Townsend et al. 2007, Hättenschwiler et al. 2008). However, there is little
information on how this variation influences plant-soil-microbial feedbacks related to
nutrient limitation (Townsend et al. 2008). To explore how nutrient additions impact
different ecological groups within the ecosystem, I compared the effect of fertilization on
foliar N and P concentrations in trees from two size classes (trees with diameter at
113
breast height (DBH) between 5-10cm were considered “small trees” and trees with DBH
>10 cm were considered “large trees”). For small trees, I expected an increase in foliar
nutrient concentrations with fertilization because trees in this subcanopy size class are
usually light limited, which prevents them from incorporating added nutrients as new
growth (Lambers et al. 1998). In contrast, I expected that large trees, which have
access to light in the canopy, would incorporate added nutrients as new growth
immediately after each addition (probably as leaves or fruits). I therefore did not expect
an increase in foliar nutrient concentrations in large trees.
I investigated six taxa-specific responses to fertilizer additions including the palm,
Socratea exohrriza and the legume Pentaclethra macroloba (Table 3-1). Overall, I
expected that nutrient concentration responses would be consistent with tree growth
responses (e.g., Vitousek 2004). Thus, I expected Socratea to increase foliar P
concentrations after P additions because this species showed a significant increase in
growth after P additions (Chapter 3). In contrast, I did not expect legumes in general
and Pentaclethra (the most abundant species) in particular, to increase foliar N or P
concentrations after nutrient additions because neither “legumes” nor Pentaclethra
showed a RGR response after N or P additions.
Methods
Experimental Design
The study was conducted at the Forest Reserve of the EARTH University (Escuela
de Agricultura de la Región del Trópico Húmedo), in Guácimo, Limón, Costa Rica (10°
11’ N and 84° 40’ W). This private reserve is located approximately 30 m above sea
level and consists of 900 ha of mature and regenerating wet forest and wetlands. Mean
annual temperature is 25.1 °C and mean annual precipitation (MAP) is 3,464 mm. A
114
complete site description is provided in Chapter 3. In May 2007, I established 24 30 x
30 m plots and assigned them randomly to three fertilizer treatments or a control in a
complete block design (n = 6). Three blocks were located at the Rancho site and three
at the Rio site (Figure 3-3). Besides the control plots, the three treatments included +P
(47 kg ha-1yr-1 of P as super triple phosphate), +N (100 kg ha-1yr-1 of N applied as
ammonium nitrate and urea), and +NP (N and P added together in quantities as in +N
and +P plots). Fertilizer was broadcast by hand twice a year on the surface of the 900
m2 plots. All measurements were restricted only to the central 400 m2 of each plot (20 x
20 m) to reduce edge effects. In this plots, various soil measurements and
measurements of tree stem growth, litterfall productivity, root productivity and biomass,
were conducted (Chapter 3).
Foliage, Litterfall and Root Collection
I collected foliage, litterfall and root samples between July 2007 and September
2009 in the experimental plots at EARTH Forest (described in Chapter 3). In each plot I
collected fully expanded, sun-leaves from common tree species using a pole pruner, or
crossbow with a bolt affixed with monofilament line. One sample was composed of a
group of at least ten leaves per tree placed in a bag. Whenever possible, I attempted to
collect samples from the same taxa as described in Chapter 3 (Dendropanax arboreus,
Goethalsia meiantha, Pentaclethra macroloba, Socratea exohrriza, Inga sp., Protium sp.
and Virola sp.). Overall, I collected foliar samples from 286 trees representing 36
genera and 46 species. When possible, samples were collected from the same 4-8
large trees (>10cm DBH) and 4-8 small trees (4-9 cm DBH) per plot on three occasions:
prior to fertilization, 1 yr, and 2 yrs after initial fertilization between the months of July
and September. When it was not possible to collect leaves from the same trees
115
(because a tree had died, lost its leaves or showed significant herbivory), foliage was
collected from another tree in the same plot and size category.
Litterfall was collected using polyvinyl chloride (PVC) traps as described in
Chapter 3. From the litterfall collected during the 2 yr study, three representative
subsamples were used to conduct chemical analyses. The first subsample (pre-
fertilization) included foliar litterfall collected between 23 August and 9 September 2007.
The second subsample (1 yr after fertilization) included foliar litterfall collected between
8 October and 5 November 2008 and the third subsample (2 yrs after fertilization)
included foliar litterfall collected between 9 July and 10 August 2009. Each of these
subsamples was separated into species, when possible, and the rest of the leaves
analyzed collectively.
Roots were collected in each plot using a pound core, as described in Chapter 3,
in three occasions: prior to fertilization, 1 yr and 2 yrs after initial fertilization. In addition,
roots obtained from ingrowth cores installed at the beginning of the experiment and
retrieved after two years in the field (Chapter 3), were also included in the chemical
analyses.
Chemical Analysis
To conduct chemical analyses, foliar, litterfall and root samples were dried at 60
°C and ground using a Wiley Mill (Thomas Scientific, Swedesboro, New Jersey, U.S.A.)
passed through a #40 screen or a coffee grinder. I measured total percent N and C with
an elemental analyzer (ECS 4010, Costech Analytical, Valencia, California,U.S.A.). I
measured P using an ash digestion (Jones and Case 1996) followed by colorimetric
determination of ortho-phosphate using a spectrophotometer microplate reader
116
(PowerWave XS Microplate Reader, Bio-Tek Instruments Inc., Winooski, Vermont,
U.S.A.).
To include nutrient concentrations in an area basis, and to test for changes in
leaf physical properties with fertilization, I calculated specific leaf area (SLA) for foliar
samples collected 1yr and 2 yrs after initial fertilization. I scanned five leaves from each
foliage sample using an image scanner and measured leaf area using Beta 4.0.3, Scion
Image software. I then obtained leaf dry weight for each sample and calculated SLA (g
cm-2) as dry weight (g)/leaf area (cm2).
Statistical Analysis
To test the effect of fertilization treatments on foliar, litter and root chemistry, I
used plot-averaged values (n = 24) in repeated measures MANOVAs with time specific
measurements as dependent variables, and treatment and block as independent
variables. I selected this approach over univariate repeated measures ANOVA because
in some cases the sphericity assumption was not met (Field 2009). When there was a
treatment*time interaction, I conducted a one-way ANOVA on each time point to
determine when there was a significant treatment effect. To test for treatment effects on
foliar nutrients from individual species, I used individual trees as the sampling unit and
did not include a block effect because there were not enough replicates from each
species in all blocks. To test the relationship between plot-averaged foliar nutrients and
soil parameters described in Chapter 3, I used simple linear regressions.
When analyzing litterfall data, it was challenging to find samples from a given
species in each plot and across the three time points. Therefore, I analyzed separately
only Pentaclethra macroloba (the most common species) and a species group (termed
“Tiliaceae”), which included primarily Goethalsia meiantha with traces of Apeiba
117
membranaceae or Luehea seemannii. Because of small sample sizes for these two
groups (Pentaclethra and Tiliaceae), I conducted T-tests for each time point and
grouped the treatments in +N (+N and +NP plots) versus -N (control and +P plots),
when comparing litterfall %N and in +P (+P and +NP plots) versus -P (control and +N
plots), when comparing litterfall P concentrations.
When analyzing the response of root P to fertilization (using repeated measures
MANOVA), I excluded data from plot 19. This plot was an outlier that had root P values
76% higher than the total average (mean root P including controls and fertilized plots =
0.56 + 0.02 mg g-1versus plot 19 P = 0.97 mg g-1). In all figures, tables, and text, means
(+ standard errors) are shown. All figures were constructed using Sigmaplot 12.0 and
analyses were conducted in JMP 8.0(SAS Institute Inc., Cary, NC, USA).
Results
Foliar Nutrients
Two years after initial fertilization, there was no difference among treatments in
plot-averaged foliar N or P concentrations but there was a significant difference among
blocks for both N and P (Table 4-1 and Figure 4-1F). Block effects were mainly driven
by differences among the two study sites. Mean foliar N was higher in the Rio site than
in the Rancho site before (T = 3.31, P < 0.01) and 2 yrs after fertilization (T = 2.89, P
<0.01). Mean foliar P was similar between sites before fertilization (t = 1.93, P = 0.07)
but higher in the Rio site 2 yrs after fertilization (T = 2.32, P = 0.03). Interestingly, this
difference disappears when considering foliar P by leaf area (T = 0.57, P = 0.57),
probably because SLA was slightly higher at the Rancho site (SLA Rancho = 89.72 g m
-2 versus SLA Rio = 87.27 g m-2). Finally, there was no significant difference among
treatments for foliar N (treatment F3,23 = 0.38, P = 0.77; block F5,23 = 0.76, P = 0.60) or P
118
(treatment F3,23 = 0.89, P = 0.47; block F5,23 = 0.72, P = 0.63) in the change of foliar
nutrients 2 yrs after initial fertilization.
Overall, large trees (>10cm DBH) had higher foliar P (T = 2.28, P = 0.02) but not
N (T = 0.082, P = 0.41) concentrations than small trees (5-10 cm DBH), both before and
2 yrs years after initial fertilization. In addition, 2 yrs after initial fertilization, large trees
had drastically higher N and P per unit leaf area than small trees (N: T = 6.89, P <0.01,
P: T = 6.72, P <0.01). I did not measure foliar nutrients by unit area prior to fertilization.
Two years after initial fertilization, large trees had higher mean foliar N in the +NP
treatment, although this result was not consistent across blocks (Table 4-1 and Figure
4-1C). Mean Foliar P, on the contrary, differed significantly among treatments only for
small trees, although this effect was likely driven by pre-fertilization differences among
treatments (Table 4-1 and Figure 4-1B). Overall, there was high variability in foliar N:P
ratios (Figure 4-2) and there was no difference across treatments, tree sizes, or sites
after fertilization (Table C-2).
I compared foliar nutrients among four species and two genera where replicates
were sufficient to conduct statistical analyses (Table 3-1). Foliar N, P and N:P ratios
differed significantly among taxa (Table 4-2) but only one species (Pentaclethra
macroloba) and one genus (Protium) differed among treatments in foliar N or P. Two
years after initial fertilization, the species Pentaclethra macroloba had higher foliar N in
the +NP treatment, (Table 4-3, Figure 4-3D; one-way ANOVA at 2 yrs post fertilization:
F3,41 = 3.08, p = 0.04). One year after initial fertilization, Protium sp. trees had higher
foliar N in the +N treatment (Table 4-3, Figure 4-3E) but this effect disappeared 2 yrs
after initial fertilization, probably due to the destruction of two +N plots in the Rio site
119
during a strong wind event (see methods section, Chapter 3). Although not statistically
significant, 2 yrs after initial fertilization the palm, Socratea exohrriza, had 15% and 19%
higher mean foliar P relative to the control in the +P and +NP treatments, respectively
(Table 4-4, Figure 4-4F). Foliar N:P ratios differed among treatments for Dendropanax
but this result was based on one individual and should therefore be interpreted with
caution (Figure 4-5A). For the rest of the study species N:P ratios did not differ among
treatments, blocks or sampling times (Figure 4-5; Table C-3).
Because community-averaged foliar nutrients were not strongly influenced by
fertilization additions, I conducted linear regressions on several soil parameters
measured before and 2 yrs after initial fertilization against foliar N and P to evaluate
which soil conditions were good predictors of foliar nutrient concentrations. Overall,
foliar nutrients were best predicted by soil pH, net nitrification rates and soil P (both
Melich and total P; Figure 4-6 and 4-7). Foliar N was negatively related to net
nitrification (Figure 4-6C) but this relationship disappeared with the fertilization
treatments (Figure 4-6D). Interestingly, soil P was not a good predictor of foliar %N
before fertilization (Figure 4-6E) but became positively correlated 2 yrs after fertilization
(Figure 4-6F). In contrast, soil P was a good predictor of foliar P before fertilization
(Figure 4-7E) but this relationship disappeared 2 yrs after initial fertilization (Figure 4-
7F). None of the other measured soil parameters were good predictors for foliar N or P.
Finally, mean foliar N:P ratios were not related to any of the soil parameters measured
(Figure C-1).
Litterfall Nutrients
Overall, litterfall nutrient concentrations were high (Figure 4-8) and varied over
time. This time effect probably indicates differences in the relative contribution of
120
species to the litter samples at different collection times, or differences in environmental
conditions among years. Mean litterfall N did not differ among treatments or blocks;
therefore I eliminated “block” from the model. When the non-significant block effect was
removed, the treatment effect became significant (Table 4-5). Plots where N was added
had higher litterfall N concentrations than plots where no N was added (Figure 4-8).
There was no significant treatment effect on mean litterfall P. However, there was a
significant block effect, which was caused by higher litterfall P in the Rio side (MANOVA
by side F1,19 = 13.34, P<0.01). There were no differences among treatments or blocks
over time in litterfall N:P ratios (Table 4-5), and mean foliar N:P ratios were not related
to mean litterfall N:P ratios (r2 = 0.01, P = 0.65). In addition, mean foliar nutrients (N and
P) were not related to mean litterfall nutrients either before or 2 yrs after initial
fertilization. Both Pentaclethra and Tiliaceae increased their litter N concentrations with
N additions but there was no increase in litter P after P additions (Figure 4-9).
Root Nutrients
There was no significant difference among treatments in root N but roots in the
+P and +NP treatments had higher P concentrations 2 yrs after initial fertilization (Table
4-6, Figure 4-10). There was also a significant block effect, which reflects the higher
root P in the Rio site, mainly 2 yrs after fertilization (mean P Rancho = 0.50 + 0.02 mg g-
1 and mean P Rio = 0.62+ 0.03mg g-1; T = 3.13 P< 0.01). This higher root P in the Rio
side also resulted in a significant block effect when comparing N:P ratios (Table 4-6,
Figure 4-10C). Roots that had grown into ingrowth cores over the course of the
experiment did not show differences among treatments or blocks in N (treatment F3,22 =
0.52, P = 0.67, block F5,22 = 0.85, P = 0.54) or P (treatment F3,22 = 3.08, P = 0.42, block
F5,22 = 0.45, P = 0.80) concentrations.
121
Prior to fertilization, there was a positive relationship between root N and P
concentrations (r2 = 0.41, d.f. = 23, P < 0.01) but one (r2 = 0.10, d.f. = 23, P = 0.10) and
two years (r2 = 0.01, d.f. = 23, P = 0.82) after fertilization this relationship disappeared.
However, in ingrowth cores installed at the beginning of the experiment and extracted 2
yrs after fertilization, there was a positive relationship between N and P concentrations
(r2 = 0.20, d.f. = 23, P = 0.03). Overall, root N could not be predicted by any soil
parameter (data not shown) but root P was strongly positively correlated to both Melich
P and total P before (Melich P vs. root P r2 = 0.45, d.f. = 23, P < 0.01; total P vs. root P
r2 = 0.31, d.f. = 23, P < 0.01) and 2 yrs after fertilization (Melich P vs. root P r2 = 0. 70,
d.f. = 23, P < 0.01; total P vs. root P r2 = 0.56, d.f. = 22, P < 0.01). In addition, root P
was positively correlated with foliar P before (r2 = 0.22, d.f. = 23, P = 0.02) and 2 yrs
after fertilization (r2 = 0.24, d.f. = 23, P = 0.02).
Discussion
Effects of Fertilization on Foliar Nutrients
Overall, both N and P concentrations at EARTH Forest were high relative to other
tropical forests where responses to fertilization were observed. In a P limited forest in
Hawaii (Kokee), mean foliar P was 0.55µg g-1 and increased to 1.88µg g-1 after long-
term fertilization. In my study plots, mean foliar P before fertilization was 1.25µg g-1.
This value is even higher than foliar P after long-term P fertilization at a N- limited forest
in Hawaii (Foliar P = 0.95µg g-1; Harrington et al. 2001). Foliar N concentrations at
EARTH Forest were high as well, with a mean of 2.65% and values up to 5%. High
foliar nutrient concentrations (reported here), in addition to high soil N and P values, and
a lack of response in total DBH increase or litterfall production after fertilization (Chapter
122
3), suggest that aboveground net primary productivity (ANPP) at EARTH forest is
probably not limited strongly by N or P availability.
Although most NPP components did not significantly change with N or P additions,
I observed a 37.4 % increase in mean tree RGR with +NP additions, which could be
interpreted as NP co-limitation (Chapter 3). Therefore, I expected to observe a mean
foliar N:P ratio between 14 and16 (mass based). However, mean foliar N:P ratio was
22.13, which is considered indicative of P limitation (Sterner and Elser 2002, Güsewell
2004, Ågren 2004, 2008). In fact, this value is higher than mean N:P ratio at Osa
Peninsula (16.4 + 4.7), a low P forest in the south of Costa Rica (Townsend et al. 2007).
Likely, the high N:P ratio at EARTH Forest reflects a high legume abundance (which are
high in N), and suggests that in this diverse forest other factors, such as species identity
are more important controls on leaf stoichiometry than relative nutrient supply.
Consistent with other tropical forests, there was great variation across species in
foliar nutrient concentrations (Townsend et al. 2007, Hattënschwiler et al. 2008; Figure
4-2). Species mean foliar P, for example, ranged from 0.67 to 2.16µg g-1, which is 57%
of the range of values for five lowland tropical forests combined (~ 0.25 – 2.75µg g-1;
Townsend et al 2007). This high variability reduces statistical power to detect
differences in foliar nutrient concentrations after fertilization. However, I conducted
measurements before, 1 yr and 2 yrs after fertilization, mostly in the same trees, and
thus it is likely that a strong and consistent across-species response to fertilization
would have been evident regardless of the variability (e.g., Ostertag 2010).
Alternatively, if there were strong but contrasting responses among species or even
among trees, these could have veiled any community response. In fact, my data
123
suggest that although at the community-level there may be no apparent response of
mean foliar nutrient concentrations to fertilization and therefore no apparent nutrient
limitation, there may be subtle and contrasting responses from individual taxa within the
community, which could have important consequences for ecosystem processes, such
as litter decomposition and nutrient turnover rate (Campo and Dirzo 2003, Vitousek
2004).
Influence of Tree Size on Foliar Nutrients
Large trees responded to fertilization by increasing their foliar nutrient
concentrations but not their RGR, and small trees responded by increasing their RGR
but not their foliar nutrients (Figure 4-1 and Chapter 3). These results indicate that,
contrary to expectations, small trees allocated the extra nutrients to growth while large
trees showed “luxury consumption” after fertilization and reveal a difference in nutrient
use strategies among age groups (Wright et al. 2011). In large trees, there was a 6.7%
increase in foliar N with +NP additions but no foliar P response. The significant increase
in foliar N after fertilization was unexpected given that high foliar N concentrations have
been associated with high levels of herbivory (Huberty and Denno 2006 but see Campo
and Dirzo 2003). Furthermore, after long term (> 10 yrs) fertilization, foliar P is more
likely to respond to changes in nutrient availability than foliar N, and it has been argued
that concentrations of foliar P are more plastic than concentrations of foliar N (Vitousek
2004, Ostertag 2010). I suspect that this apparent paradox is driven by the species-
specific response by the most abundant tree species in this forest, Pentaclethra
macroloba, which responded to fertilization by increasing foliar N after +NP additions.
124
Influence of Taxa on Foliar nutrients
Individual trees of Pentaclethra macroloba (mainly >10 cm DBH individuals or
“large trees”), had approximately 10% higher foliar N concentrations in the +NP
treatment, relative to the control. Interestingly, there was no difference in foliar N
between the +N treatment and the control. This response could be interpreted as N and
P co-limitation (Elser et al. 2007, Townsend et al. 2008); although in that case I would
have expected a simultaneous increase in foliar P in the +NP treatment. Possibly, this
species is allocating nutrients to different tissues, such as N to leaves and P to roots. I
did not separate roots by species, but there was a significant increase in mean root P
both in the +P and +NP treatments. To test this hypothesis, species-specific analyses of
nutrient concentrations in roots, leaves, and reproductive structures, should be
conducted.
The most abundant palm species at EARTH forest, Socratea exohrriza, was the
only taxon (from those observed individually) to show a strong response to fertilization
by increasing both RGR (Chapter 3) and foliar P concentrations in the +P and +NP
treatments. This can be interpreted as: (1) from the observed taxa, Socratea was the
only species clearly limited by P (sensu Chapin et al. 1986) or (2) Socratea has specific
life-history traits that influenced its responsiveness. For example, the response of
Socratea could have been influenced by its single stem architecture with no branches
(where a growth response would be missed by DBH measurements), relatively high
growth rate, and the formation of dense, superficial root mats (which would be beneficial
for fast nutrient uptake after fertilization (Henderson et al. 1995).It is likely that a
combination of both P limitation and life history traits, influenced the response of this
palm to fertilization. Interestingly, if Socratea was limited by P, as explained by the RGR
125
foliar P response, I would expect this species to have a high N:P ratio, at least relative
to other taxa in the same site. However, Socratea’s mean foliar N:P ratio was the lowest
among the studied taxa (16.93 + 0.37). This inconsistency suggests that in this system,
foliar N:P ratios may be more influenced by inherent species traits than by resource
availability (Townsend et al. 2007). Most importantly, these contrasting responses from
two of the most abundant tree species at EARTH Forest illustrate the importance of
considering species composition and life history traits when making interpretations
concerning nutrient limitation in diverse tropical forests (Townsend et al. 2007, Vitousek
et al. 2010, Cleveland et al. 2011).
Effects of Fertilization on Litterfall and Root Nutrients
Consistent with foliar nutrient responses to fertilization, there was an increase in
litterfall N concentrations in the +NP treatment. However, in contrast with foliar N
concentrations, which only increased when N and P were added together, litterfall N
also increased in the +N treatment. As postulated above, it is likely that the foliar
response was driven by the species-specific response of Pentaclethra. The litterfall
response, on the other hand, was determined by all leaves that fell in litter traps, and is
probably more representative of the community response to fertilization. This 13%
increase in litterfall N is equivalent to ~30 Kg N ha-1 yr-1 and is likely to have an impact
on decomposition and nutrient cycling (Aerts 1997, Hobbie and Vitousek 2000 but see
Kaspari et al. 2008). However, as for foliar N concentrations, most studies in the tropics
have reported a larger response in litter P than litter N after fertilization (Mirmanto et al.
1999, Hobbie and Vitousek 2000, Kaspari et al. 2008) and therefore, this result
deserves further investigation.
126
Contrary to foliage and litterfall, roots showed a significant increase in P
concentrations after P (and NP) fertilization. There was a 12% increase in root P with P
additions, which is equivalent to 0. 09 kg P ha-1 yr-1, calculated using root productivity
data obtained from root ingrowth cores (Chapter 3). Therefore, although the magnitude
of the response in litterfall N was similar to the response in root P (13% and 12%
increase respectively), lower root P concentrations and productivity make the additional
P input in roots significantly smaller than the additional N input in litterfall. To predict the
impact that these changes in nutrient concentrations will have on ecosystem dynamics,
further investigation that addresses the differences in decomposition rates between
litterfall and roots and potential seasonality of these responses, should be conducted
(e.g., Ostertag and Hobbie 1999, Cusack et al. 2009).
Inconsistent responses to fertilization between root and foliar concentrations can
be indicative of a decoupling between aboveground and belowground nutrient limitation
(Ostertag 2001, Kaspari et al. 2008, Wright et al 2011, Yavitt et al 2011). Further
measurements of belowground processes, including soil respiration, enzyme activities,
and microbial biomass and community composition, would be necessary to test this
hypothesis. Moreover, it has been shown that multiple nutrients regulate belowground
processes, such as decomposition rates, in a seasonal forest in Panama (Kaspari et al.
2008). Thus, future research focusing on belowground dynamics would ideally include
fertilization with other nutrients, in addition to N and P (Chapter 5).
Total Soil P as a Driver of Tissue Nutrient Concentrations
Traditionally, total soil P, or P that can only be extracted by strong acid assays
(Hedley et al. 1982), has been considered of no practical importance in tropical soils
(Sanchez 1976). Because these soils are generally highly weathered with high iron and
127
aluminum content, most P is found in crystalline or occluded forms associated with
secondary minerals (Cross and Schlesinger 1995, Tiessen 1998) and is considered
unavailable for plant use. Although this might be true in the short term, the tight
correlation between total soil P and foliar and root P observed across plots in this study,
suggests that total soil P is a strong predictor of foliar and root P at local (200 m2)
spatial scales and that in the long term some of this “occluded P” becomes available to
plant use (Turner and Engelbrecht 2011). Moreover, there was a strong correlation
between total P and available P (Chapter 3), and across a precipitation gradient
available P was a good predictor of mean foliar P (Chapter 2). In both the precipitation
gradient study and the fertilization study, foliar N could not be predicted by any soil
parameter and there was no strong relationship between soil N and P. Thus, at least for
these forests in Costa Rica, it could be stated that total P is a good indicator of the P
available for plant uptake. The lack of relationship between foliar P and tree growth,
however, deserves further investigation.
Conclusions
Results from this chapter support NPP results reported in Chapter 3, which
demonstrates that at EARTH Forest there is no strong limitation by N or P but that there
is some evidence of co-limitation by these two nutrients. However, the mechanism
driving this pattern is probably not a simultaneous scarcity of N and P but rather
“heterogeneous nutrient limitation” within the tree community. This is illustrated by the
foliar N increase in large Pentaclethra trees in the +NP treatment, and the RGR and
foliar P increase in Socratea palms in the +P and +NP treatments. Contrary to results
obtained in monospecific forests in Hawaii (Vitousek 2004), in this diverse tropical
forest, responses to experimental nutrient additions appear to be driven by the
128
interaction of functional traits (or at least species-specific traits) and resource availability
(Chapin, Vitousek and Van Cleve 1986). For example, N:P ratios at EARTH forest
reflect the high abundance of legumes (Pentaclethra in particular) more than the
nutritional status of the plant community. This highlights the importance of considering
both species identity and relative species abundance when making predictions related
to nutrient limitation in this forest. Moreover, the relative importance of a species-
specific response for the community-wide signal is going to depend not only on the
strength of the response but also on the relative abundance of that species in the
community.
Although it is clear that N:P ratios can provide insight into the relative availability of
these nutrients at larger scales (Koerselmann and Meuleman 1996, Güsewell 2004,
Reich and Oleskyn 2004, McGroddy et al. 2004, Ågren 2008, Cleveland et al. 2011), it
has been noticed that at local scales, variation in N:P ratios may reflect more
differences in intrinsic physiology than in the external environment (Townsend et al.
2007, Ågren 2008). This appears to be the case in my study; therefore it is advisable
that for diverse tropical forests, nutrient limitation assertions that result from
interpretation of N:P ratios are made at the functional group, genera, or species level,
rather than generalizing nutrient limitation from a single mean value.
Finally, further research is needed to explore how the contrasting responses to N
versus P additions observed in this experiment will affect nutrient cycling and carbon
storage in the long-term. For example, increased litter N and root P concentrations
could have drastically different effects by differentially changing nutrient turnover times,
microbial community composition, soil respiration, and/or greenhouse gas emissions. A
129
mechanistic understanding of how changes in nutrient availability affect nutrient uptake
and storage in plant tissues is critical for predicting the role that tropical forests will have
in mitigating increased N and P fluxes, which are caused by human alterations of global
biogeochemical cycles (Galloway and Cowling 2002, Galloway 2004, Okin et al. 2004,
Mahowald et al. 2008).
130
Table 4-1. Results from repeated measures MANOVAs for foliar chemistry by tree size class. F-values for treatment, block and time were obtained from exact tests but time*treatment and time*block interactions are F-value approximations resulting from Wilk’s lambda multivariate tests. In these cases degrees of freedom (df) are approximated as well. Significant effects are signaled with an asterisk.
dfn,d F Probability
Foliar %N 5-10 cm Treatment 3,15 2.60 0.09 Block 5,15 0.81 0.56 Time 2,14 0.80 0.47 Time*treatment 6,28 0.28 0.94 Time*block 10,28 0.57 0.82 > 10 cm Treatment 3,15 3.35 0.04* Block 5,15 8.22 <0.01* Time 2,14 0.52 0.61 Time*treatment 6,28 3.42 0.01* Time*block 10,28 0.94 0.51 Total Treatment 3,15 1014 0.41 Block 5,15 5.57 <0.01* Time 2,14 0.17 0.85 Time*treatment 6,28 1.73 0.15 Time*block 10,28 0.78 0.65 Foliar P (mg g-1) 5-10 cm Treatment 3,15 4.02 0.03* Block 5,15 2.58 0.07 Time 2,14 2.46 0.12 Time*treatment 6,28 2.25 0.07 Time*block 10,28 1.16 0.36 > 10 cm Treatment 3,15 0.72 0.56 Block 5,15 2.35 0.09 Time 2,14 1.66 0.23 Time*treatment 6,28 0.50 0.80 Time*block 10,28 0.65 0.76 Total Treatment 3,15 0.42 0.74 Block 5,15 3.79 0.02* Time 2,14 3.94 0.04* Time*treatment 6,28 1.36 0.27 Time*block 10,28 1.06 0.43
131
Table 4-2. Results from repeated measures MANOVAs comparing foliar %N, P (mg g-1) and N:P ratios among the six most common taxa (see table 3-1 for details). F-values for taxa and time were obtained from exact tests but time*taxa interactions are F-value approximations resulting from Wilk’s lambda multivariate tests. In these cases degrees of freedom (df) are approximated as well. Significant effects are signaled with an asterisk.
dfn,d F Probability
Foliar %N Taxa 5,107 78.50 < 0.01* Time 2,106 0.22 0.81 Time*taxa 10,212 1.82 0.06 Foliar P (mg g-1) Taxa 5,99 5.68 < 0.01* Time 2,98 2.78 0.07 Time*taxa 10,196 2.98 < 0.01* Foliar N:P ratios Taxa 5,98 52.67 < 0.01* Time 2,97 3.41 0.04* Time*taxa 10,194 2.70 < 0.01*
132
Table 4-3. Results from repeated measures MANOVAs for foliar %N by species. F-values for treatment and time were obtained from exact tests but time*treatment interactions are F-value approximations resulting from Wilk’s lambda multivariate tests. In these cases degrees of freedom (df) are approximated as well. Significant effects are signaled with an asterisk.
dfn,d F Probability
Dendropanax arboreus Treatment 3,6 3.71 0.08 Time 2,5 2.76 0.15 Time*treatment 6,10 2.46 0.10 Goethalsia meiantha Treatment 2,3 0.53 0.64 Time 2,2 0.91 0.52 Time*treatment 4,4 0.90 0.54 Inga Treatment 3,10 0.68 0.58 Time 2,9 0.18 0.84 Time*treatment 6,18 0.83 0.56 Pentaclethra macroloba Treatment 3,22 0.62 0.61 Time 2,21 2.16 0.14 Time*treatment 6,42 2.65 0.03* Protium Treatment 3,19 0.25 0.86 Time 2,18 0.76 0.48 Time*treatment 6,36 2.27 0.05* Socratea exohrriza Treatment 3,30 0.44 0.72 Time 2,29 3.73 0.04* Time*treatment 6,58 0.30 0.94
133
Table 4-4. Results from repeated measures MANOVAs for foliar P by species. F-values for treatment and time were obtained from exact tests but time*treatment interactions are F-value approximations resulting from Wilk’s lambda multivariate tests. In these cases degrees of freedom (df) are approximated as well. Significant effects are signaled with an asterisk.
dfn,d F Probability
Dendropanax arboreus Treatment 3,6 2.24 0.18 Time 2,5 1.51 0.31 Time*treatment 6,10 1.61 0.24 Goethalsia meiantha Treatment 2,3 2.43 0.24 Time 2,2 0.59 0.63 Time*treatment 4,4 3.26 0.18 Inga Treatment 3,9 0.20 0.90 Time 2,8 1.33 0.32 Time*treatment 6,16 1.41 0.27 Pentaclethra macroloba Treatment 3,20 0.80 0.51 Time 2,19 1.48 0.25 Time*treatment 6,38 0.66 0.68 Protium Treatment 3,19 2.23 0.11 Time 2,18 1.15 0.34 Time*treatment 6,36 0.53 0.77 Socratea exohrriza Treatment 3,25 0.79 0.51 Time 2,24 3.18 0.05* Time*treatment 6,48 2.01 0.08
134
Table 4-5. Results from repeated measures MANOVAs for litterfall chemistry. F-values for treatment, block and time were obtained from exact tests but time*treatment and time*block interactions are F-value approximations resulting from Wilk’s lambda multivariate tests. In these cases degrees of freedom (df) are approximated as well. Significant effects are signaled with an asterisk.
dfn,d F Probability
Litterfall %N Treatment 3,15 2.43 0.111
Block 5,15 2.16 0.11 Time 2,14 15.13 <0.01* Time*treatment 6,28 1.97 0.10 Time*block 10,28 01.16 0.36 Litterfall P (mg g-1) Treatment 3,15 2.66 0.09 Block 5,15 2.98 0.04* Time 2,14 11.68 <0.01* Time*treatment 6,28 0.78 0.72 Time*block 10,28 0.56 0.83 Litterfall N:P Treatment 3,15 1.23 0.34 Block 5,15 0.63 0.68 Time 2,14 8.05 <0.01* Time*treatment 6,28 0.30 0.93 Time*block 10,28 0.85 0.59
1 Treatment effect becomes significant if non-significant block effect is removed from analysis
(Time*treatment F6,38 = 2.37 p= 0.04).
135
Table 4-6. Results from repeated measures MANOVAs for root chemistry. F-values for treatment, block and time were obtained from exact tests but time*treatment and time*block interactions are F-value approximations resulting from Wilk’s lambda multivariate tests. In these cases degrees of freedom (df) are approximated as well. Significant effects are signaled with an asterisk.
dfn,d F Probability
Root %N Treatment 3,15 0.81 0.51 Block 5,15 1.14 0.38 Time 2,14 5.56 0.02* Time*treatment 6,28 1.47 0.22 Time*block 10,28 1.15 0.36 Root P (mg g-1) Treatment 3,14 2.29 0.12 Block 5,14 3.35 0.03* Time 2,13 31.9 <0.01* Time*treatment 6,26 4.32 <0.01* Time*block 10,26 2.11 0.06 Root N:P Treatment 3,15 0.93 0.45 Block 5,15 3.20 0.04* Time 2,14 11.38 <0.01* Time*treatment 6,28 1.31 0.29 Time*block 10,28 1.34 0.26
136
%N
2.3
2.4
2.5
2.6
2.7
2.8
2.9
3.0A 5-10cm
P (
mg
g-1
)
1.0
1.1
1.2
1.3
1.4
1.5Control
+N+P+NP
B 5-10cm
%N
2.3
2.4
2.5
2.6
2.7
2.8
2.9
3.0C >10cm
P (
mg
g-1
)
1.0
1.1
1.2
1.3
1.4
1.5
Years after fertilization
Pre Fert. 1 yr 2 yr
%N
2.3
2.4
2.5
2.6
2.7
2.8
2.9
3.0
D >10cm
E All (f) All
Years after fertilization
Pre Fert. 1yr 2 yr
P (
mg
g-1
)
1.0
1.1
1.2
1.3
1.4
1.5F All
Figure 4-1. Mean (+ SE) foliar N and P for trees with DBH between 5-10 cm (A, B),
larger than 10cm (C, D) and both size classes combined (E, F), for the four nutrient addition treatments measured before, 1yr and 2 yrs after initial fertilization.
137
All t
empe
rate
tree
s
All t
ropi
cal tre
es
Res
erva
Sam
uel
Cau
axi
EARTH
Pre
Fer
t.
EARTH
2yrs
Fo
liar
N:P
0
10
20
30
40
50
Figure 4-2. Box Plot comparing variability in N:P ratios at EARTH forest before
fertilization and 2yrs after fertilization with other sites. Redrawn with permission from Townsend et al. (2007). Boxes denote median with 50th and 75th percentiles.
138
Fo
liar
%N
1.5
2.0
2.5
3.0
3.5
4.0
4.5A Dendropanax
arboreus
Control
+N+P+NP
B Goethalsia meianthaF
olia
r %
N
1.5
2.0
2.5
3.0
3.5
4.0C Inga
Years after fertilization
Prefert 1 yr 2 yr
Fo
liar
%N
1.5
2.0
2.5
3.0
3.5
4.0
4.5
D Pentaclethra macroloba
E Protium (f) All
Years after fertilization
Prefert 1 yr 2 yr
F Socratea exohrriza
Figure 4-3. Mean (+ SE) foliar %N for six common tree species for the four nutrient
addition treatments measured before, 1yr and 2 yrs after initial fertilization.
139
Fo
liar
P (
mg
g-1
)
0.8
1.0
1.2
1.4
1.6
1.8
2.0A Dendropanax
arboreus
Control
+N+P+NP
B Goethalsia meianthaF
olia
r P
(m
g g
--1)
0.8
1.0
1.2
1.4
1.6
1.8
2.0
C Inga
Years after fertilization
Prefert 1 yr 2 yr
Fo
liar
P (
mg
g--1
)
0.8
1.0
1.2
1.4
1.6
1.8
2.0
D Pentaclethra macroloba
E Protium (f) All
Years after fertilization
Prefert 1 yr 2 yr
F Socratea exohrriza
Figure 4-4. Mean (+ SE) foliar P for six common tree species for the four nutrient
addition treatments measured before, 1yr and 2 yrs after initial fertilization.
140
Fo
liar
N:P
10
15
20
25
30
35 A Dendropanax
arboreus
Control
+N+P+NP
B Goethalsia
meianthaF
olia
r N
:P
10
15
20
25
30
35 C Inga
Years after fertilization
Prefert 1 yr 2 yr
Fo
liar
N:P
10
15
20
25
30
35
D Pentaclethra macroloba
E Protium (f) All
Years after fertilization
Prefert 1 yr 2 yr
F Socratea exohrriza
Figure 4-5. Mean (+ SE) foliar N:P ratios for six common tree species for the four
nutrient addition treatments measured before, 1yr and 2 yrs after initial fertilization.
141
Soil pH
3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8
Folia
r %
N
2.3
2.4
2.5
2.6
2.7
2.8
2.9
3.0
2
7
8
14
19
24
1
10
1213
1722
35
11
18
20
23
4
6
9
15
1621
Soil pH
3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8
2
7
8
14
19
24
1
10
12
131722
3
511
18
2023
46 15
21
Net Nitrification (ug N g-1 d
-1)
-2 -1 0 1 2 3 4 5
Folia
r %
N
2.3
2.4
2.5
2.6
2.7
2.8
2.9
3.0
2
7
8
14
19
24
1
10
1213
1722
35
11
18
20
236
9
15
1621
-2 -1 0 1 2 3 4 5
2
7
8
14
19
24
1
10
12
13 1722
3
511
18
2023
46 915
21
Net Nitrification (ug N g-1 d
-1)
Melich P (ug g-1)
0 1 2 3 4 5 6 7
Folia
r %
N
2.3
2.4
2.5
2.6
2.7
2.8
2.9
3.0
2
7
8
14
19
24
1
10
1213
1722
35
11
18
20
23
4
6
9
15
1621
0 1 2 3 4 5 6 7
2
7
8
1424
1
10
12
13 1722
3
511
18
2023
469 15
Melich P (ug g-1)
A R2 = 0.45, P < 0.001
B R2 = 0.17 P = 0.04
C R2 = 0.16 P = 0.05 D R
2 = 0.01 P = 0.89
E R2 = 0.10 P = 0.13 F R
2 = 0.21 P = 0.02
Figure 4-6. Relationship between several soil variables and plot-averaged foliar %N
before fertilization (A, C, E) and 2 yrs after initial fertilization (B, D, F). Symbols represent each plot and colors each treatment (Control= grey, +N = pink, +P= green, +NP = Cyan).
142
Soil pH
3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8
Folia
r P
(m
g g
-1)
1.0
1.1
1.2
1.3
1.4
1.5
1.6
2
7
814
1924
1
10
12
1317
223511
1820
234
69
15
16
21
Soil pH
3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8
2
78
14
19
24110
12
13
17223
5
11
18
20
23
4
615
16
21
Net Nitrification (ug N g-1 d
-1)
-2 -1 0 1 2 3 4 5
Folia
r P
(m
g g
-1)
1.0
1.1
1.2
1.3
1.4
1.5
1.6
2
7
814
1924
1
10
12
1317
2235
11
1820
23
6 9
15
16
21
-2 -1 0 1 2 3 4 5
2
7
814
1924
1
10
12
1317
223511
1820
234
6 9
15
16
21
Net Nitrification (ug N g-1 d
-1)
Melich P (ug g-1)
0 2 4 6 8 10 12 14
Folia
r P
(m
g g
-1)
1.0
1.1
1.2
1.3
1.4
1.5
1.6
2
7
814
1924
1
10
12
1317
223511
1820
234
69
15
16
21
0 2 4 6 8 10 12 14
2
78
14
19
24110
12
13
17223
5
11
18
20
23
4
6915
16
21
Melich P (ug g-1)
A R2 = 0.14, P = 0.07
B R2 = 0.01 P = 0.82
C R2 = 0.05 P = 0.29 D R
2 = 0.02 P = 0.49
E R2 = 0.19 P = 0.04 F R
2 = 0.15 P = 0.07
Figure 4-7. Relationship between several soil variables and plot-averaged foliar P (mg
g-1) before fertilization (A, C, E) and 2 yrs after initial fertilization (B, D, F). Symbols represent each plot and colors each treatment (Control= grey, +N = pink, +P= green, +NP = Cyan).
143
Litte
rfall
%N
2.0
2.1
2.2
2.3
2.4
2.5
2.6
2.7
Litte
rfa
ll P
(m
g g
-1)
0.7
0.8
0.9
1.0
1.1
1.2
Control
+N
+P
+NP
(a)
Years after fertilization
Pre Fert. 1 yr 2 yr
Litte
rfa
ll N
:P
20
22
24
26
28
30
32
34
(b)
(c)
Figure 4-8. Mean (+ SE) litterfall chemistry, including N concentration (A), P
concentration (B), and N:P ratios for the four nutrient addition treatments measured before, 1yr and 2 yrs after initial fertilization.
A
B
C
144
Figure 4-9. Mean litterfall nutrient concentrations by taxa, for the most common species
(A and C) and a common family (B and D) found in litterfall samples collected before, 1yr and 2 yrs after initial fertilization. Empty bars represent plots where no N (in A and B) or no P (in C and D) was added and filled bars represent plots where N (in A or B) or P (C and D) was added.
145
Ro
ot %
N
1.3
1.4
1.5
1.6
1.7
1.8
1.9
2.0
Ro
ot
P (
mg g
-1)
0.3
0.4
0.5
0.6
0.7
Control
+N
+P
+NP
(a)
Years after fertilization
Pre Fert. 1 yr 2 yr
Ro
ot
N:P
20
24
28
32
36
40
44
(b)
(c)
Figure 4-10. Mean (+ SE) root chemistry, including N concentration (A), P concentration
(B), and N:P ratios (C) for the four nutrient addition treatments measured before, 1yr and 2 yrs after initial fertilization. For root P analyses, an outlier (plot 19) was excluded (see methods for details).
C
B
A
146
CHAPTER 5 CONCLUSIONS AND LESSONS LEARNED
Conclusions
The purpose of this study was to investigate how environmental and biological
factors, such as mean annual precipitation (MAP) and tree species composition, can
influence nutrient limitation and availability in Costa Rican lowland tropical forests. I
used both large-scale observational and fine-scale experimental approaches to explore
patterns and mechanisms related to nutrient use. In the observational gradient study
(Chapter 2), I described patterns of soil and foliar Nitrogen (N) and Phosphorus (P)
across nine forest sites in Costa Rica. Overall, the results suggest that across the
precipitation gradient, soil N is relatively more abundant than P but it is also more
sensitive to changes in MAP. In addition, foliar P is a better predictor of soil P than foliar
N of soil N.
In the experimental fertilization study (Chapters 3 and 4), I tested N or P
limitation in a lowland tropical wet forest. After two years of fertilization treatment, no
significant effect of either N or P fertilization was detected on tree diameter increase,
litterfall productivity, or root biomass or productivity. However, there were interesting
and contrasting responses among tree species and size classes. For example, the
legume species, Pentaclethra macroloba, the most abundant tree in this forest, did not
grow more with N or P additions. Instead, this species increased mean foliar N
concentration when N and P were added collectively, indicating luxury consumption of N
and resulting in increased litterfall-N concentrations. In contrast, individuals from
Socratea exohrriza, the most abundant palm species in this forest, showed larger mean
diameter increase and higher mean foliar P concentration after P additions, which
147
suggest that this species is P limited (results are summarized in Table 5-1). Several
conclusions can be drawn from these results:
The Myth of P Limitation in the Tropics
Multiple lines of evidence suggest that tropical forests occurring in old soils are
limited by P (e.g., Walker & Syers 1976, Vitousek 2004, Reich & Oleksyn 2004,
Kerkhoff et al. 2004). However, tropical forests occur on multiple soil types (Townsend
et al. 2008), some of which are not old. Thus, the generalization that “tropical forests are
limited by P” is misleading. Though N was relatively more available than P, both in the
nine forests included in the observational gradient study and at EARTH Forest (where
the fertilization experiment was conducted), there was no distinct indication of P
limitation. Phosphorus limitation, as operationally described (Chapin et al. 1986), would
have required an increase in net primary productivity after P additions, but this was not
the case in my experiment. Moreover, other indicators, such as foliar nutrient
concentrations and stoichiometry (Güsewell 2004, Reich & Oleskyn 2004, McGroddy et
al. 2004, Ågren 2008, Cleveland et al. 2011), in most cases were more influenced by
species identity or by other environmental variables, than by nutrient availability. Thus,
either these forests are not strongly limited by P, or they are P limited but the research
approaches used in this study could not detect P limitation. If the latter is true, we need
to devise better techniques for accurately identifying nutrient limitation in a specific
system, within a reasonable time frame.
Heterogeneous Nutrient Limitation
Second, and probably the most relevant conclusion from this study, is the
assertion that in lowland tropical diverse forests a plant community is not a black box of
trees reacting in tandem to nutrient availability. On the contrary, “community nutrient
148
limitation” (Vitousek et al. 2010, Harpole et al. 2011) is a heterogeneous process,
resulting from the competing and conflicting responses of different biological and
biochemical processes. In the case of this study, I would expand to say that at EARTH
Forest there is “heterogeneous nutrient limitation”, not only driven by variability in
nutrient responses among tree species but also among size classes. This heterogeneity
highlights the importance of considering different aspects of the plant community, such
as forest structure and species composition, when making predictions concerning
nutrient limitation in these forests. Furthermore, the differential response of size classes
and species suggests that changes in nutrient availability could lead to changes in
forest structure or even diversity in the long term, and could have important implications
for plant-soil-microbial feedbacks as they relate to nutrient limitation.
Environmental and Biological Processes Influence Nutrient Limitation and Carbon Cycling
Finally, this study has shown that complex feedbacks and interactions among
environmental and biological factors make it difficult to predict how long-term changes in
climate (e.g., MAP) or plant resources (e.g., nutrient availability) will influence nutrient
limitation in lowland tropical forests. However, results from the gradient study (Chapter
2) suggest that a decrease in MAP in certain forests of Costa Rica, which has been
predicted by several climate models (Neelin et al. 2006), could lead to an increase in
available N from mineralization in the wetter sites of the gradient. Although I did not
measure P mineralization, it could be expected that results would be similar for P
availability (Schuur 2001). Thus, in wet sites (such as EARTH Forest), a decrease in
MAP could result in an increase in available nutrients (i.e., N and P).
149
Results obtained in the fertilization experiment (Chapters 3 and 4) suggest that at
EARTH Forest, an increase in N and P could cause an increase in foliar and litter N in
Pentaclethra macroloba (the dominant species). Returning to the model used in Chapter
1 (Figure 1-1) these results could be incorporated into a positive feedback where an
increase in available nutrients feeds back positively through an increase in foliar and
litter N that could further increase N availability at EARTH forest. This feedback,
however, would not happen for P availability because I did not record an increase in
foliar P in Pentaclethra after N or P additions. However, I did record an increase in foliar
P in the most abundant palm, Socratea (but did not measure litter P for this species).
Assuming that an increase in foliar P in Socratea resulted in an increase in litter P, then
the relative rate (or strength) of the feedback would be dependent on specific life history
traits of the two contrasting species, such as leaf life span, litter quality, and litterfall
productivity. If these rates differ between species, this could lead to a temporary
decoupling of N and P cycling, caused by heterogeneous responses within the tree
community (Figure 5-1).
Lessons Learned
To date, full factorial fertilization experiments have been viewed as the best way to
directly test nutrient limitation in a land ecosystem (Chapin et al. 1986, Vitousek and
Howarth 1991, Tanner et al. 1992, Vitousek 2004, Elser et al. 2007, LeBauer and
Treseder 2008, Cleveland et al. 2011). However, probably due to the methodological
and logistical challenges associated with these types of experiments, only a handful of
fertilization experiments have been conducted in the tropics (Table 1-1). After facing
many challenges during the course of my doctorate research, I have several
150
suggestions for improving study design and expanding working hypotheses in
fertilization experiments in diverse tropical forests.
Though funding and labor may be prohibitive, I suggest including as many
experimental treatments as possible, depending on the hypotheses for a particular site.
Given recent data from Panama (Kaspari et al. 2008, Wright et al. 2011), at least
potassium (K), and a treatment with micronutrients, should be included in a fertilization
experiment. An initial complete commercial soil analysis would provide a baseline to
formulate hypotheses regarding which nutrient may limit productivity at a given site.
When establishing a fertilization experiment in a tropical forest, I suggest planning
for a long-term experiment, acknowledging that although several biological processes
(e.g., soil respiration) may respond in a short term, other processes (e.g., tree stem
diameter increase) may exhibit limited responses during the first years. During the initial
2-3 years of study, I suggest collecting background data on as many parameters as
possible, and then running power analyses on all variables to verify that samples sizes
are sufficient to capture the immense heterogeneity characteristic of these forests
(Townsend et al. 2008, Field 2009). If certain parameters, such as soil nitrate, are too
variable to obtain reliable estimates, then they should be eliminated or improvements
made to the methodology in order to reduce the variability. During the initial
experimental stage, I suggest conducting P fractionations and P sorption curves (Lajtha
et al. 1999) to evaluate how much P fertilizer has to be added to increase available P by
a significant proportion (e.g., 10%). At EARTH forest, I suspect that a larger proportion
of “total P”, than captured by Melich extractions, was available to the plant community
thereby making the amount of fertilizer added in this study a smaller fraction of the
151
“available pool”. Thus, although time consuming, I recommend extensive P
measurements prior to fertilization (e.g., Sato and Comerford 2006).
I suggest including measurements of belowground processes, such as soil
respiration and microbial community ecology, because these processes are likely to
respond faster to changes in nutrient availability than tree growth or litterfall production
(e.g., Cleveland and Townsend 2006). To capture a representative sample of the litter
productivity in plots, I suggest installing both permanent and portable litter traps in each
plot. The portable traps should be moved to different sectors of the plot monthly or bi-
monthly to capture litter from trees not included in the footprint of the permanent traps.
Although excruciatingly slow and difficult, it would be greatly beneficial to separate
litterfall by species, at least two months per year, or at least from focus species. Ideally,
these months would coincide with green-leaf collection for chemical analysis. During
these months, the portable litter traps can be installed under target trees where foliar
samples are being collected. This design would enable calculation of retranslocation
measurements (Chapin et al. 2002). In my experiment, green-leaf N and P values were
obtained from specific trees within the study plots. However, litter N and P values were
obtained from samples collected in permanent litter traps. Because variability in foliar
nutrients between trees and between collection times was so large, and the litter
samples were not from the same trees or times as the green-leaf samples, I did not feel
comfortable calculating retranslocation values from these samples.
Finally, one of the most interesting and challenging aspects of this study were the
differential responses of tree species and size classes to N and P additions, or
“heterogeneous limitation” (Chapter 3). It is widely recognized that the traditional
152
definition of nutrient limitation can break down when applied across wide differences in
phylogeny, or climate (Chapin et al. 1986, Vitousek and Howarth 1991, Vitousek et al.
2010). Yet, there have been limited recommendations on how to address the situation.
Originally, Chapin et al. (1986) stated that “community nutrient limitation” should be
tested by comparing relative growth rates (RGR) of same species across a nutrient
availability gradient. For example, one could measure the maximum possible growth of
a group of species in a high P system. Then, one could measure RGR of the same
group of species in lower P sites after adding P fertilizer. If RGRs reached the levels of
the high P system, those species (or community) were P limited in the lower P sites.
This approach was successfully used in Hawaii, where valuable research on this topic
has been conducted (summarized in Vitousek 2004). However, this approach is
impractical in the continental tropics because of incredibly high species diversity, and
the difficulty in identifying natural nutrient gradients. Thus, the question remains: what is
the best way to test nutrient limitation in a diverse tropical forest?
Based on my experience at EARTH Forest, there is no simple solution to advice of
the optimal design for studying nutrient limitation in a diverse tropical forest. However,
there are several options depending on the objective of the individual study. If the main
objective is to explore how carbon (C) storage and cycling is affected by nutrient
availability in a forest, then not all tree species are weighted equally. At least in the short
term, only the most “important” species (species with higher importance value index
(IVI) values; Chapter 3) will control net primary productivity (NPP). For example at
EARTH Forest, although there were more than 100 species of trees within my study
plots (Table C-1), large trees (>10cm DBH) from the four most-important species
153
(Pentaclethra, Socratea, Goethalsia, and Dypterix) comprised 40% of the total IVI of the
forest. Thus, to test how NPP is influenced by N and P availability in this forest, one
could focus on these trees instead of on the entire tree community. In contrast, if the
objective of the study was to explore how different life history strategies interact with
nutrient availability to control C cycling, then it is advisable to focus on species that are
representative of those groups such as N fixers versus non-N fixers, light demanding
versus shade tolerant, etc. In any case, my study highlights that regardless of the
objective of the study, to understand nutrient limitation in a diverse tropical forest, it is
crucial to acknowledge diversity and select methodologies and hypotheses that
incorporate heterogeneity of responses within the community.
154
Table 5-1. Summary of f responses ratios (RR) of treatments relative to the control in the fertilization experiment. Measurements were taken two years (or 2.7 yrs for RGR) after fertilization. Only RRs of statistically significant tests are shown. Refer to chapters 3 and 4 for measurement details and statistical tests. RR were calculated as ln (control/treatment).
Response Group
Variable Treatment
+N +P +NP
All trees RGR 0.34 Foliar N Foliar P Litterfall N 0.12 0.10 Litterfall P Root N Root P 0.22 0.15 Root N:P Large trees RGR Foliar N Foliar P Small Trees RGR 0.62 Foliar N Foliar P Pen Mac RGR Foliar N 0.09 Foliar P
Large RGR
PenMac Foliar N 0.09 Foliar P Small RGR
PenMac Foliar N Foliar P 0.12 Soc Exo RGR 1.04 0.99 Foliar N Foliar P 0.09 0.05
Large RGR Soc Exo Foliar N Foliar P 0.13 0.11 Small RGR 1.65 1.31
Soc Exo Foliar N Foliar P
155
Figure 5-1. Diagram representing how environmental and biological factors could interact to influence nutrient cycling in a diverse tropical forest. Environmental factors are shown in blue, biological factors in green, processes measured during this experiment in black and hypothesized processes in gray. MAP = mean annual precipitation, and RGR = relative growth rate. For a full description of the diagram refer to the text and Chapter 1. Model adapted from Vitousek 2004.
156
APPENDIX A
SPATIAL AND TEMPORAL VARIATION OF LIGHT REACHING THE UNDERSTORY
IN A WET TROPICAL FOREST IN COSTA RICA
Understanding how light and nutrient availability interact to affect primary
productivity is crucial to predict future effects of climate change on carbon storage
(Denslow et al. 1990, Wright et al. 2011). In this study, we explored the spatial and
temporal variability of transmitted diffuse light reaching the understory in 24, 20 x 20 m
plots that had been previously fertilized with nitrogen (N) phosphorous (P), N and P, or
left as controls in a lowland tropical wet forest in Costa Rica (see Chapter 3 for detailed
description of the experimental design). We utilized hemispherical photography to
analyze canopy structure and light reaching the understory, which can give insight into
below canopy photosynthetically active radiation (Frazer et al. 1997). Specifically, we
used transmitted diffuse light because calculations take the entire sky into account, and
this allows for the comparison of different sites (Madgwick and Brumfield 1969). The
specific objectives of this study were (1) to describe the spatial and temporal variation of
transmitted diffuse light to the understory in plots established in a wet tropical forest in
Costa Rica, and (2) to test the effect of increased nutrient availability by fertilization on
transmitted diffuse light to the understory, an indicator of canopy cover.
In February 2008, December 2008, and August 2009, we took four hemispherical
photographs in each of the 24 plots using a Nikon Coolpix 950 Limited Edition 2.11
megapixel digital camera with a fisheye lens set on a tripod, one meter above the forest
This study was conducted in collaboration with an undergraduate student, Laura Morales, and was funded through a SEAGEP- REU Minority Fellowship awarded to L. M
157
floor. All photographs were taken before 9 am or after 4 pm, to avoid distortions caused
by reflection of the sun on canopy leaves. When there was a large leaf from an
understory palm obstructing the canopy view, we removed the leaf before taking the
photograph. As an index of available light in the understory we used “transmitted diffuse
light” (which we refer to as light from here on). We used Gap Light Analyzer (GLA) free
software (http://www.ecostudies.org/gla/) to analyze our pictures. We conducted a
repeated measures MANOVA to test the effect of treatment, block and time on light. If
an interaction was significant we conducted a 1-way ANOVA to identify the time of the
effect. Statistical analyses were conducted in JMP 8.0(SAS Institute Inc., Cary, NC,
USA).
There was no significant difference in light levels among fertilizer treatments
through the course of the study. However, there was a significant increase in light with
time (Table A-1, Figure A-1). This was mainly due to the opening of several light gaps
after a particularly strong storm in August 2008. Figure A-2 demonstrates the difference
in canopy cover before and after the storm. It is not certain how this event could have
impacted the growth response of trees to fertilization. On one hand, trees exhibited
mechanical damage and a significant loss of leaves (personal observation), with an
evident peak in coarse litterfall after the event (Figure 3-12). This damage could have
repressed any fertilization effect because the trees would have probably used the extra
resources to repair damage and produce new leaves instead of increasing stem
diameter. On the other hand, the disturbance significantly increased light availability in
certain plots, which could have enhanced the response to fertilization. This is supported
158
by the significant contribution of light in explaining variation in relative growth rates of
trees larger than 10 cm DBH (Chapter 3).
Table A-1. Results from repeated measures MANOVAs for “transmitted diffuse light. F-values for treatment, block and time were obtained from exact tests but time*treatment and time*block interactions are F-value approximations resulting from Wilk’s lambda multivariate tests. In these cases degrees of freedom (df) are approximated as well. Significant effects are signaled with an asterisk. Plots 17 and 22 were excluded from the analysis (see text).
Parameter dfn,d F Probability
Treatment 3,13 2.68 0.09
Block 5,13 1.60 0.23
Time 2,12 18.33 <0.01
Time*treatment 6,24 0.52 0.78
Time*block 10,24 2.29 0.05
Years after fertilization
6 mo 1.5 yr 2 yr
Ln (
tra
nsm
itte
d d
iffu
se
lig
ht (m
ol m
-2 d
-1 )
)
0.0
0.2
0.4
0.6
0.8
1.0
Control
+N+P+NP
Figure A-1. Mean (+ SE) transmitted diffuse light for the four nutrient addition treatments
measured before, 1 yr and 2 yrs after initial fertilization.
159
Figure A-2. Hemispherical canopy photograph taken in the same position at three
successional dates. Intact forest is observed in (A), but during August 2008 there was a strong storm, which resulted in multiple gaps within the study plots, such as the one observed in (B) and (C).
A February 2008
2008
B December 2008
2008
C August 2009
160
APPENDIX B BIOTIC AND ABIOTIC FACTORS INFLUENCING TREE GROWTH IN A LOWLAND
TROPICAL WET FOREST: A MIXED MODEL APPROACH
The objective of this study was to explore how nutrient availability interacts with
other biotic and abiotic factors to control tree growth in a lowland tropical wet forest in
Costa Rica. A site description, experimental design, and methodology used to conduct
measurements are detailed in Chapter 3. Specifically, we tested three hypotheses:
(1) At the community level, nutrient availability limits tree “growth” in a lowland
tropical wet forest. Growth refers to relative growth rate (RGR) and is calculated as the
slope from a line traced from six log-transformed measurements of diameter at breast
height (DBH) of each tree bole, conducted over a 2.5 yr period (Chapter 3).
(2) There is an important effect of tree species on RGR response to
fertilization.
(3) There is a difference in the RGR response to fertilization depending on
tree size. We measured RGR in ten trees per plot with DBH between 5-10 cm DBH and
all trees with DBH larger than 10 cm (Chapter 3).
To test our hypotheses we fitted a series of mixed models and compared them
using AIC criteria. We used the lmer package and ML method in R (ver. 2.10.1; R
Corporation, Vienna, Austria) to calculate the AIC values and the REML method in lme
to calculate the parameter estimates for the best fitting model. The delta AIC (∆AIC)
provides a relative measurement of how good a model fits the data relative to other
The analysis included in this section was conducted in collaboration with Mollie Brooks and
included in an oral presentation at the Ecological Society of America (ESA) Meeting, Pittsburg, 2010.
161
models; the best model is assigned a value of zero and then the other models are
evaluated relative to that one, the lower the ∆AIC, the better the fit (Anderson 2008).
The specific parameters we used in our models are summarized in Table B1. We
treated both block and tree species as random effects and the rest of the parameters as
fixed effects.
To test our first hypothesis we compared model M5 (null hypothesis) against M2
and M3 (Table B-2). Delta AIC values revealed that there was a significant treatment
effect, but only in the Rio side (because there was a treatment * side interaction; Table
B-3; L-ratio test between M5 and M3 P = 0.01). The two sides represent two areas in
the forest where plots were established. Although they were separated by less than 1
km, the Rio side had more evidence of having been disturbed in the past. For example,
the Rio side was dominated by Goethalsia meinatha, a fast growing tree species
characteristic of disturbed areas (Hartshorn and Hammel 1994), which was not found in
the Rancho side. Therefore, we suspect that the stronger treatment effect in the Rio
side was caused by faster RGRs of species occurring in this area, rather than by
differences in nutrient availability among sides (Chapter 3).
To test our second hypothesis, which concerned the importance of tree species
identity to control RGRs, we compared model M8 (null hypothesis) with model M4.
There was a large decrease in ∆AIC when species was included in the model (Table B-
3; L-ratio test between M8 and M4 P <0.01). Thus, species identity contributes an
important portion of the variation in RGR at EARTH forest. However, because of the
high species diversity and low replicate trees from each species, it was difficult to detect
species-specific patterns (Figure 3-2). The canopy palm Socratea exorrhiza was one of
162
the few species where a clear fertilizer effect was observed (Figure B-2). This species
responded to fertilization by increasing RGR with P additions in the Rancho side and
with both P and NP additions in the Rio side. Differences among sides are likely due to
variation in other environmental controls, such as light or co-limitation with other soil
nutrients, potentially potassium (Wright et al. 2011).
To test our third hypothesis, which posed that initial tree size (represented by
DBH) was an important control on RGR, we compared models M4 (null hypothesis) and
M7. Initial DBH was, in fact, an important effect in determining RGR (Table B-3; L-ratio
test between M4 and M7 p <0.001). Overall, large trees grew less (relative to their size)
than small trees, regardless of fertilizer treatment. This pattern has frequently been
observed in other tropical forests (Losos and Leigh 2004, Wright et al. 2011).
From all the models tested, our best fit model was M1, which explained 18% of the
variation in RGR (Table B-3). This was a 15% improvement from our initial null model
(M8), which explained only 3% of the variation in RGR. In fact, the fit improved slightly
more if we include a three way interaction among treatment, initial DBH, and species in
the model. However, parameter estimates become non-significant due to a loss of
degrees of freedom caused by the three way interaction. Therefore we report as our
best fit model M1. From the parameter estimates (Table B-4) we conclude that RGR of
small trees was influenced by +NP additions in the Rio side. Although this analysis
showed no effect of light on tree RGR, this was probably due to methodological
constraints on light measurements (Appendix A). Integrated and accurate light
measurements are extremely difficult to obtain in tropical forests (Lambers et al. 1994).
163
Nonetheless, we suspect that these would probably contribute an important percentage
of the variability in RGR observed in this forest.
Results obtained with this analysis were consistent with those presented in
Chapter 3, where data were analyzed using repeated measures MANOVAs. One
difference is that here, we separated the “block effect” in two components, the random
block parameter and the “side” parameter. In the analyses presented in Chapter 3, all
the variation caused by this spatial heterogeneity was consolidated in the block effect.
Another difference is that because of the nature of the analysis, we could no incorporate
a block by treatment interaction in the MANOVA analyses. Here, the equivalent
treatment *side interaction proved to be important.
Table B-1. Parameters used in the mixed models. For measurement details refer to Chapter 3.
Parameter Description n
Growth Response variable. Refers to tree RGR ln( cm yr-1)
760
DBH 1 Initial DBH (cm) 760 Treatment fertilizer additions (C, N, P, NP) 4 Side side of the forest (Rio, Rancho) 2 (1 | block) Block. Treated as a random effect 6 (1| Sp.Code) Tree species. Treated as a random effect 130
Average Gap Qualitative measurement of % gap for each tree
760
Average trans. diffuse Index of available light (Appendix A) 24
164
Table B-2. Models used to test the hypotheses. For a description of each parameter refer to Table B-1 and Chapter 3.
Models/ Hypotheses
Growth ~ (1|block)
Growth ~ (1 | block) + (1 | Sp.Code)
Growth~ side+(1|block)+(1 | Sp.Code)
Growth ~ treatment+(1|block)+ (1 | Sp.Code)
Growth ~ treatment + side + (1 | block) + (1 | Sp.Code)
Growth ~ treatment * side + (1 | block) + (1 | Sp.Code)
Growth ~ DBH1+(1|block)+(1 | Sp.Code)
Growth ~ DBH1+treatment * side + (1 | block) + (1| Sp.Code)
Table B-3. Models used to test the hypotheses organized by increasing AIC values. AIC
values were obtained using the lmer package in R and the maximum likelihood estimates (ML). For comparison purposes the smallest AIC value is converted to zero and the rest of the values relative to it (∆AIC ) (Anderson 2008).
# Model/Hypothesis AIC
(lmer, ML) ∆
AIC
M1 Growth ~ DBH1+treatment * side + (1 | block) + (1| Sp.Code) -4539 0
M7 Growth ~ DBH1+(1|block)+(1 | Sp.Code) -4529 10
M2 Growth ~ treatment * side + (1 | block) + (1 | Sp.Code) -4515 24
M3 Growth ~ treatment + side + (1 | block) + (1 | Sp.Code) -4510 29
M5 Growth~ side+(1|block)+(1 | Sp.Code) -4510 29
M4 Growth ~ (1 | block) + (1 | Sp.Code) -4504 35
M6 Growth ~ treatment+(1|block)+ (1 | Sp.Code) -4504 35
M9 Growth ~ Average.Gap.2 + (1|block) + (1 | Sp.Code) -4504 35
M8 Growth ~ (1|block) -4469 70
M10 Growth ~ Average.trans.diffuse + (1 | block) + (1 | Sp.Code) -4451 88
165
Table B-4. Maximum likelihood estimates, their standard error, and T-value for parameters included in model M1 (Table B3), the best fitting model from those included in the analysis. A parameter estimate is considered significant if the confidence interval (estimate + std. error) does not include zero. Significant parameters are labelled with an asterisk.
Estimate Std. Error T
(Intercept) 1.08 e-02 1.77 e-03 6.10*
DBH1 -1.95 e-04 3.80 e-05 -5.13*
Treatment N: side Rancho 2.14 e-03 2.06 e-03 1.04
Treatment NP: side Rancho -9.43 e-04 2.16 e-03 -0.44
Treatment P: side rancho 1.46 e-03 2.12 e-03 0.69
Side Rio 1.37 e-03 2.14 e-03 0.64
Treatment N:side Rio -9.78 e-04 2.80 e-03 -0.35
Treatment NP: side Rio 7.92 e-03 2.93 e-03 2.71*
Treatment P: side Rio 2.47e-03 2.835e-03 0.873
Figure B-1. Box plots showing stem diameter increase (calculated as relative growth
rate, see methods) for the four fertilizer treatments, separated by “side”. There were three blocks on each of these two areas of the EARTH forest, which are separated by less than 1 km.
166
Figure B-2. Box plots showing stem diameter increase (calculated as relative growth
rate, see methods) for the most common canopy palm, Socratea exorrhiza in the four fertilizer treatments, separated by “side”.
Figure B-3. Relationship between initial diameter at breast height (DBH) and stem
diameter increase (calculated as relative growth rate, see methods).
167
APPENDIX C ADDITIONAL TABLES AND FIGURES
Table C-1. Species of trees found in the study plots. Trees were identified by Enrique
Rojas and Orlando Vargas.
Family Genus Species
Anacardiaceae Tapirira guianensis Annonaceae Anona subnubila Annonaceae Rollinia pittieri Annonaceae Unonopsis pittieri Annonaceae Xylopia sericophylla Anonaceae Guatteria sp1 Anonaceae Guatteria sp2 Anonaceae Guatteria sp3 Apocynaceae Tabernaemontana arborae c.f Aquifoliaceae Ilex skutchii Araliaceae Dendropanax arboreus Arecaceae Astrocarium alatum Arecaceae Chrysophylla warscewiczia Arecaceae Cryosophila warscewiczii Arecaceae Socratea exorrhiza Bignoniaceae Jacaranda Copaia Boraginaceae Cordia bicolor Boraginaceae Cordia lucidula Boraginaceae Cordia porcata Burseraceae Protium confusum Burseraceae Protium panamense Burseraceae Protium pittieri Burseraceae Protium ravenii Burseraceae Tetragastris panamensis Burseraceae Trattinnickia aspera Capparaceae Capparis pittieri Carycaceae Jacaratia Dolichaula Cecropicaceae Cecropia insignis Cecropicaceae Pourouma bicolor Chrysobalanaceae Licania manicarpa Clusiaceae Garcinia intermedia Dilleniaceae Doliocarpus dentatus Dilleniaceae Pinzona coriaceae Elaeocarpaceae Sloanea guianensis Euphorbiaceae Croton schiedeanus Euphorbiaceae Hyeronima alchorneoides
Table C-1. Continued
168
Family Genus Species
Fabaceae/Cae Bauhinia guinanensis Fabaceae/Mim Balizia elegans Fabaceae/Mim Inga alba Fabaceae/Mim Inga leiocalycina c.f Fabaceae/Mim Inga pezizifera Fabaceae/Mim Inga Thiboudiana Fabaceae/Mim Inga umbilifera Fabaceae/Mim Inga venusta Fabaceae/Mim Pentaclethra macroloba Fabaceae/Mim Stryphodendron microstachyum Fabaceae/Pap Dipteryx panamensis Fabaceae/Pap Dussia macroprophyllata Flacourtiaceae Cassearia arborea Flacourtiaceae Casearia commersoniana Flacourtiaceae Cassearia tacanensis Flacourtiaceae Laetia procera Flacourtiaceae Ryania especiosa Lauraceae Beilschmiedia sp Lauraceae Nectandra reticulata Lauraceae Ocotea laetevirens Lauraceae Ocotea leucoxylon Lecythidaceae Eschweilera costarricensis Lecythidaceae Eschweilera longirachis Malpighiaceae Byrsonima arthropoda Malpighiaceae Spachea correae Malvaceae Hampea appendiculata Melastomataceae Miconia affinis Melastomataceae Miconia elata Meliaceae Carapa guianensis Meliaceae Guarea guidonia Meliaceae Guarea rhopalocarpa Meliaceae Trichilia septentrionalis Moraceae Castilla elastica Moraceae Ficus maxima Moraceae Nauclopsis naga Myrcinaceae Ardisia fimbrillifera Myristicaceae Compsoneura Mexicana Myristicaceae Otoba novogranatensis Myristicaceae Virola koschnyi Myristicaceae Virola multiflora Myristicaceae Virola sebifera Myrsinaceae Parathesis trichogyne
169
Table C-1. Continued.
Family Genus Species
Myrtaceae Myrcia splendens Nyctaginaceae Neea laetevirens Piperaceae Piper colonense Quiinaceae Lacunaria panamensis Rubiaceae Coussarea hondensis Rubiaceae Faramea parvibractea Rubiaceae Miconia multispicata Rubiaceae Posoqueria latifolia Rubiaceae Warsewicsia coccinea Rutaceae Zanthoxylum panamensis Sapinadaceae Cupania sp Sapinadaceae Talisia nervosa Sapotaceae Chrysophyllum venezuelanense Sapotaceae Micropholis crotonoides Sapotaceae Pouteria calistophylla Simaroubaceae Simarouba amara Sterculiaceae Sterculia costaricana Tiliaceae Apeiba membranacea Tiliaceae Goethalsia meiantha Tiliaceae Luehea seemannii Ulmaceae Ampelocera macrocarpa Violaceae Rhinorea hummeli
170
Table C-2. Results from repeated measures MANOVAs for foliar N:P ratios by tree size class. F-values for treatment, block and time were obtained from exact tests but time*treatment and time*block interactions are F-value approximations resulting from Wilk’s lambda multivariate tests. In these cases degrees of freedom (df) are approximated as well. Significant effects are signaled with an asterisk.
dfn,d F Probability
5-10 cm Treatment 3,15 0.45 0.72 Block 5,15 0.93 0.49 Time 2,14 0.52 0.61 Time*treatment 6,28 0.59 0.74 Time*block 10,28 0.48 0.89
> 10 cm Treatment 3,15 0.53 0.67 Block 5,15 0.90 0.51 Time 2,14 1.79 0.20 Time*treatment 6,28 0.23 0.96 Time*block 10,28 0.40 0.94
Total Treatment 3,15 0.47 0.71 Block 5,15 0.57 0.72 Time 2,14 0.27 0.77 Time*treatment 6,28 0.87 0.91 Time*block 10,28 0.88 0.99
171
Table C-3. Results from repeated measures MANOVAs for foliar N:P ratios by species. F-values for treatment and time were obtained from exact tests but time*treatment interactions are F-value approximations resulting from Wilk’s lambda multivariate tests. In these cases degrees of freedom (df) are approximated as well. Significant effects are signaled with an asterisk.
dfn,d F Probability
Dendropanax arboreus
Treatment 3,6 5.17 0.04* Time 2,5 0.06 0.94 Time*treatment 6,10 0.40 0.57
Goethalsia meiantha Treatment 2,3 0.10 0.90 Time 2,2 14.77 0.06 Time*treatment 4,4 6.85 0.04*
Inga Treatment 3,9 0.60 0.63 Time 2,8 1.45 0.29 Time*treatment 6,16 0.79 0.59
Pentaclethra macroloba Treatment 3,19 1.20 0.34 Time 2,18 1.71 0.21 Time*treatment 6,36 1.40 0.25
Protium Treatment 3,19 2.01 0.14 Time 2,18 0.71 0.50 Time*treatment 6,36 1.24 0.31
Socratea exohrriza Treatment 3,25 0.12 0.95 Time 2,24 0.29 0.76 Time*treatment 6,48 0.70 0.65
172
Soil pH
3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8
Folia
r N
:P
18
20
22
24
26
28
30
2
7
8 141924
1
10
12
13172235
11
18
20
23
46
9 15
16
21
Soil pH
3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8
27 814
1924
1
10
12
131722
3 5
11
18
20
23
46
15
1621
Net Nitrification (ug N g-1 d
-1)
-2 -1 0 1 2 3 4 5
Folia
r N
:P
18
20
22
24
26
28
30
2
7
8141924
1
10
12
13 17 22 35
11
18
20
23
6
9 15
16
21
-2 -1 0 1 2 3 4 5
2781419
241
10
12
131722
35
11
18
20
23
46 9
15
1621
Net Nitrification (ug N g-1 d
-1)
Total P (ug g-1)
500 1000 1500 2000 2500 3000 3500
Folia
r N
:P
18
20
22
24
26
28
30
2
7
8141924
1
10
12
13 172235
11
18
20
23
46
9 15
16
21
500 1000 1500 2000 2500 3000 3500
27 814
1924
1
10
12
131722
35
11
18
20
23
469 16
21
Total P (ug g-1)
(a) R2 = 0.04, P = 0.36
(b) R2 = 0.04 P = 0.36
(c) R2 = 0.01 P = 0.63 (d) R
2 = 0.04 P = 0.38
(e) R2 = 0.01 P = 0.56 (f) R
2 = 0.14 P = 0.77
Figure C-1. Relationship between several soil variables and plot-averaged foliar N:P ratios before fertilization (a, c, e) and 2 yrs after initial fertilization (b, d, f). Symbols represent each plot and colors each treatment (Control= grey, +N = pink, +P= green, +NP = Cyan).
173
Me
lich
P (
ug
g-1
)
0
1
2
3
4
5
6
7
2
7
8
14
19
24
1 10
1213
17
22
35
11
18
20
23
46
9
15
16
21
Total P (ug g-1)
500 1000 1500 2000 2500 3000 3500
Me
lich
P (
ug
g-1
)
-2
0
2
4
6
8
10
12
14
27
814
19
24
110
12131722
35
11
18
20
23
4 6
9
16
21
A R2 = 0.50, P < 0.001
B R2 = 0.73, P < 0.001
Figure C-2. Relationship between soil Total P and Melich P before fertilization (A) and 2
yrs after initial fertilization (B). Symbols represent each plot and colors each treatment (Control= grey, +N = pink, +P= green, +NP = Cyan).
174
LIST OF REFERENCES
Aerts, R. 1997. Climate, leaf litter chemistry, and leaf litter decomposition in terrestrial ecosystems: a triangular relationship. Oikos 79:439–49.
Aerts, R., and F. S. Chapin. 2000. The mineral nutrition of wild plants revisited: a re-evaluation of processes and patterns. Advances in Ecological Research 30:1–67.
Ågren, G. I. 2004. The C:N:P stoichiometry of autotrophs – theory and observations. Ecology Letters 7:185–191.
Ågren, G. I. 2008. Stoichiometry and nutrition of plant growth in natural communities. Annual Review of Ecology Evolution and Systematics 39:153–170.
Alvarez-Clare, S., and M. C. Mack. 2011. Influence of precipitation on soil and foliar nutrients across nine Costa Rican forests. Biotropica 43:433–441.
Amundson, R., A. T. Austin, E. A. G. Schuur, K. Yoo, V. Matzek, C. Kendall, A. Uebersax, D. Brenner, and W. T. Baisden. 2003. Global patterns of the isotopic composition of soil and plant Nitrogen. Global Biogeochemical Cycles 17:1031–1041.
Andersen, K. M., M. D. Corre, B. L. Turner and J. W. Dalling. 2011. Plant–soil associations in a lower montane tropical forest: physiological acclimation and herbivore mediated responses to Nitrogen addition. Functional Ecology 24:1171–1180.
Anderson, D. R. 2008. Model Based Inference in the Life Sciences: A Primer on Evidence. Springer, New York, USA.
Arnold, J., M. D. Corre, and E. Veldkamp. 2008. Cold storage and laboratory incubation of intact soil cores do not reflect in-situ Nitrogen cycling rates of tropical forest soils. Soil Biology and Biochemistry 40:2480–2483.
Austin, A. T. and P.M. Vitousek. 1998. Nutrient dynamics on a precipitation gradient in Hawai’i. Oecologia 113:519–529.
Baraloto, C., D. Bonal and D. E. Goldberg. 2006. Differential seedling growth response to soil resource availability among nine Neotropical tree species. Journal of Tropical Ecology 22: 487–497.
Bazzaz, F. A. 1998. Tropical forests in a future climate: Changes in biological diversity and impact on the global carbon cycle. Climatic Change 39:317–336.
Bouyoucos, C. Y. 1950. Recalibration of hydrometer method for making mechanical analysis of soil. Agronomy Journal 43:434¬–439.
175
Bowen, G. D. and E. K. S. Nambiar. 1984. Nutrition of Plantation Forests. Academic Press, London, UK.
Bunker, D. E., and W. P. Carson. 2005. Drought stress and tropical forest woody seedlings: effect on community structure and composition. Journal of Ecology 93:794–806.
Burslem D. F. R. P., P. J. Grubb, and I. M. Turner. 1995. Responses to nutrient addition among shade-tolerant tree seedlings of lowland tropical rain forest in Singapore. Journal of Ecology 83: 113–122.
Campo, J., and R. Dirzo. 2003. Leaf quality and herbivory responses to soil nutrient addition in secondary tropical dry forest of Yucatan, Mexico. Journal of Tropical Ecology 19: 525–530.
Ceccon, E., S. Sanchez, and J. Campo. 2004. Tree seedling dynamics in two abandoned tropical dry forests of differing successional status in Yucatan, Mexico: a field experiment with N and P fertilization. Plant Ecology 170:277–285.
Cernusak, L. A., K. Winter and B. L.Turner. 2010. Leaf Nitrogen to phosphorus ratios of tropical trees: experimental assessment of physiological and environmental controls. New Phytologist 185:770–779.
Chandler, G. 1985. Mineralization and nitrification in three Malaysian forest soils. Soil Biology and Biochemistry 17:347–353.
Chapin, F. S. 1980. The mineral-nutrition of wild plants. Annual Review of Ecology and Systematics 11:233–260.
Chapin, F. S. III, A. J. Bloom, C. B. Field, and R. H. Waring. 1987. Plant response to multiple environmental factors. Bioscience 37:49–57.
Chapin, F. S. III, P. A. Matson, H. A. Mooney. 2002. Principles of Terrestrial Ecosystem Ecology. Springer, New York, USA.
Chapin III, F. S., P. M. Vitousek, and K. Van Cleve. 1986. The nature of nutrient limitation in plant communities. The American Naturalist 127: 48–58.
Clark, D. A. 1994. Plant Demography. Pages 90-105 in L. A. McDade, K. S. Bawa, H. A. Hespenheide, and G. S. Hartshorn, editors. La Selva: Ecology and natural history of a neotropical rain forest. The University of Chicago Press, Chicago, Illinois, USA.
Clark, D. A., S. Brown, D. W. Kicklighter, J. Q. Chambers, J. R. Thomlinson, J. Ni, and E. A. Holland. 2001a. Net primary production in tropical forests: An evaluation and synthesis of existing field data. Ecological Applications 11:371–384.
176
Clark, D. A., S. Brown, D. W. Kicklighter, J. Q. Chambers, J. R. Thomlinson, and J. Ni. 2001b. Measuring net primary production in forests: Concepts and field methods. Ecological Applications 11: 356–370.
Clark, D. A. and D. B. Clark. 1992. Life history diversity of canopy and emergent trees in a neotropical rain forest. Ecological Monographs 62: 315–344.
Clark, D. A. and D. B. Clark. 2011. Assessing tropical forests’ climatic sensitivities with long-term data. Biotropica 43: 31–40.
Clark, D. B., M. W. Palmer, and D. A. Clark. 1999. Edaphic factors and the landscape-scale distributions of tropical rain forest trees. Ecology 80:2662–2675.
Clark, D. A., S. C. Piper, C. D. Keeling, and D. B. Clark. 2003. Tropical rain forest tree growth and atmospheric carbon dynamics linked to interannual temperature variation during 1984–2000. Proceedings of the National Academy of Sciences of the United States of America 100: 5853–5857.
Cleveland, C. C., and A. R. Townsend. 2006. Nutrient additions to a tropical rain forest drive substantial soil carbon dioxide losses to the atmosphere. Proceedings of the National Academy of Sciences of the United States of America 103:10316–10321.
Cleveland, C. C., A. R. Townsend, P. Taylor, S. Alvarez-Clare, M. M. C. Bustamante, G. Chuyong, P. Grierson, K. Harms, B. Z. Houlton, A. Marklein, W. Parton, S. Porder, S. C. Reed, C. A. Sierra, W. L. Silver, E. J. R. Tanner, W. R. Wieder. 2011. Relationships among net primary productivity, nutrients and climate in tropical rain forest: a pan-tropical analysis. Ecology Letters 14:939–947.
Cornwell, W. K., W. K. Cornwell, J. H. C. Cornelissen, K. Amatangelo, E. Dorrepaal, V. T. Eviner, O. Godoy, S. E. Hobbie, B. Hoorens, H. Kurokawa, N. Pérez-Harguindeguy, H. M. Quested, L. S. Santiago, D. A. Wardle, I. J. Wright, R. Aerts, S. D. Allison, P. van Bodegom, V. Brovkin, A. Chatain, T. V. Callaghan, S. Díaz, E. Garnier, D. E. Gurvich, E. Kazakou, J. A. Klein, J. Read, P. B. Reich, N. A. Soudzilovskaia, M. V. Vaieretti, and M. Westoby. 2008. Plant species traits are the predominant control on litter decomposition rates within biomes worldwide. Ecology Letters 11:1065–1071.
Crews, T. E. 1999. The presence of nitrogen fixing legumes in terrestrial communities: Evolutionary vs. ecological considerations. Biogeochemistry 46: 233–246.
Crews,T. E., K. Kitayama, J. H. Fownes, R. H. Riley, D. A.Herbert, D. Mueller-Dombois, P. M. Vitousek. 1995. Changes in soil phosphorus fractions and ecosystem dynamics across a long chronosequence in Hawaii. Ecology 76:1407–1424.
Cross A. F. and W. H. Schlesinger. 1995. A literature review and evaluation of the Hedley fractionation scheme: applications to the biogeochemical cycle of soil phosphorus in natural ecosystems. Geoderma 64:197–214.
177
Cuevas, E. and E. Medina. 1988. Nutrient dynamics within Amazonian forests: fine root-growth, nutrient availability and leaf litter decomposition. Oecologia 76: 222–235.
Cusack, D. F., W. W. Chou, W. H. Yang, M. E. Harmon, W. L. Silver, and The LIDET Team. 2009. Controls on long-term root and leaf litter decomposition in Neotropical forests. Global Change Biology 15:1339–1355.
Davidson, E. A., C. J. R. de Carvalho, I. C. G. Vieira, R. D. Figueiredo, P. Moutinho, F. Davidson, E. A., C. J. R. de Carvalho, I. C. G. Vieira, R. D. Figueiredo, P. Moutinho, F. Y. Ishida, M. T. P. dos Santos, J. B. Guerrero, K. Kalif, and R. T. Saba. 2004. Nitrogen and Phosphorus limitation of biomass growth in a tropical secondary forest. Ecological Applications 14:S150–S163.
Denslow, J. S., and G. S. Hartshorn. 1994. Tree-fall gap environments and forest dynamic processes. Pages 120–127 in L. A. McDade, K. S. Bawa, H. A. Hespenheide, and G. S. Hartshorn, editors. La Selva: Ecology and natural history of a neotropical rain forest. The University of Chicago Press, Chicago, Illinois, USA.
Denslow, J. S., J. C. Schultz, P. M. Vitousek, and B. R. Strain. 1990. Growth responses of tropical shrubs to tree fall gap environments. Ecology 71: 165–179.
Denslow, J. S., and G. S. Hartshorn. 1994. Tree-fall gap environments and forest dynamic processes. Pages 120–127 in L. A. McDade, K. S. Bawa, H. A. Hespenheide, and G. S. Hartshorn, editors. La Selva: Ecology and natural history of a neotropical rain forest. The University of Chicago Press, Chicago, Illinois, USA.
Denslow, J. S., J. C. Schultz, P. M. Vitousek, and B. R. Strain. 1990. Growth responses of tropical shrubs to tree fall gap environments. Ecology 71:165–179.
Dixon, R. K., S. Brown, R. A. Houghton, A. M. Solomon, M. C. Trexler, and J. Wisniewski. 1994. Carbon pools and flux of global forest ecosystems. Science 263:185–190.
Dominy, N. J., P. W. Lucas, and S. J. Wright. 2003. Mechanics and chemistry of rain forest leaves: canopy and understorey compared. Journal of Experimental Botany 54:2007–2014.
Douville, H., D. Salas-Melia, and S. Tyteca. 2006. On the tropical origin of uncertainties in the global land precipitation response to global warming. Climate Dynamics 26:367–385.
Drechsel, P., and W. Zech. 1991. Foliar nutrient levels of broad-leaved tropical trees- a tabular review. Plant and Soil 131:29–46.
178
Eklund, T. J., W. H. Mcdowell, and C. M. Pringle. 1997. Seasonal variation of tropical precipitation chemistry: La Selva, Costa Rica. Atmospheric Environment 31:3903–3910.
Elser, J. J., M. E. S. Bracken, E. E. Cleland, D. S. Gruner, W. S. Harpole, H. Hillebrand, J. T. Ngai, E. W. Seabloom, J. B. Shurin, and J. E. Smith. 2007. Global analysis of Nitrogen and Phosphorus limitation of primary producers in freshwater, marine and terrestrial ecosystems. Ecology Letters 10:1135–1142.
Elser, J. J., W. F. Fagan, A. J. Kerkhoff, N. G. Swenson, and B. J. Enquist. 2010. Tansley review: Biological stoichiometry of plant production: metabolism, scaling and ecological response to global change. New Phytologist 186:593–608.
Engelbrecht, B. M. J., L. S. Comita, R. Condit, T. A. Kursar, M. T. Tyree, B. L. Turner, and S. P. Hubbell. 2007. Drought sensitivity shapes species distribution patterns in tropical forests. Nature 447:80–82.
Enquist, C. A. F. 2002. Predicted regional impacts of climate change on the geographical distribution and diversity of tropical forests in Costa Rica. Journal of Biogeography 29:519–534.
Espeleta, J. F., and D. A. Clark. 2007. Multi-scale variation in fine-root biomass in a tropical rain forest: a seven-year study. Ecological Monographs 77:377–404.
Fetcher, N., S. F. Oberbauer, and R. L. Chazdon. 1994. Physiological Ecology of Plants. Pages 198-141 in L. A. McDade, K. S. Bawa, H. A. Hespenheide, and G. S. Hartshorn, editors. La Selva: Ecology and natural history of a neotropical rain forest. The University of Chicago Press, Chicago, Illinois, USA.
Field, A. 2009. Discovering statistics using SPSS. SAGE Publications Ltd. California, USA.
Field C., and H. A. Mooney. 1986. The photosynthesis-nitrogen relationship in wild plants. Pages 22–55 in T. J. Givnish, editor. On the economy of plant form and function. Cambridge University Press, NewYork, USA.
Frazer, G. W., J. A. Trofymow, and K. P. Lertzman. 1997. A method for estimating canopy openness, effective leaf area index, and photosynthetically active photon flux density using hemispherical photography and computerized image analysis techniques. Report BC-X-373. Forest Ecosystem Processes Network, Pacific Forestry Center. Canadian Forestry Service, Canada.
Galloway, J. N., and E. B. Cowling. 2002. Reactive nitrogen and the world: 200 years of change. Abstracts of Papers of the American Chemical Society 31: 64–71.
179
Galloway, J. N., F. J. Dentener, D. G. Capone, E. W. Boyer, R. W. Howarth, S. P. Seitzinger, G. P. Asner, C. C. Cleveland, P. A. Green, E. A. Holland, D. M. Karl, A. F. Michaels, J. H. Porter, A. R. Townsend, C. J. Vorosmarty. 2004. Nitrogen cycles: past, present, and future. Biogeochemistry 70: 153–226.
Garten, C. T. 1993. Variation in foliar N-15 abundance and the availability of soil-nitrogen on walker branch watershed. Ecology 74:2098–2113.
Gleeson, S. K., and D. Tilman 1992. Plant allocation and the multiple limitation hypothesis. American Naturalist 139:1322–1343.
Grace, J., Y. Malhi, N. Higuchi, P. Meir. 2001. Productivity of tropical rainforests. Pages 401-426 in Mooney, H., J. Roy, and B. Saugier, editors. Terrestrial global productivity: past, present and future. Academic Press, San Diego, California, USA.
Grace J, San Jose J, Meir P, Miranda HS, Montes RA (2006) Productivity and carbon fluxes of tropical savannas. Journal of Biogeography, 33, 387–400.
Güsewell, S. 2004. N:P ratios in terrestrial plants: variation and functional significance. New Phytologist 164:243¬–266.
Haggar J. P., and J. J. Ewel 1997. Primary productivity and resource portioning in model tropical ecosystems. Ecology 78:1211–1221.
Handley, L. L., and J. A. Raven. 1992. The use of natural abundance of nitrogen isotopes in plant physiology and ecology. Plant and Cell Environment 15:965–985.
Harpole, W. S., J. T. Ngai, E. E. Cleland, E. W. Seabloom, E. T. Borer, M. E. S. Bracken, J. J. Elser, D. S. Gruner, H. Hillebrand, J. B. Shurin, and J. E. Smith. 2011. Nutrient co-limitation of primary producer communities. Ecology Letters 14:852–862.
Harrington, R. A., J. H. Fownes, and P. M. Vitousek. 2001. Production and resource use efficiencies in N- and P-limited tropical forests: A comparison of responses to long-term fertilization. Ecosystems 4:646–657.
Hartshorn, G. S. 1983. Plants. Pages 118–350 in D. H. Jansen, editor. Costa Rican natural history. The University of Chicago Press, Chicago, USA.
Hartshorn, G. S., and B. E. Hammel. 1994. Vegetation types and floristic patterns. Pages 73–89 in L. A. McDade, K. S. Bawa, H. A. Hespenheide, and G. S. Hartshorn, editors. La Selva: Ecology and Natural History of a Neotropical Rain Forest. The University of Chicago Press, Chicago, Illinois, USA.
Hattenschwiler, S., B. Aeschlimann, M. M. Couteaux, J. Roy, and D. Bonal. 2008. High variation in foliage and leaf litter chemistry among 45 tree species of a Neotropical rainforest community. New Phytologist 179:165–175.
180
Hedin, L. O. 2004. Global Organization of Terrestrial Plant-Nutrient Interactions. Proceedings of the National Academy of Sciences of the United States of America 30:10849–10850.
Hedin, L. O., P. M. Vitousek, and P. A. Matson. 2003. Nutrient losses over four million years of tropical forest development. Ecology 84:2231–2255.
Hedley, M. J., J. W. Stewart, and B. S. Chauhan.1982. Changes in inorganic and organic soil phosphorus fractions induced by cultivation practices and by laboratory incubations. Soil Science Society of America Journal 46:970–976.
Henderson, A., G. Galeano, and R. Bernal. 1995. Field Guide to the Palms of the Americas. Princeton University Press, New Jersey.
Herbert, D. A., and J. H. Fownes. 1995. Phosphorus limitation of forest leaf-area and net primary production on a highly weathered soil. Biogeochemistry 29:223–235.
Hiremath, A. J., and J. J. Ewel. 2001. Ecosystem nutrient use efficiency, productivity, and nutrient accrual in model tropical communities. Ecosystems 4:669–682.
Hobbie S. E. 1992. Effects of plant species on nutrient cycling. Trends in Ecology and Evolution 7:336–339.
Hobbie, S. E., and P. M. Vitousek. 2000. Nutrient limitation of decomposition in Hawaiian forests. Ecology 81:1867–77.
Hogberg, P. 1997. Tansley review no 95 - N-15 natural abundance in soil-plant systems. New Phytologist 137:179–203.
Holdridge, L. R., W. C. Grenke, W. H. Hatheway, T. Liang, and J. A. Tosi. 1971. Forest environments in tropical life zones: A pilot study. Pergamon Press. Oxford, UK.
Holste, E. K., R. K. Kobe, and C. V. Vriesendorp. 2011. Seedling growth responses to soil resources in the understory of a wet tropical forest. Ecology 92: 1828–1838.
Hooper, D. U., and P. M. Vitousek. 1998. Effects of plant composition and diversity on nutrient cycling. Ecological Monographs 68:121–149.
Houlton, B. Z., Y. Wang, P. M. Vitousek and C. B. Field. 2008. A unifying framework for dinitrogen fixation in the terrestrial biosphere. Nature 454: 327–331.
Huberty, A. F. and R. F. Denno. 2006. Consequences of nitrogen and phosphorus limitation for the performance of two plant hoppers with divergent life-history strategies. Oecologia 149: 444–455.
Idol, T., P. J. Baker, and D. Meason. 2007. Indicators of forest ecosystem productivity and nutrient status across precipitation and temperature gradients in Hawaii. Journal of Tropical Ecology 23:693–704.
181
IPCC (Intergovernmental Panel for Climate Change). 2007. IPCC climate change 2007 working group I to the fourth assessment report of the IPCC 2007: the physical science basis. Cambridge University Press, New York, USA.
Jenny, H. 1941. Factors of soil formation. McGraw-Hill, New York, USA.
John, R., J. W. Dalling, K. E Harms, J. B. Yavitt, R. F. Stallard, M. Mirabello, S. P. Hubbell, R. Valencia, H. Navarrete, M. Vallejo, R. B. Foster. 2007. Soil nutrients influence spatial distributions of tropical tree species. Proceedings of the National Academy of Sciences of the United States of America 104:864–869.
Jones, J. B., and B. W. Case. 1996. Soil testing and plant analysis no. 3. Pages 389–415 in D. L. Sparks, editor. Methods of soil analysis. Part 3: Chemical methods. Soil Science Society of America, Madison, Wisconsin.
Kaspari, M., M. N. García, K. E. Harms, M. Santana, S. J. Wright, and J. B. Yavitt. 2008. Multiple nutrients limit litterfall and decomposition in a tropical forest. Ecology Letters 11:35–43
Keeland, B. D., and P. J. Young. 2004. Construction and installation of dendrometer bands for periodic tree-growth measurements. U.S. Geological Survey, National Wetlands Research Center. http://www.nwrc.usgs.gov/Dendrometer/index.htm
Kerkhoff, A. J., B. J. Enquist, J. J. Elser, and W. F. Fagan. 2005. Pleant allometry, stoichiometry and the temperature-dependence of primary productivity. Global Ecology and Biogeography 14:585–598.
Kerkhoff, A. J., W. F. Fagan, J. J. Elser, and B. J. Enquist 2006. Phylogenetic and functional variation in the scaling of nitrogen and phosphorus in the seed plants. American Naturalist 168: E103–E122.
Killingbeck, K. T. 1996. Nutrients in senesced leaves: keys to the search for potential resorption and resorption proficiency. Ecology 77:1716–27.
Kleber, M., L. Schwendenmann, E. Veldkamp, J. Rossner, and R. Jahn. 2007. Halloysite versus gibbsite: Silicon cycling as a pedogenetic process in two lowland Neotropical rain forest soils of La Selva, Costa Rica. Geoderma 138:1–11.
Kobe, R. K, C. A. Lepczyk, and M. Iyer 2005. Resorption efficiency decreases with increasing green leaf nutrients in a global data set. Ecology 86:2780–2792.
Koerselmann, W., and A. F. M. Meuleman. 1996. The vegetation N:P ratio: a new tool to detect the nature of nutrient limitation. Journal of Applied Ecology 33:1441–1450.
Kuo, S. 1996. Phosphorus. Pages 869–919 in D. L. Sparks, editor. Methods of soil analysis. Part 3: Chemical methods. Soil Science Society of America, Madison, Winsconsin, USA.
182
Lambers, H., F. S. Chapin III, and T. L. Pons. 1998. Plant physiological ecology. Springer-Verlag, New York, USA.
Lajtha, K., C. T. Driscoll, W. M. Jarrell, and E. T. Elliot. 1999. Soil Phosphorus. Characterization and Total Element Analysis. Pages 115–142 in G. P. Robertson, D. C. Coleman, C. S. Bledsoe, and P. Sollins, editors. Standard soil methods for long-term ecological research. Oxford University Press, New York, USA.
Lawrence, D. 2003. The response of tropical tree seedlings to nutrient supply: meta-analysis for understanding a changing tropical landscape. Journal of Tropical Ecology 19:239–250.
LeBauer, D. S., and K. K. Treseder. 2008. Nitrogen limitation of net primary productivity in terrestrial ecosystems is globally distributed. Ecology 89:371–379.
Lieberman, M., and D. Lieberman. 1994. Patterns of density and dispersion of forest trees. Pages 106-119 in L. A. McDade, K. S. Bawa, H. A. Hespenheide, and G. S. Hartshorn, editors. La Selva: Ecology and natural history of a Neotropical rain forest. The University of Chicago Press, Chicago, Illinois, USA.
Liebig, J. 1842. Animal chemistry or organic chemistry in its application to physiology and Pathology. Johnson Reprint Corporation, New York, USA.
Losos, E. C., and E. G. Leigh, Jr. 2004. Tropical forest diversity and dynamism. The University of Chicago Press, Chicago, Illinois, USA.
Madgwick, H. A. I., and G. L. Brumfield. 1969. The use of hemispherical photographs to assess light climate in the forest. Journal of Ecology 57:537–542.
Mahowald N., T. D. Jickells, A. R. Baker, P. Artaxo, C. R. Benitez-Nelson, G. Bergametti, T. C. Bond, C. Tami, Y. Chen, D. D. Cohen, B. Herut, N. Kubilay, R. Losno, C. Luo, W. Maenhaut, K. A. McGee, G. S. Okin, R. L. Siefert, S. Tsukuda. 2008. Global distribution of atmospheric phosphorus sources, concentrations and deposition rates, and anthropogenic impacts. Global Biogeochemical Cycles 22:GB4026.
Marklein, A. R., and B. Z. Houlton. 2011. Nitrogen inputs accelerate phosphorus cycling rates across a wide variety of terrestrial ecosystems. New Phytologist 193:696–704.
Marschner, H. 1995. Mineral Nutrition of higher plants. Second edition. Elsevier, USA.
Martinelli, L. A., S. Almeida, I. F. Brown, M. Z. Moreira, R. L. Victoria, S. Filoso, C. A. C. Ferreira, and W. W. Thomas. 2000. Variation in nutrient distribution and potential nutrient losses by selective logging in a humid tropical forest of Rondonia, Brazil. Biotropica 32:597–613.
183
Martinelli, L. A., M. C. Piccolo, A. R. Townsend, P. M. Vitousek, E. Cuevas, W. Mcdowell, G. P. Robertson, O. C. Santos, and K. Treseder. 1999. Nitrogen stable isotopic composition of leaves and soil: tropical versus temperate forests. Biogeochemistry 46:45–65.
McGroddy, M. E., T. Daufresne, and L. O. Hedin. 2004.Scaling of C:N:P stoichiometry in forests worldwide: Implications of terrestrial Redfield-type ratios. Ecology 85:2390 –2401.
McKey, D. 1994. Legumes and nitrogen: the evolutionary ecology of a nitrogen-demanding lifestyle. Pages 211–228 in: J. L. Sprent, and D. McKey, editors. Advances in legume systematics: Part 5 - The Nitrogen factor. Royal Botanic Gardens, Kew, England.
Meijer E. L. and P. Buurman. 2003. Chemical trends in perhumid soil catena on the Turrialba volcano (Costa Rica). Geoderma 117:185–201.
Miller, A. J., E. A. G. Schuur, and O. A. Chadwick. 2001. Redox control of phosphorus pools in Hawaiian montane forest soils. Geoderma 102: 219–237.
Mirmanto, E., J. Proctor, J. Green, L. Nagy, and Suriantata. 1999. Effects of nitrogen and phosphorus fertilization in a lowland evergreen rainforest. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences 354:1825–1829.
Murphy, J., and J. P. Riley. 1962. A modified single solution method for the determination of phosphate in natural waters. Annual Chemical Acta 27: 31–36.
Neelin, J. D., M. Münnich, H. Su, J. E. Meyerson, and C. E. Holloway. 2006. Tropical drying trends in global warming models and observations. Proceedings of the National Academy of Sciences USA 103:6110–6115.
Nepstad, D. C., P. Moutinho, M. B. Dias–Filho, E. A. Davidson, G. Cardinot, D. Markewitz, R. Figueiredo, N. Vianna, J. Q. Chambers, D. G. Ray, J. B. Guerreiros, P. A. Lefebvre, L. Da S. L. Sternberg, M. Z. Moreira, L. Barros, F. Y. Ishida, I. Tohlver, E. Belk, K. Kalif, and K. Schwalbe. 2002. The effects of partial throughfall exclusion on canopy processes, aboveground production, and biogeochemistry of an Amazon forest. Journal of Geophysical Research 107(D20):8085.
Newbery, D. M., G. B. Chuyong, J. J. Green, N. C. Songwe, F. Tchuenteu, and L. Zimmermann. 2002. Does low phosphorus supply limit seedling establishment and tree growth in groves of ectomycorrhizal trees in a central African rainforest. New Phytologist 156: 297–311.
Niklas, K. J., T. Owens, P. B. Reich, and E. D. Cobb. 2005.Nitrogen/phosphorus leaf stoichiometry and the scaling of plant growth. Ecology Letters 8:636–642.
184
Okin, G. S., N. Mahowald, O. A. Chadwick, and P. Artaxo. 2004. Impact of desert dust on the biogeochemistry of phosphorus in terrestrial ecosystems. Global Biogeochemical Cycles 18:GB2005.
Olander, L. P., and P. M. Vitousek 2000. Regulation of soil phosphatase and chitinase activity by N and P availability. Biogeochemistry 49: 175–191.
Olander, L. P., and P. M. Vitousek. 2004. Biological and geochemical sinks for phosphorus in soil from a wet tropical forest. Ecosystems 7:404–419.
Ostertag, R. 2001. Effects of nitrogen and phosphorus availability on fine-root dynamics in Hawaiian montane forests. Ecology 82:485–499.
Ostertag, R. 2010. Foliar nitrogen and phosphorus accumulation responses after fertilization: an example from nutrient-limited Hawaiian forests. Plant Soil 334:85–98.
Ostertag, R. and S. E. Hobbie. 1999. Early stages of root and leaf decomposition in Hawaiian forests: effects of nutrient availability. Oecologia 121:564–573.
Palow, D. T. and S. F. Oberbauer. 2009. Soil type affects seedling shade response at low light for two Inga species from Costa Rica. Plant Soil 319:25–35.
Pandey, C. B., R. C. Srivastava, and R. K. Singh. 2009. Soil nitrogen mineralization and microbial biomass relation, and nitrogen conservation in humid-tropics. Soil Science Society of America Journal 73:1142–1149.
Porder, S., D. A. Clark and P. M. Vitousek. 2006. Persistence of rock derived nutrients in the wet tropical forests of La Selva, Costa Rica. Ecology 87:594–602.
Powers, J. S., R. A. Montgomery, E. C. Adair, F. Q. Brearley, S. J. Dewalt, C. T. Castanho, J. Chave, E. Deinert, J. U. Ganzhorn, M. E. Gilbert, J. A. Gonzalez-iturbe, S. Bunyavejchewin, H. R. Grau, K. E. Harms, A. Hiremath, S. Iriarte-vivar, E. Manzane, A. A. De oliveira, L. Poorter, J. B. Ramanamanjato, S. Salk, A. Varela, G. D. Weiblen, and M. T. Lerdau. 2009. Decomposition in tropical forests: a pan-tropical study of the effects of litter type, litter placement and mesofaunal exclusion across a precipitation gradient. Journal of Ecology 97:801–811.
Proyecto Atlas Digital de Costa Rica. 2008. Instituto Tecnológico de Costa Rica (ITCR), editors. Escuela de Ingeniería Forestal. Cartago, Costa Rica. 1 DVD.
Radulovich R. and P. Sollins 1991. Nitrogen and Phosphorus leaching in zero-tension drainage from a humid tropical soil. Biotropica 23:84–87.
Rastetter, E. B., and G. R. Shaver. 1992. A model of multiple-element limitation for acclimating vegetation. Ecology 73:1157–1174.
185
Reed, S. C., C. C. Cleveland, and A. R. Townsend. 2007. Controls over free-living nitrogen fixation in a lowland tropical rain forest. Biotropica 39:585–592.
Reed, S. C, C. C. Cleveland, and A. R. Townsend. 2008. Tree species control rates of free-living nitrogen fixation in a tropical rain forest. Ecology 89:2924–2934.
Reich, P. B., and J. Oleksyn. 2004. Global patterns of plant leaf N and P in relation to temperature and latitude. Proceedings of the National Academy of Sciences (USA) 101:11001–11006.
Richardson, S. J., R. B. Allen, and J. E. Doherty. 2008. Shifts in leaf N:P ratio during resorption reflect soil P in temperate rainforest. Functional Ecology 22:738–745.
Riley, R. H., and P. M. Vitousek. 1995. Nutrient dynamics and nitrogen trace gas flux during ecosystem development in montane rain-forest. Ecology 76:292–304.
Rinklebe, J., and U. Langer. 2006. Microbial diversity in three floodplain soils at the Elbe River (Germany). Soil Biology and Biochemistry 38:2144–2151.
Robertson, G. P., D. Wedin, P. M. Groffman, J. M. Blair, E. A. Holland, K. J. Nadelhoffer, and D. Harris. 1999. Soil carbon and nitrogen availability, nitrogen mineralization, nitrification, and soil respiration potentials. Pages 115–142 in G. P. Robertson, D. C. Coleman, C. S. Bledsoe, and P. Sollins, editors. Standard soil methods for long-term ecological research. Oxford University Press, New York, USA.
Sack, L., and P. J. Grubb. 2002. The combined impacts of deep shade and drought on the growth and biomass allocation of shade-tolerant woody seedlings. Oecologia 131:175–185.
Sanchez, P. A. 1976. Properties and management of soils in the tropics. Wiley-Interscience Publications, New York.
Sancho, F., R. Mata, E. Molina, and R. Salas. 1990. Estudio de suelos finca de la Escuela de Agricultura de la Región Tropical Húmeda. Reporte para la Escuela de Agricultura de la Región Tropical Húmeda (EARTH), Limón, Costa Rica.
Santiago, L. S, K. Kitajima, S. J. Wright, and S. S. Mulkey. 2004. Coordinated changes in photosynthesis, water relations and leaf nutritional traits of canopy trees along a precipitation gradient in lowland tropical forest. Oecologia 139:495–¬502.
Santiago, L. S., E. A. G. Schuur, and K. Silvera. 2005. Nutrient cycling and plant-soil feedbacks along a precipitation gradient in lowland Panama. Journal of Tropical Ecology 21: 461–470.
Sato, S., and N. B. Comerford. 2006. Assessing methods for developing phosphorus desorption isotherms from soils using anion exchange membranes. Plant and Soil 279:107–117.
186
Schuur, E. A. G. 2001. The effect of water on decomposition dynamics. Ecosystems 4:259–273.
Schuur, E. A. G. 2003. Productivity and global climate revisited: The sensitivity of tropical forest growth to precipitation. Ecology 84: 1 165–1170.
Schuur, E. A. G., O. A. Chadwick, and P. A. Matson. 2001. Carbon cycling and soil carbon storage in mesic to wet Hawaiian montane forests. Ecology 82:3182–3196.
Schuur, E. A. G., and P. A. Matson. 2001. Net primary productivity and nutrient cycling across a mesic to wet precipitation gradient in Hawaiian montane forest. Oecologia 128:431–442.
Sheil, D. Growth assessment in tropical tree: large daily diameter fluctuations and their concealment by dendrometer bands. 2003. Canadian Journal of Forest Research 33: 2027–2035.
Smil, V. 2000. Phosphorus in the environment: Natural flows and human interferences. Annual Review of Energy in the Environment 25: 53–88.
Sollins, P., M. F. Sancho, Ch. R. Mata, and R. L. Sanford, Jr.1994. Soils and soil process research. Pages 34–53 in L. A.McDade, K. S. Bawa, H. A. Hespenheide, and G. S. Hartshorn, editors. La Selva: ecology and natural history of a neotropical rain forest. University of Chicago Press, Chicago, Illinois, USA.
Sterner, R. W., and J. J. Elser. 2002. Ecological stoichiometry: the biology of elements from molecules to the biosphere. Princeton, NJ, USA: Princeton University Press.
Stone, E. L. 1968. Microelement nutrition of forest trees: A review. Pages 132–175 in Forest fertilization, theory and practice. Tennessee Valley Authority. N.F.D.C., Muscle Shoals, AL.
Tanner, E. V. J., and V. Kapos. 1992. Nitrogen and phosphorus fertilization effects on Venezuelan montane forest trunk growth and litterfall. Ecology 73:78–86.
Tanner, E. V. J., P. M. Vitousek, and E. Cuevas. 1998. Experimental investigation of nutrient limitation of forest growth on wet tropical mountains. Ecology 79: 10-22.
Tiessen, H. 1998. Resilience of phosphorus transformations in tropical forest and derived ecosystems. Pages 92–98 in A. Schulte and D. Ruhiyat, editors. Soils of tropical forest ecosystems: characteristics, ecology and management. Springer, New York.
Tilman, D., J. Fargione, B. Wolff, C. D'Antonio, A. Dobson, R. Howarth, D. Schindler, W. H. Schlesinger, D. Simberloff, and D. Swackhamer. 2001. Forecasting agriculturally driven global environmental change. Science 292:281–284.
187
Tosi, J. A. 1969. Mapa Ecológico, República de Costa Rica: Según la clasificación de zonas de vida del mundo de L.R. Holdridge. Tropical Science Center, San José, Costa Rica.
Townsend, A. R., G. P. Asner, and C. C. Cleveland. 2008. The biogeochemical heterogeneity of tropical forests. Trends in Ecology and Evolution 23:424–431.
Townsend, A. R., G. P. Asner, C. C. Cleveland, M. E. Lefer and M. M.C.Bustamante. 2002. Unexpected changes in soil phosphorus dynamics along pasture chronosequences in the humid tropics. Journal of Geophysical Research 107(D20): 8067.
Townsend, A. R., C. C. Cleveland, G. P. Asner, and M. M. C. Bustamante. 2007. Controls over foliar N:P ratios in tropical rain forests. Ecology 88:107–118.
Townsend, A. R., C. C. Cleveland, B. Z. Houlton, C. B. Alden, and J. W. C White. 2011. Multi-element regulation of the tropical forest carbon cycle. Frontiers in Ecology and the Environment 9:9–17.
Treseder, K. K., and P. M. Vitousek. 2001. Effects of soil nutrient availability on investment in acquisition of N and P in Hawaiian rain forests. Ecology 82:946–954.
Turner, B. L. and B. M. J. Engelbrecht. 2011. Soil organic Phosphorus in lowland tropical rain forests. Biogeochemistry 103:297–315.
Vance, C. P., C. Uhde-Stone, and D. L. Allen. 2003. Phosphorus acquisition and use: critical adaptations by plants for securing a nonrenewable resource. New Phytologist 157:423–447.
Vandecar , K. L., D. Lawrence, T. Wood, S. F. Oberbauer, R. Das, K. Tully, and L. Schwendenmann. 2009. Biotic and abiotic controls on diurnal fluctuations in labile soil phosphorus of a wet tropical forest. Ecology 90: 2547–2555.
Vitousek, P. M. 1984. Litterfall, nutrient cycling, and nutrient limitation in tropical forests. Ecology 65: 285–298.
Vitousek, P. M. 1998. Foliar and litter nutrients, nutrient resorption, and decomposition in Hawaiian Metrosideros polymorpha. Ecosystems 1:401–407.
Vitousek, P. M. 2004. Nutrient cycling and limitation. Hawai'i as a model system. Princeton University Press, New Jersey, USA.
Vitousek, P. M. and J. S. Denslow. 1986. Nitrogen and Phosphorus availability in tree fall gaps of a lowland tropical rainforest. Journal of Ecology 74: 1167–1178.
Vitousek P. M. and J. S. Denslow. 1987. Source differences in extractable phosphorus among soils of the La Selva Biological Station, Costa Rica. Biotropica 19:167–170.
188
Vitousek, P. M., and H. Farrington. 1997. Nutrient limitation and soil development: Experimental test of a biogeochemical theory. Biogeochemistry 37:63–75.
Vitousek, P. M, and R. W. Howarth. 1991. Nitrogen limitation on land and in the sea: How can it occur? Biogeochemistry 13:87–115.
Vitousek, P.M. R. Lawrence, R. Walker, L. D. Whiteaker and P. M. Matson. 1993. Nutrient limitations to plant growth during primary succession in Hawaii Volcanoes National Park. Biogeochemistry: 197–215.
Vitousek, P. M., S. Porder, B. Z. Houlton, and O. A. Chadwick. 2010. Terrestrial phosphorus limitation: mechanisms, implications, and nitrogen-phosphorus interactions. Ecological Applications 20:5–15.
Walker, T. W., and J. K. Syers. 1976. Fate of phosphorus during pedogenesis. Geoderma 15:1–19.
Wieder, W. R., C. C. Cleveland, and A. R. Townsend. 2008. Tropical tree species composition affects the oxidation of dissolved organic matter from litter. Biogeochemistry 88:127–138.
Wieder, W. R., C. C. Cleveland, and A. R. Townsend. 2009. Controls over leaf litter decomposition in wet tropical forests. Ecology 90: 3333–3341.
Wood, T. E., D. Lawrence, D. A. Clark and R. L. Chazdon. 2009. Rain forest nutrient cycling and productivity in response to large-scale litter manipulation Ecology 90: 109-121.
Wood, T. E., D. Lawrence, and J. A. Wells. 2011. Inter-specific variation in foliar nutrients and resorption of nine canopy tree species in a secondary neotropical rain forest. Biotropica 43:544–551.
Wright, R. B., B. G. Lockaby, and M. R. Walbridge. 2001. Phosphorus availability in an artificially flooded southeastern floodplain forest soil. Soil Science Society of America Journal 65:1293–1302.
Wright, S. J., J. B., Yavitt, N. Wurzburger, B. L. Turner , E. V. Tanner , E. J. Sayer, L. S. Santiago, M. Kaspari, L. O. Hedin, K. E. Harms, M. N. Garcia, and M. D. Corre. 2011. Potassium, Phosphorus or nitrogen limit root allocation, tree growth and litter production in a lowland tropical forest. Ecology 92:1616–1625.
Wu, C. C., C. C. Tsui, C. F. Hseih, V. B. Asio, and Z. S. Chen. 2007. Mineral nutrient status of tree species in relation to environmental factors in the subtropical rain forest of Taiwan. Forest Ecology and Management 239:81–91.
189
Yavitt, J. B., K. E. Harms, M. N. Garcia, M. J. Mirabello, and S. J. Wright. 2011. Soil fertility and fine root dynamics in response to 4 years of nutrient (N, P, K) fertilization in a lowland tropical moist forest, Panama. Austral Ecology 36:433–445.
Yavitt, J. B., and S. J. Wright. 2008. Seedling growth responses to water and nutrient augmentation in the understorey of a lowland moist forest, Panama. Journal of Tropical Ecology 24:19–26.
Yuan, Z. Y., Y. H. H. Chen, and P. B. Reich. 2011. Global-scale latitudinal patterns of plant fine-root nitrogen and phosphorus. Nature Communications 2:344.
Zarin, D., R. Janaki, R. Alavalapati, F. E. Putz, and M. Schmink. 2004. Working Forests in the Neotropics: Conservation through sustainable management. Columbia University Press, New York, USA.
190
BIOGRAPHICAL SKETCH
Silvia Alvarez Clare was born in San José, Costa Rica in 1977. She attended Saint
Francis High School and upon graduation, decided to explore the world by backpacking
through Asia for a year. After her return, she studied biology at the Universidad de
Costa Rica (UCR), where she obtained a bachelor’s degree in July 2001. She briefly
taught biology at Lincoln High School, in San José but decided to continue her studies
in Tropical Plant Ecology. Silvia became a Gator in 2002, when she started a master’s
degree with Dr. Kaoru Kitajima. She conducted her master’s research on seedling
biomechanics in Barro Colorado Island, Panama. Silvia obtained her MSc degree in
botany with a minor in statistics in May 2005. Silvia then decided to expand the scope of
her research from plant ecophysiology to ecosystems ecology, and started a PhD in
interdisciplinary ecology with Dr. Michelle Mack, in the Department of Biology at the
University of Florida. During her PhD, Silvia conducted her research at EARTH
(Escuela de Agricultura de la Región del Trópico Húmedo) University Forest Reserve, in
the Caribbean Slope of Costa Rica. During these years, Silvia also got married, moved
to Chicago, and became a mom. Silvia obtained her PhD with a concentration in soil
and water science in May 2012 and will continue her work at EARTH as an NSF
Postdoctoral Fellow, measuring the effect of nutrient additions on greenhouse gas
emissions from tropical forest soils.