<%BANNER%>

Plant Growth and Soil Responses to Simulated Nitrogen Deposition and Dry Season Precipitation in a Neotropical Savanna

Permanent Link: http://ufdc.ufl.edu/UFE0025076/00001

Material Information

Title: Plant Growth and Soil Responses to Simulated Nitrogen Deposition and Dry Season Precipitation in a Neotropical Savanna
Physical Description: 1 online resource (75 p.)
Language: english
Creator: Copeland, Stella
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: brazil, cerrado, grass, growth, nitrogen, phosphorus, precipitation, reproduction, soil
Wildlife Ecology and Conservation -- Dissertations, Academic -- UF
Genre: Wildlife Ecology and Conservation thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Anthropogenic increases in nitrogen deposition and precipitation change could alter plant growth and biomass allocation, foliar nutrient concentration, and the influence of plant species on soil nutrient composition. The potential responses of Neotropical savanna plant species and soils to global change factors are unknown, despite the possibility of climate feedbacks and reduced biodiversity. We tested how simulated precipitation change and nitrogen deposition would affect the growth and reproduction of the native grasses Loudetiopsis chrysothrix and Tristachya leiostachya and characteristics of their associated soils in the Brazilian Cerrado. The two species responded differently to water, nitrogen, and their interaction. Tristachya was more likely to flower with water addition, whereas Loudetiopsis individuals were more likely to flower with both water and nitrogen. Tristachya decreased genet diameter growth whereas Loudetiopsis increased genet diameter growth with added nitrogen. Loudetiopsis dry season leaf senescence decreased with water addition however, none of the treatments affected Tristachya leaf senescence. Loudetiopsis individuals' root:shoot ratios decreased with added water, but Tristachya individuals' root:shoot ratios did not change in any treatment. The foliar phosphorous of Tristachya individuals increased, while Loudetiopsis foliar phosphorus decreased with the water and nitrogen treatment. None of the treatments significantly affected foliar nitrogen for either species. Plant - available phosphorus concentration increased in Loudetiopsis associated soils with nitrogen addition and increased in Tristachya associated soils with the nitrogen and water treatment. These results suggest that interspecific differences in phosphorus acquisition and use could influence Neotropical plant and soil responses to global change factors. Nitrogen addition alone did not affect most reproductive variables, increased the growth of only one species, and did not increase foliar nitrogen. In contrast, water addition affected a wide variety of traits, especially in combination with nitrogen. These results imply that the impacts of nitrogen deposition and precipitation change in the Cerrado may significantly interact, and vary by species.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Stella Copeland.
Thesis: Thesis (M.S.)--University of Florida, 2009.
Local: Adviser: Bruna, Emilio M.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0025076:00001

Permanent Link: http://ufdc.ufl.edu/UFE0025076/00001

Material Information

Title: Plant Growth and Soil Responses to Simulated Nitrogen Deposition and Dry Season Precipitation in a Neotropical Savanna
Physical Description: 1 online resource (75 p.)
Language: english
Creator: Copeland, Stella
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: brazil, cerrado, grass, growth, nitrogen, phosphorus, precipitation, reproduction, soil
Wildlife Ecology and Conservation -- Dissertations, Academic -- UF
Genre: Wildlife Ecology and Conservation thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Anthropogenic increases in nitrogen deposition and precipitation change could alter plant growth and biomass allocation, foliar nutrient concentration, and the influence of plant species on soil nutrient composition. The potential responses of Neotropical savanna plant species and soils to global change factors are unknown, despite the possibility of climate feedbacks and reduced biodiversity. We tested how simulated precipitation change and nitrogen deposition would affect the growth and reproduction of the native grasses Loudetiopsis chrysothrix and Tristachya leiostachya and characteristics of their associated soils in the Brazilian Cerrado. The two species responded differently to water, nitrogen, and their interaction. Tristachya was more likely to flower with water addition, whereas Loudetiopsis individuals were more likely to flower with both water and nitrogen. Tristachya decreased genet diameter growth whereas Loudetiopsis increased genet diameter growth with added nitrogen. Loudetiopsis dry season leaf senescence decreased with water addition however, none of the treatments affected Tristachya leaf senescence. Loudetiopsis individuals' root:shoot ratios decreased with added water, but Tristachya individuals' root:shoot ratios did not change in any treatment. The foliar phosphorous of Tristachya individuals increased, while Loudetiopsis foliar phosphorus decreased with the water and nitrogen treatment. None of the treatments significantly affected foliar nitrogen for either species. Plant - available phosphorus concentration increased in Loudetiopsis associated soils with nitrogen addition and increased in Tristachya associated soils with the nitrogen and water treatment. These results suggest that interspecific differences in phosphorus acquisition and use could influence Neotropical plant and soil responses to global change factors. Nitrogen addition alone did not affect most reproductive variables, increased the growth of only one species, and did not increase foliar nitrogen. In contrast, water addition affected a wide variety of traits, especially in combination with nitrogen. These results imply that the impacts of nitrogen deposition and precipitation change in the Cerrado may significantly interact, and vary by species.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Stella Copeland.
Thesis: Thesis (M.S.)--University of Florida, 2009.
Local: Adviser: Bruna, Emilio M.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0025076:00001


This item has the following downloads:


Full Text

PAGE 1

1 PLANT GROWTH AND SOIL RESPONSES TO SIMULATED NITROGEN DEPOSITION AND DRY SEASON PRECIPITATION IN A NEOTROPICAL SAVANNA By STELLA M. COPELAND A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Stella M. Copeland

PAGE 3

3 To the Cerrado To the hope of a future with wild plac es and intact ecosystems To my grandparents, who collectively encouraged intellectual curiosity, compassion for my fellow human beings and a passion for biodiversity

PAGE 4

4 ACKNOWLEDGMENTS I am eternally grateful to my committee, Dr. Heraldo Vasconcelos, Dr. Michelle Mack, and my adviser Dr. Emilio Bruna, for their support and advice. A special note of thanks to Dr. John (Jack) Ewel, who graciously agreed to participate in my defense in replacement for an off campus member. I am indebted to the students of the Laboratrio de Inse to s Socias, Universidade Federal de Uberlndia, Braz il, for their help and kindness as I navigated Brazilian culture, Portuguese language, and Cerrado ecology in pursuit of my Masters data I thank the staff of the Laboratrio de Anlises de Solos e Calcri os Universidade Federal de Uberlndia, who cheerfully guided me through a litany of soil and foliar analyses. I am grateful to numerous professors, staff, and fellow graduate students of the Department of Wildlife Conservation and Ecology and Department of Botany at University of the Florida who were invaluable throughout the process of thesis development, implementation, and completion. Meg han Brennan and James Colee, of the UF Institute of Food and Agricultural Sciences Statistics Consulting Unit provided crucial research design and statistical analysis support. Special thanks to the Bruna Lab and the Mack Lab for feedback, moral support, and encouragement, particularly through the dark period known as learning how to use the autoanalyzer I would neve r have arrived, or succeeded, at the University of Florida without the exemplary training I received Colorado College where I completed my BA. Last, but never least, to my friends and family who have been steadfast in their support for the last two years, and always This research was supported by the National Science F oundation (Grant DEB 0542287 and the Graduate Research Fellowship Program ).

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF TABLES ...........................................................................................................................7 LIST OF FIGURES .........................................................................................................................9 ABSTRACT ...................................................................................................................................10 CHAPTER 1 INTRODUCTION ..................................................................................................................12 2 METHODS .............................................................................................................................17 Study Site ................................................................................................................................17 Species ....................................................................................................................................17 Experimental Design ..............................................................................................................18 Treatments ..............................................................................................................................19 Light Availability ....................................................................................................................20 Plant Response ........................................................................................................................20 Soil and Foliar Nutrients ........................................................................................................22 Statistical Analysis ..................................................................................................................23 Reproductive Variables ...................................................................................................23 Growth and Foliar Nutrient Variables .............................................................................23 Soil Variables ..................................................................................................................24 Model Assumptions and Variable Transformations ........................................................24 3 RESULTS ...............................................................................................................................26 Reproductive Response ..........................................................................................................26 Growth, Allocation, and Foliar Nutrients ...............................................................................27 Soil Response ..........................................................................................................................29 4 DISCUSSION .........................................................................................................................31 Reproductive Respons e ..........................................................................................................31 Growth Response ....................................................................................................................32 Leaf Senescence ......................................................................................................................33

PAGE 6

6 Foliar Nitrogen and Phosphorus .............................................................................................34 Soils Response ........................................................................................................................36 5 CONCLUSION .......................................................................................................................38 APPENDIX: TABLE OF UNIVARIATE STATISTICS ............................................................64 LIST OF REFERENCES ...............................................................................................................69 BIOGRAPHICAL SKETCH .........................................................................................................75

PAGE 7

7 LIST OF TABLES Table page 31 Flowering probability in response to water and nitrogen by species .................................50 32 Flowering likelihood in response to resin available nitrate and ammonium by species ....50 33 Number of flowering tillers per flowering individual by treatment and species. ..............51 34 Total number of spikelets by treatment and species ..........................................................51 35 Number of spikelets per flowering tiller by treatment and species ...................................52 36 Difference in diameter between beginning and end of experiment by species and treatment ............................................................................................................................53 37 Root:shoot ratios by species and treatment ........................................................................54 38 Live leaf (LL), and dead aboveground biomass (DAB), and total aboveground biomass (TAB, g dry weight), each divided by diameter by treatment and species ..........55 39 Number of live leaves per area (cm2) by treatment and species ........................................56 310 Foliar nitrogen (g/Kg) by treatment and species ................................................................56 311 Foliar phosphorus (g/Kg) by treatment and species ..........................................................57 312 N:P ratios by treatment and species ...................................................................................57 313 Response of soil phosphorus concentration (g/Kg of dry soil) to treatments and species ................................................................................................................................58 314 Response of soil phosphorus concentration (g/Kg of dry soil) to treatments by species ................................................................................................................................58 315 Analysis of pH, aluminum, potassium, and calcium response to treatments. ....................59 316 species ................................................................................................................................60 317 species ................................................................................................................................60

PAGE 8

8 318 treatments and species ........................................................................................................61 319 species ................................................................................................................................61 320 Results of soil ammoniu .........................62 321 ..................................62 41 Table of significant results .................................................................................................63

PAGE 9

9 LIST OF FIGURES Figure page 11 Nitrogen deposition responses in nitrogen limited ecosystems (adapted from Aber et al., 1998). ...........................................................................................................................40 12 Nitrogen deposition responses in phosphorus limited ecosystems (adapted from Matson et al. 1999). ..........................................................................................................41 21 Map of Estao Ecolgica do Panga and location of the study area .................................42 22 Photos of focal species and habitat ....................................................................................43 31 Flowering likelihood by treatment for Loudetiopsis and Tristachya ................................44 32 Number of flowering tillers by treatment for Loudetiopsis and Tristachya .....................45 33 Diameter difference by treatment for Loudetiopsis and Tristachya .................................46 34 Foliar nitrogen by treatment for Loudetiopsis and Tristachya ...........................................47 35 Foliar phosphorus by treatment for Loudetiopsis and Tristachya .....................................48 36 Soil phosphorus (g/Kg) by treatment for Loudetiopsis and Tristachya ............................49

PAGE 10

10 Abstract of Thesis Presented to the Graduate School of the University of Florida in Par tial Fulfillment of the Requirements for the Degree of Master of Science PLANT GROWTH AND SOIL RESPONSES TO SIMULATED NITROGEN DEPOSITION AND DRY SEASON PRECIPITATION IN A NEOTROPICAL SAVANNA By Stella M. Copeland August 2009 Chair: Emilio M. Bruna Maj or: Wildlife Ecology and Conservation Anthropogenic increase s in nitrogen deposition a nd precipitation change could alter plant growth and biomass allocation, fol iar nutrient concentration and the influence of plant species on soil nutrient composition. T he potential responses of Neotropical savanna plant species and soils to global change factors are unknown, despite the possibility of climate feedbacks and reduced biodiversity. We tested how simulated precipitation change and nitrogen deposition would a ffect the growth and reproduction of the native grass es Loudetiopsis chrysothrix and Tristachya leiostachya a nd characteristics of their associated soi ls in the Brazilian Cerrado. T he two species responded differently to water, nitrogen, and their interaction Tristachya was more likely to flower with water addition, whereas Loudetiopsis individuals were more likely to flower with both water and nitrogen. Tristachya decreased genet diameter growth whereas Loudetiopsis increased genet diameter growth with added nitrogen. Loudetiopsis dry season leaf senescence decreased with water addition however, none of the treatments affected Tristachya leaf senescence. Loudetiopsis individuals r oot:shoot ratios decreased with added water, but Tristachy a i ndividuals root:shoot ratios did not change in any treatment The foliar phosphorous

PAGE 11

11 of Tristachya individuals increased, while Loudetiopsis foliar phosphorus decreased with the water and nitrogen treatment. None of the treatments significantly affected f oliar nitrogen for either species Plant available phosphorus concentration increased in Loudetiopsis associated soils with nitrogen addition and increased in Tristachya associated soils with the nitrogen and water treatment. These results suggest that int erspecific differences in phosphorus acquisition and use could influenc e Neot ropical plant and soil response s to global change factors Nitrogen addition alone did not a f fect most reproductive variables increased the growth of only one species, and d id not increase foliar nitrogen. In contrast, water addition affected a wide variety of traits, especially in combination with nitrogen. These results impl y that the impacts of nitrogen deposition and precipitation change in the Cerrado may significantly intera ct and var y by species

PAGE 12

12 CHAPTER 1 INTRODUCTION Anthropogenic nitrogen addition has more than doubled pre industrial nitrogen inputs to terrestrial environments (reviewed in Schlesinger, 2009) In addition to a suite of soil biogeoche mical changes such as increased soil acidity and ca tion availability (Aber et al. 1998, Vitousek et al. 1997) nitrogen deposition also affects plants by causing them to alter their patterns of growth biomass allocation, phenology, fitness, and their interactions with other members of the local plant community (Clark & Tilman, 2008, Cleland et al. 2006, Lau et al. 2008) However, t he dire ction and magnitude of these plant response s can also be influenced by other factors including cli mate and species identity (Craine et al. 2002, Zavaleta et al. 2003) The responses of plants and soil to nitrogen deposition can ultimately change aboveground net primary production (ANPP ) decreas e plant diversity, and impact climate via negative or positive feedback s to the carbon cycle (Gruber, 2008, Vitousek et al. 1997) H uman forced climate change is altering the pattern and abundance of precipitation across the worlds surface (Zhang et al. 2007) Plants are particularly sen sitive to changes in precipitation patterns because the amount of annual rainfall and its variability exert strong effects on plant growth and phenology (Fay et al. 2003, Kochy & Wilson, 2004, Zavaleta et al. 2003) Precipitation change can affect the nitrogen cycle in x eric systems because nitrogen mineralization and plant demand for nitrogen are often tightly synchronized with seasonal precipitation (Austin et al. 2004, Knapp et al. 2006, Yahdjian et al. 2006) The few experimental studies to date combining nitrogen enrichment and precipitation change (Cleland et al., 2006, Siemann et a l. 2007) have documented idiosyncratic responses additive, dampened, and multiplicative interactive effects on plant flowering and growth (Cleland et al. 2006, Henry et al. 2006, Zavaleta et al. 2003)

PAGE 13

13 M ost experimental research evaluating the ecological consequences of global change phenomena has been conducted in the temperate zone (Carrera et al. 2003, Cleland et al. 2006, Fis her & Whitford, 1995, Kochy & Wilson, 2004, Zavaleta et al. 2003) Consequently, the responses of tropical ecosyst ems which account for a large proportion of the globes net primary production and are reservoirs of biodiversity to global change factors are relatively unknown (Matson et al. 1999) Th e lack of experimental evidence exists despite the increase of nitrogen deposition in several tropical hotspots (Phoenix et al. 2006) and the recognition that even moist tropical forests coul d be highly sensitive to precipitation change (Phillips et al., 2006) The research and theory that does address the response of tropical ecosystems to global change focuses primarily on tropical wet forests (Davidson et al. 2007, Matson et al. 1999, Phillips et al. 2009, Vitousek, 1984) However, the tro pics include a range of biomes including savannas, grasslands, and dry forests and the responses of these ecosystems to nitrogen deposition and altered precipitation regimes are likely to be markedly different Vegetation in t ropical savannas and grass lands may be particularly sensitive to either increases or decreases in rainfall patterns because of ecosystem dependence on seasonal precipitation (Austin et al. 2004, O'Connor, 1994, Pandey & Singh, 1992, Seagle & McNaughton, 1993) Because nitrogen can limit plant growth in tropical grasslands and savannas (Augustine, 2003, Barger et al. 2002, Sarmiento et al. 2006) nitrogen deposition could increase aboveground productivity Precipitation changes and nitrogen deposition could interact t o influence plant growth because seasonal rainfall patterns can determine the magnitude and timing of nitrogen mineralizatio n (Austin et al. 2004, Seagle & McNaughton, 1993, Yahdjian et al., 2006) T he relative importance of precipitation and nitrogen as controls over plant growth can vary with extraneous factors such as successional state, soil fertility and fire fr equency

PAGE 14

14 (Bustamante et al. 2006, Davidson et al. 2004, Harrington et al. 2001, Sarmiento et al. 2006, Vitousek, 1984) The potential ecosystem consequences of the combination of altered precipitation and nitrogen deposition in tropical savannas are unclear, since to my knowledge no research ha s addressed their interactive effects on savanna plants and soils The global significance of tropical savanna response to global change factors is well represented by their conservation value and substantial extent in the Neotropics. Savannas are the second largest Neotropical biome after lowland Neotropical forests, comprising 45 % of the area of South America (Huntley & Walker, 1982) The largest and most biodiverse Neotropical savanna region is the Brazilian Cerrado which occupies 2 million km2 and harbors an estimated 10,000 plant species (Oliveira & Marquis, 2002, Ratter et al. 1997) Cerrado species and ecosystem function are increasingly threatened by mecha nized agriculture, introduction of nonna tive grasses, and urbanization (Ratter et al., 1997) The increasing use of nitrogen fertilizer and rising rates of fossil fuel combustion in Brazils ex tensive urban areas are increasing nitrogen deposition over wide swathes of the Cerrado from approximately an average of 5 13 kg N ha 1 yr1 in the 1990s to a projected 14 38 kg N ha1 yr1 by the year 2050 (Bustama nte et al. 2006, Phoenix et al. 2006) The effects of nitrogen enrichment on Cerrado plant species is uncertain as the relative importance of nitrogen and phosphorus limitation can vary by time since fire, species characteristics, and insect herbivore activity (Bucci et al. 2006, Bustamante et al. 2004, Nardoto et al. 2006, Sternberg et al. 2007) F or a number of reasons, the Cerrado nitrogen cycle is quite distinc t from other tropical savannas For example, t he Cerrado lacks the migrating large ungulates that seasonally increase nitrogen availability (Augustine et al. 2003) and counteract nitrogen loss from frequent grassland fires (Cech et al. 2008, Holdo et al. 2007) in African

PAGE 15

15 savannas. The Cerrado equi va lent may be insect h erbivory: leaf cutting ants have been tied to increased soil nitrogen availability (Sternberg et al. 2007) and foliar nutrient content (Mundim et al. 2009) While herbivores like leaf cutting ants may harvest a significant amount of biomass (Costa et al. 2008) their enriching effects on soil nutrients may be offset by fire (Sousa Souto et al., 2008) A substantial proportion of aboveground nitrogen and sulfur can be lost in Cerrado fires (Boone Kauffman et al. 1994) though soil nitrogen and other nutrient values may actually be higher in areas with higher fire frequency (da Silva & Batalha, 2008) Native nitrogen fixing plant species and free living microbes may ad d 1644 kg N ha1 yr1 to the Cerrado biome as a whole however their effectiveness may be limited by other macronutrients (Bustamante et al. 2006) Nitrogen availability is also linked to plant water uptake in the Cerrado, as there is some evidence that nitrogen fertilizatio n can affect plant water uptake (Bucci et al. 2006, Scholz et al. 2007) but the possible interactive effects of nitrogen and water on native plant species have not been addressed experimentally Because of the uncertainty regarding plant nutrient limitation in the Cerrado, I propose two opposing hypotheses for the effects of dry season water addition and nitrogen addition on plant growth and allocation. P revious research has suggested that Cerrado species may be nitrogen limited (Bustamante et al. 2006) I therefore predicted that nitrogen addition c ould have a fertilizing effect on plant growth ( referred to hereafter as the Fertilization Hypothesis ). Specifically, I expected that growth and reproduction would increase in response to increasing plant a vailable nitrogen resulting from nitrogen enrichment Whereas, t he effects of p recipitation change effects woul d depend on the direction of rainfall change (increase or decrease) and the timing of rain events (Austin et al. 2004) A majority of IPCC m odels predict increasing precipitation in the dry season for the Cerrado ; I predict ed that this could lead to reduced leaf

PAGE 16

16 senescence due to water limitation, higher rates of photosynthesis and carbon storage. Water addition could also increase growth and reproduction indirectly b y mimic king the stimulating effect of precipitation on nitrogen mineralization (Seagle & McNaughton, 1993) Although nitrogen and water could limit Cerrado plant growth and reproduction, nitrogen deposition could actually exacerbate soil conditions that inhibit g r owth (e.g., low cation and phosphorus availability, high acidity, and high aluminum levels (Matson et al., 1999) The result could be growth limitation by macronutrients other than nitrogen (e.g., phosphorus ) or aluminum toxicity Such as response could diminish or reverse the fertilizing effect of added nitrogen particularly if phosphorus currently limits plant growth (referred to hereafter as the Limitation Hypothesis) I did not expect precipitation change to have direct negative effects on soil fertility However, if plants increased growth with precipitation change, increasing plant demand for macronutrients could diminish their concentrations in soil (Sardans & Penuelas, 2007) Here I report the results of a year long experiment testing the effects of simulated nitrogen deposition and dry season precipitation increase on two dominant C 4 grass species and their associated soils, in the Brazilian Cerrado. I tested the Fertilization and Limitation Hypotheses by experimentally adding water during the dry season in amounts and intervals consistent with IPCC predictions for this area of South America (Magrin et al. 2007) and adding nitrogen in amounts and composition (ammonium nitrate) relevant to future nitrogen deposition predictions (Phoenix et al. 2006) As my objectives were to disentangle the direct and indirect effects of my treatments, I measured soil fertility factors in addition to plant nutrient and growth characteristics .

PAGE 17

17 CHAPTER 2 METHODS Study Site This study wa s conducted at the Estao Ecolgica do Panga, a 404 ha preserve 40 km from Uberlndia, Minas Gerais, Brazil (19 S, 48 23 W Fig. 21 ). Panga contains most of the m ajor Cerrado vegetation physiognomies (mata semidecdua, mata galeria, vereda, cerrado, cerrado sensu stricto, de Oliveira Filho & Ratter, 2002) The climate is subtropical, with approximately 1600 mm rainfall per year, monthly average temperatures between 20 25 n almost rainless dry season between May and September (Instituto de Geografia, 2008) The soil is a highly weathered O xisol with a high clay content and low pH ; it is classified as an Anionic Acrustoxe (Soil Survey Staff, 2003) according to US soil taxonomy, or a Latossolo VermelhoAmarelho (EMPRAPA, 1999) in the Brazili an soil classification system. I conducted my research in a vegetation physiognomy called campo ralo. Campo ralo vegetation is typified by dense grass cover interspersed by small stature trees and shrubs (Ottmar et al. 2001) The preserve is generally protected from grazing or other agricultural activities, but is subject to occasional anthropogenic fires originating on adjoining roads and f arms. The most recent fire in the study area occurred in 2006, 2 years before the beginning of the experiment. Species The focal species for this study were two native caespitose C 4 perennial grasses, Tristachya leiostachya Nees and Loudetiopsis chrysothrix (Nees) Conert (Poaceae, Tribe: Arundinelleae Fig. 22) T ristachya is generally larger than L oudetiopsis the average Tristachya genet (individual bunchgrass) is 25 cm in diameter whereas Loudetiopsis genets are 10 cm in average diameter Tristachya vegetative tillers are 90 cm tall on average and Loudetiopsis vegetative tillers are 70 cm tall on average. B oth species flower between February

PAGE 18

18 and April and are co dominant in the campo sujo physiognomy where the experiment took pla ce. Combined, they accounted for 69% of the aboveground biomass in the study area ( Loudetiopsis = 12%, Tristachya = 57%, all other species < 5 %, E. Bruna and H. Vasconcelos unpublished data) In addition to being locally common, both species have wide geographic ranges : Tristachya ranges from southern Brazil to Paraguay and Loudetiopsis is known from eastern Boliv ia to southern Brazil, and Paraguay (Missouri Botanical Garden, 2009) Experimental Design To select grass individuals for the experiment I first established 6 transects 50 meters apart and 150 meters (5 transects) or 50 meters (1 transect) in long in a relatively homogeneous area ( e.g., similar aspect, slope, vegetation) I chose l ocations and species along transect s by randomly generating 32 numbers from 0149 and randomly ass igning species to each location before going into the field. Approximately 32 individuals ( N= 16 of each species) were located along each transect, separated by a minimum of 2 meters and sometimes as much as 10 meters In total I established 80 plots for each species within an area of approximately 150 m x 200 m duri ng a one week period in May 2008. With a few exceptions i ndividuals were from intermediate size class es for each species (18.3 Tristachya 9 Loudetiopsis based on a preliminary random sample of individuals in the study area ) I t hen delineated plot s of 50 x 50 cm around each focal individual Within each plot I clipped a ll above ground biomass surrounding the focal individual ; T his left all roots and other below ground biomass as well as leaf litter undisturbed. I recorded the s pecies identity and total biomass (after drying for 48 hours at 60 degrees C) of all collected biomass. I cl i pped re growth at 2 3 week intervals throughout the experiment to reduce aboveground competition.

PAGE 19

19 Treatments I randomly assigned treatments to individuals for a total of N = 20 plants per species per treatment ( i.e., control, nitrogen, water, nitrogen x water). I added nitrogen in the form of ammonium nitrate (NH4NO3) fertilizer a pplied at the rate of 25 kg ha1 yr1 nitrogen (2.5 g m1 y r1 N ) divi ded into 4 applications (June 2008, Sept. 2008, Dec. 2008, Feb. 2009) to simulate atmospheric de position throughout the year. Previous work estimated a total N deposition rate of 9.5 kg ha1 yr1 N near Uberlndia ( between 19971999) with the w et deposition component made up of 48% ammonium and 38% nitrate Given development trends in the region, including urbanization and intensifying agricultural production (Cavalcanti & Joly, 2002) it is unlikely that deposition rates decreased in the ten year period between th e estimate and my study I added water in the dry season in amounts consistent with climate models th at predict increasing rainfall in the Cerrado region (Lilienfein & Wilcke, 2004) Thes e models predict that rainfall will continue to increase, following the trend observed over the last 40 years in the Cerrado (Haylock et al. 2006) In my water addition I applied two liters of water over 24 hours (8 mm/day) with drip irrigation. These water addition pulses were divided by alternate dry periods of 2 and 9 days between June and August. This is about half of the a verage rainfall per 24 hours on days with rain, during the wet season (Oct ober April 14.2 mm/day or 3.55 liters over a 50 x 50 cm plot area ; data collected at field station, 20032004, unpublished). However, since dry season rainfall averages just 4.5 mm/month (May Sept.) the addition represented a substanti al rainfall increase for this part of the year (data collected at field station, 20032004, unpublished). In total I added 72 mm of water or approximately 5% of the average annual rainfall in addition to the ambient precipitation 9.5 mm between June Aug ust 2008 (data from Uberlndia, 40 km from site, 20042008, Instituto de Geografia, 2008)

PAGE 20

20 I measured volumetric water content with Soil Moisture Smart Sensors ( Onset Computer Corp., Bourne, MA, USA ) in watered, control, and nitrogen treatments from July 2008 to August of 2008. In total I monitored soil moisture during 6 watering treatments with 4 or 5 sensors in different plots (varied by treatment) to confirm the efficacy of my water addition Dataloggers recorded data every 5 minutes and were moved approximately every 10 days to new s ites. Average, maximum, and minimum daily soil moisture was significantly different (p < 0.0001 for minimum, maximum, and average, F value: avg .: 66.10, m ax .: 105.17, m in .: 32.82) between watered and unwatered plots during this period (general linear model with day as a random factor ). Average daily average volumetric water content (m3/m3) in watered plots was 6 times water content in unwatered plants during the study period. Light Availability Because light availability can mediate plant and soil responses to my simulated global change factors (Cruz, 1997, Ludwig et al. 2001) I measured p hotosynthetic ally acti ve radiation (PAR) with a quantum line sensor ceptometer (Accupar LP80, Decagon Devices, Pullman WA) on a cloudless day between 11am and 2pm on Mar. 3, 2009. I averaged three 80 cm measurements taken by bisecting the plot at different points at approximately the height of the tallest leaves of each in dividual Plant R esponse I quantified reproductive output by measuring number of flowering tiller s and number of florets per flowering tiller for all individuals before (May 2008) and afte r the treatments were applied (March 2009). I estimated senescence b y counting all green leaves on the plant and dividing the number by genet area in August of 2008. A larger value of this index corresponded to lower leaf senescence during the dry season. Growth was quantified by measuring the

PAGE 21

21 diameter a round the base of e ach genet (individual bunchgrass ) before and after the treatments (10 month period) to quantify growth. To control for the effects of original size and shading on growth, I modeled the diameter change in response to treatments with the original diameter and average PAR as covariates in the analyses. Bunch diameter to was used to quantify growth because diameter significantly correlates with total biomass for both species ( Loudetiopsis : R2 = 0.23, p = 0.003, Tristachya : R2 = 0.25, p = 0.014). I collect ed all plants at peak biomass ( the end of wet season) over a 4 week period in March April 2009. I randomized plant collections by treatment and species to avoid any systematic ef fect s of time on biomass estimation over the collection period. I trench ed around the perimete r of the plot and excavated the plant to a depth of approximately 15 cm for Loudetiopsis and 20 cm for Tristachya I attempted to recover as much of the root biomass connected to the plant as possible. To quantify remaining root biomass I collected fiv e 6 cm deep cores in a line bisecting the soil remaining in the plot area afte r plant removal. I sieved the cores for roots with a 2 mm sieve and dried the material at 55 (until reaching constant weight ) Root collections may have included grass roots of other neighboring grass individuals. Remaining coarse root mass averaged 0.0015 g/ cm2 ( se + 0.0009), while average dr y root mass per plant was 62.33 g ( se + 38.28) for Loudetiopsis and 399.69 g ( se + 255.83) for Tristachya I separated plants into dead vegetative tillers live vegetative tillers and flowering tillers I also sorted all biomass into live leaves, dead leaves, flowering tiller s, flowering parts (florets and seeds), and roots and dried t he material at 55 re ached constant weight ). I calculated r oot :shoot ratios by dividing the total aboveground biomass including dead leaves and reproductive parts by the root biomass.

PAGE 22

22 Soil and Foliar Nutrients I measured p lant available nitrogen by extracting ammonium and nitrate from mixed bed resin bags buried in the top 10 c m of 76 plots (32 for Loudetiopsis 44 for Tristachya ) in February March 2009. I first charged resins with 1 M NaCl and extracted with 2 M KCl after 28 days of incubation in the plots I also collected soil samples from the first 10 cm of the soil (March 2009), homogenized the material, and extracted 10g of fieldmoist soil with 2 M KCl within 24 hours of collection. We incubated 10 g of the same soil sample for 7 days at room temperature in the lab and extracted with 2 M KCl to quantify mineralization rates. We measured gravimetric water content in subsamples of field moist soil by drying them for 48 hours at 110 Soil and resin extracts were analyzed colorimetrically with an Astoria Autoanalyzer (Astoria Pacific, Inc. Clackamas, OR, USA) for nitrate and ammonium concentration. Soil nitrogen values were adjusted for bulk density and gravimetric soil moisture. I dri ed separate soil samples at 55 C for 48 hours for pH and macronutrient analysis pH was measured in deionized water (ratio: 1:2.5 soil:H2O) I extracted p otassium and phosphorous with Mehlich (HCl H2SO4) solution (K ratio: sample:solution 10:1, P sample:solution 20:1) and analyzed nutrient concentrations with flame emission spectrometry for potassium (B 462, Micronal, So Paulo, SP, Brazil ) and atomi c absorption spectrometry for phosphorus (Cary 50 Conc UV Vis Varian Inc., Palo Alto, CA, USA ). I used 1 M KCl to extract aluminum, calcium, and magnesium (Ca & Mg ratio sample:solution 100:1, Al ratio: sample:solution 10:1) I analyzed aluminum content by titration with NaOH in the presence of bromothymol blue. I a nalyzed calcium and magnesium with flame emission spectrometry (GBC Scientific Equipment 932 A Dandenong, VIC, Australia )

PAGE 23

23 I washed a subsample of green undamaged leaves from each individual with deionized water, dried the material for 48 hours at 60 C, and ground the samples in a plant mill (Marconi Equipamentos, MA 048, Piricicaba, SP, Brazil ) I digested tissue samples with the Kjeldahl method, steam distilled the digest into boric acid, and titrated wit h sulfuric acid. For foliar phosphorus, I digest ed tissue in nitric and perchloric acid and analyzed solution P concentration with atomic absorption spectrometry (Cary 50 Conc UV Vis Varian Inc., Palo Alto, CA, USA ). Statistical Analysis R eproductive Variables R eproductive output variables were analyzed separately for each species I analyzed the lik e lihood an individual would flower with my treatments with a binomial model. The response variable was f lowering or nonflowering state while nitrogen and water were fixed effects. I conducted the same analysis in a separate regression model, with resin available nitrate and ammonium as fixed effects, to evaluate the direct effects of nitrogen on flowering. I analyzed treatment effects on the n umber of flowering tiller s per flowering individual and number of total spikelets with a negative binomial generalized linear model with nitrogen and water as fixed effects. I used a general linear model to evaluate treatment effects on number of spikelets per tiller. G rowth and Foliar Nutrient Variables I calculated the difference in di ameter between year one and year two and analyzed the effects of the treatments with a general linear model with nitr ogen water original diameter, and average PAR as fixed effect s I analyzed treatment effects on senescence (numbe r of green leaves by area ) with an ANCOVA with original diameter as the continuous variable.

PAGE 24

24 To explore the effects of treatments on root:shoot ratio, I evaluated the response of root:shoot ratios treatments with a gamma distribution in a generalized line ar model. I used the gamma distribution because goodness of fit tests and analysis of residual deviance showed that it improved model fit over the normal distribution and models with transformed variables. I analyzed the effects of my treatments on total a boveground biomass, total dead leaf biomass, and total live leaf biomass with general linear models. I evaluated change in foliar nitrogen and phosphorus concentrations and nitrogen: phosphorus (N:P) ratios to treatments with an ANCOVA with live leaf mass as the continuous covariate. All plant response variables were analyzed with separate models for each species. Soil Variables I tested the effects of treatments and species on pH, nitrogen, potassium, calcium, and aluminum concentrations with general linea r models with nitrogen, water, and species as fixed effects. I used a generalized linear model with the gamma distribution to evaluate treatment effects and species effects on soil phosphorus resin available ammonium and nitrate and ammonium mineralization rates Because species significantly interacted with the water treatment effects on soil phosphorus I included that interaction into the model. I tested treatment effects on resin available nitrate, soil ammo nium and nitrate and nitrate and ammonium m ineralization rates with a general linear model. M odel A ssumptions and V ariable T ransformations I tested for homogeneity of variance with Levenes test and applied l og transformations when necessary to meet the assumption of normally distributed resi duals for general linear models. For all transformations and nonnormal distributions (gamma binomial, and negative binomial ) I reported back transformed means and standard errors or 95% c onfidence intervals I

PAGE 25

25 used D unnetts test for comparison of treatment means to control values All analys e s w ere performed using SAS 9.2 (SAS Institute Inc., Cary, NC USA).

PAGE 26

26 CHAPTER 3 RESULTS Reproductive Response There was no statistically significant nitrogen treatment or water treatment effect on the likelihood that an individual would flower with nitrogen or water for Loudetiopsis However, there was a significant positive interactive effect of the combination o f nitrogen and water on plant flowering for this species (95% of individuals flowered with nitrogen x water treatment, 89% for nitrogen, and 65% for water Table 3 1, Fig. 3 1). However, when these values were compared to controls, these differences were not significant (Dunnetts test, Table 3 1). With Tristachya w ater alone increased the percentage of flowering Tristachya individuals from 45% to 70% ( Table 3 1, Fig. 3 1). Contrary to expectations, higher resin available ammonium and nitrate values decreased the likelihood that a Tristachya individual would flower. There was no significant effect of resin available nitrogen on the flowering likelihood for Loudetiopsis (Table 32) While only the interaction of water and nitrogen treatments increased the likelihood that an individual Loudetiopsis would flower, those individuals that did flower produced significantly more flowering tillers in response to both water alone and the combined water and nit rogen treatment Individuals of a ll three treatments produce d significantly more flowering tillers than controls ( Table 3 3, Fig. 3 2). Flowering Tristachya individuals did not produce more flow ering tiller s with treatments or in contrast to control values ( Table 3 3, Fig. 3 2 ). For the plants that did flower, I tested whether the nitrogen and water treatment s affected reproductive output defined as the total number of spikelets produced. The water treatment for Loudetiopsis was the only species and treatment combination w ith a significa nt effect on total spikelet number. However, the total number of spikelets for Loudetiopsis for w ater ed plants

PAGE 27

27 (mean=57.15) and plants receiving the nitrogen and water treatment ( mean=60.31) was significantly greater when compared to control plant values ( Dunnetts test Water: p=0.09, Nitrogen: p=0.03, Table 3 4) None of the treatments were significant for Tristachya ( Table 3 4). The number of spikelets per flowering tiller marginally increased with nitrogen addition for Loudetiopsis There was n o significant change in number of spikel ets per tiller with either water or the water and nitrogen treatment in contrast to this species response to water with other reproductive variables There was no significant difference in spikelets per flowering tiller with treatments for Tristachya ( Table 3 5). Growth, Allocation and Foliar Nutrients Tristachya bunchgrass diameter increased significantly while Loudetiopsis diameter growth decreased w ith added nitrogen ( Table 3 6, Fig. 3 3). For both species t he larger the individual ( the greater the original diameter ) the lower the gr owth rate ( decreased diameter difference). Greater photosynthetic active radiation (PAR) or less shading, correlated with high er growth rate for both species ( Table 3 6) B oth wa ter and nitrogen significantly reduced root:shoot ratios for Loudetiopsis but only plants in the combined treatment (water x nitrogen) had significantly lower ratios than control plants ( Table 3 7). There was no significant diffe rence with any treatment f or Tristachya The significant de crease of Loudetiopsis root:shoot ratio s with water and nitrogen seemed to suggest increased aboveground growth while the decreasing effects of nitrogen on diameter suggested lower growth rates. To test for treatment effect s on the general relationship between aboveground biomass and diameter with this species I analyzed the effects of treatments on the amount of aboveground biomass by diameter, the amount of dead biomass by diameter, and the amount of live leaf bioma ss by diameter (Table 3 8) Loudetiopsis had signi fi cantly more

PAGE 28

28 aboveground biomass by diameter (p=0.03) with the water treatment. This was probably due to th e significant increase in dead aboveground material by diameter (p=0.01) with water addition In o ther words, the lower diameter growth rate for Loudetiopsis with the water treatment was combined with a significant decrease in root:shoot ratio with water addition due to increased dead aboveground biomass with the water treatment. There were no si gnific ant treatment effects on biomass by diameter for Tristachya Loudetiopsis individuals had significantly more green leaves per bunch area by August of the dry season with the water treatment ( Table 3 9 ). Because of the water e ffect on dead aboveground biomass for Loudetiopsis I tested whether or not the number of senesced leaves might correlate with the dead aboveground biomass/diameter variable. There was a significant positive linear relationship ( p=0.04, F value = 4.34, R2 = 0.05) between the number of green leaves per area and the amount of dead biomass at the end of the experiment with Loudetiopsis but no hint of a significant relationship for Tristachya (p = 0.99, F value <0.01, R2 = < 0.01). None of the treatments had significant effects on number of green leaves per area for Tristachya With increasing d iameter both species had significant ly fewer green leaves per area ( Table 3 9 ) N itrogen concentration and N:P ratios for Loudetiopsis decreased with increasing live leaf biomass (which was not significantly affected by treatments) ( Table 3 10 & Table 3 12). Tristachya nutrient concentrations did not vary significantly with live leaf biomass. Treatments did not significantly affect n itrog en concentration for either species, contrary to my prediction that nitrogen addition would increase foliar nitrogen content ( Table 3 10, Fig. 34) Phosphorus concentration marginally decreased with nitrogen addition for Loudetiopsis when compared with co ntrol values. However, phosphorus concentration s increased with both nitrogen and water treatments for Tristachya and their combination resulted in a significant difference compared to

PAGE 29

29 controls ( Table 3 11, Fig. 35) Nitrogen addition significantly incre ased nitrogen: phosphorus ratios (N:P) for Loudetiopsis ( Table 3 12) Tristachya N:P ratios with the combined nitrog en and water treatment were significantly lower than control values (Dunnetts test, Table 312) W ater rather than ni trogen, or an interactive effect, appeared to be driving this result as it was the only factor significant in the model ( Table 3 12 ). Soil Response Plant available soil phosphorous significantly increased with nitrogen addition and differed by species (Table 3 13) The water treatment significantly affected phosphorus concentrations for both species, but these effects were in opposite directions (Table 3 14) In Loudetiopsis plots, water addition lowered soil phosphorous concentrations by about 10% (p=0.07, Table 3 14, Fig. 36) However, Loudetiopsis plots with nitrogen addition marginally increased phosphorous concentrations ( Table 3 14, Fig. 3 6) For Tristachya nitrogen and combined water and nitrogen treatments significantly increased soil phosphorous concentra tions, but only the combined nitrogen and water values signif icantly differed from controls (Dunnetts contrasts, p = 0.03, Table 3 14). Though I expected that nitrogen addition might change pH or aluminum mobility, my data did not support this hypothesis. However, potassium availability marginally decreased and calcium marginally increased with nitrogen addition. None of the soil factors changed with water addition, and there were no significant differences for pH, Al, Ca, or K between control and treatment values ( Table 3 15). Resin extracted nitrate and ammonium concentrations significantly increased with water addition, but not with nitrogen addition. Species did not significantly affect either resin available nitrogen form ( Table 3 16, ammonium, Table 3 17, nitrate). The mineralization rate for nitrate and ammonium did not vary by treatment or species

PAGE 30

30 (Table 3 18 ammonium, Table 319 nitrate). Soil ammonium did not vary by treatment or species, but soil nitrate marginally decreased with water addition (Table 320 ammonium, Table 321, nitrate).

PAGE 31

31 CHAPTER 4 DISCUSSION This research demonstrates that global change factors have the potential to alter growth and reproduction of two codominant Cerrado bunchgrasses, and that these eff ects vary by species (Table 4 1) As both focal species are C 4 grasses the divergence in species response I observed is not due to known functional group differences in nutrient demand (Craine et al. 2002) The data did not uniformly support the original expect ation that added water and nitrogen would result in increase d vege tative and reproductive biomass. Reproductive Response In general, Tristachya reproductive variables were unresponsive to treatments, whereas Loudetiopsis altered a suite of flowering traits in response to water and nitrogen addition. Contrary to the fertilization hypothesis, increased nitrate and ammonium were not correlated with flowering of Loudetiopsis and decreased flowering of Tristachya While the experiment was not designed to determine how ammonium and nitrate levels directly affect plant phenology, the data suggest that additional nitrate and ammonium will not necessarily increase grass flowering as they do in some temperate grasslands (Silletti et al. 2004) Though the results suggest that the combination of nitrogen deposition and precipitation could increase flowering with these t wo species, the lack of positive correlation with resin available nitrogen concentration does not confirm a fertilization effect of nitrogen addition. However, since water addition was associated with increased resin available nitrogen, water addition might have increased flowering by increasing microbial nitrogen mineralization. Conclusions based on these results are qualified by the large reduction in flowering between year one (before the fertilization began) and year two of the study for all plants, inc luding controls In the reproductive season prior to the experiment (Feb. March. 2008) 94 %

PAGE 32

32 of Tristachya individuals flower ed while only 58% flowered in 2009. Loudetiopsis phenology followed the same pattern, diminishing fro m 83 % to 58% flowering individuals between 2008 and 2009. This effect was not due to the experimental setup ; plants in the study area outside of the experiment were as likely to flower and had on average the same numbers of flowering tillers as control plants. This highly signif icant difference in years suggests that both species are capable of higher flowering rates then recorded during the experiment It is possible that a fire that in the study area in 2006 is related to the higher reproduction I observed in 2008. South Americ an savanna grass species as a group tend to respond to fire by increasing flowering though dependence on fire for reproduction varies by species (Baruch & Bilbao, 1999, Sarmiento, 1992) If so, my results show that variability in annual climate and nitroge n deposition can increase reproductive out put, but may pale in comparison to the effects of other environmental changes like fire Growth R esponse While both water and nitrogen affected the species reproductive output, only nitrogen significantly affe ct ed diameter growth. H owever, the nitrogen addition had opposite effects on the two species Loudetiopsis grew less with the simulated nitrogen deposition treatment, whereas Tristachya grew more a res ponse consis tent with nitrogen induced tiller production for bunchgrasses in general (Tomlinson & O'Connor, 2004) Water did not seem to increase genet diameter growth, contrary to predictions The result is especially puzzling with Loudetiopsis because water decreased the root:shoot ratio and reduced leaf senescence. Loudetiopsis genets increased aboveground biomass by diameter with water, but the increase was primarily due to an increase in dead tissue This response correlated with this species decrease in senescence with water addition I t is possible that the additional water resulted in either decreased leaf life span or

PAGE 33

33 increased leaf production during the dry season, which did not translate into increased live biomass or diameter by the end of the study T he responses I observed diverge from results in to similar experiment s in other ecosystems. Research focusing on annual grass species in California found that more grass species responded to nitrogen than precipitation change (Zavaleta et al. 2003) While biomass generally increased with water addition, nitrogen led to even higher values, and the combined water and nitrogen treatment led to the highest biomass accumulation (Zavaleta et al. 2003) In this experiment, n itrogen affected both species, but only one species increased growth ( Tristachya ) whereas the other increased reproduction ( Loudetiopsis ) The species that decreased growth with nitrogen ( Loudetiopsis ) increased dead biomass accumulation instead Finally, I did not obs erve any significant interactive effects of nitrogen and water on growth variables for either species Leaf S enescence Species drought tolerance traits probably influenced the differential leaf senescence response to water addition in the study Even in control plants, the two species differed in senescence patterns : Tristachya maintained a higher density of green leaves in the dry season than Loudetiopsis in control plants ( Copeland, unpublished data). Neither species completely senesced in the dry season, a typical behavior for Neotropical perennial savanna grasses (Sarmiento, 1992) The species differences in senescence may be r elated to root characteristics as Tristachya had deeper and larger roots than Loudetiopsis The complementary nature of root structure between these two co dominant grass species mirrors a pattern found elsewhere (Fargione & Tilman, 2005) that al lows co occurring perennial grasses to exploit different nutrient and water sources

PAGE 34

34 A comparable set of experiments also found dramatic differences in species growth response between the response s of two codominant perennial C 4 grasses in the Great Plains to altered precipitation (Silletti & Knapp, 2001, Silletti et al. 2004) They identified a s uite of traits, including decreased leaf senescence and higher root:shoot ratio, which confer red greater drought resistance on the dominant species ( Andropogon gerardii ) (Swemmer et al. 2006) These same traits may explain the lack of response of Tristachya to water addi tion, as this species also demonstrated lower leaf senescence in the dry season and higher root:shoot ratio. Foliar Nitrogen and P hosphorus F oliar nutrient concentrations changed in response to the treatments, but these responses di ffered by species Both species ha d generally low nitrogen: phosphorus ratios ( control means Loudetiopsis : 9.43 ( + 1.16), Tristachya : 9.74 ( + 1.17) ) which could indicate nitrogen limitation (Tessier & Raynal, 2003) However, r ecen t examin ations of N:P ratios in tropical forests (Townsend et al. 2007) and grasslands (Craine et al. 2008) suggest that N:P ratios may not predict absolute limitation In one study in a s early successional tropical forest, tree species had higher nitrogen: phosphorus ratios (13:1 and 15:1) then the grass species (9:1) yet the tree species increased both biomass and foliar nitrogen concentrations with added nitrogen, whereas neither factor increased with the grass species (Davidson et al. 2004) In t his experiment one species, Loudetiopsis had significantly higher N:P ratios w ith nitrogen whereas Tristachya had lower N:P ratios in the presence of water. The measurements of phosphorus and nitrogen c oncentrations in response to treatments suggest that foliar phosphorus, rather than nitrogen, drove N:P changes in both species I had predicted that foliar nitrogen not phosphorus, would change in response to my treatments via the dir ect effects of nitrogen fertilization on plant available nitrogen or indirect

PAGE 35

35 stimulation of microbial nitrogen mineralization with dry season water addition What might explain change in foliar phosphorus in lieu of nitrogen? Plant N:P ratios vary based on a wide variety of characteristics, including functional group. It may be that Neotropical grass species in general simply maintain lower leaf N:P ratios relative to other savanna plant species (such as trees) a suggestion borne out by data from a similar study in the neotropics (Davidson et al. 2004) The same pattern may exist in the preserve where this study took place. The average nitrogen concentration for 93 tree species from Estao Ecolgica do Panga was 25 g kg1 with phosphorus concentrations of 2 g kg 1 on average (Hardisan, 2005) The tree species average N:P ratio of 12.5:1 is higher than both of the grass species ratios (8.5 10.5 N:P) suggesting that the C 4 grass species N: P ratios may be lower than those of other functional groups in the study area Species diff erences in foliar nutrient changes i n response to treatments could be due to a variety of species traits other than functional group identity While m y data are insufficient to explore all possible characteristics that could be responsible for differences in foliar chemistry, they do suggest do suggest some traits correlated with the foliar responses to treatments Loudetiopsis increased N:P ratios with increasing live leaf biomass suggesting nitrogen limitation, whereas the significant downward trend in N :P ratios with increasing live leaf tissue suggest that the species was capable of diluting nitrogen in leaf tissue with increased living biomass The significant decrease in N:P for Tristachya is harder to explain since t he decrease was not associated wit h any of the species few growth and reproductive responses to water and nitrogen. Another grassland study observed that water addition simultaneously increased foliar phosphorus while reducing phosphatase activity (Menge & Field, 2007) and suggested that water might stimulate phosphorus mineralization in tandem with increased nitrogen mineralization This mechanism still does not explain the differences between species in this experiment

PAGE 36

36 however some combination of plant capacity to increase phosphotase activity with nitrogen treatments or increased uptake of mineralized phosphorus could have led to the increased foliar phosphorus concentrations in Tristachya Since neither species significantly increased foliar nitrogen, the experiment did not support the proposed nitrogen fertilization effect with nitrogen deposition and increased precipitation. My alternative limitation hypothesis that nitrogen could decrease growth by decreasing phosphorus availability, was not supported either, as one species decreased phosphorus concentrations with increasing live leaf biomass, whereas the other increased foliar phosphorus concentration Soils Response The suggestion that the two perennial grass species differed in their phosphorus use and uptake is consistent with the soils data. P hosphorous increased with water in soil under Tristachya and decreased in s oil under Loudetiopsis Irrespective of species, soil phosphorus concentrations increased with added nitrogen. Tristachya maintained higher soil phosphorus concentrations than Loudetiopsis for all treatments including controls. This evidence combine d with the foliar data suggest that Loudetiopsis may be limited by phosphorus when water and water and nitrogen are added I did not expect significant increases in soil phosphorus with added water or the divergence in species response with either the Fer tilization or Limitation H ypothesis. Even though water addition decreased Loudetiopsis soil phosphorus, the decrease was only back to contr ol levels. This result did not support my Limitation H ypothesis which predicted that phosphorus and other macronutri ent s would decreas e with added nitrogen and remain unchanged with added water The only significant change in r esin available nitrogen forms, both ammonium and nitrogen, was the increase in resin nitrate and the decrease in soil nitrate with added water The

PAGE 37

37 increase with water could be due to increased mineralization with added water during the dry season, as observed in other savanna ecosystems (Augustine & McNaughton, 2004) It is possible that the decrease in soil nitrate is related to increased plant uptake of soil nitrogen or leaching. In aggregate, the increase in total availabl e nitrate for 28 days with a decrease in available nitrogen at one time point suggests increased nitrate turnover in soil with added water. Though I had predicted that soil nitrate and ammonium would increase with added nitrogen, but the results do not support this expectation It is possible that plant uptake, soil immobilization, gaseous losses, or some other me chanism explain the lack of significant change in available nitrogen with fertilization Other soil nutrient factors did not seem as sens itive to simulated global change factors as phosphorus and nitrogen. As predicted in the Limitation H ypothesis, pH decreased and extractable aluminum increased with added nitrogen, however these differences were not significant. Calcium increased margin ally and potassium decreased marginally with my simulated nitrogen deposition. These changes do not particularly support the theory of general cation de pletion predicted for phosphorus limited tropical ecosystems (Matson et al. 1999) The fact that I obs erved changes in pH, Al, and cation concentrations in the extremely short time per iod of this study suggest s that significant change in these soil factors with chronic nitrogen deposition is possible However, there w as no evidence that soil macronutrient changes result ed i n decreased plant productivity The decreased soil pH, increased Al, increased Ca, and decreased K, did not translate into decreased plant growth or allocation in general, though Loudetiopsis did gro w less with added nitrogen. That fact that phosphorus did not decreas e along with pH, as predicted, might explain the lack of support for my Limitation H ypothesis f or plant growth and allocation.

PAGE 38

38 CHAPTER 5 CONCLUSION The results from this experiment indicate that nitrogen deposition and precipitation change have the potenti al to alter the growth and reproduction of two co dominant grass species, and via plant effects on soils, available phosphorus and nitrogen. Since this experiment was conducted with mature native individuals at realistic levels of nitrogen and water addition, the results should accurately represent the short term effects of these global changes on native species. This evidence also suggests the possibility of future cation depletion and increasing acidity, both factors which could eventually limit plant gro wth and reproduction. M ost of the species responses occurred with precipitation change rather than nitrogen addition, but this outcome could have been affected by the high levels of nitrogen deposition at the study site (Lilienfein & Wilcke, 2004) T he increase of one species N:P ratio following nitrogen addition suggests that at leas t one species is still limited by nitrogen. Soil phosphorus concentration also seems to be related to water addition under the influence of one of the grass species, suggesting that plant phosphorus ac quisition could be affected by precipitation change T he paradigm that tropical species on highly weathered soils are primarily limited by phosphorus (Vitousek, 1984) cannot be proven or disprove n with this experiment, but it is suggestive that the responses observed were in foliar phosphorus rather than nitrogen, yet growth response to nitrogen addition was also recorded. This corresponds with a body of research that qualifies Neotropical ecosystem level nutrient limitation by phosphorus or n itrogen by factors like successional stage (Davidson et al. 2004) and fir e (Nardoto et al. 2006) The similarities between water treatment responses of C 4 grass species here and C 4 gras s species in the temperate zone (Fargione & Tilman, 2005, Fay et al. 2003, Silletti et al. 2004) suggest that species drought tolerance traits may determine response to altered precipitation. However, in

PAGE 39

39 general, the responses to treatments we observed for foliar and growth variables did not greatly resemble biomass and flowering responses observed in temperate global change grassland experiments (Siemann et al. 2007, Silletti & Knapp, 2001, Zavaleta, 2002) Water and nitrogen drove different plant and soil responses T heir combination had simple additive effects on some variables, and interacti ve effects on others Overall my results imply that we cannot predict the responses of species to global change if the effects of nitrogen deposition and precipitation change are addressed independently The abbreviated nature of this study suggests a number of avenues for future research. Future global change research in Neotropical savannas should address the interactive effects of fire and herbivory on water and nitroge n deposition responses, given their known impact on nitrogen cycling and productivity. The relationship of both the water and nitrogen cycles to the phosphorus cycles needs to be examined, particularly from the standpoint of Cerrado plant species which ma y have differential responses to global change driven phosphorus limitation.

PAGE 40

40 Figure 1 1. Nitrogen deposition responses in nitrogen limited ecosystems (adapted from Aber et al., 1998) Chronic Nitrogen Deposition Plant Soil 3 + Leaching Ca2+ Mg2+ Na+ K+ NO3 NOx NHy Plant Soil 3 + Nitrification NO NOx NHy Initial State Final State

PAGE 41

41 Figure 1 2. Nitrogen deposition responses in phosphorus limited ecosystems (adapted from Matson et al. 1999) Chronic Nitrogen Deposition Leaching Ca2+ Mg2+ Na+ K+ NO3 Plant Biomass P Soil 3 + Nitrification NO NOx NHy

PAGE 42

42 Fig ure 2 1. Map of Estao Ecolgica do Pang a and location of the study area. Legend vegetation physiognomies refer to standard vegetation classifications in Cerrado literature without English translation (Oliveira & Marquis, 2002, Ottmar et al. 2001)

PAGE 43

43 A) B) C) Figure 2 2. Photos of focal species and habitat A) T ristachya leiostachya B) Loudetiopsis chrysothrix C) Species habitat, campo sujo vegetation physiognomy near plots.

PAGE 44

44 Figure 3-1. Flowering lik elihood by treatment for Loudetiopsis and Tristachya Bars represent 95% confidence intervals. 0.00.20.40.60.81.0 LoudetiopsisFlowering Likelihood (+/95% CI) ControlNitrogenWaterWater x Nitrogen 0.00.20.40.60.81.0 Tristachya ControlNitrogenWaterWater x Nitrogen

PAGE 45

45 Figure 3-2. Number of flower ing tillers by treatment for Loudetiopsis and Tristachya Significant differences from controls (Dunnetts test, Table 3-3) are indicated by for p < 0.10, ** for p < 0.05. ControlNitrogenWaterWater x Nitrogen 0 510152025LoudetiopsisNo. Flowering Tillers ControlNitrogenWaterWater x Nitrogen 0 510152025Tristachya

PAGE 46

46 Figure 3-3. Diameter diffe rence by treatment for Loudetiopsis and Tristachya No significant differences between controls and treatments for either species. Reported values are adjusted by covariates in model (Avg. PAR & original diameter). 02468 LoudetiopsisDiameter Difference (cm) ControlNitrogenWaterWater Nitrogen 02468 Tristachya ControlNitrogenWaterWater Nitrogen

PAGE 47

47 Figure 3-4. Foliar nitrogen by treatment for Loudetiopsis and Tristachya No significant differences between treatments for either species. Reported valu es are raw values not adjusted by live leaf bioma ss (covariate in ANCOVA model). ControlNitrogenWaterWater x Nitrogen 567891011LoudetiopsisNitrogen (g/Kg) ControlNitrogenWaterWater x Nitrogen 567891011Tristachya

PAGE 48

48 Figure 3-5. Foliar phosphorus by treatment for Loudetiopsis and Tristachya Significant differences from controls (Dunnetts test, Table 3-11) are indicated by for p < 0.10, ** for p < 0.05. Reported values are raw values not adjusted for live leaf biomass (covariate in ANCOVA model). ControlNitrogenWaterWater x Nitrogen 0.2 0.4 0.6 0.8 1.0 1.2LoudetiopsisPhosphorus (g/Kg) ControlNitrogenWaterWater x Nitrogen 0.2 0.4 0.6 0.8 1.0 1.2Tristachya

PAGE 49

49 Figure 3-6. Soil phosphorus (g/Kg) by treatment for Loudetiopsis and Tristachya. Significant differences from controls (Dunnetts test, Table 3-14) are indicated by for p < 0.10, ** for p < 0.05. Bars represent 95% confidence intervals. 0 00 10 20 30 4 LoudetiopsisSoil P g/Kg (+/95% CI) ControlNitrogenWaterWater Nitrogen 0 00 10 20 30 4 Tristachya ControlNitrogenWaterWater Nitrogen

PAGE 50

50 Table 3 1. Flowering probability in response t o water and nitrogen by species Significance values are from logistic model with two fixed effects, nitrogen and water, and the interaction. Means are probabilities of flowering, for plots with nitrogen or water and plot s with water and nitrogen. Model fit: AIC = 76.20, n = 80 ( Loudetiopsis ) ; AIC = 111.92, n = 79 ( Tristachya). Arrows indicate direction of difference from control values Species Treatment df 2 p Lower Confidence Interval (95%) Mean (Probability) Upper Confidence Interval (95%) Dunnetts test Loudetiopsis Nitrogen 1 1.0715 0.30 0.73 0.89 0.97 0.70 Water 1 0.0002 0.99 0.66 0.86 0.95 0.18 Nitrogen x Water 1 4.5955 0.03 0.71 0.95 0.99 0.87 Tristachya Nitrogen 1 0.000 1.00 0.42 0.58 0.73 1.00 Water 1 4.988 0.03 0.54 0.70 0.82 0.28 Nitrogen x Water 1 0.000 1.00 0.47 0.70 0.86 0.28 Table 3 2. Flowering likelihood in response to resin available nitrate and ammonium by species Significance values are from a logistic model with resin nitrate and resin ammonium as fixed effects, without interaction. Model fit: AIC = 29.74, n = 32 ( Loudetiopsis ) ; AIC = 61.53, n = 44 ( Tristachya ). Arrows indicate direction of effect on flowering probability. Species Treatment df 2 p Loudetiopsis Resin Nitrate 1 0.1840 0.67 Resin Ammonium 1 0.7637 0.38 Tristachya Resin Nitrate 1 2.5481 0.11 Resin Ammonium 1 5.436 0.02

PAGE 51

51 Table 3 3. Number of flowering tillers per flowering individual by treatment and species Significance values are from a generalized linear model with negative binomial response variable distribution with nitrogen, water, and their interaction as fixed effects. Means are probabilities of flowering, for plots with nitrogen or water and plot s with water and nitrogen. Model fit: AIC = 364.49, 2 = 69.12 /df = 1.11, n = 66 ( Loudetiopsis ) ; AIC = 137.02, 2 = 27.12 /df = 0.65, n = 46 ( Tristachya ). Arrows indicate direction of differenc e from control values Species Treatment df F value p Mean ( + se) Dunnetts test Loudetiopsis Nitrogen 62 1.41 0.24 6.67 ( + 0.79) 0.08 Water 62 5.07 0.03 7.33 ( + 0.90) 0.02 Nitrogen x Water 62 3.85 0.05 6.10 ( + 1.89) 0.04 Tristachya Nitrogen 42 0.77 0.38 1.42 ( + 0.26) 0.85 Water 42 <0.00 0.97 1.59 ( + 0.24) 0.54 Nitrogen x Water 42 0.81 0.37 1.59 ( + 0.34) 0.86 Table 3 4. Total number of spikelets by treatment and species. Statistics are from a generalized model with negative binomial response variable distribution with nitrogen, water, and water x nitrogen as fixed effects. Means are mean number of total spikelets for plots with nitrogen or water and plots wit h water and nitrogen. Model fit AIC = 644.87, 2 = 76.56 /df = 1.23, n = 66, ( Loudetiopsis ) ; AIC = 376.61, 2 = 61.39/df = 1.57, n =4 6 ( Tristachya ). Arrows indicate direction of effect relative to controls. Species Treatment df F value p Mean ( + se) Dunnetts test Loudetiopsis Nitrogen 62 1.71 0.20 53.03 ( + 8.39) 0.03 Water 62 4.77 0.03 58.71 ( + 9.84) 0.09 Nitrogen x Water 62 1.16 0.29 60.32 ( + 12.89) 0.01 Tristachya Nitrogen 39 0.01 0.92 26.89 ( + 4.73) 0.97 Water 39 0.91 0.35 29.96 ( + 4.87) 0.60 Nitrogen x Water 39 2.14 0.15 35.58 ( + 8.31) 0.87

PAGE 52

52 Table 3 5. Number of spikelets per flowering tiller by treatment and species Statistics are from a general linear model with nitrogen, water, and nitrogen x water as fixed effects. Means are the average number of spikelets per tiller of flowering, for all plots with nitrog en or water and plots with water and nitrogen. Model fit: R2 = 0.07, pvalue = 0.24, n = 66 ( Loudetiopsis ) ; R2 = 0.05, pvalue = 0.51, n = 4 3 ( Tristachya ). Arrows indicate direction of difference related to controls. Species Treatment df F value P Mean ( + se) Dunnetts test Loudetiopsis Nitrogen 1 3.32 0.07 7.01 ( + 1.08) 0.77 Water 1 0.11 0.74 6.11 ( + 1.08) 0.65 Nitrogen x Water 1 0.88 0.35 7.10 ( + 1.10) 0.65 Tristachya Nitrogen 1 0.42 0.52 13.54 ( + 1.16) 0.56 Water 1 1.75 0.19 14.66 ( + 1.15) 0.52 Nitrogen x Water 1 0.19 0.67 16.31 ( + 1.22) 0.15

PAGE 53

53 Table 3 6. Difference in diameter between beginning and end of experiment by species and treatment Statistics are from a general linear model with nitrogen, water, nitrogen x water (treatments) original diameter, and average PAR (covariates) as fixed effects. Interactions between treatments and covariates were included in the model when significant. Means are the centimeters of diameter difference for treatments from model with covariates Model fit for treatment effects : R2 = 0.23, p = 0.003, n = 77, ( Loudetiopsis ) ; R2 = 0. 13, p = 0.07, n = 77 ( Tristachya ) Species Treatment df F value p Dunnetts test Mean ( + se) Loudetiopsis Nitrogen 1 1.62 0.20 0.88 1.25 ( + 1.12) Water 1 0.13 0.72 0.90 1.42 ( + 1.12) Nitrogen x Water 1 1. 30 0.26 0.22 1.11 ( + 1.18) Original Diameter 1 8.08 <0.01 Original Diameter x Nitrogen 1 3.74 0.06 Average PAR 1 5.70 0.02 Tristachya Nitrogen 1 3.76 0.06 0.13 3.22 ( + 1.12) Water 1 0.06 0.80 0.96 2.77 ( + 1.12) Nitrogen x Water 1 0.68 0.41 0.53 2.96 ( + 1.18) Original Diameter 1 3.73 0.06 Average PAR 1 4.08 0.05

PAGE 54

54 Table 3 7. R oot :shoot ratio s by species and treatment Statistics are from a general linear model. Means are for root:shoot ratio for all plants with nitrogen or water, and plants with just nitrogen and water. Model fit: R2 = 0.09, p = 0.07, n = 80 ( Loudetiopsis ) ; R2 = 0.003, p = 0.98, n = 79 ( Tristachya ) Species Treatment df F value p Mean ( + se) Dunnetts test Loudetiopsis Nitrogen 1 3.06 0.08 0.93 ( + 1.09) 0.39 Water 1 4.16 0.04 0.91 ( + 1.09) 0.52 Nitrogen x Water 1 0.01 0.91 0.82 ( + 1.12) 0.02 Tristachya Nitrogen 1 0.18 0.67 1.20 ( + 1.08) 0.99 Water 1 0.02 0.88 1.17 ( + 1.11) 0.98 Nitrogen x Water 1 <0.01 0.98 1.19 ( + 1.11) 0.99

PAGE 55

55 Tab le 3 8. Live leaf (LL) and dead aboveground biomass (DAB) and total aboveground biomass (TAB, g dry weight), each divided by diameter by treatment and species. Model is a general linear model. Fit statistics are embedded in the table. Loudetiopsi s : n = 79; Tristachya : n = 79 Species Treatment df DAB LL TAB (LL + DAB) F value p Dunnetts test F value p Dunnetts test F value p Dunnetts test Loudetiopsis Nitrogen 1 0.08 0.78 1.00 = 2.12 0.15 0.90 1.92 0.17 0.81 Water 1 6.92 0.01 0.23 1.72 0.19 0. 94 5.12 0.03 0.40 Nitrogen x Water 1 0.09 0.76 0.11 0.42 0.52 0.13 0.13 0.72 0.03 78 2.36 0.08 R2 = 0.09 1.44 0.2 4 0.05 2.42 0.09 R2 = 0.09 Tristachya Nitrogen 1 0.48 0.49 0.42 1.13 0.29 0.47 0.98 0.33 0.41 Water 1 0.91 0.33 0.31 0.15 0.70 0.79 0.45 0.50 0 .56 Nitrogen x Water 1 1.33 0.25 0.49 0.35 0.50 0.61 0.78 0.38 0.50 Model Fit 78 0.93 0.43 R2 = 0.06 0.59 0.63 R2 = 0.02 0.74 0.50 R2 = 0.05

PAGE 56

56 Table 3 9. Number of live leaves per area ( cm2) by treatment and species Statistics derive from an analysis of covariance with nitrogen and water and their interaction as fixed effects and diameter as a continuous covariate. Model fit: R2 = 0.14, p = 0.02, n = 80 ( Loudetiopsis ) ; R2 = 0.06 p =0.23, n = 80 ( Tristachya). Species Treatment d f F value p Mean ( + se) Dunnetts test Loudetiopsis Nitrogen 1 0.48 0.49 0.27 ( + 0.88) 0.74 Water 1 4.80 0.03 0.31 ( + 1.13) 0.98 Nitrogen x Water 1 0.32 0.95 0.33 ( + 1.19) 0.50 D iameter 1 6.99 0.01 Tristachya Nitrogen 1 0.02 0.88 0.10 ( + 1.11) = 0.99 Water 1 0.23 0.63 0.10 ( + 0.89) 0.99 Nitrogen x Water 1 0.03 0.87 0.10 ( + 1.16) 0.99 D iameter 1 24.93 <0.0001 Table 3 10. Foliar nitrogen (g/Kg) by treatment and species Model is an analysis of covariance with live leaf bioma ss as a covariate in the model, and treatments as fixed effects. Loudetiopsis : n = 80; Tristachya : n=79. Model fit statistics are embedded in the table. Species Treatment /Factor F value p Means ( + se) Dunnetts test Loudetiopsis Nitrogen 1.62 0.21 6.87 ( + 0.10) 0.99 Water 0.39 0.53 6.82 ( + 0.10) 0.97 Water x Nitrogen 1.35 0.25 7.00 ( + 0.15) 0.41 Live Leaves 4.77 0.03 Model Fit 1.91 0.12 R 2 = 0.09 Tristachya Nitrogen 0.40 0.53 7.74 ( + 0.14) 0.77 Water 0.69 0.41 7.59 ( + 0.14) 0.99 Water x Nitrogen 0.24 0.63 7.61 ( + 0.20) 0.99 Live Leaves 0.07 0.79 Model Fit 0.36 0.84 R 2 = 0.02

PAGE 57

57 Table 3 11. Foliar phosphorus (g/Kg) by treatment and species Model is an analysis of covariance with live leaf biomass as a covariate in the model, and treatments as fixed effects. Loudetiopsis : n=80; Tristachya : n=79. Species Treatment F value p Means ( + se) Dunnetts test Loudetiopsis Nitrogen 1.78 0.19 0.66 ( + 0.018 ) 0.09 Water 1.17 0.28 0.67 ( + 0.018 ) 0.13 Water x Nitrogen 3.01 0.09 0.67 ( + 0.025) 0.22 Live Leaves 1.96 0.17 Model Fit 1.71 0.16 R 2 = 0.08 Tristachya Nitrogen 2.93 0.09 0.87 ( + 0.020) 0.40 Water 3.12 0.08 0.87 ( + 0.020) 0.39 Water x Nitrogen 0.02 0.88 0.90 ( + 0.028) 0.04 Live Leaves 0.21 0.65 Model Fit 1.70 0.16 R 2 = 0.08 Table 3 12. N:P ratios by treatment and species. Model is an analysis of covariance with live leaf biomass as a covariate in the model, and treatments as fixed effects. Loudetiopsis : n = 80; Tristachya : n = 79. Species Treatment F value p Means ( + se) Dunnetts test Loudetiopsis Nitrogen 3.74 0.06 10.40 ( + 1.03) 0.09 Water 1.70 0.20 10.28 ( + 1.03) 0.23 Water x Nitrogen 1.18 0.28 10.44 ( + 1.03) 0.07 Live Leaves 7.34 0.01 Model Fit 2.91 0.03 R 2 = 0.13 Tristachya Nitrogen 1.57 0.21 8.89 ( + 1.03) 0.71 Water 5.87 0.02 8.70 ( + 1.03) 0.23 Water x Nitrogen <0.01 0.99 8.51 ( + 1.04 ) 0.03 Live Leaves 0.02 0.90 Model Fit 1.90 0.12 R 2 = 0.09

PAGE 58

58 Table 3 13. Response of s oil phosphorus concentration (g/Kg of dry soil) to treatments and species. Model is a generalized linear model with gamma distribution. Means presented are for comparisons of water ed and unwatered, plots with and without nitrogen, plots with nitrogen and water species effects, and significant interactions between species and treatment Arrows indicate difference relative to controls. Model fit: AIC = values: 251.26 2 = 11.94 /df = 0.16, n=80. Treatment df F value p Means ( + se ) Dunnetts test Nitrogen 1 4.16 0.04 0.14 ( + < 0.01) 0.12 Water 1 0.04 0.84 0.13 ( + <0.01) 0.82 Nitrogen x Water 1 0.64 0.43 0.14 ( + 0.01) 0.27 Species 1 2.68 0.06 Loudetiopsis : 0.12 ( + < 0.0 1 ) Tristachya : 0.14 ( + < 0.0 1 ) Water Species 1 7.37 <0.01 Loudetiopsis : 0.11 ( + < 0.01) Tristachya : 0.16 ( + 0.01 ) Table 3 14. Response of soil phosphorus concentration (g/Kg of dry soil) to treatments by species. Model is a generalized linear model with gamma response variable distribution. Means presented are for comparisons of water ed and un watered, plots with and without nitrogen, and plots with nitrogen and water. Arrows indicate difference relative to controls. Model fit: AIC = 131.42, 2 = 5.75 /df = 0.16, n = 40 ( Loudetiopsis ) ; AIC = 114.17, 2 = 6.18 2/df =0.17, n=40 ( Tristachya ) Species Treatment df F value p Mean ( + se) Dunnetts test Loudetiopsis Nitrogen 36 2.97 0.11 0.13 ( + 0.011) 0.09 Water 36 1.87 0.07 0.11 ( + 0.009) 0.47 Nitrogen x Water 36 0.63 0.43 0.11 ( + 0.013) 0.89 Tristachya Nitrogen 36 1.56 0.22 0.15 ( + 0.013 ) 0.27 Water 36 3.99 0.05 0.16 ( + 0.014 ) 0.11 Nitrogen x Water 36 0.12 0.73 0.17 ( + 0.02 0 ) 0.03

PAGE 59

59 Table 3 15. Analysis of pH, aluminum, potassium, and calcium response to treatments. Statistics are for a general linear model with nitrogen and water as fixed effects (n = 80 for all variables) Means presented are for comparisons of water ed and unwatered nitrogen fertiliz ed and unfertilized plots, and nitrogen x water plots. Arrows in dicate direction of difference relative to controls. Model fit statistics are embedded in the table. Treatment d f pH (in H 2 O) Aluminum (g/Kg dry soil) F value p Means ( + se) Dunnetts test F value p Means ( + se) Dunnetts test Nitrogen 1 0.38 0.54 4.93 ( + 0.04) 0.85 1.60 0.21 0.064 ( + 0.002) 0.28 Water 1 0.05 0.83 4.94 ( + 0.04) 0.96 1.09 0.30 0.061 ( + 0.002) 0.99 Nitrogen x Water 1 0.10 0.75 4.93 ( + 0.05) 0.88 0.89 0.35 0.062 ( + 0.003) 0.99 Species 1 2.24 0.14 L: 4.90 ( + 0.04) T: 4.80 ( + 0.04) 0.74 0.39 L: 0.064 ( + 0.002) T: 0.061 ( + 0.002) Model Fit 79 0.69 0.60 R2 = 0.06 1.08 0.37 R2 = 0.06 Potassium (g/Kg dry soil) Calcium (g/Kg dry soil) F value p Means ( + se) Dunnetts test F value p Means ( + se) Dunnetts test 3.00 0.09 8.25 ( + 1.01) 0.23 3.15 0.08 0.012 ( + 1.11) 0.23 0.08 0.77 8.25 ( + 1.01) 0.85 0.09 0.76 0.011 ( + 1.11) 0.86 0.42 0.52 8.25 ( + 1.01) 0.35 0.36 0.55 0.012 ( + 1.16) 0.32 2.49 0.12 L: 8.33 ( + 1.01) T: 8.25 ( + 1.01) 0.01 0.93 L: 0.011 ( + 1.11) T: 0.010 ( + 1.11) 1.29 0.28 R2 = 0.07 0.90 0.47 R2 = 0.05

PAGE 60

60 Table 3 16. Results of resin available ammonium ( /ml in 28 days) response to treatments and species Statistics are for a generalized linear model with a gamma response variable distribution with nitrogen, water, their interaction, and species as fixed effects. Means presented are for species and treatments. Model f it: AIC = 164.53, n = 76. Treatment df F value p Mean ( + se) Dunnetts tes t Nitrogen 1 <0.001 0.98 1.10 ( + 0.13) 0.48 Water 1 4.72 0.03 1.31 ( + 0.15) 0.02 Nitrogen x Water 1 1.16 0.28 1.20 ( + 0.20) 0.14 Species 1 0.18 0.67 Loudetiopsis : 1.14 ( + 0.15) Tristachya : 1.06 ( + 0.12) Table 3 17. Results of resin available nitrate ( /ml in 28 days) response to treatments and species Statistics are for a general linear model with nitrogen, water their interaction, and species as fixed effects. Means presented are for treatments and species. Model f it: R2 = 0.06, F value: 1.13, p = 0.35, n = 76. Treatment df F value p Means ( + se) Dunnetts test Nitrogen 1 <0.001 0.99 0.98 ( + 1.13) 0.86 Water 1 3.70 0.06 1.14 ( + 1.14) 0.12 Nitrogen x Water 1 0.83 0.37 0.90 ( + 0.83) 0.30 Species 1 0.01 0.93 Loudetiopsis : 0.81 ( + 0.82 ) Tristachya : 0.83 ( + 1.13 )

PAGE 61

61 T able 3 18. Results of ammonium mineralization rate N /g over 7 days) response to treatments and species. Statistics are for a general linear model with nitrogen, water, their interaction, and species as fixed effects. Mean s presented are for treatments and species. Model Fit: R2 = 0.02, F value: 0.36, p = 0.84, n = 99. Treatment df F value p Means ( + se) Dunnetts test Nitrogen 1 0.06 0.81 0 72 ( + 0 41 ) 0. 63 Water 1 <0.01 0.96 0.78 ( + 0.41 ) 0. 72 Nitrogen x Water 1 1.36 0.25 1.05 ( + 0.58 ) 0. 99 Species 1 0.03 0.86 Loudetiopsis : 0.84 ( + 0.41 ) Tristachya : 0.74 ( + 0.41 ) T able 3 19. Results of nitrate mineralization rate ( N /ml in 7 days) response to treatments and species Statistics are for a general linear model with nitrogen, water their interaction, and species as fixed effects. Means presented are for treatments and species. Model Fit: R2 = 0.001, F value: 0.03, p = 0.99, n = 99. Treatment df F value p Means ( + se) Dunnetts test Nitrogen 1 <0.01 0.99 0.05 ( + 0.04) 0.99 Water 1 0.03 0.87 0.06 ( + 0.04 ) = 1.00 Nitrogen x Water 1 0.03 0.86 0.06 ( + 0.06 ) 0.99 Species 1 0.07 0.80 Loudetiopsis : 0.06 ( + 0.04 ) Tristachya : 0.04 ( + 0.04 )

PAGE 62

62 T able 3 20. Re sults of soil ammonium ( N/g ) response to treatments and species Statistics are for a general linear model with nitrogen, water their interaction, and species as fixed effects. Means presented are for treatments and species. Model Fit: R2 = 0.02, F value: 0.38, p = 0.82, n = 98. Treatment df F value p Means ( + se) Dunnetts test Nitrogen 1 0.21 0.65 2.15 ( + 1.14) 0.74 Water 1 0.72 0.40 2.43 ( + 1.14) 0.99 Nitrogen x Water 1 0.12 0.73 2.31 ( + 1.20 ) 0.99 Species 1 0.53 0.47 Loudetiopsis : 2.17 ( + 1.14 ) Tristachya : 2.49 ( + 1.14 ) T able 3 21. Results of soil nitrate N /g) response to treatments and species. Statistics are for a general linear model with nitrogen, water, their interaction, and species as fixed effects. Means presented are for treatment and species. Model Fit: R2 = 0.05, F value: 1.08, p = 0.37, n = 98. Treatment df F value p Means ( + se) Dunnetts test Nitrogen 1 0.52 0. 47 0.08 ( + 1.23 ) 0. 77 Water 1 3.53 0.06 0.05 ( + 1.22 ) 0. 59 Nitrogen x Water 1 0.16 0. 69 0.06 ( + 1.33 ) 0. 75 Species 1 0.09 0.76 Loudetiopsis : 0.07 ( + 1.22 ) Tristachya : 0.07 ( + 1.23 )

PAGE 63

63 Table 4 1. Table of s ignificant results The significance level is represented by 0.05 < p and ** p Treatment Loudetiopsis Tristachya Soil Reproductive Vegetative Reproductive Vegetative Nitrogen Spikelets/ flowering tiller Water biomass biomass likelihood resin NH4 + resin NO3 soil NO3 L. ) T. ) Water x Nitrogen likelihood

PAGE 64

64 APPENDIX A TABLE OF UNIVARIATE STATISTICS Treatment Loudetiopsis Tristachya Min Value Mean Median M ax. Value Min Value Mean Median Max Value Live Leaf Biomass (g) Control 5.50 16.8 16.8 29. 2 38.2 118.1 96.7 298.6 Nitrogen 1.9 20.4 18.7 50.1 25.9 94.1 84.0 187.0 Water 7.9 21.0 20.3 49.6 0.0 91.1 87.8 214.7 Water x Nitrogen 6.9 22.0 20.2 39.2 22.1 92.2 72. 3 267.2 Dead Leaf Biomass (g) Control 11.7 25.4 22.5 51.1 63.5 291.4 238.7 749.2 Nitrogen 2.92 28.4 28.1 71.2 43.2 229.6 201.8 648.1 Water 13.0 43.2 31.7 145.7 60.3 189.9 140.2 716.8 Water x Nitrogen 13.4 36.1 28.3 82.8 38.0 226.2 142.9 810.8 Reproductive Tiller Biomass (g) Control 0.0 4.0 2.4 15.9 0.0 7.0 0.0 63.5 Nitrogen 0.0 6.1 3.0 25.4 0.0 6.9 0.0 35.8 Water 0.0 5.4 2.3 34.8 0.0 13.9 12.0 48.3 Water x Nitrogen 0.0 7.1 3.7 21.7 0.0 11.1 5.9 81.7 Root Biomass (g) Control 20.5 59.7 54.4 154.7 160.4 441.8 336.1 778.8 Nitrogen 11.31 59.6 55.4 134.0 71.5 418.7 336.4 1342.0 Water 14.7 71.8 66.6 229.1 56.5 362.6 325.7 814.9 Water x Nitrogen 22.0 53.8 50.3 132.9 102.0 363.3 274.4 1020.2 Total Aboveground Biomass (g) Control 22.1 46.2 43.0 93.7 116.9 416.5 374.9 1047.7 Nitrogen 11.5 54.9 56.3 111.8 69.1 330.6 280.5 835.1 Water 23.6 69.6 56.7 209.2 97.6 294.8 247.8 979.8 Water x Nitrogen 27.2 65.2 56.3 143.7 60.1 329.5 223.4 1090.1 Dead Tillers Control 0 14.9 14.0 36 6 17.4 15.0 61 Nitrogen 2 14.6 8.0 72 3 14.2 12.0 41 Water 4 22.2 12.5 88 3 15.1 13.0 65 Water x Nitrogen 1 17.9 14.5 45 1 15.1 9.0 45 Live Flowering and Non flowering Tillers Control 24 63.1 67.5 109 17 47.5 38.0 112 Nitrogen 21 65.4 62.0 261 16 46.0 42.0 84 Water 31 81.5 75.0 149 12 46.7 40.0 110 Water x Nitrogen 34 60.7 60.0 112 13 46.9 35.5 118

PAGE 65

65 Treatment Loudetiopsis Tristachya Min Value Mean Median M ax. Value Min Value Mean Median Max Value Average PAR Control 285 1532 1742 1998 338 1626 1719 2068 Nitrogen 193 1525 1749 2041 274 1438 1695 2943 Water 222 1518 1556 2012 271 1581 1767 1963 Water x Nitrogen 762 1546 1754 1968 391 1746 1890 2089 Number of Flowering Individuals (2009) Control 18 9 Nitrogen 16 9 Water 13 14 Water x Nitrogen 19 14 Number of Flowering Tillers per Flowering Individual Control 1 3.7 3.0 12 1 2.0 2.0 4 Nitrogen 1 6.5 5.0 16 1 1.2 1.0 2 Water 1 7.8 6.0 24 1 1.6 1.0 4 Water x Nitrogen 1 6.8 7.0 16 1 1.6 1.0 6 Total Number of Spikelets Control 3 26.7 15.5 121 2 28.7 24.0 97 Nitrogen 4 46.6 37.5 129 1 19.3 16.0 51 Water 2 57.2 39.0 281 5 25.2 18.0 79 Water x Nitrogen 4 60.3 34.0 183 8 35.6 25.5 184 Number of Spikelets per Flowering Tiller Control 3 6.6 6.8 11.8 2 12.8 12.0 24.3 Nitrogen 4 7.4 6.7 13.3 1 14.9 14.0 25.5 Water 2 5.9 5.3 11.7 5 15.0 17.0 23 Water x Nitrogen 3.5 7.8 8.1 17.0 8 18.1 14.0 34 Original Diameter (cm) Control 5.2 9.1 8.9 14.8 14.6 25.0 23.6 42.8 Nitrogen 5.9 10.5 9.6 20.3 12.5 24.4 23.5 38.6 Water 6.2 11.0 10.1 17.4 11.1 22.7 20.4 46.7 Water x Nitrogen 4.7 9.4 9.1 17.8 10.8 21.7 21.9 36.2 Final Diameter (cm) Control 6.1 10.9 10.9 17.6 18.2 28.1 27.6 45.9 Nitrogen 4.8 11.9 10.5 20.8 15.2 28.1 26.4 43.6 Water 6.4 13.0 12.9 20.7 13.7 26.1 23.2 50.5 Water x Nitrogen 6.7 10.3 10.2 15.8 13.2 25.3 25.4 39.15.6

PAGE 66

66 Treatment Loudetiopsis Tristachya Min Value Mean Median M ax. Value Min Value Mean Median Max Value Root:shoot Ratio Control 0.50 1.56 1.07 4.67 0.37 1.31 1.29 2.32 Nitrogen 0.46 1.24 0.98 4.79 0.24 1.35 1.38 2.55 Water 0.47 1.16 1.03 2.97 0.53 1.24 1.20 2.49 Water x Nitrogen 0.42 0.90 0.87 2.31 0.51 1.27 1.21 1.92 Foliar Nitrogen (g/Kg) Control 5.60 6.78 6.86 7.70 6.30 7.64 7.63 8.82 Nitrogen 5.60 6.73 6.65 8.26 5.60 7.88 8.05 10.01 Water 5.46 6.63 6.65 7.56 6.02 7.58 7.42 10.01 Water x Nitrogen 5.60 6.97 7.00 8.40 6.16 7.61 7.39 10.36 Foliar Phosphorus (g/Kg) Control 0.50 0.73 0.72 1.02 0.48 0.79 0.79 1.07 Nitrogen 0.40 0.65 0.67 0.88 0.64 0.85 0.88 1.00 Water 0.49 0.66 0.64 0.87 0.70 0.85 0.84 1.15 Water x Nitrogen 0.56 0.67 0.69 0.98 0.68 0.90 0.89 1.14 Foliar N:P Control 6.5 9.5 9.3 12.4 7.4 9.9 9.4 14.3 Nitrogen 8.2 10.5 10.0 15.6 5.6 9.4 9.4 13.3 Water 6.6 10.2 10.4 13.6 7.4 9.0 8.7 12.2 Water x Nitrogen 6.6 10.5 10.3 13.7 6.9 8.6 8.1 11.9 Soil Phosphorus (g/Kg) Control 0.08 0.11 0.09 0.18 0.06 0.12 0.11 0.19 Nitrogen 0.08 0.15 0.14 0.32 0.08 0.14 0.12 0.29 Water 0.05 0.10 0.10 0.16 0.08 0.15 0.16 0.26 Water x Nitrogen 0.06 0.11 0.11 0.24 0.08 0.17 0.17 0.34 pH (in water ) Control 4.61 4.95 4.97 5.32 4.50 4.92 4.90 5.29 Nitrogen 4.56 4.83 4.83 4.99 4.75 4.97 4.97 5.24 Water 4.42 4.98 5.04 5.36 4.49 4.98 5.01 5.58 Water x Nitrogen 4.70 4.93 4.96 5.11 4.49 4.94 5.10 5.31 Calcium (g/Kg) Control 0.0045 0.0099 0.0072 0.0280 0.0027 0.0106 0.0090 0.0208 Nitrogen 0.0063 0.0203 0.0154 0.0814 0.0036 0.0108 0.0090 0.0208 Water 0.0054 0.0097 0.0099 0.0154 0.0036 0.0162 0.0099 0.0515 Water x Nitrogen 0.0045 0.0122 0.0104 0.0235 0.0045 0.0199 0.0099 0.0787

PAGE 67

67 Treatment Loudetiopsis Tristachya Min Value Mean Median M ax. Value Min Value Mean Median Max Value Potassium (g/Kg) Control 8.07 8.36 8.45 8.52 8.06 8.35 8.37 8.61 Nitrogen 7.51 8.19 8.23 8.50 7.63 8.26 8.36 8.53 Water 8.16 8.41 8.41 8.65 7.75 8.21 8.18 8.49 Water x Nitrogen 7.82 8.36 8.40 8.59 7.43 8.13 8.11 8.51 Aluminum (g/Kg) Control 0.043 0.063 0.063 0.081 0.041 0.059 0.063 0.0678 Nitrogen 0.054 0.068 0.068 0.081 0.047 0.065 0.063 0.101 Water 0.045 0.063 0.060 0.081 0.032 0.059 0.065 0.072 Water x Nitrogen 0.043 0.063 0.061 0.081 0.041 0.060 0.059 0.086 Resin Available Nitrate ( N/28 days) Control 0.125 0.351 0.255 0.961 0.010 0.164 0.141 0.344 Nitrogen 0.016 0.441 0.293 1.033 0.007 0.240 0.213 0.950 Water 0.006 0.336 0.161 0.828 0.008 0.217 0.161 0.503 Water x Nitrogen 0.112 0.493 0.355 1.338 0.027 0.127 0.098 0.232 Resin Available Ammonium ( N/28 days) Control 0.178 1.03 0.554 2.732 0.239 0.707 0.651 1.447 Nitrogen 0.293 0.786 0.538 1.911 0.311 1.117 0.703 3.619 Water 0.136 1.222 1.171 2.493 0.364 1.680 1.530 4.705 Water x Nitrogen 0.644 1.685 1.401 4.068 0.128 0.939 0.639 2.032 Soil Nitrate ( g N/g ) Control 0.008 0.129 0.114 0.302 <0.001 0.080 0.086 0.186 Nitrogen <0.001 0.165 0.105 0.823 0.058 0.247 0.105 1.211 Water <0.001 0.067 0.078 0.163 0.014 0.145 0.097 0.502 Water x Nitrogen 0.019 0.138 0.111 0.329 <0.001 0.069 0.091 0.119 Soil Ammonium ( N /g ) Control 1.474 2.433 2.236 4.107 1.380 2.832 2.572 4.949 Nitrogen <0.001 2.729 2.181 4.993 1.286 3.139 2.918 6.335 Water 1.191 2.622 2.186 5.442 1.059 3.133 2.541 6.633 Water x Nitrogen 1.568 2.961 2.923 4.905 1.131 2.068 1.984 3.209 Nitrate Mineralization Rate ( N/7 days) Control 0.141 0.045 0.033 0.579 0.186 0.057 0.050 1.078 Nitrogen 0.485 0.121 0.089 0.989 1.211 0.038 0.019 0.551 Water 0.163 0.019 0.007 0.407 0.418 0.088 0.028 0.961 Water x Nitrogen 0.330 0.059 0.039 1.139 0.114 0.064 0.025 0.629

PAGE 68

68 Treatment Loudetiopsis Tristachya Min Value Mean Median M ax. Value Min Value Mean Median Max Value Ammonium Mineralization Rate ( N/7 days) Control 3.827 0.606 1.571 6.631 4.315 1.831 1.943 2.214 Nitrogen 3.395 0.434 0.418 6.753 4.320 1.225 1.934 9.751 Water 4.652 1.473 1.300 1.976 4.395 0.604 0.266 14.518 Water x Nitrogen 4.791 1.581 1.623 2.602 3.034 0.552 0.917 5.052

PAGE 69

69 LIST OF REFERENCES Aber J, McDowell W, Nadelhoffer K et al. (1998) Nitrogen saturation in temperate forest ecosystems Hypotheses revisited. Bioscience, 48 921934. Augustine DJ (2003) Spatial heterogeneity in the herbaceous layer of a semi arid savanna ecosystem. Plant Ecology, 167, 319332. Augustine DJ, McNaughton SJ (2004) Temporal asynchrony in soil nutrient dynamics and plant production in a semiarid ecosystem. Ecosystems, 7, 829840. Augustine DJ, McNaughton SJ, Frank DA (2003) Feedbacks between soil nutrients and large herbivores in a managed savanna ecosystem. Ecological Applications, 13, 13251337. Austin AT, Yahdjian L, Stark JM et al. (2004) Water pulses and biogeochemical cycles in arid and semiarid ecosystems. Oecologia, 141, 221235. Barger NN, D'Antonio CM, Ghneim T, Brink K, Cuevas E (2002) Nutrient limitation to primary productivity in a secondary savanna in Venezuela. Biotropica, 34, 493501. Baruch Z, Bilbao B (1999) Effect s of fire and defoliation on the life history of native and invader C 4 grasses in a Neotropical savanna. Oecologia, 119, 510520. Boone Kauffman J, Cummings DL, Ward DE (1994) Relationships of fire, biomass, and nutrient dynamics along a vegetation gradie nt in the Brazilian Cerrado. Journal of Ecology, 82, 519531. Bucci SJ, Scholz FG, Goldstein G et al. (2006) Nutrient availability constrains the hydraulic architecture and water relations of savannah trees. Plant Cell and Environment, 29, 21532167. Busta mante MMC, Martinelli LA, Silva DA, Camargo PB, Klink CA, Domingues TF, Santos RV (2004) N 15 natural abundance in woody plants and soils of central Brazilian savannas (cerrado). Ecological Applications, 14, 200213. Bustamante MMC, Medina E, Asner GP, Nar doto GB, Garcia Montiel DC (2006) Nitrogen cycling in tropical and temperate savannas. Biogeochemistry, 79, 209237. Carrera AL, Bertiller MB, Sain CL, Mazzarino MJ (2003) Relationship between plant nitrogen conservation strategies and the dynamics of soil nitrogen in the arid Patagonian Monte, Argentina. Plant and Soil, 255, 595604. Cavalcanti RB, Joly CA (2002) Biodiversity and Conservation Priorities in the Cerrado Region. In: The Cerrados of Brazil: Ecology and Natural History of a Neotropical Savanna. (eds Oliveira PS, Marquis RJ) pp 351367. New York, Columbia University Press. Cech PG, Kuster T, Edwards PJ, Venterink HO (2008) Effects of herbivory, fire and N 2fixation on nutrient limitation in a humid African savanna. Ecosystems, 11, 9911004.

PAGE 70

70 Clar k CM, Tilman D (2008) Loss of plant species after chronic low level nitrogen deposition to prairie grasslands. Nature, 451, 712715. Cleland EE, Chiariello NR, Loarie SR, Mooney HA, Field CB (2006) Diverse responses of phenology to global changes in a gras sland ecosystem. Proceedings of the National Academy of Sciences of the United States of America, 103, 1374013744. Costa AN, Vasconcelos HL, VieiraNeto EHM, Bruna EM (2008) Do herbivores exert topdown effects in Neotropical savannas? Estimates of biomas s consumption by leaf cutter ants. Journal of Vegetation Science, 19, 849U814. Craine JM, Morrow C, Stock WD (2008) Nutrient concentration ratios and colimitation in South African grasslands. New Phytologist, 179, 829836. Craine JM, Tilman D, Wedin D, R eich P, Tjoelker M, Knops J (2002) Functional traits, productivity and effects on nitrogen cycling of 33 grassland species. Functional Ecology, 16, 563574. Cruz P (1997) Effect of shade on the growth and mineral nutrition of a C 4 perennial grass under fi eld conditions. Plant and Soil, 188, 227237. da Silva DM, Batalha MA (2008) Soil vegetation relationships in cerrados under different fire frequencies. Plant and Soil, 311, 8796. Davidson EA, de Carvalho CJR, Figueira AM et al. (2007) Recuperation of nit rogen cycling in Amazonian forests following agricultural abandonment. Nature, 447, 995996. Davidson EA, de Carvalho CJR, Vieira ICG et al. (2004) Nitrogen and phosphorus limitation of biomass growth in a tropical secondary forest. Ecological Applications, 14, 150163. de Oliveira Filho AT, Ratter JA (2002) Vegetation Physiognomies and Woody Flora of the Cerrado Biome. In: The Cerrados of Brazil. (eds Oliveira PS, Marquis RJ). New York, Columbia University Press. EMPRAPA (1999) Sistema brasilei ro de classifio de solos., Rio de Janeiro, Centro Nacional de Pesquisa de Solos, Embrapa Solos. Fargione J, Tilman D (2005) Niche differences in phenology and rooting depth promote coexistence with a dominant C 4 bunchgrass. Oecologia, 143, 598606. Fay PA, Carlisle JD, Knapp AK, Blair JM, Collins SL (2003) Productivity responses to altered rainfall patterns in a C 4dominated grassland. Oecologia, 137, 245251. Fisher FM, Whitford WG (1995) Field simulation of wet and dry years in the Chihuahuan Desert Soil moisture, N mineralization, and ionexchange resin bags. Biology and Fertility of Soils, 20, 137146.

PAGE 71

71 Gruber NG, J. N. (2008) An Earthsystem perspective of the global nitrogen cycle. Nature, 45, 293296. Hardisan MA, G.M. (2005) Perfil nutricional de espcies lenhosas de duas florestas semidecduas em Uberlndia, MG. Revista Brasileira de Botanica, 28, 295 303. Harrington RA, Fownes JH, Vitousek PM (2001) Production and resource use efficiencies in N and P limited tropical forests: a comparison of responses to long term fertilization. Ecosystems, 4, 646657. Haylock MR, Peterson TC, Alves LM et al. (2006) Trends in total and extreme South American rainfall in 19602000 and links with sea surface temperatur e. Journal of Climate, 19, 14901512. Henry HAL, Chiariello NR, Vitousek PM, Mooney HA, Field CB (2006) Interactive effects of fire, elevated carbon dioxide, nitrogen deposition, and precipitation on a California annual grassland. Ecosystems, 9, 10661075. Holdo RM, Holt RD, Coughenour MB, Ritchie ME (2007) Plant productivity and soil nitrogen as a function of grazing, migration and fire in an African savanna. Journal of Ecology, 95, 115128. Huntley BJ, Walker BH (eds) (1982) Ecology of Tropical Savannas, Berlin, Springer Verlag. Instituto de Geografia (2008) Alturas Pluviomtricas Mensais Uberlndia, MG 20042008. pp 1, Uberlndia, MG, Universidade Federal de Uberlndia. Knapp AK, Burns CE, Fynn RWS, Kirkman KP, Morris CD, Smith MD (2006) Convergence and contingency in productionprecipitation relationships in North American and south african C 4 grasslands. Oecologia, 149, 456464. Kochy M, Wilson SD (2004) Semiarid grassland responses to short term variation in water availability. Plant Ecology, 174, 197203. Lau JA, Peiffer J, Reich PB, Tiffin P (2008) Transgenerational effects of global environmental change: long term CO2 and nitrogen treatments influence offspring growth response to elevated CO2. Oecologia, 158, 141150. Lilienfein J, Wilcke W (2004) Water and element input into native, agri and silvicultural ecosystems of the Brazilian savanna. Biogeochemistry, 67, 183212. Ludwig F, de Kroon H, Prins HHT, Berendse F (2001) Effects of nutrients and shade on tree grass in teractions in an East African savanna. Journal of Vegetation Science, 12, 579588.

PAGE 72

72 Magrin G, Gay Garca C, Cruz Choque D et al. (2007) 2007: Latin America. In: Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II t o the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. (eds Parry ML, Canziani OF, Palutikof JP, van der Linden PJ, Hanson CE) pp 581615. Cambridge, UK, Cambridge University Press. Matson PA, McDowell WH, Townsend AR, Vitousek PM (1999) The globalization of N deposition: ecosystem consequences in tropical environments. Biogeochemistry, 46, 6783. Menge DNL, Field CB (2007) Simulated global changes alter phosphorus demand in annual grassland. Global Change Biology, 13, 25822591. M issouri Botanical Garden (2009) w3 TROPICOS Accessed June 19, 2009 http://www.tropicos.org Mundim FM, Costa AN, Vasconcelos HL (2009) Leaf nutrient content and host plant selection by leaf cutter ants, Atta laevigata, in a Neotropical savanna. Entomologia Experimentalis Et Applicata, 130, 4754. Nardoto GB, Bustamante MMD, Pinto AS, Klink CA (2006) Nutrient use efficiency at ecosystem and species level in savanna areas of Central Brazil and impacts of fire. Journal of T ropical Ecology, 22, 191201. O'Connor TG (1994) Composition and population responses of an African savanna grassland to rainfall and grazing. Journal of Applied Ecology, 31, 155171. Oliveira PS, Marquis RJ (eds) (2002) The cerrados of Brazil: ecology and natural history of a neotropical savannas, New York, Columbia University Press. Ottmar RD, Vihnanek RE, Miranda HS, Sato MN, Andrade SMA (2001) Stereo photo series for quantifying Cerrado fuels in central Brazil Volume 1 General Technical Report PNW GTR 519. pp 87. Portland, OR, U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. Pandey CB, Singh JS (1992) Rainfall and grazing effects on net primary productivity in a tropical savanna, India. Ecology, 73, 20072021. Phillip s OL, Aragao L, Lewis SL et al. (2009) Drought Sensitivity of the Amazon Rainforest. Science, 323, 1344 1347. Phillips OL, Rose S, Mendoza AM, Vargas PN (2006) Resilience of southwestern Amazon forests to anthropogenic edge effects. Conservation Biology, 20, 16981710. Phoenix GK, Hicks WK, Cinderby S et al. (2006) Atmospheric nitrogen deposition in world biodiversity hotspots: the need for a greater global perspective in assessing N deposition impacts. Global Change Biology, 12, 470476.

PAGE 73

73 Ratter JA, Ribeiro JF, Bridgewater S (1997) The Brazilian cerrado vegetation and threats to its biodiversity. Annals of Botany, 80, 223230. Sardans J, Penuelas J (2007) Drought changes phosphorus and potassium accumulation patterns in an evergreen Mediterranean forest. Functional Ecology, 21, 191201. Sarmiento G (1992) Adaptive strategies of perennial grasses in SouthAmerican savannas Journal of Vegetation Science, 3, 325336. Sarmiento G, da Silva MP, Naranjo ME, Pinillos M (2006) Nitrogen and phosphorus as limiting f actors for growth and primary production in a flooded savanna in the Venezuelan Llanos. Journal of Tropical Ecology, 22, 203212. Schlesinger WH (2009) On the fate of anthropogenic nitrogen. Proceedings of the National Academy of Sciences of the United States of America, 106, 203208. Scholz FG, Bucci SJ, Goldstein G, Meinzer FC, Franco AC, Miralles Wilhelm F (2007) Removal of nutrient limitations by long term fertilization decreases nocturnal water loss in savanna trees. Tree Physiology, 27, 551559. Seagle SW, McNaughton SJ (1993) Simulated effects of precipitation and nitrogen on Serengeti grassland productivity. Biogeochemistry, 22, 157178. Siemann E, Rogers WE, Grace JB (2007) Effects of nutrient loading and extreme rainfall events on coastal tall grass prairies: invasion intensity, vegetation responses, and carbon and nitrogen distribution. Global Change Biology, 13, 21842192. Silletti AM, Knapp AK (2001) Responses of the codominant grassland species Andropogon gerardii and Sorghastrum nutans to l ong term manipulations of nitrogen and water. American Midland Naturalist, 145, 159167. Silletti AM, Knapp AK, Blair JM (2004) Competition and coexistence in grassland codominants: responses to neighbour removal and resource availability. Canadian Journal of Botany Revue Canadienne De Botanique, 82, 450460. Soil Survey Staff (2003) Keys to Soil Taxonomy USDA Natural Resources Conservation Service. Sousa Souto L, Schoereder JH, Schaefer C, Silva WL (2008) Ant nests and soil nutrient availability: the nega tive impact of fire. Journal of Tropical Ecology, 24, 639646. Sternberg LD, Pinzon MC, Moreira MZ, Moutinho P, Rojas EI, Herre EA (2007) Plants use macronutrients accumulated in leaf cutting ant nests. Proceedings of the Royal Society B Biological Science s, 274, 315321. Swemmer AM, Knapp AK, Smith MD (2006) Growth responses of two dominant C4 grass species to altered water availability. International Journal of Plant Sciences, 167, 10011010.

PAGE 74

74 Tessier JT, Raynal DJ (2003) Use of nitrogen to phosphorus rati os in plant tissue as an indicator of nutrient limitation and nitrogen saturation. Journal of Applied Ecology, 40, 523534. Tomlinson KW, O'Connor TG (2004) Control of tiller recruitment in bunchgrasses: uniting physiology and ecology. Functional Ecology, 18, 489496. Townsend AR, Cleveland CC, Asner GP, Bustamante MMC (2007) Controls over foliar N : P ratios in tropical rain forests. Ecology, 88, 107118. Vitousek PM (1984) Litterfall, nutrient cycling, and nutrient limitation in tropical forests Ecology, 65, 285298. Vitousek PM, Aber JD, Howarth RW et al. (1997) Human alteration of the global nitrogen cycle: Sources and consequences. Ecological Applications, 7, 737750. Yahdjian L, Sala O, Austin A (2006) Differential controls of water input on litter dec omposition and nitrogen dynamics in the Patagonian steppe. Ecosystems, 9, 128141. Zavaleta ES (2002) Influences of climate and atmospheric changes on plant diversity and ecosystem function in a California grassland. Unpublished PhD Stanford University, Ca lifornia,United States 151 pp. Zavaleta ES, Shaw MR, Chiariello NR, Thomas BD, Cleland EE, Field CB, Mooney HA (2003) Grassland responses to three years of elevated temperature, CO2, precipitation, and N deposition. Ecological Monographs, 73, 585604. Zhan g XB, Zwiers FW, Hegerl GC et al. (2007) Detection of human influence on twentiethcentury precipitation trends. Nature, 448, 461464.

PAGE 75

75 BIOGRAPHICAL SKETCH Stella Copeland grew up in the Bay Area for the first few years of her life but the formative ones were spent in the mountains and streets of small town Ashland, Oregon. A senior year in a public school program called Wilderness Charter School gave her the freedom to explore botany and organic gardening. Taking a leap east she ended up at Colorado College for an amazing four years of a liberal arts education with lots of field science a nd a dash of everything from ethnography to sociology of i mmigration. To make a buck, and follow her muse, she quit her cushy job as a n ushe r at the Oregon Shakespeare Festival, and began working for the Bureau of Land Management, in Medford OR during summers off from CC. A season of rare plant surveys in one of the botanical hotspots of the US, and she was hooked on plant ecology. A study abr oad in Ecuador clinched the matter, especially after a month in the cloud forest collecting rare orchids. Plant s now on the brain, she worked for the Colorado Natural Heritage Program in the imposing, wild Sangre de Cristo mountain range then headed south for the winter to Peru for another orchid oriented position. Northward she spent a year as head of a small botany and forestry monitoring crew for the Nature Conservancy, t hen to Argentina to be a tourist and teach English for three months A fter a brief stint as a organic farm intern and a rare plant contractor in Ashland, sh e headed off to Florida where she has spent the last two years chasing academic dreams, learning Portuguese, delving into ecosystem ecology, and generally coming to the realization t hat she may not know much, but the trick is to make peace with ignorance. In the spirit of learning, and going forward in the field of conservation ecology, she is now headed to the University of California, Davis, in search of a PhD, and the secrets of de clining rare plants