Citation
Competitive Interactions in a Pecan (Carya illinoensis K. Koch)--Cotton (Gossypium hirsutum L.) Alleycropping System in the Southern United States

Material Information

Title:
Competitive Interactions in a Pecan (Carya illinoensis K. Koch)--Cotton (Gossypium hirsutum L.) Alleycropping System in the Southern United States
Creator:
ZAMORA, DIOMIDES SANTOS ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Biomass ( jstor )
Cotton ( jstor )
Crops ( jstor )
Modeling ( jstor )
Monoculture ( jstor )
Parametric models ( jstor )
Photosynthetically active radiation ( jstor )
Plant roots ( jstor )
Plants ( jstor )
Soil science ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Diomides Santos Zamora. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
5/31/2006
Resource Identifier:
436098741 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

COMPETITIVE INTERACTIONS IN A PECAN ( Carya illinoensis K. Koch) – COTTON (Gossypium hirsutum L.) ALLEYCROPPING SYSTEM IN THE SOUTHERN UNITED STATES By DIOMIDES SANTOS ZAMORA A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Diomides Santos Zamora

PAGE 3

This work is dedicated to my family especi ally to my mother, Alejandra Santos-Zamora. Also this is dedicated in the memory of father, Saturnino Zamora.

PAGE 4

iv ACKNOWLEDGMENTS I would like to express my sincere thanks to my graduate committee chair, Dr. Shibu Jose, whose help, encouragement and financial support through a graduate assistantship made this study a reality. Thanks also go to my committee members, Drs. P. K. R. Nair, James W. Jones, Wendell Cropper and Barry Brecke, for their contribution to this study. Each of them provided me with input and support at di fferent times and in different capacities. Their effo rts are greatly appreciated. I am also deeply grateful to the administ rative staff and farm crew of the West Florida Research and Education Center (W FREC) headed by Dr. Jeffrey Mullahey and Mr. Doug Hatfield, respectively. Special thanks to Dr. Craig Ramsey, my patient friend and mentor, and to my fellow graduate student s at UF Milton for their support and help throughout the course of my graduate program. Thanks are also extended to the staff of IFAS-Statistics Division for their help in data analyses; and to faculty and staff at the School of Forest Resources and Conservation, led by Dr. Tim White, and to the staff of the Center for Tropical and S ub-tropical Agroforestry. I would also like to express my appreciation to my friends Sam Allen, Rico Gazal, Kent Apostol, Ryan Curran, George and Eva Reyes and others in Gainesville and Milton for their support and encouragement. I thank my family in the Philippines, particularly my mother and sisters, for their love and support and for allowing me to be away from them in order to complete this work and fulfill my dream of earning a Ph.D. Most

PAGE 5

v importantly, I give all credit and honor to God, Author of all creation, and to Jesus Christ, who provided me wisdom. This study was funded in part by two gran ts; one from the USDA Southern Region Sustainable Agriculture Research and Edu cation (SARE) program (# LS02-136) and another from the USDA-IFAFS (# 00-52103-9702) through the Center for Subtropical Agroforestry at the University of Florida.

PAGE 6

vi TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES.............................................................................................................ix LIST OF FIGURES...........................................................................................................xi ABSTRACT.....................................................................................................................xi ii CHAPTER 1 INTRODUCTION........................................................................................................1 Statement of the Problem..............................................................................................1 Review of Literature.....................................................................................................2 Temperate Alleycropping......................................................................................2 Major Factors and Mechanisms in Aboveground Interactions..............................3 Root Dynamics......................................................................................................5 Modeling in Agroforestry......................................................................................7 2 INTERSPECIFIC COMPETITION IN A PECAN-COTTON ALLEYCROPPING SYSTEM IN THE SOUTHERN UNITED STATES: IS LIGHT THE LIMITING FACTOR?...................................................................................................................11 Introduction.................................................................................................................11 Materials and Methods...............................................................................................12 Study Area...........................................................................................................12 PAR and Radiation Use Efficiency.....................................................................13 Biomass and Lint Yield.......................................................................................15 Data Analysis.......................................................................................................16 Results........................................................................................................................ .16 Incident and Absorbed PAR................................................................................16 Biomass and Lint Yield.......................................................................................17 Radiation Use Efficiency.....................................................................................18 Discussion...................................................................................................................19 Conclusions.................................................................................................................22

PAGE 7

vii 3 INTERSPECIFIC COMPETITION IN A PECAN-COTTON ALLEYCROPPING SYSTEM IN THE SOUTHERN UN ITED STATES: PRODUCTION PHYSIOLOGY...........................................................................................................30 Introduction.................................................................................................................30 Materials and Methods...............................................................................................33 The Study Site.....................................................................................................33 Gas Exchange Measurements..............................................................................34 Specific Leaf Area (SLA) and Foliar N..............................................................35 Aboveground Biomass and Lint Yield................................................................35 Data Analysis.......................................................................................................35 Results........................................................................................................................ .36 SLA, SLN and A ..................................................................................................36 CNPI vs. Biomass and Lint Yield.......................................................................37 Discussion...................................................................................................................38 Conclusions.................................................................................................................41 4 MORPHOLOGICAL PLASTICITY OF COTTON ROOTS IN RESPONSE TO INTERSPECIFIC COMPETITION WITH PECAN IN AN ALLEYCROPPING SYSTEM IN THE SOUTHERN UNITED STATES.................................................51 Introduction.................................................................................................................51 Materials and Methods...............................................................................................53 Study Site.............................................................................................................53 Treatments and Sample Collection......................................................................54 Data Analysis.......................................................................................................56 Results........................................................................................................................ .56 Biomass and Root: Shoot Ratio...........................................................................56 Total Root Length................................................................................................57 Root Length Density............................................................................................58 Specific Root Length...........................................................................................59 Discussion...................................................................................................................59 Below and Aboveground Carbon Allocation......................................................59 Root Morphological Plasticity.............................................................................60 Root Length Density............................................................................................61 Specific Root Length...........................................................................................63 Implications................................................................................................................64 5 MODELING COTTON PRODUCTION IN A PECAN ALLEYCROPPING SYSTEM USING CROPGRO...................................................................................75 Introduction.................................................................................................................75 Materials and Methods...............................................................................................78 Field and Experiment Description.......................................................................78 The CROPGRO model in DSSAT......................................................................79 CROPGRO Model Inputs....................................................................................80 Weather data inputs......................................................................................80

PAGE 8

viii Soil parameter inputs....................................................................................80 Model Execution.................................................................................................81 Model parameterization................................................................................81 Model calibration.........................................................................................81 Results and Discussion...............................................................................................83 Assessing Model Capability................................................................................83 Model Evaluation................................................................................................84 Evaluating SLA and LAI..............................................................................84 Assessment of predicted aboveground biomass...........................................84 Evaluating Relationships.....................................................................................85 Effects of leaf size (SIZELF) on LAI...........................................................85 Effects of SLA on LAI and biomass prediction...........................................85 Effects of partitioning on LAI and biomass prediction................................86 Spatial Prediction of Biomass.............................................................................88 Conclusions.................................................................................................................89 6 SUMMARY AND CONCLUSION.........................................................................100 APPENDIX A. MODEL INTERFACE.............................................................................................105 B. SAMPLE GRAPHICAL OUTPUT OF THE CROPGRO COTTON MODEL.......106 C. SAMPLE OVERVIEW OUTPUT OF THE CROPGRO-COTTON MODEL........107 LIST OF REFERENCES.................................................................................................109 BIOGRAPHICAL SKETCH...........................................................................................122

PAGE 9

ix LIST OF TABLES Table page 2-1 Leaf area index (LAI), light extinction coefficient (k) and mean absorbed photosynthetically active radiation (P AR) of cotton in non-barrier, barrier and monoculture treatments.....................................................................................23 2-2 Aboveground biomass production of co tton in barrier, non-barrier and monoculture treatments in 2001 and 2002 growing seasons....................................24 2-3 Radiation use efficiency of cotton in barrier, non-barrier and monoculture treatments in 2001 and 2002 growing seasons.........................................................25 3-1 Specific leaf area (SLA ) of cotton in non-barrier, barrier, and monoculture treatments.................................................................................................................42 3-2 Canopy net photosynthetic index of co tton measured under ambient condition (PAR and CO2) in a pecan-cotton alleycropping system.........................................43 3-3 Leaf level transpiration ( E ) and stomatal conductance ( g ) of cotton measured under ambient condition (PAR and CO2 concentration) in a pecan-cotton alleycropping system................................................................................................44 4-1 Growth parameters of cotton grown in non-barrier, barr ier and monoculture treatments in 2002 and 2003 growing seasons.........................................................66 4-2 Total root length of cotton in non-barrie r, barrier and monoculture treatments at physiological maturity..............................................................................................67 4-3 Root length density (RLD, cm cm-3) of cotton in barrier, non-barrier and monoculture treatments across soil depth................................................................68 4-4 Root length density of cott on under different soil depths........................................69 4-5 Percent amount of root biomass in different layers of soil for the non-barrier, barrier and monoculture treatments..........................................................................70 5-1 Initial soil conditions in Jay, Florid a in 2001 growing seas on used in the CROPGRO-cotton model.........................................................................................90

PAGE 10

x 5-2 Definition of parameters used during the parameterization stage of employing the CROPGRO-cotton model to simulate production of cotton under shaded environment..............................................................................................................90 5-3 Computed coefficient parameter values based on actual data of cotton collected in a pecan-cotton alleycroppi ng system in Jay, Florida..........................................90 5-4 Calibrated genetic coefficients of cotton grown in a pecan-cotton alleycropping system in Jay, Florida...............................................................................................91 5-5 Comparison of RMSE and d-statistics of simulated and observed LAI based on calibrated and default model values.........................................................................91 5-6 Effects of varying size of leaf of cotton (SIZELF) on LAI prediction....................91 5-7 Effects of varying SLA on LAI and biomass production.........................................92 5-8 Effects of partitioning on LAI and biomass production...........................................93 5-9 Effects of changing LFM AX in biomass prediction................................................94 5-10 Comparison of simulated and obser ved biomass of cotton in 2001 and 2002.........94

PAGE 11

xi LIST OF FIGURES Figure page 2-1 Average diurnal variation of light tran smittance to cotton A) in different rows and B) in different treatments in a pecan-cotton alleycropping system in northwest Florida......................................................................................................26 2-2 Mean photosynthetically active radiati on transmitted to cotton in barrier, non-barrier and monoculture treatment s in a pecan-cotton alleycropping system in northwest Florida.....................................................................................27 2-3 Relationship between LAI and light extin ction coefficient (k) in a pecan-cotton alleycropping system in northwest Florida..............................................................27 2-4 Relationship between LAI and absorb ed photosynthetically active radiation in a pecan-cotton alley cropping system in northwest Florida.................................28 2-5 Relationship between LAI and cotton lin t yield in a pecancotton alleycropping system in northwest Florida.....................................................................................28 2-6 Relationship between aboveground biomass production and cumulative absorbed PAR in a pecan-cotton alley cropping system in northwest Florida.........29 2-7 Relationship between lint yield produ ction and cumulative absorbed PAR in a pecan-cotton alleycropping system in northwestern Florida.................................29 3-1 Specific leaf nitrogen (SLN) conten t of cotton in non-ba rrier, barrier and monoculture treatments............................................................................................45 3-2 Light response curve of cotton in non -barrier, barrier and monoculture in a pecan-cotton alleycropping system.......................................................................46 3-3 Intercellular CO2 (A-Ci) curve of cotton in non-barrier , barrier and monoculture in a pecan-cotton alleycropping system...................................................................47 3-4 Relationship between specific leaf nitrogen (SLN) and leaf level net photosynthesis of cotton in non-barrier, barrier and monoculture treatments.........48 3-5 Relationship between canopy net ph otosynthetic index aboveground biomass production of cotton in non-barrier, barrier and monoculture treatments................49

PAGE 12

xii 3-6 Relationship between canopy net photos ynthetic index and lint yield production of cotton in non-barrier, barrier and monoculture treatments..................................50 4-1 Root and shoot biomass production of cotton in 2002 and 2003 growing seasons.71 4-2 Root length density of cotton in barrie r, non-barrier and monoculture treatments..72 4-3 Temporal changes in root length of cotton in barrier, non-barrier and monoculture treatments during 2003 growing season..............................................73 4-4 Relationship between root dry weig ht and root length of cotton in 2003 growing season.........................................................................................................74 5-1 Conceptual framework showing shad ing as it affects biomass production under the CROPGRO-cotton model.........................................................................95 5-2 Simulated SLA of cotton in each treatment after model calibration........................96 5-3 Simulated LAI of cotton in each treatment after model calibration.........................97 5-4 Simulated aboveground biomass of co tton in each treatment in 2001 and 2002 growing seasons.......................................................................................................98 5-5 Relationship between simulated and observed aboveground biomass of cotton in 2001 and 2002 growing seasons...............................................................99 6-1 Model showing the competitive vectors and their influence on biomass production in a pecan-cotton alleycropping system in Jay, Florida.......................104

PAGE 13

xiii Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy COMPETITIVE INTERACTIONS IN A PECAN ( Carya illinoensis K. Koch) – COTTON (Gossypium hirsutum L.) ALLEYCROPPING SYSTEM IN THE SOUTHERN UNITED STATES By Diomides Santos Zamora May 2005 Chair: Shibu Jose Major Department: Forest Resources and Conservation A research project was conducted at the West Florida Research and Education Center Research Farm of the University of Florida (Jay, Florida) to examine the competitive interactions involving light in a pecan-cotton alleycropping system. Polyethylene root barriers we re used to prevent belowgr ound interaction between pecan and cotton in half the numbers of test plots. Light distribution (int erception and absorption) was gr eatly affected by leaf area index of both pecan and cotton. Interspecifi c competition resulted in varying leaf morphology (e.g., specific leaf area) resulting in variati ons in canopy net photosynthesis ( Pnet). Eliminating belowground competition via the barrier treatment resu lted in a tri-fold increase in Pnet over the non-barrier and was comparable to the Pnet in monoculture. Despite 50% shading in the alleys, the barrie r treatment had similar biomass to that of monoculture. Although cotton yield was si milar for the monoculture and barrier treatments during the first year of the study, a slight decrease was noted in cotton yield in

PAGE 14

xiv the barrier treatment in the second year. Biom ass and lint yield always remained lower in the non-barrier treatment compared to the other two treatments. Cotton root morphology was also affected by interspecific competition. Shading resulted in the allocation of more carbon to the abovegr ound components at the expense of the root systems, resulting in lower root :shoot ratio compared to the monoculture. Plants in the non-barrier treatment exhibited 25% and 33% reduction in total root length compared to barrier and monoculture, respec tively. Similar reduction in root length density was observed in comparison to the othe r two treatments. Results also revealed significant curvilinear relationships between root length and root biomass regardless of treatment, but the magnitude of relationshi p varied, with non-barrier plants producing significantly lower root length compared to the barrier and monoculture for the same amount of carbon. Results of simulation modeling indicated that the CROPGRO-cotton model can be used to predict cotton biomass under varying light levels. Corr elation analyses indicated a significant relationship between measur ed and simulated aboveground biomass (R2=0.95 and R2=0.92, respectively for 2001 and 2002). Results from this study can be used to improve system design and management techniques of pecan-cotton and similar alleyc ropping systems in the temperate region.

PAGE 15

1 CHAPTER 1 INTRODUCTION Statement of the Problem In the United States, interest by resear chers and landowners in agroforestry has recently begun to escalate. However, adop tion is hampered by the apprehension of landowners, whose cultural pract ices do not typically employ tree-crop systems, and who are unfamiliar with the management skills n eeded for such complex systems (Zinkhan and Mercer, 1997). Agroforestry, which is perceived as a pr actice that can provide optimum production while maintaining ecological sustainability, can enhan ce many of the biophysical cornerstones of ecologically-sound agri cultural production (Gordon et al., 1997). However, much work is needed to docume nt such potentials for greater adoption in temperate regions. Agroforestry research in the past has focused on identifying advantages and limitations of agroforestry sy stems, and mechanisms of system dynamics. Many of the mechanisms in agroforestry ha ve been identified but not quantified and hence remain only partially understood. Although agroforestry research in the temperate region is growing, only a few attempts have been made to examine aboveground production dynamics of such systems. The present study was conducted to explore competitive interactions involvi ng light in a temperate alle ycropping system with pecan ( Carya illinoensis K. Koch) and cotton ( Gossypium hirsutum L.) in the southern United States. The specific objectives were the following: 1. Determine light distribution in pecan a lleys and its effect on cotton production,

PAGE 16

2 2. Determine the production physiology of cotton as influenced by interspecific interactions, 3. Examine the morphological plas ticity of cotton roots in response to interspecific competition; and 4. Apply a process-level mode l to examine cotton production as affected by varying levels of light. Review of Literature Temperate Alleycropping The practice of alleycropping involves the cultivation of woody perennials, usually for nut or timber production, in rows, while crops are cultivated between the rows (United State Department of Agriculture [U SDA], 1999). In the southern United States, pines ( Pinus spp.) have been intercropped w ith row crops such as cotton ( Gossypium spp.), maize ( Zea mays L.), soybean ( Glycine max L. (Merr)), wheat ( Triticum spp.) and oats ( Avena spp.) (Zinkhan and Mercer, 1997; Ramsey and Jose, 2001). Pecan ( Carya illinoensis K. Koch), an important nut-bearing sp ecies, has been intercropped with cotton, soybean, squash ( Cucurbitaceae spp.), potatoes ( Solanum tuberosum ) and various grains and other crops (Zinkhan and Mercer, 1997; Ramsey and Jose, 2001). Generally, species components and management in alleycropping sy stems can be varied to achieve a variety of objectives, primarily, but not limited to, soil management and increases in system productivity (Nair, 1993). As an association of plant communities, alleycropping is deliberately designed to optimize use of spatial, temporal a nd physical resources by maximizing positive interactions (facilitation) and minimizing th e negative ones (competition) between trees and crops (Jose et al., 2000a, 2000b; van Noordwijk and Luciana, 1999, 2000). More often, alleycropping is considered by many to hold potential as a vi able and profitable land-use system in the United States. Much of the research today in temperate

PAGE 17

3 alleycropping has centered around systems in th e southern and midwestern U.S. Research results from these systems show that prof itable agronomic crop production can occur for 10 or more years for shade-intolerant sp ecies, depending on alley width, and longer for shade tolerant species (Garrett and Buck, 1997). The southern United States alone ma y have greater potential to practice agroforestry in general and alleycropping in particular, than othe r regions. For one, the longer growing season in the South presents more possibilities for agroforestry adoption and promotion. However, much work is need ed not only to addre ss the apprehensions over its adoption due to uncerta inties and risks, but also to showcase the potentials of such systems for environmental protec tion and sustainable production of good and services. Major Factors and Mechanisms in Aboveground Interactions Productivity of any agroforestry system is to a large degree the net result of positive and negative interactions among the tree and agronomic crop components. Interactions occur as compone nt species strive to capture growth resources above and belowground (Ong et al., 1996). The likelihood a nd intensity of inters pecific interactions decline with decreases in or ganism density until a maximu m yield is reached (Kropff, 1994). Beyond any maximum-yield density, inter actions among plants occur when two or more organisms attempt to capture resources from the same location (temporally or spatially) (Monteith, 1994). The physical and phenological differences of system components can lead to an intensified interact ion for capture of the limiting resource(s) of a particular agro-ecological system (Ong et al., 1996). Greater capture of the limiting resources would be accompanied by an increas ed ability to utilize nonlimiting resources, which by definition are available but underutilized (Canne ll et al., 1996).

PAGE 18

4 Understanding of the biophysical proce sses and mechanisms of capture of resources in alleycropping systems is, thus, necessary. One such mechanism is the aboveground competition for light. Availability of light and its effect upon aboveground production are equally important to agricultura l system sustainability. When plant growth is not limited by water or nutrients, producti on is limited by the amount of radiant energy that the foliage can intercept (Monteith et al., 1991; Kropff, 1994). The effect of light on dry matter production has been studied extensively in alleycropping systems (Ramakrishna and Ong, 1994; Koslowski and Pallardy, 1997; Lambers et al., 1998; Gillespie et al., 2000), and most of these studies revealed strong linear relationship between biomass of the understory species a nd the amount of light intercepted. Biomass growth is dependent upon the fraction of incident PAR (400-700 nm) that each species intercepts, and the efficiency with which the intercepted radiation is converted by photosynthesis (Ong et al., 1996). These factors, in turn, ar e influenced by time of day, aspect, temperature, CO2 level, species combination, photosynthetic pathway (C3 vs. C4), canopy structure, plant age and height, l eaf area and angle, and transmission and reflectance traits of the plant canopy (Mont eith, 1978; Brenner, 1996; Kozlowski and Pallardy, 1997). Trees change the understory environment in such ways as reduced radiation availability (Beer et al., 1998), thus affecting the co mpanion crops. As PAR passes through a tree canopy, it is s ubject to changes in quanti ty (energy) and quality (wavelength) (Kozlowski and Pallardy, 1997) . Plants growing under reduced levels of PAR often show different growth responses in low light than at hi gher levels, though the nature and extent of adaptation varies among species. Reduced radiation poses stresses on

PAGE 19

5 plants because irradiance limits photosynthe sis and thus net carbon gain and plant growth. On the other hand, high light intens ities may also be a stressor for plants, particularly if other factors are not optimal. Optimal levels of PAR to maximize the net assimilation rate are still unclear in alleycropping systems. As indicated earlier, lo w light availability re duces photosynthesis with consequences on production. Chirko et al. ( 1996) noted that production of wheat grown at farther distance from the tree (2.5 m) in a temperate alleycropping in China generated 45.7 kg ha-1 more yield than wheat planted 2.0 m away from the tree line. Droppelmann et al. (2000) also reported that yield of sorghum ( Sorghum bicolor L.) in row position closest to the tree row was lower than that in rows farther from the tree. A similar observation was also made by Ni ssen et al. (1999), who fou nd that shading reduced the yield of cabbage ( Brassicas oleracae L.) under a Eucalyptus camaldulenses alleycropping system in the Philippines. Th ese authors concluded that the amount of radiation intercepted caused row yield diffe rences. However, Gillespie et al. (2000) reported that irrespective of high correl ation between PAR and net photosynthesis, reduced light level did not have a ma jor influence on the yield of corn ( Zea mays L.) planted with black walnut ( Juglans nigra L.). Root Dynamics A fundamental hypothesis in agroforestry systems is that different plants occupy different soil strata with their respectiv e root systems. Knowledge of the spatial distribution and density of tr ee-crop root systems is necessa ry in order to understand both above-and-belowground pro cesses including assessmen t of the degree of complementarity and competition among sy stem components (Schroth, 1999). The number of roots or the root length density is an important factor that determines the

PAGE 20

6 dynamics of belowground competition. The potential of plants to compete for soil nutrients and water resources from the same soil is proportional to their root length density (Bowen, 1985; Schroth and Z ech, 1995; Livesley et al., 2000). Belowground competition is most likely to occur when two or more species have developed a specialized root sy stem that directs them to explore the same soil strata for growth resources (Van Noordw ijk et al., 1996). This can be problematic even in mixedspecies systems. Researchers in the temperat e zone, humid tropics and semiarid tropics have reported observing the grea test concentration of tree root density within the top 30 cm of soil, the region predominantly explor ed by crop rooting systems (e.g., Itimu and Giller, 1997; Lehmann et al., 1995; Nissen et al., 1999; Immo and Timmer, 2000; Jose et al., 2000a, 2000b) to about 50 cm soil depth (R ao et al., 1998) where severe water and nutrient competition exist. Even in a system where complementary interactions are taking place, some amount of competition can be expected, since each species is vying for resources from the same finite pool (Ong et al., 1996). The root syst ems of all components in an intercropping study in semiarid Kenya, consisting of maize and tree species Grevillea robusta (Cunn.) and Gliricidia sepium (Jacq.), were found to occupy most heavily the top 20 cm of soil and decreased in density with depth. Howe ver, there was some degree of temporal separation in rooting patter ns since the tree roots decreased 71% and 54% (for Gliricidia and Grevillea , respectively) by the end of the rain y season, when maize root density was at its highest. Chirwa et al. (1994) and Le hmann et al. (1998) al so observed spatial separation of roots of Acacia saligna and sorghum in their individual studies. Acacia developed a greater root system in a d eeper layer of soil to minimize or avoid

PAGE 21

7 competition for water with sorghum, which deve loped greater root density at the upper soil layer. Interspecific competition for resources can greatly affect partitioning of biomass between shoot and roots. Root-shoot ratio tends to respond to environmental growth conditions. When light becomes a limiting f actor, plants partition more carbon to aboveground component resulting in lower r oot-shoot ratio. However, when water is limiting, photosynthates are diverted belowgr ound for root development, resulting in higher root-shoot ratio, in order to enhance cap ture of resources, as observed for maize grown with Leucaena leucocephala (Lam.) in semiarid India (Klepper, 1991; Govindajaran et al., 1996). This mechanism, however, can occur differently in temperate regions where tree pruning is not a practi ce and where shady environments may thus ensue. Water competition can lead the crop co mponent to allocate substantial amounts of carbon, not to root, but to shoot in respons e to limiting light (Smith and Huston, 1989; Sack and Grubb, 2002; Jose et al., 2002). Modeling in Agroforestry Modeling is becoming an integral part of agro forestry research as scientists seek to understand the complexities of agroforestry sy stems. As a complex system, agroforestry inevitably experiences intera ctions between and among syst em components, as earlier presented, and the effects of these interac tions should be thoroughl y quantified. Most of the research conducted on tree-crop functioni ng in agroforestry systems has revealed significant relationships (eit her positive or negative) among system components with regard to their use of available growth res ources (i.e., light, water, and nutrients). Only few attempts, however, have been made to quantify the relationshi ps holistically through model application.

PAGE 22

8 Agroforestry models, when used appropriate ly, can assist in evaluating agroforestry alternatives, testing research hypotheses and understanding the processes and interactions of system components (Jagtap and Ong, 1997; Mu etzelfeldt and Taylor , 1997). In recent years, several integrated computer-based ag roforestry models have been developed in response to the need for handling the many social, economic and ecological variables encountered in dealing with complex agrofo restry systems. Among others, these models include WaNuLCAS (water, nutrients, light capture in agroforestry systems; Van Noordwijk and Lucian, 1999, 2000); SCUAF (soil changes in agroforestry; Young and Muraya, 1990); and HyCAS (simulating comp etition for light, water and nutrients in Cassava; Matthew and Lawson, 1997). The WaNu LCAS model, for instance, is an integrated model of tree-cr op interactions based on above and below-ground resource capture and competition for water, nutrient s and light under different management scenarios in agroforestry systems. Integr ated modeling accounts for the numerous combinations possible between pl ants and environments. It is, thus, necessary that models should support experimentation as much as possible, by (or through) testing and predicting the most suita ble plant associations. The use of models to predict production ha s been used in several environments. Mayus et al. (1999) found a reasonable agr eement between simulated and measured soil water content and dry matter production of millet ( Pennisetum glaucum (L.) R. Br.) planted under a windbreak system ( Bauhinia rufescens Lam) in Sahel, Niger, using the WIMISA (Winbreak Millet Sahel) model. Mayu s et al. (1999) concluded that the model was appropriate for analyzing competition for light and water between windbreaks and crops. Further, in another study, after calib ration of the Hydrus-2D model applied in

PAGE 23

9 Amazonian agroforestry system, Schlegel et al. (2004) showed a good conformity between simulated and measured water uptake by Pueraria phaseoloides (Roxb) and Bactris gasipeas H. B. K intercropped together . Also, application of the APSIM (agricultural production system simulator) mode l in a hedgerow intercropping system in the Philippines showed that the model gave a cceptable prediction of maize yields and soil loss when compared to the actual data (Nel son et al., 1998). APSIM accurately predicted the fluctuation in maize yield associated with seasonal climatic variation and environmental soil conditions. Traditional agroforestry a nd agronomic experiments are conducted at particular time and space, making results siteand season-specific, as well as time consuming and expensive. Unless new data and research findi ngs are put into formats that are relevant and accessible, they may not be used eff ectively. The DSSAT (decision support system for agrotechnology transfer) (Tsuji, 1994; Jones et al., 1998, 2003) model, which is widely used in pure agricultural systems, can also be applied in agroforestry systems. DSSAT, along with the agrofore stry models earlier mentione d, can be used to integrate knowledge about systems’ biophysical char acteristics and necessary management regimes for making better decisions for tr ansferring production technology. The DSSAT model has been widely adapted to better cont rol and manage the pa rticular system of interest. An improved understanding of tree-crop interactions on cap ture and use of resources would provide a greater scien tific basis for developing appropriate recommendations and strategies to improve or enhance productivit y. While the adoption of agroforestry in temperate regions remains critical, there is not much understanding of

PAGE 24

10 the interactive dynamics of light in such systems and the resulting impacts on system yield and sustainability. Fr om an agronomic standpoint, the effects of tree-crop interactions must ultimately be considered in light of the efficiency of the component species to yield under limiting light environments.

PAGE 25

11 CHAPTER 2 INTERSPECIFIC COMPETITION IN A PECAN-COTTON ALLEYCROPPING SYSTEM IN THE SOUTHERN UNITED STATES: IS LIGHT THE LIMITING FACTOR? Introduction The manner in which light is intercep ted by crop canopies and converted to structural dry matter can signi ficantly affect primary producti on at a given site. A number of authors have investigated plant performance under diffe rent environmental conditions, including different levels of light, in alle ycropping and similar agroforestry systems (Azam-ali et al., 1991; Monteith et al., 1991; Rosenthal and Gerik, 1991; Heitholt et al., 1992; Gilliespie et al., 2000; Jose et al ., 2000a, 2000b). These studies have revealed strong linear relationships between phot osynthetically active radiation (PAR, 400-700 nm) and dry matter production. Plants that develop under low levels of PA R such as in agroforestry systems grow and develop differently than plants grown under full sun (Monteith et al., 1991; Lambers et al., 1998). The amount of intercepted PA R becomes the major determinant of biomass production when belowground resources are not limiting. This relationship has been conceptualized as the time in tegrated product of three f actors (Monteith et al., 1991): W = .i.Q. dt equation (1) where W is crop biomass (Mg ha-1), is the radiation use effi ciency (RUE), amount of biomass produced per absorbed light, i is the incident PAR in tercepted by the canopy (MJ m-2) and Q is the PAR incident at the top of the canopy (MJ m-2) . Light interception by plants has been shown to be affected by several factors (Beer et al., 1998; Bellow and

PAGE 26

12 Nair, 2003). These factors includ e leaf area, spatial distribu tion of leaves, crown height and diameter among others. Crop growth and development in alleycropping systems depend on the intensity and availability of light. As such, how much light is captured and how efficiently it is used to create dry matter must be considered in the design and management of alleycropping systems . Understanding the temporal and spatial variations in light transmittance and subsequent crop production is of great importance in this context. Hence, the objectives of this study were to (1) quantify the spatial and temporal distribution of light in an alleycropping system involving pecan ( Carya illinoensis K. Koch) and cotton ( Gossypium hirsutum L . ) and (2) determine its effect on the productivity of cotton. Our primary hypothesis was that cotton with its characteristic C3 photosynthetic pathway would perf orm well under shade if light levels in the alleys were above the light saturation point and belowg round competition for water and nutrients was alleviated. We further hypothesized that co tton grown in alleycropping might exhibit higher RUE than that of monoculture cotton due to competition for light between system components. Materials and Methods Study Area The study was conducted in a 50 -yr old pecan orchard converted into an alleycropping system, located in Jay, Flor ida, USA (30 N, 87 W). The climate is considered temperate with moderate wint ers and hot humid summers. The soil is classified as a Red Bay sandy loam and described as a fine-loamy, siliceous, thermic Rhodic Paleudult.

PAGE 27

13 The pecan trees were planted at a uni form spacing of 18.3 m and remained under grass cover for 29 years until the initiati on of the current study. Ten plots were established within the orchard and arranged into five blocks using a randomized complete block design in spring 2001. Each plot, which co nsisted of two rows of trees oriented in a north-south direction, was 27.4 m long and 18.3 m wide, with a practical cultivable width of 16.2 m, and was separated from its adjacent plot by a buffer of the same dimensions. Each block was randomly divided into a barrie r plot and a non-barrie r plot. Barrier plots were subjected to a root pruning treatment in which a trenching machine was used to dig a 0.2 m wide and 1.2 m deep trench along both sides of the plot at a distance of 1.5 m from the trees to separate root system s of pecan and cotton. A double layer of 0.15 mmthick polyethylene sheeting was used to line th e ditch prior to mechanical backfilling. The barrier plots (referred to as barrier treatment or barrier pl ants) thus served as the tree root exclusion treatment, preventing inter action of tree and cotton roots, while the nonbarrier plots (referred to as non-barrier treat ment or non-barrier plants), which did not receive this treatment, served as the treecrop competition treatment. Monoculture plots (referred to as monoculture treatment or m onoculture plants) were also established to compare production with barrier and non-barr ier treatments (Allen, 2003; Wanvestraut et al., 2004). Sixteen rows of cotton, one meter apar t, were planted in each alley. Cotton (DP458/RRvariety) was plante d in a north-south orientat ion on 16 May 2001 and 13 May 2002 after disking the alleys. PAR and Radiation Use Efficiency Two 0.8 m Decagon Ceptometers (Decagon, De vices, Inc., Model SF-80, Pullman, WA), consisting of 80 PAR sensors with each sens or placed at a 1 cm interval, were used

PAGE 28

14 to measure incoming, transmitted, and reflected PAR (400-700 nm) in the alleys. Incoming PAR (Qi) was measured right above the cotton canopy. Diurnal transmission of incoming radiation to cotton plants at rows 1, 4, 8, 13, and 16 was measured every hour from 7:00 a.m. to 6:00 p.m. Measurements were made twice a month from June to October 2001. Light measurement started immediately , two weeks after the cotton plants emerged. The two Ceptometers were used simultaneously to measure Qi in external Rows (Row 1 and 16), the intermediate rows (Rows 4 and 13) and then the middle row (Row 8) to ensure minimal variation in light readi ngs among the rows for the specific time of measurement. Ten random sample light r eadings along each row were recorded and averaged in each plot. Incoming radiation outside the orchard was also measured one meter above the ground before and after measuring Qi for each row. An inverted Decagon Ceptometer locate d 1.0 m above the cotton also measured canopy reflected radiation. Reflect ed radiation in each row at the time of measurement was taken and then averaged. Light transmittance and reflection were measured on clear sunny days. The transmission coefficient, k , for cotton growing in rows 1, 4, 8, 13 and 16 was calculated based on the Beer-Lambert law. Absorbed PAR by the cotton canopy in each row of both alleys was then determined from the calculated k , reflected PAR, and calculated LAI values: APAR = (Qireflected PAR) x (1 – E xp (-k* LAI) equation (2). Litterfall was collected using 1 m x 0.5 m 2 mm screen litter traps. Four litter traps were randomly placed in each pl ot, of which one litter trap was installed per row. Litter

PAGE 29

15 traps were also installed under pecan tr ees to collect pecan foliage. Litter was collected twice a month from August to November 2002. Leaves collected from litter traps were separated by species (pecan and cotton) and were stored, oven-dried at a constant temperature (70oC) , and then weighed. The litterfall and specific leaf area (SLA; leaf area per unit weight; describe d below) data were used to calculate the LAI of cotton plants in each row. SLA of cotton was determined monthl y in 2001 and 2002 by collecting six fully expanded leaves in each row. SLA of pecan was determined by harvesting 20 leaves each of sun and shade in August 2001 and August 2002, during the peak of pecan growth. Twenty-four pecan trees in the orchard and three trees in the m onoculture pecan were sampled for SLA. Leaf area was determined using a leaf area meter (Li-Cor, Lincoln Nebraska), oven-dried for three days at 70oC and weighed. RUE (g MJ-1) of cotton was determined for 2001 and 2002. Daily Absorbed PAR by cotton, measured twice a month, was determined based on the diurnal ( 7:00 am to 6:00 pm) readings of PAR. Biomass and Lint Yield In 2001, aboveground biomass of cotton was harvested at physiological maturity. In 2002, aboveground biomass was quantified m onthly, from July to October. Whole plants (separated into leaves, stem, and bolls) were harvested in 1m x 1m subplots in each row in each plot. Harvested plan ts were dried for 72 hours at 70oC, and weighed. Biomass was expressed on a per area (m2) basis. Lint yield of cotton in each row (rows 1, 4, 8, 9, 13 and 16) in each treatment as well as in the sole stand (monoculture) was quantified by harvesting two random strips of

PAGE 30

16 0.61 m. x 6.1 m. in each row. Lint dry we ight was determined following oven drying (70oC) for 48 hours. Data Analysis Statistical analyses were performed usi ng the proc-mixed procedure within the framework of Split block design (SAS Institute, Cary, NC). The Shapiro-Wilk’s test was used to test for normality of distribution. A logarithmic (log(x+1)) transformation was performed to improve normality when necessary. Least square mean differences were performed to determine significant differences of the means at = 0.05. Results Incident and Absorbed PAR Light availability inside the pecan alley was affected by the LAI of pecan, which varied by treatment. Mean LAI of pecan in the barrier treatment (3.64) was 17% lower than that in the non-barrier treatment (4.39) (Table 2-1), resulting in 25% higher average growing season daily incident light transmittan ce for the barrier cotton plants (Figures 21 and 2-2). In general, the pecan trees caused about 50% reduction of incoming incident light to cotton plants compared to the da ily average light receiv ed by the monoculture plants. Diurnal changes in spatial variation (resul ting from row location) of incident PAR is illustrated in Figure 2-1a. Irrespective of the barrier treatment, incident light transmittance by row changed with time of the day, with rows situated on the eastern part of the alley (Rows 16 and 13) receiving more light during the morning hours while rows located on the western side (Rows 1 and 4) were shaded. However, this pattern was reversed in the afternoon, with eastern rows being shaded and western rows receiving greater amount of PAR. At midday, incident PAR was high in Row 8 (middle row) and

PAGE 31

17 remained high until mid-afternoon while all other rows also received high levels of PAR (Figure 2-1a). LAI of cotton differed significantly among treatments. Cotton LAI values ranged from 1.72 for the non-barrier treatment to 3.15 for the barrier treatm ent. Lower LAI in non-barrier plants resulted in less light absorption. Differences in light extinction coefficients and absorbed PAR were also noted among treatments (Table 2-1). Cotton in the barrier treatment had greater amount of light absorbed and hi gher light attenuation compared to the non-barrier plants. The mean light extinction coefficient in monoculture plants was 17.9% lower than in barrier plan ts, but 30.1% higher than that in the nonbarrier plants (Table 2-1). Although LAI of monoculture cotton was 15.2% lower than that of the barrier treatment, monoculture cotton exhibited gr eater light absorption due to higher incident PAR. Light extinction coefficient showed a si gnificant, but weak negative correlation with LAI (R2 = 0.43) (Figure 2-3). PAR absorbed by cotton also exhibited significant curvilinear relationships with LAI (R2 = 0.61 and R2 = 0.78 for cotton growing in the orchard and in monoculture, respectively) (Figure 2-4). Biomass and Lint Yield Cotton in the barrier treatment produced 60% higher biomass compared to nonbarrier treatment, but was statistically sim ilar to biomass produced in monoculture both years of the study (Table 2-2). In 2002, there was an average 45% decline in aboveground dry matter across all treatments. Biomass in non-barrier treatment was 39.5% and 36.2% lower than that in the barrie r and monoculture treatme nts, respectively. In 2001, inter-row difference in abov eground biomass was significant ( P = 0.0038) in the non-barrier trea tment. Aboveground biomass increased in Row 8 ( P = 0.0014) by

PAGE 32

18 39% over Row 1 and by 15% compared to Row 4 ( P = 0.0091). Inter-row variation in 2002 was not significant in either the barrier or the non-barrier treatments (Table 2-2). Restricting belowground competition had an impact on cotton lint yield both years and lint yield differed significantly betw een 2001 and 2002. In 2001, lint yield in the barrier treatment (70.04 g m-2) was higher than that of th e non-barrier treatment (51.54 g m-2) ( P = 0.0324), but was not different from the monoculture treatment (69.01 g m-2). In 2002, lint yield in barrier treatment was agai n higher than the non-barrier treatment but lower than monoculture. Inter-ro w variation in lint yield wa s not significant for the nonbarrier treatment. However, the presence of the barrier had the grea test impact on plants in row 1 resulting in greater yield compared to the intermediate and middle rows. LAI and lint yield showed a significant curvilinear relationship in our experiment (R2 = 0.45; P < 0.0001) (Figure 2-5). Maximum lint yield was obtained when LAI was between 3.0 and 4.0. Increase in LAI beyond 4.0 di d not result in an increase in lint yield. Radiation Use Efficiency Cotton aboveground biomass and lint yield were both influenced by levels of cumulative absorbed PAR. Although R2 values (R2 = 0.44 and R2 = 0.41, respectively for 2001 and 2002) were low, the relationship between aboveground biomass and PAR was significant and linear (Figure 2-6). Similarly, lint yield of cotton e xhibited significant and strong curvilinear relationship w ith cumulative absorbed PAR (R2 = 0.61, R2= 0.58) (Figure 2-7). Apparently, maximum lint yield (90 g m-2 for barrier treatment and 70 g m-2 for non-barrier) was achieved at approximately 500 MJ m-2 and 400 MJ m-2, respectively, for the barrier and non-barrier treatments (Figure 2-7).

PAGE 33

19 Significantly higher leaf area of the barrier plants that captured more light resulted in 31% and 52% higher RUE than that in the non-barrier and monoculture plants (Table 2-3). RUE in 2001 did differ significantly among treatments ( P = 0.0017). However, there was an average 47% re duction in RUE in all treatme nts in 2002 compared to 2001 (Table 2-3). Monoculture plants had the lowe st RUE in both years and were statistically similar to that of the non-b arrier plants in 2002. Inter-row variation in RUE was nonsignificant both years in all the treatments except for non-barrier plants in 2001. In the non-barrier treatment, RUE of plants in Row 1 was significantly lower than that in rows 4 and 8 ( P = 0.0002) (Table 2-3). Discussion Light has been identified as one of the major limiting factors influencing production in many agroforestry systems (Mont eith et al., 1991; Corlette et al., 1992; Nair, 1993; Chirko et al., 1996; Jose et al., 2004) incl uding temperate (Gordon and Newman, 1997; Gillespie et al., 2000) and tropical (Lawson and Kang, 1990; Karim et al., 1993; Nissen et al., 1999) alleycropping. In all these studies, decrease in incident light resulted in lower crop production. In our study, aboveground biomass and yiel d of cotton were strongly affected by the amount of light absorbed by cotton. The am ount of light absorbed, in turn, was a function of both the amount of incident light and cotton leaf area. Although the absorbed PAR was 42% lower for the barrier plants co mpared to the monoculture plants in 2001 (Table 2-1), lint yield was similar for bot h treatments. This clearly supports our hypothesis that cotton can grow and yield reasonably well under moderate shade (50% shade in the barrier compared to monoculture, Figure 2-2). However, as hypothesized, if belowground competition for water and nutrients existed (as in the non-barrier treatment),

PAGE 34

20 PAR capture was lower because of reduction in cotton LAI, hence resulting in lower yield. LAI of plants in the non-barrier trea tment was 45% lower than in the barrier treatment. Barrier plants outperformed non-barr ier plants in both y ears with nearly 40% and 60% higher biomass and yield. Lower p ecan leaf production and self-shedding that took place in both years (personal observat ion) also resulted in slightly higher transmission of incident light to the barrier plants compar ed to the non-barrier plants. LAI has long been recognized as an indicator of plant productivity. Although, regression analysis showed a weak relationshi p between LAI and yield in our experiment (R2 = 0.45), the relationship was still significant ( P < 0.0001) (Figure 2-5). Rosenthal and Gerik (1991) reported a simila r, but stronger relationship (R2 = 0.90) between absorbed PAR and lint yield for cotton grow n under irrigated conditions. Cotton plants in our system attained maximum yield (approximately 65 g m-2) between LAI values of 3.0 and 4.0, which is in agreement with He itholt et al. (1992) who observed maximum yield between the same range of LAI. Heithol t et al. (1992) further concluded that this range of LAI provided the optimum absorption of incident light by cotton, which is also in agreement with our results (Figure 2-4). The trade-off hypothesis (Smith and Hust on, 1989) states that plants grown under shade tend to preferentially allocate carbon in building larger canopies, for greater capture of light, at the expense of root syst ems (Kozlowski and Palardy, 1997; Jose et al., 2002). Despite shading, there was no such incr ease in leaf area in the non-barrier treatment compared to the monoculture tr eatment. However, eliminating belowground competition resulted in larger canopy (h igher aboveground biomass (Chapter 4)) and higher LAI for the barrier plan ts in response to shading. Similar results have been

PAGE 35

21 reported before. For example, Zhao and Ooster hius (1998) noted in th eir experiment that cotton under shade expanded their leaves resulting in larger leaves and higher LAI. Increasing leaf area by the plants enhances th e ability to capture more light under light limiting conditions. As expected, distance from tree rows had an impact on the growth of non-barrier plants, affecting their biomass and yield. For the non-barrier plants, any benefit from the edge effect (increased development due to lack of intraspecific competition on one side) was not detected. In stead, there was a trend of decreasing LAI and yield with closer proximity to the tree ro w. This reaffirms the earlier findings from the same study site that competition for wate r is perhaps intense in the non-barrier treatment compared to the barrier treatment (Wanvestraut et al., 2004). RUE is an indirect expression of the photos ynthetic capacity of pl ants at the whole plant level (Muchow and Sincla ir, 1993; Bennett et al., 1993). The barrier plants had 30% higher efficiency in utilizing light and conve rting it into biomass in both years compared to the non-barrier plants. In 2002, RUE in non-ba rrier plants was statistically similar to that of monoculture plants. Lower light interception, coupled with competition for belowground resources in the non-barrier tr eatment, affected biomass production and consequently RUE by the non-barrier plants (Table 2-3). With high levels of light available for growth, monoculture plants ex hibited about 50% lower RUE compared to the barrier plants. The values we observed for RUE (0.71 to 2.37 g MJ-1) are within the range of published values for C3 plants. Kiniry et al. ( 1989) found RUE ranging from 2.0 to 3.0 g MJ-1 while Rosenthal and Gerik (1991 ) found RUE values of 1.3 1.5 g MJ-1 for cotton grown in a narrow-row planting configuration.

PAGE 36

22 In addition to light, competition for wa ter and nutrients can also affect RUE through their effect on plant growth. For exam ple, Bange and Milroy (1998) showed that cotton fertilized with 150 kg ha-1 of N had higher RUE (1.07 g MJ-1) than cotton receiving only 113 kg ha-1 N (0.89 g MJ-1). Sinclair and Horie (1983) found that foliar nitrogen was positively corre lated to RUE of cotton grow n under open field conditions. The decline in RUE from 2001 to 2002 in our syst em could also be attributed to decline in soil nutrient status. Allen (2003) reporte d a significant decreas e in soil nitrogen mineralization rate in our system from 2001 to 2002 growing season, which was caused by a declining fallow effect. Conclusions Despite having lower light transmittance (about 50% of outside PAR) in the alleys, cotton aboveground biomass was comparable to monoculture in both years. It is reasonable to assume that light is not a lim iting factor in the production of cotton in our alleycropping system. Cotton tolerated mode rate shade and provided acceptable yield when belowground competition was alleviate d. Results also revealed a curvilinear relationship between light absorbance and lint yield. Light absorb ance, in turn, was influenced by LAI, which varied signifi cantly among treatments. The optimum LAI (3.0 to 4.0) for maximum light absorbance and lint yield was observed in both the monoculture and the barrier treatments, indicating that competition for belowground resources played a major role than competition for light in this particular system. The results offer promise for establishing alleycr opping systems in new or existing nut or fruit orchards by planting C3 crops in the alleys. However, management strategies such as early root training or root pruning need to be explored so that belowground competition for resources could be alleviated.

PAGE 37

23 Table 2-1. Leaf area index (LAI), light exti nction coefficient (k) and mean absorbed photosynthetically active radiation (PAR ) of cotton in non-barrier, barrier and monoculture treatments. Treatment LAI k APAR (mol m-2 s-1) Cotton Pecan Non-Barrier 1.72 4.39 0.51 541.44 Barrier 3.15 3.64 0.89 765.00 Monoculture 2.67 0.73 1330.65

PAGE 38

24 Table 2-2. Aboveground biomass production of cotton in barrier, non-barrier and monoculture treatments in 2001 and 2002 growing seasons. Treatment Row Aboveground biomass % change (kg ha-1) Year 2001 Year 2002 No Barrier 1 284.17bc1194.65a 32 (31.22)2(34.61) 4 336.27b201.11a 40 (34.38)(34.06) 8 468.27a192.17a 59 (37.92)(23.78) Mean3 362.91B195.98B 46 (27.71)(17.47) p value5 0.00380.9403 Barrier 1 526.09a294.35a 44 (70.91)(30.70) 4 576.53a292.54a 49 (44.99)(49.27) 8 622.45a396.44a 36 (56.10)(73.08 Mean3 575.02A323.74A 44 (32.89)(30.02) p value5 0.30210.1825 Monoculture Mean3 545.82A307.09A 44 (21.23)(11.81) p value4 0.0020 0.0303 1 Within-treatment values followed by the same lo wercase letter are not significantly different at the 0.05 level of probability 2 standard error of the mean are given in parenthesis 3 Mean indicates the treatment means 4 p value indicated significance between treatment means 5 p value indicated significance among rows in specific treatment

PAGE 39

25 Table 2-3. Radiation use efficiency of co tton in barrier, non-barrier and monoculture treatments in 2001 and 2002 growing seasons. Treatment Row Radiation Use Efficiency (g MJ-1) % change Year 2001 2002 No Barrier 1 1.07c10.71a 34 (0.16)2(0.20) 4 1.53b0.92a 40 (0.09)(0.12) 8 2.13a0.9a 58 (0.13)(0.14) Mean3 1.57B0.84B 46 (0.19)(0.12) p value5 0.00020.4115 Barrier 1 1.99a1.09a 45 (0.22)(0.10) 4 2.41a1.11a 54 (0.40)(0.14) 8 2.37a1.38a 42 (0.27)(0.20) Mean3 2.26A1.19A 47 (0.20(0.12) p value5 0.32010.1928 Monoculture Mean3 1.09C0.58B 49 (0.04)(0.02) p value4 0.0017 0.0031 1 Within-treatment values followed by the same lo wercase letter are not significantly different at the 0.05 level of probability 2 standard error of the mean are given in parenthesis 3 Mean indicates the treatment means 4 p value indicated significance between treatment means 5 p value indicated significance among rows in specific treatment

PAGE 40

26 Figure 2-1. Average diurnal variation of light transmittance to cotton A) in different rows and B) in different treatments in a pecan-cotton alleycropping system in northwest Florida. Photosnthetically Active Radiation (mol m-2 s-1) 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Row 1 Row 4 Row 8 Row 13 Row 16 Outside PAR 0 200 400 600 800 1000 1200 1400 1600 1800 2000 789101112123456 Time (hour) No-Barrier Barrier Outside PARA B

PAGE 41

27 0 200 400 600 800 1000 1200 1400 1600 1800 2000 BarrierNo-BarrierOutside PARTreatmentPhotosynthetically Active Radiation (mol m-2 s-1 ) Figure 2-2. Mean photosynthetically active radia tion transmitted to cotton in barrier, nonbarrier and monoculture tr eatments in a pecan-cotton alleycropping system in northwest Florida. Figure 2-3. Relationship between LAI and light extinction coefficient (k) in a pecancotton alleycropping system in northwest Florida. Figure 3 y = 1.1058e-0.203xR2 = 0.43 p < 0.00010.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.60.00.51.01.52.02.53.03.54.04.55.0 Leaf Area Index Light extinction coefficient (k) Barrier Non-Barrier Monoculture

PAGE 42

28 Figure 4 y = -29.946x2 + 234.43x + 163.62 R2 = 0.61 P<0.0001 y = 0.0984 X2+105.14x +1076.4 R2 0.78 P<0.00010 200 400 600 800 1000 1200 1400 1600 0.00.51.01.52.02.53.03.54.04.55.0 Leaf Area IndexAbsorbed PAR (mol m-2 s-1) Barrier Non-Barrier Monoculture Figure 2-4. Relationship between LAI and abso rbed photosynthetically active radiation in a pecan-cotton alley cropping system in northwest Florida. y = -6.2525x2 + 45.501x 13.384 R2 = 0.57 P<0.00010 20 40 60 80 100 120 0.00.51.01.52.02.53.03.54.04.55.0 Leaf Area IndexLint Yield (g m-2) Barrier Non-Barrier Monoculture Figure 2-5. Relationship between LAI and cotton lint yield in a pecan-cotton alleycropping system in northwest Florida.

PAGE 43

29 Figure 2-6. Relationship between above ground biomass production and cumulative absorbed PAR in a pecan-cotton alley cropping system in northwest Florida. Figure 2-7. Relationship between lint yield production and cu mulative absorbed PAR in a pecan-cotton alleycropping system in northwestern Florida. 2002 y = 1.3558x 75.276; R2 = 0.4118 P <0.0001 2001 y = 1.5176x 63.008; R2 = 0.4372 P<0.00010 100 200 300 400 500 600 050100150200250300350400 Cumulative Absorbed PAR (MJ m-2)Aboveground biomass (g m-2) 2001 Non-Barrier, 2001 Barrier 2002 Non-Barrier, 2002 Barrier Figure 7 2001 y = -0.0005x2 + 0.5697x 58.184 R2 = 0.59 P <0.0001 2002 y = -0.0009x2 + 0.777x 102.95 R2 = 0.61 P<0.00010 20 40 60 80 100 120 0100200300400500600700 Cumulative Absorbed PAR (MJ m-2)Lint Yield (g m-2) 2001 Non-Barrier, 2001 Barrier 2002 Non-Barrier, 2002 Barrier

PAGE 44

30 CHAPTER 3 INTERSPECIFIC COMPETITION IN A PECAN-COTTON ALLEYCROPPING SYSTEM IN THE SOUTHERN UNITED STATES: PRODUCTION PHYSIOLOGY Introduction It is well known that canopy level mechan isms influence growth and yield in plants. Leaf level traits such as specific l eaf area (SLA), specific l eaf nitrogen (SLN) and net photosynthesis (Pnet) have all been explored in e xplaining growth and yield in agronomic and forestry systems (Reich et al., 1998a, 1998b, 1999; Zhao and Oosterhius, 1998; Gillespie et al., 2000). These traits, which influence carbon fixation and allocation patterns in plants (Evans, 1989; Sinclair et al., 1993; Muchow and Sinclair, 1993; Pettigrew et al., 2000; Milroy and Bange, 2003) , are greatly influenced by resource (light, water and nutrients) competition and availability. Plants develop and grow differently unde r different environmental conditions. All plants respond morphologically and physiologica lly to shade and vary considerably in regard to their shade tolerance. Plants th at grow in low-light environment invest relatively more of the products of photosynthe sis and other resources in building greater leaf surface area, resulting in thinner leaves and higher specif ic leaf area (SLA). This, in turn, is associated with relatively fewer and smaller palisade and mesophyll cells and that may affect the photosynthetic capacity per unit le af area. In contrast, plants grown in full sun develop thicker leaves, which contain more photosynthe tic apparatus and thereby exhibit a higher rate of Pnet per unit leaf area (Pettigrew et al., 1993; Lambers et al., 1998; Taiz and Zeiger, 2000). Thus, low light in tensity limits photosynthesis and thereby net

PAGE 45

31 carbon gain and plant growth. High light in tensity may also limit photosynthesis, particularly if other factors are not optimal (Lambers et al., 1998). In alleycropping systems, the photosynthe tic response of understory plants to shading may also depend on the carbon fixation pathways of the associated crop species. It is well known that the photosynthetic rate of C3 plants increases sharply as PAR increases from deep shade up to approximately 25% to 50% of full sunlight, then peaks and remains constant with increasing light. However, C4 species do not become lightsaturated and the photosynthetic rate continues to increase up to full sunlig ht (Monteith, 1978; Kozlowski and Pallardy, 1997; Lamb ers et al., 1998). As a result, C3 crop plants may be better suited for alleycropping than C4 plants. For example, a study by Gillespie et al. (2000) in the midwestern United States showed subs tantial reduction in photosynthetic rates of maize ( Zea mays L.), a C4 species, in a black walnut ( Juglans nigra L.) alleycropping. A 45 % reduction in PAR resulted in 40% decrease in Pnet. However, cotton, a C3 species, was light satu rated at 50 % of the full sun and hence was not affected by shading as much in another study (Milroy and Bange, 2003). The importance of foliar N in photosynthe sis is indicated by the well-known positive correlation between foliar N (either %N or SLN) and photosynthetic activity (Gulmon and Chu, 1981; Evans, 1989; Field and Mooney, 1986; Harrington et al., 1989; Egli and Schmid, 1999). Generally, sun leaves te nd to have higher SLN than shade leaves (Hollinger, 1996; Bond et al., 1999). SLA and SL N are also often negatively correlated across the canopy light gradient (Ellsworth and Reich, 1993; Bond et al., 1999; Grassi and Minota, 2000; Stenberg et al ., 2001). It has been demonstr ated that this N gradient results in efficient use of canopy N in car bon fixation (Field, 1983; Werger and Hirose,

PAGE 46

32 1991; Chen et. al., 1993). Other feedback mech anisms (i.e., allocati on of carbon to roots) also function to increase belowground biomass in order to enhance resource capture when nutrients such as N are limiting for photosynthe tic processes. For in stance, cotton planted in a monoculture in Arkansas with limiting N increased its root:shoot ratio to enhance resource capture for growth. Competition for water between system co mponents can also affect agronomic productivity in alleycropping (Jose et al. 2000; Miller and Pallar dy 2003; Wanvestraut et al., 2004). It is well known that intake of CO2 decreases as water availability decreases due to decreased stomatal conductance (Lam bers et al., 1998; Taiz and Zeiger, 2000). Decreased water availabili ty can thus restrict Pnet on a leaf area and weight basis, which often translates into reduced aboveground bi omass (Periere et al., 1992; Davis et al., 1999; Samuelson, 2000). The physiological mechanisms affecting pr oduction in agroforestry systems have received limited attention both in the tropical and temperate regions of the world. This study was designed to examine how SLA and SLN would respond to aboveand belowground competition for resources and how these mechanisms affect foliar and canopy level photosynthesis and thereby above ground production. We hypothesized that SLA would be higher under shade in the alleycropping system compared to the monoculture. We also hypothesi zed that this would result in lower SLN for plants in alleycropping. A further reduction in SLN in the non-barrier treatment compared to the barrier treatment was expected if competiti on for N existed. Changes in SLA and SLN would influence the overall canopy net photosyn thesis, which, in turn, would affect biomass and lint yield.

PAGE 47

33 Materials and Methods The Study Site The study was conducted at the West Florid a Research and Education Center farm of the University of Florida located in Jay, Florida, U.S.A. (30 N, 87 W). C limate is considered temperate with moderate winters and hot humid summers. The soil is classified as Red Bay sandy loam (fine-loam y, siliceous, thermic Rhodic Paleudult) with an average water table depth of 1.8 m. Aver age rainfall and air temperature during the 2002 growing season (June to October) was 34.85 cm and 24.31C, respectively. For the current study, a pecan-cotton alleyc ropping system was established in 2001 from an existing orchard of pecan trees planted in 1954. The orchard had remained under non-intensive clover ( Trifolium spp.) and rye grass ( Lolium spp.) production for 29 years prior to the initiation of th e current study. Twelve plots we re demarcated within the orchard and arranged into si x blocks using a randomized block design. Each plot, which consisted of two rows of trees oriented in a north-south direction, was 27.4 m long and 18.3 m wide, with a practical cultivable wi dth of 16.2 m and was separated from each adjacent plot by a buffer zone of the same dimensions. To assess tree root competition, each bloc k was randomly divided into a barrier and non-barrier plot. Barrier plots were subjected to a root pr uning treatment in March 2001 in which a trenching machine was used to di g a 0.2 m wide and 1.2 m deep trench along both sides of the plot at a distance of 1.5 m from the tree line to separate root system of pecan and cotton. Trenche s were lined with 0.15 mm-thick polyethylene sheets prior to mechanical backfilling. The ba rrier plots served as the tr ee-root exclusion treatment (referred to as barrier treatment) preventing interaction of tree and cotton roots, while the non-barrier (referred to as non-barrier treatm ent) served as the tree-root competition

PAGE 48

34 treatment. Sixteen rows of cotton, one meter ap art, were established in each plot. Cotton (DP458/RRvariety) seeds were planted in e ach row along a north-south orientation on 13 May 2002 using conventional tillage (15-20 cm deep) at a rate of 23,600 seeds per hectare in the alley between the pecan tree rows. For control purposes, three plots in an ad jacent field were maintained as cotton monoculture (referred to as monoculture tr eatment). All treatments received standard fertilizer and pesticide application for cott on in the southern U.S. No irrigation was applied. Gas Exchange Measurements Net photosynthetic rate ( A ) (mol CO2 m-2 s-1), transpiration rate ( E ) (mm H2O m-2 s-1) and stomatal conductance ( g ) (mm m-2 s-1) were measured using a LICOR 6400 (LICOR, Lincoln, Nebraska) portable infrared gas analyzer (IRGA). Measurements were made four times on a monthl y interval from June to September 2002, on the uppermost and fully expanded main-stem leaves of three co tton plants in the first, fourth and eighth rows. All measurements were taken between 10:00 am and 3:00 pm central daylight savings time under ambient conditions on cl ear sunny days. Instantaneous water use efficiency (WUE) defined as the ratio of A and E was calculated for each sampled leaf. To determine whether light was a limiting factor in our system, maximum light saturated photos ynthetic rate ( Amax) and intercellular CO2 were also measured under constant light (i.e., 2000 mol m-2 s-1) at the peak of cotton growth in August 2002. Soil gravimetric water content was determined at the time of measurement. Photosynthetic light response curves were also generate d for each treatment at the same time under constant air temperature (30C), relative humidity (60%), and CO2 (370 ppm). Ambient light was used to generate A -Ci curves by measuring A under different CO2

PAGE 49

35 concentrations. Canopy net photosynthetic index (CNPI), an index of the canopy photosynthesis was calculated as the product of A or Amax and canopy LAI. Canopy LAI was calculated from canopy biomass and specifi c leaf area (described below) for each treatment. Specific Leaf Area (SLA) and Foliar N Leaves subjected to gas exchange measurements (six leaves total in each row per plot) were collected immediately after m easurement to determine SLA. To avoid desiccation, leaves were placed in polyet hylene bags and placed in a cooler and transported to the lab where leaf area was measured usi ng a LICOR-3000 leaf area meter (LICOR, Lincoln, Nebraska). Leaves were then oven-dried at 70C for a minimum of 72 hours, ground using a coffee grinder and anal yzed for total Kjeldahl Nitrogen. Leaf nitrogen concentration was multiplied by specific leaf weight to determine SLN. Aboveground Biomass and Lint Yield Cotton plant parts such as leaves, stems, bolls, and flowers within 1 x 1 m sub-plots were carefully harvested in each main plot. Harvested cotton plants were placed in paper bags, dried for 72 hours at 70C, and then weighed. Lint yield (devoid of seeds) of cotton in each row (rows 1, 4 and 8) in each plot and in the monoculture was quantified from two 0.61 m x 6.1 m sections in each row. Data Analysis Analysis of variance (ANOVA) within the framework of a randomized split block design was used to test for st atistical differences in measur ed parameters using the mixed procedure of the SAS statistical software p ackage (SAS Institute, Cary, North Carolina). The Shapiro-Wilk’s test was used to te st all data for normality of distribution. Logarithmic (log(x+1)) transformation wa s performed to improve normality when

PAGE 50

36 necessary. Differences between means were determined using Least Square Means procedure. Treatment effects were considered significant at = 0.05. Regression analysis was used to define relationships betw een measured variables when necessary. Results SLA, SLN and A There were variations in leaf morphology (leaf area and weight) of cotton among treatments resulting in differences in SLA (Table 3-1). Cotton in barrier treatment had higher SLA than those of the non-barrier pl ants. Monoculture plants, which exhibited higher leaf weight (0.47 g) and lower SLA, contained significantly higher SLN (2.30 g m-2) compared to those plants in the barrier (1.92 g m-2) and non-barrier (1.99 g m-2) treatments (Figure 3-1). No spatial varia tions in SLA among rows were found in the nonbarrier treatment, but SLA in row 1 of the ba rrier treatment was lower from rows 4 and 8 ( P = 0.0038). While the barrier and monoculture plants had similar light response curves with light saturation occurring at about 50% of th e full sun, light saturation was observed at about 30% of the full sun in the non-barrier treatment (Figure 32). Light saturated maximum photosynthetic rate al so varied accordingly. While Amax was 23 and 25 mol CO2 m-2 s-1 for the barrier and monoculture plan ts, respectively, it was only 18 mol CO2 m-2 s-1 for the non-barrier treatment. A positive curvilinear relationship between A and SLN was observed, with peak photosynthesis observed between 2.2 to 2.4 mg N m-2 (Figure 3-4). Since SLN did not differ significantly between the barrier and nonbarrier plants, it is reasonable to assume that the 28% reduction in Amax in the non-barrier plants comp ared to barrier plants was not a result of decreased foliar N.

PAGE 51

37 Both E and g were higher in the monoculture compared to the barrier and nonbarrier treatments (Table 3-3). The non-exclus ion of cotton roots from pecan resulted in lower E and g in non-barrier plants (34.0% and 40.7% , respectively), than those of the barrier plants. WUE (2.71 mmol mol-1) of the monoculture plants was statistically similar to non-barrier plants (2.59 mmol mol-1). WUE in barrier plants (2.13 mmol mol-1), however, was much lower than those of the non-barrier and monoculture plants. CNPI vs. Biomass and Lint Yield CNPI showed significant di fferences among treatments (Table 3-3). Monoculture treatment had the highest canopy A , which was 34.2% higher than the barrier treatment. Average canopy A in the barrier treatment was 65.3% higher compared to the non-barrier treatment ( P = 0.0429). The barrier treatment resulted in 39.4% increase in aboveground biomass (exclusive of lint yield) of cotton plants compared to the non-barrier (323.7 g m-2 vs. 195.9 g m-2, respectively) treatment. A boveground biomass production in the monoculture plants was 307.01 g m-2 and was statistically similar with barrier but different from the non-barrier ( P = 0.0303) treatment. There were no inter-row variations in aboveground biomass among sampled rows. Alleviating belowground competition resulted in differences in cotton lint yield among treatments ( P = 0.0001). Lint yield in the barrier treatment (51.02 g m-2) was 66.5% higher than that of the non-barrier treatment (17.06 g m-2). Lint yield wa s highest for the monoculture treatment (58.1 g m-2). Mean aboveground biomass showed a str ong and significant relationship with CNPI under ambient condition (R2 = 0.63, P < 0.0001). The relationships was further improved (R2 = 0.77) when Amax (measured under a constant PAR of 2000 mol m-2 s-1)

PAGE 52

38 was used in the analysis (Fi gure 3-5). Similar relationships were also observed between CNPI and lint yield (Figure 3-6). Maximum lin t yield was obtained at an optimum CNPI of 65 to 70 mol m-2 s-1. Discussion Competition for resources in our system has affected leaf morphology of cotton. The barrier plants outperformed the non-barrier plants both in their leaf development and photosynthetic rates. Large differences were observed in SLA among these treatments that affected their gas exchange capacity. Co tton in the barrier and non-barrier treatments had higher SLA than the monoculture. Foliage formed in low light often has lower A than foliage formed in high light and generally has different morphological and bioc hemical characteristics, such as higher SLA and lower leaf N (Cregg et al., 1993; Sten berg et al., 1994; Gr oninger et al., 1996; Jose et al., 2003; Nippert and Marshall, 2003). Accordingly, many studies have demonstrated significant decline in A for plants growing under low light levels. Campbell et al. (1990) found th at leaves of soybean ( Glycine max L.) grown in full sun were capable of higher photosynthesis and becam e light saturated at high light intensities than leaves grown in lower light intensit ies. This reduction in photosynthesis was attributed to a decrease in leaf thickness (higher SLA) that led to lower chloroplast (Campbell et al., 1990; Paul and Foyer, 2001) on a leaf area basis for plants under shade. It is well established that foliar N and chlorophyll content are strongly correlated (Evans 1989). As a result, strong positive correlations between foliar N and net photosynthesis have been observed in a numbe r of species (Evans, 1989; Mitchell and Hinckley, 1993; Bond et al., 1999; Egli a nd Schmidt, 1999). Many process models utilize foliar N as a scalar for integrati ng photosynthetic processes from leaves to

PAGE 53

39 canopies (Leuning et al., 1995; Kull and Jarvis, 1995; Lai et al., 2002; Milroy and Bange, 2003). The net photosynthetic rate of cott on in this study showed significant positive correlations with SLN (R2 = 0.44 and R2 = 0.39, for constant and ambient PAR, respectively). Although monoculture plants ha d lower SLA, they exhibited 20.9% higher SLN than those of the barrier and non-barrie r plants resulting in greater photosynthetic rates. Despite higher SLA, lower SLN in th e leaves of barrier and non-barrier plants resulted in lower photosynthetic rates (lower cluster in Figure 3-4) compared to the monoculture plants. Although reductions in SLN were observed in both the barrier and non-barrier treatments compared to the monoc ulture, SLN was similar for the former treatments. As a result, the variation in A among the barrier and non-barrier treatments could not be explained based on SLN alone. Difference in foliar N or SLN can result not only from competition for light, but also from competition for belowground resources (Mooney et al., 1981; Traw and Ackerly, 1995). Since our study showed no grad ient in SLN for cotton grown in barrier and non-barrier treatments (Figure 3-1), co mpetition for N can be ruled out. A companion study by Allen (2003) showed that competition for N was not a major factor affecting productivity of the barrier and non-barrier treatme nts in our system. This indicates that water was the major belowgr ound limiting factor affecting photosynthetic rates of cotton in the non-barrier treatment. This was supported by the Amax and A -Ci curves generated for the study. Providing equal amounts of light and di fferent levels of intercellular CO2 resulted in varying Amax of cotton among treatments (Figures 3-2 and 33). Less soil moisture content resulted in lower Amax in the non-barrier plants. Net photosynthesis diminished with increasing levels of irradiance. Diminished net

PAGE 54

40 photosynthetic rate of shade leaves is often caused by phot oinhibition and damage of photosynthetic apparatuses (Lambers et al., 1998 ). Higher soil moisture in the barrier treatment (Wanvestraut et al ., 2003) has resulted in greater Amax and higher light saturation point, which indicated that car boxylation rate limitation took place at higher light levels for the barrier and monoculture plants than for the non-barrier plants. Higher WUE in non-barrier plants indicat ed drier soil conditions, which implied less water available for absorption due to competition with pecan trees. Wanvestraut (2004) also noted 36% lower water uptake by cotton in non-barri er (0.56 kg plant-1 day-1) compared to barrier treatment (0.88 kg plant-1 day-1). Consistent with these findings, the non-barrier plants in our study also showed lower E than the barrier plants (Table 3-3) as indicated by having only 14.8% gravimetric mois ture content compared with the barrier treatment (18.8%). Non-exclus ion of cotton roots from pecan resulted in lower E and g in non-barrier plants (34.0% and 40.7%, respectively) than those of the barrier plants. Although the monoculture treatment had th e lowest moisture content (10.4%), E of the monoculture plants was highe r in response to high air temperature and high vapor pressure deficit (data not shown) at the time of measurement. Zhenmin et al. (1998) found that higher E is needed for cotton growing in harsher environments in order to provide evaporative cooling effects on the leaves to prevent cell cavitation. The observation of lower E of cotton in barrier and non-barrie r plants is consistent with that of Zhao and Oosterhius (1998) who found lower E rate in cotton grown under shaded conditions due to the relatively cooler envir onment, which indicated lower heat stress. The non-barrier treatment that exhibited the lowest canopy photosynthesis had the lowest biomass and lint yield. We observe d significant relations hips for biomass (R2 =

PAGE 55

41 0.76) and lint (R2 = 0.68) with CNPI (Figures 3-5 and 3-6). Optimum lint yield in our study was attained at a CNPI of approximately 65 to 70 mol CO2 m-2 s-1 (Figure 3-6). This indicated that any further increase in A would not result in an increase in yield. Cotton in the barrier treatment had a CNPI of 52.7 mol CO2 m-2 s-1, which was slightly lower than the optimum CNPI necessary fo r maximum yield. CNPI of cotton in the monoculture treatment was within the op timum range. This explains the yield differences between the two treatments. Our result was in agreement with other studies of cotton in Australia (e.g., Milroy and Ba nge, 2003) and in the southern USA (Muchow and Sinclair, 1993) showing increased production with increase in net photosynthesis, but only up to a certain threshold. Conclusions Results of the study showed that mor phological differences in cotton leaves resulted in varying photosynthetic rates of cotton among treatments with consequences on biomass and lint yield of cotton. The barr ier plants, where belowground competition for water was eliminated, not only increased their photosynthetic rate s compared to nonbarrier plants, but also exhi bited photosynthetic rates comp arable to the monoculture. Competition for water, coupled with shadi ng in the non-barrier treatment, has lowered biomass and lint yield of the non-barrier plan ts. Since foliar nitrogen (SLN) was similar between the barrier and nonbarrier plants, it is reas onable to assume that the performance of cotton in our alleycroppi ng system was influenced mainly by the availability of water and light . Maximum lint yield in our study was obtained at a CNPI of approximately 65 to 70 mol CO2 m-2 s-1. This study demonstrat es that interspecific competition between trees and crops can regulate leaf traits such as SLA and SLN of the associated crop species, which in turn influence CNPI and yield in alleycropping systems.

PAGE 56

42 Table 3-1. Specific leaf area (SLA) of cott on in non-barrier, barri er, and monoculture treatments. Specific Leaf Area (SLA) Treatment Row Non-Barrier Barrier Monoculture 1 267.2 a1 270.6 b (8.4)2 (8.1) 4 277.8a 303.4a (9.0) (10.6) 8 269.4a 314.1a (10.4) (13.6) Overall Mean3 271.5 B 296.6 A 184.6C (5.8) (6.4) (10.6) 1 Within-treatment values followed by the same lo wercase letter are not significantly different at the 0.05 level of probability. Computed P va lues for non-barrier and barrier were 0.597 and 0.0038 2 Standard error of the mean are given in parenthesis 3 Within-treatment values followed by the same lo wercase letter are not significantly different at the 0.05 level of significance. (P<0.0001)

PAGE 57

43 Table 3-2. Canopy net photosynthetic index of cotton measured under ambient condition (PAR and CO2) in a pecan-cotton alleycropping system. Canopy Net Photosynthetic Index (CNPI) (mol CO2 m-2 s-1) Treatment Row 1 Row 4 Row 8 Overall Mean Mean Mean Mean1 Non-Barrier 17.43 a3 18.10a 19.44a 18.32C (0.50)2(0.70)(1.80) (0.70) Barrier 23.79b 65.67a 68.61a 52.69B (1.80)(2.20)(2.30) (2.10) Monoculture 70.72A (1.90) 1 The uppercase letters are for treatment comparisons. Mean followed by the same capital letters are not significantly different at 0.05 level of significance (P<0.0001) 2 Standard error of the mean are given in parenthesis 3 The lowercase letters are for within-treatment comparison among rows. Within-treatment values across the rows followed by the same lowercase le tter are not significantly different at the 0.05 level of significance. (P<0.0001)

PAGE 58

44 Table 3-3. Leaf level transpiration ( E ) and stomatal conductance ( g ) of cotton measured under ambient condition (PAR and CO2 concentration) in a pecan-cotton alleycropping system. Leaf Level Transpiration ( E ) Leaf Level Stomatal Conductance ( g ) (mmol m-2 s-1) (mmol m-2 s-1) Treatment Row 1 Row 4 Row 8 Row 1 Row 4 Row 8 Mean MeanMeanOverall Mean Mean MeanOverall Mean1 Mean1 Non-Barrier 4.73a3 5.03a5.09a4.95C 0.31a 0.33a 0.30a0.31C (0.19)2 (0.21)(0.20)(0.34) (0.04) (0.04) (0.01)(0.01) Barrier 6.54b 7.7a8.13a7.48B 0.43b 0.54ab 0.57a0.52B (0.31) (0.30)(0.34)(1.13) (0.03)(0.03) (0.02)(0.02) Monoculture 12.65A 1.12A (1.28) (0.07) 1 The uppercase letters are for treatment comparisons. Means followed by the same capital letters are not significantly different at the 0.05 level of significance. 2 Standard error of the mean in parenthesis 3 The lowercase letters are within-treatment comparison among rows. Within-treatment values across the rows followed by the same lowercase le tter are not significantly different at the 0.05 level of significance.

PAGE 59

45 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 No-BarrierBarrierMonocultureSpecific Leaf Nitrogen (g m-2) Row 1 Row 4 Row 8 M Figure 3-1. Specific leaf nitrog en (SLN) content of cotton in non-barrier, barrier and monoculture treatments.

PAGE 60

46 -10 -5 0 5 10 15 20 25 30 35 40 0500100015002000 PhotosyntheticallyActive Radiation (molm-2s-1)Net Photosynthesis (molCO2m-2s-1) Non-Barrier Barrier Monoculture -10 -5 0 5 10 15 20 25 30 35 40 0500100015002000 PhotosyntheticallyActive Radiation (molm-2s-1)Net Photosynthesis (molCO2m-2s-1) Non-Barrier Barrier Monoculture Figure 3-2. Light response curve of cotton in non-barrier, barrier and monoculture in a pecan-cotton alleycropping system.

PAGE 61

47 -10 0 10 20 30 40 50 60 70 0500100015002000 Intercellular CO2 (mol m-2 s-1)Net Photosynthesis (mol CO2 m-2 s-1) Non-Barrier Barrier Monoculture Figure 3-3. Intercellular CO2 (A-Ci) curve of cotton in non-barrier, barrier and monoculture in a pecan-co tton alleycropping system.

PAGE 62

48 Ambient PAR y = -33.314x2 + 159.93x 166.07 R2 = 0.3916 P=0.0013 Constant PAR y = -34.287x2 + 162.42x 166.05 R2 = 0.4416 P=0.00070 5 10 15 20 25 30 35 1.61.82.02.22.42.62.83.0 Specific Leaf Nitrogen ( g m-2)Net Photosynthesis (mol CO2 m-2 s-1) Ambient PAR Constant PAR Figure 3-4. Relationship betw een specific leaf nitrogen (SLN) and leaf level net photosynthesis of cotton in non-barrier, barrier and monoculture treatments.

PAGE 63

49 Constant PAR y = 149.88Ln(x) 312.93 R2 = 0.7666 P<0.0001 Ambient PAR y = 126.28Ln(x) 203.97 R2 = 0.6259 P<0.00010 100 200 300 400 500 600 020406080100120 Canopy Net Photosynthetic Index ( mol CO2 m-2 s-1)Aboveground Biomass (g m-2) Constant PAR Ambient PAR Constant PAR y = 149.88Ln(x) 312.93 R2 = 0.7666 P<0.0001 Ambient PAR y = 126.28Ln(x) 203.97 R2 = 0.6259 P<0.00010 100 200 300 400 500 600 020406080100120 Canopy Net Photosynthetic Index ( mol CO2 m-2 s-1)Aboveground Biomass (g m-2) Constant PAR Ambient PAR Figure 3-5. Relationship between canopy net photosynthetic index aboveground biomass production of cotton in non-barrier, barrier and monoculture treatments.

PAGE 64

50 Constant PAR y = -0.0077x2 + 1.6478x 15.074 R2 = 0.6819 P<0.0001 Ambient PAR y = -0.0114x2 + 1.9167x 12.919 R2 = 0.6815 P<0.00010 10 20 30 40 50 60 70 80 90 100 020406080100120 Canopy Net Photosynthetic Index (mol CO2 m-2 s-1) Lint Yield (g m-2) Constant PAR Ambient PAR Constant PAR y = -0.0077x2 + 1.6478x 15.074 R2 = 0.6819 P<0.0001 Ambient PAR y = -0.0114x2 + 1.9167x 12.919 R2 = 0.6815 P<0.00010 10 20 30 40 50 60 70 80 90 100 020406080100120 Canopy Net Photosynthetic Index (mol CO2 m-2 s-1) Lint Yield (g m-2) Constant PAR Ambient PAR Figure 3-6. Relationship between canopy ne t photosynthetic index and lint yield production of cotton in non-barrier, barrier and monoculture treatments.

PAGE 65

51 CHAPTER 4 MORPHOLOGICAL PLASTICITY OF COTTON ROOTS IN RESPONSE TO INTERSPECIFIC COMPETITION WITH PECAN IN AN ALLEYCROPPING SYSTEM IN THE SOUTHERN UNITED STATES Introduction Minimizing resource competition between trees and crops, while maximizing the use of available resources, is central to improving yield and overall productivity in alleycropping and similar agroforestry systems (Ong and Black, 1995; Cannell et al., 1996). The realization that productivity of many agroforestry systems is limited by belowground competition for resources (Singh et al., 1989; Ong et al., 1991; Govindjaran et al., 1996, Scroth, 1999; Jose et al., 2000a , 2000b, 2001) has made root research an integral part of agroforestry experiments in recent years. As the physical, chemical and biological interface for resource uptake, root systems of the component species and their associated rhizosphere play major roles in belowground resource competition. Therefore, knowledge on spatial distribution and density of tree -crop root systems is necessary to gauge the degree of competitive or comp lementary resource sharing among system components (Gregory, 1994, 1996; Van Noordwijk et al., 1995). The degree to which interspecific competition influences root deve lopment and morphology is of great interest in this context. Previous studies have shown that interspecific competition has been shown to reduce root length density. For example, Live sley et al. (2000) observed that greater proximity to a tree row reduced maize root length and therefore reduced its ability to compete for resources. Induced spatial separa tion of tree and crop root systems has also

PAGE 66

52 been reported in the tropics. In alleycr opping research in arid northern Kenya, Lehmann et al. (1998) reported that sorghum ( Sorghum bicolor L.) developed more roots in the topsoil because of competition with Acacia saligna compared to that of sorghum grown in a monoculture stand. Similarly, Jose et al . (2001) found higher c oncentration of maize roots in the upper layer of soil in response to competition for resources with red oak ( Quercus rubra L.) in a temperate alleycropping system in the United States. In all these studies, it was concluded that concentration of roots in the upper soil layer caused severe belowground competition for resources, which, in turn, resulted in decreased crop productivity. It is well known that envi ronmental factors such temperature (McMichael et al., 1996; Reddy et al., 1997a), soil strengt h (Bingham and Bengough, 2003), soil water content (McMichael et. al., 1996) and soil nut rient availability (Robinson, 1994)) affect root growth and development. Taylor and Klepper (1974) an d Keino (1998) reported that cotton planted under drought condition in Arka nsas, United States, increased its root growth with increasing drought stress. The de pletion of water in the upper layer of soil caused proliferation of roots deeper in the soil profile. Increased root growth during water deficit has been associated with decr eased partitioning of carbohydrates into leaves (Kasperbauer and Busscher, 1991). Howeve r, preferential car bohydrate allocation to roots will be limited when li ght also becomes a limiting fact or along with water (Smith and Huston, 1989; Jose et al., 2002; Sack and Grubb, 2002). Plant root systems have the inherent capability to adjust to prevailing environmental conditions through their mor phological and physiol ogical plasticity. Studies on morphological response of roots to localized patc hes of mineral nutrition has

PAGE 67

53 shown that lateral roots tend to proliferate in zones of local nutrie nt enrichment and the proliferation can involve an increase in root number (enhanced initiation), an increase in root mass, or both (Robinson, 1994; Bingham et al., 1997; Zhang and Forde, 1998). In agroforestry systems, the restrictions of lateral root development and formation of vertically stratified root systems in the c ontact zone of competing root systems maybe seen as a morphological mechanism by wh ich plants avoid excessive intra and interspecific competition (Schroth, 1999). Considering the growing interest in cotton as an alleycropping crop in the southern United States and the limited information on its root development and plasticity, specifically in response to both above and belowground competition, our study was conducted with the following questions: How does interspecific competition alter root:shoot ratio in cotton? How will root length density respond to resource competition? Will specific root length change in response to competition? We hypothesized that (1) root:shoot ratio w ould decrease with competition for light and belowground resources; (2) root length density would be a dversely affected by belowground competition for resources; and (3) specific root length would decrease in response to both above and belowground competition. Materials and Methods Study Site The study was conducted at the West Flor ida Research and Education Center (WFREC) farm of the University of Florida located in Jay, Florida, U.S.A. (30 N, 87 W). C limate is considered temperate with moderate winters and hot humid

PAGE 68

54 summers. The soil is classified as a Red Ba y sandy loam (fine-loamy, siliceous, thermic Rhodic Paleudult) with an average water tabl e depth of 1.8 m. Average monthly rainfall during the study period was 23.2 cm. For the current study, a pecan-cotton alle ycropping system was initiated in 2001 from an existing orchard of pecan trees plan ted in 1954. The orchard, arranged in a 5 x 20 grid pattern with trees sp aced 18.3 m apart, had remained under non-intensive clover ( Trifolium spp.) and rye grass ( Lolium spp.) production for 29 y ears prior to the initiation of the current study. For this study, 12 plots were demarcated within the orchard and arranged into six blocks using a randomi zed complete block design. Each plot, which consisted of two rows of trees oriented in a north-south direction, was 27.4 m long and 18.3 m wide, with a practical cultivable widt h of 16.2 m and was separated from adjacent plots by a buffer zone of the same dimensions. Treatments and Sample Collection For our study, five blocks were set up in the orchard. Each block was randomly divided into barrier and non-ba rrier plots. Barrier plots were subjected to a root pruning treatment in March 2001 in which a trenching machine was used to dig a 0.2 m wide and 1.2 m deep trench along both sides of the plot at a distance of 1.5 m from the tree line to separate root systems of pecan and cotton. Trenche s were lined with 0.15 mm-thick polyethylene sheets prior to mechanical backfi lling. The barrier plots served as the treeroot exclusion treatment (referred to as ba rrier treatment or barri er plants) preventing interaction of tree and cotton roots, while the non-barrier (referred to as non-barrier treatment or non-barrier plants ) served as the tree-root co mpetition treatment. Sixteen rows of cotton, one meter apart, were established in each plot. Cotton (DP458/RRvariety) seeds were planted in each row in a no rth-south orientation on May 26, 2001, May 13,

PAGE 69

55 2002, and May 20, 2003 using conventional tillag e (15-20 cm deep) at a rate of 23,600 seeds per hectare in the alley between the pecan tree rows. Three plots in an adjacent field were also maintained as cotton monoculture (referred to as monoculture treatment or m onoculture plants). All treatments received standard fertilizer and pesticide applicati on common in the southern United States. No irrigation was applied. To evaluate root morphological differences , six cotton plants in each of rows 1, 4 and 8 (18 plants per treatment) were randomly selected and harvested each month (from June to October) in 2002 and 2003 growing s easons. Cotton plants were carefully dug up to ensure that the whole root system wa s properly collected. During plant harvesting, water was supplied surrounding the plant to loosen the soil to ensure minimal destruction and loss of fine roots. After harvesting, co tton roots and shoots were separated. Roots were washed immediately and placed in polye thylene bags. Prior to oven drying of cotton shoot for 72 hours at 70oC, six fully expanded leaves were collected to determine leaf area using leaf area meter (LIC OR 1300, Lincoln, Nebraska). Root length density (RLD, cm cm-3) of cotton was determined during the middle (August) of the growing season for which soil coring was employed to sample roots in a given volume. Three cores (30 cm (length) x 5 cm (diameter)) per plot in each treatment were taken in 2001, 2002 and 2003 growing seas ons. Coring was done to a depth of 90 cm and at 30 cm intervals. Cores were taken in rows 1, 4 and 8. Soil cores were washed using a fine sieve in a gentle flow of water. Roots were stored in sealed plastic bags at 4.4oC. Cotton roots treated in this manner remained firm and fresh (Samson and Sinclair, 1994).

PAGE 70

56 RLD was measured using the line inte rcept method developed by Newman (1966) and as modified by Tennant (1975). Root samp les were distributed evenly over a 31 x 46 cm silkscreen cloth. The cloth was placed ove r a grid of lines 0.5 cm apart, and the number of intersections that roots made with the grid lines was counted. RLD was calculated from the number of inte rsections of the grid and vo lume of the original soil cores. Roots were also analyzed for mor phological parameters using WinRHIZOTM (Rgent Instruments Inc., Quebec, Canada) image analysis system. Individual root samples were spread out on a clear tray (30 cm x 48 cm glass frame) that was filled with water and placed on a flatbed scanner. Analysis of the scanned images provided sizeclass distributions and quantification of parameters including root length, surface area and average diameter. Scanned roots we re then oven-dried for 72 hours at 70oC for root biomass assessment. Data Analysis Analysis of variance (ANOVA), within the framework of a randomized split block design, was used for analysis (SAS Institute, Cary, NC). In each analysis, main effects and interactions were tested for signifi cance using the appropriate error terms. If significant treatment effe cts were revealed at = 0.05, then Least Square Means procedure was used for mean separation. Re gression analysis was used to define relationships between root length and leaf area as well as ro ot dry weight a nd root length. Results Biomass and Root: Shoot Ratio Whole plant biomass showed variation among treatments. Across growing seasons, the non-barrier plants exhi bited 47% and 57% reducti on in whole plant biomass

PAGE 71

57 compared to the barrier and monoculture plan ts (Table 4-1; Figure 4-1). Whole plant biomass for the latter treatments was statis tically similar in 20 02, but higher production was noted for monoculture in 2003. Root:shoot ratio was significantly lower in the barrier plants (0.16) compared to monoculture plants (0.18), but was higher compared to plants in the non-barri er treatment (0.12) ( P =0.0031) in 2002 (Table 4-1; Figure 4-1). A similar trend was also observed in 2003. Total Root Length Total root length of cotton plant in the non-barrier treatment (359 cm) was significantly lower compared to that of plants in the ba rrier (477 cm) a nd monoculture treatments (539 cm) (Table 4-2). Absen ce of belowground competition in the barrier treatment did not, however, re sult in any significant incr ease in total root length compared to that of plants in the monoculture treatment. Length of fine roots (< 2.0 mm diameter) also differed significantly among treatments ( P = 0.0238), with non-barrier plants (336.60 cm) having 23% and 26% lower length in fine roots than that of the barrier (437.64 cm) and monoculture (457.82 cm) treatments, respectively (Table 4-2). Results also showed no variati on between the latter treatments in length of fine roots. Although th e barrier and monoculture plants had higher total root length, the ratio of fine and coarse (>2.0 mm) roots (on a plant basis) was higher in non-barrier plants. Ninety five pe rcent of the root syst em in the non-barrier plants was composed of fine roots compared to only 85% and 90% in the monoculture and barrier treatments, re spectively (Table 4-2). First (> 4.0 mm diameter) and second order (> 2.0 to < 4.0 diameter) categories of roots also differed among treatments ( P <0.0001) (Table 4-2). Th e monoculture plants

PAGE 72

58 exhibited highest root length for both categ ories and the non-barrier plants had the lowest. The barrier plants developed 47% more coarse roots than the non-barrier plants, but the barrier plants was 48% lower in co arse roots composition than that of the monoculture plants. Root Length Density The RLD profiles of cotton at physiological maturity are presented in Figure 4-2 and Tables 4-3 and 4-4. RLD significantly varied among treatments with non-barrier plants having the lowest RLD across soil depth in each year of the study except in 2001 where it was statistically similar to monocultu re (Table 4-2). The barrier plants showed consistently higher RLD in every growing season. In all treatments, roots were concentrated in the upper 30 cm of soil (Fig ure 4-2; Table 4-4). RLD declined rapidly with depth with the exception in the non-ba rrier treatment in 2002 where RLD was higher in 30-60 cm than in 0-30 cm soil depth. Appa rently, RLD was reduced to as low as 0.11, 0.21 and 0.29 cm cm-3 for the non-barrier, monoculture and barrier plants, respectively, in the 60-90 cm depth (Table 4-4). Root distribu tion analysis also show ed that in 2003, 59% of roots of the non-barrier plants were f ound in the upper 30 cm. However, only 52% and 50% were found in the same soil depth for th e barrier and monoculture plants (Table 45). A similar trend was also observed in 2001 growing season. This trend, however, was reversed in 2002. Consistent with total root length, the non-barrier plants had the lowest root surface area (146 cm2) representing 28% and 55% reduction over barrier (204 cm2) and monoculture (327 cm2) plants, respectively.

PAGE 73

59 Specific Root Length Specific root length (SRL, the ratio of root length and the amount of biomass produced) also showed significant difference among treatments ( P= 0.0254). Cotton in non-barrier treatment had the lowest SRL (146.06 cm g-1) while cotton in monoculture treatment had the highest (179 cm g-1) representing an 18% increase over the former treatment. SRL of cotton in th e barrier treatment was 165 cm g-1. Discussion Below and Aboveground Carbon Allocation Interspecific competition for resources (i.e., water greatly affected partitioning of carbon between shoot and root systems of co tton in our study. The non-barrier plants, which experienced both aboveand belowgr ound resource competition, were shown to exhibit consistently lower whole plant biomass production compared to that of monoculture and barrier treatm ents in each growing season. Cotton in the non-barrier and barrier treatments had a signi ficantly lower root:s hoot ratio compared to that of the monoculture plants both in 2002 and 2003. This implies that more carbon was allocated in the aboveground plant components than in the roots in the shaded treatments. Root:shoot ratio, which tends to respond to environmental growth conditions, is an indicator of plant growth perf ormance in a stressed environment (Van Noordwijk et al., 1994). Plants grown in a low-light environmen t tend to have higher SLA than those under exposure to direct irradiance, resulting in greater partitioning of carbon to shoots (Lambers et al., 1998; Jose et al., 2002; S ack and Grubb, 2002). This plasticity has been the basis for the trade-off hypothesis which predicts that competition for aboveground resources has a stronger impact on individuals of a species grown in deep shade than those in higher light levels (Smith and Huston, 1989; Jose et al., 2002; Sack and Grubb,

PAGE 74

60 2002). This trade-off arises when plants have a higher SLA and leaf area ratio (leaf area per total dry mass) for effici ent light capture at the expe nse of their root allocation, resulting in greater sensitivit y to drought. This was observe d in cotton plants grown in non-barrier treatment, which produced relativel y smaller amounts of root biomass than aboveground biomass. The trade-off hypothesis has important implications for overall production of the agronomic crop in agroforestry systems. Root Morphological Plasticity The acquisition of belowground resources is based on the architecture and morphology of the plant root systems, whic h can exhibit variations resulting from adjustments to the prevailing environmen tal conditions (Lynch, 1995; Rosolem et al., 1999; Bingham and Bengough, 2003). Both total r oot length and RLD of cotton varied in response to interspecific competition with p ecan for aboveand belowground resources. Competition for water (Wanvestraut et al., 2004) and aboveground resources (Chapters 2 and 3) significantly reduced the growth of cotton roots in the non-barrier treatment compared to the barrier and monoculture treatments in our system. Monoculture and barrier plants had higher total length of fi ne roots compared to the non-barrier plants (Table 4-2). However, the average compositi on of fine roots in these treatments on a whole plant basis was lower th an that of the non-barrier pl ants. Root of the non-barrier plants were composed of 95% fine roots, compared to that of the barrier (90%) and monoculture (85%) treatments. Competition for water between pecan and cotton in non-barrier treatment significantly affected root system developm ent in the non-barrier plants, resulting in a significant reduction in total root length. In addition, more than 50% of the roots in nonbarrier plants were concentrated in the uppe r 0-30 cm layer of soil (Table 4-5), which

PAGE 75

61 implies intensity of competition. Wetter soil in barrier treatmen t provided a greater advantage for the barrier plants to grow rapidly, resulting in expansion of their root systems. A similar trend was observed for monoc ulture, but the plasticity in roots was in response to greater allocation of carbon to roots due to soil moisture deficit (Kasperbauer and Busscher, 1991; Baker and Acock, 1986). Throughout the 2003 growing season, there wa s an increasing trend in root length of cotton in barrier and monoc ulture treatments (Figure 43). A similar trend was also observed for cotton in non-barr ier treatment. However, competition for water inhibited cotton root growth at the same rate of expansion as with th e barrier plants. Wetter soil in the barrier treatment did not provide temporal va riations (plateau tren d) in root length of barrier plants during the sample collection s eason. This implies that no or minimal root growth was taking place compared to that of the monoculture and non-barrier treatments. Root Length Density Variations in root length among treatments resulted in differences in RLD, with non-barrier plants exhibiting the lowest RLD. Although cotton in monoculture and barrier treatments developed greater root length (and we re statistically simila r to each other), the monoculture treatment still cons isted of lower RLD than the barrier. Greater root length in monoculture treatment did not provide co mpensatory adjustment to increase RLD. Roots of cotton in monoculture treatment were sporadically distributed, occupying a greater volume of soil that resulted in lower RLD (Tables 4-3 and 4-4). Lesser development of cotton roots in non-barrier treatment, on the other hand, may have inhibited the uptake of water and nutrients, thus affecting the demand for resources to satisfy growth (Gregory, 1996). McMichael and Quisenberry (1993) and Bowen (1985) observed that the ability of a plant to compete for soil nutri ent and water resources from

PAGE 76

62 the same soil volume is proportional to its density of fine roots. Simila rly, Livesley et al. (2000) noted a lower root biomass production, bu t a greater fine root system, on a plant basis, of maize planted with Senna ( Senna spectabilis DC). Accordingly, this root morphological change in their study facil itated better resource uptake by maize in response to greater competition for soil re sources (Livesley et al., 2000). However, lowered RLD and surface area in our study may have positively affected cotton in nonbarrier treatment to absorb belowground resource s for growth in spite that 95% of its root system was composed of fine roots. A slight but statistically significant increas e in root density at 30-60 cm soil depth in 2002 than that of the 0-30 layer of soil indicates that non-ba rrier plants were experiencing belowground limitation for water. When competition is inevitable, plants promote greater root development in the deeper layers of soil to cap ture greater amounts of water and nutrients (Kasperbauer and Busscher, 1991; Van Noordwijk and Purnomosidhi, 1995). The increase in RLD in 30 -60 cm layer of soil in barrier treatment implies water stress in the upper layer of soil. Klepper et al. (1973) observed that depletion of water in the uppe r layers of the soil profile caused proliferat ion of cotton roots into lower soil depth regardless of cu ltivar differences. Plau t et al. (1996) also observed that changes in soil water content resulted in ch anges in rooting patterns and root activity of cotton gr own under different water application regimes, with consequences for production. Teklehaimanot and Ouedraogo (2004) noted that RLD of sorghum planted with nr ( Parkia biglobosa Benth) was reduced due to drier soil conditions caused by competition for water between these two species.

PAGE 77

63 The RLD of cotton (Table 4-3, Figure 4-2) fell within the range reported by other authors. Plaut et al. (1996) computed RLD of cotton planted unde r different irrigation regimes that ranged from 0.1 to 1.1 cm cm-3. McMichael et al. (1996) also noted differences in RLD of cotton grown in irriga ted and rainfed conditions. RLD of cotton in irrigated field reached up to 1.8 cm cm-3 while only 1.4 cm cm-3 was observed under rainfed conditions (McMichael et al., 1996). Further, Scroth and Zech (1995) reported 3.48 cm cm-3 RLD of maize at 0-10 cm so il depth grown 1 m away from Gliricidia sepium . However, 33% reduction in RLD was no ted for maize grown 2.5 m away from G. sepium (Scroth and Zech, 1995). The pattern of sp atial (distance) variations of RLD of maize is consistent with cotton in this study (data not presented). Cotton grown near pecan trees in both barrier and non-barrier tr eatments had higher RLD than those of plants grown at furthe r distances from pecan. Specific Root Length Root length: weight ratio is a plant parameter that s hows the relationship between root penetration intensity and belowg round biomass allocation. Generally, the relationship between these two variables is curvilinear as indicated in Figure 4-4 (R2 = 0.33, R2 = 0.38, and R2 = 0.24, respectively for barrier, monoculture and non-barrier treatments), where all treatments followed a similar trend, but of different magnitude. The non-barrier plants did produce significantly lo wer root length compared to that of the barrier and monoculture plants using the sa me amount of biomass (Figure 4-4). Results indicated a two-fold increase in root lengt h of monoculture plan ts over the non-barrier plants for the same amount of carbon alloca tion in the roots. A similar trend was also observed in root weight-lengt h ratio between barrier and non -barrier plants. Observed variation in specific root le ngth among treatments indicates that cotton in each treatment

PAGE 78

64 responded differently based on their growth environment. Monoculture plants had a greater advantage to allocate carbon in the ro ot system, resulting in higher specific root length, since light was not a limiting factor for their growth. Highe r carbon allocation to leaves in the barrier and non-barrier treatments (Chapter 2) resulted in lower specific root length in these treatments. Generally, th e presence of both aboveand belowground competitions in non-barrier tr eatment prevented cotton plan ts to develop and expand roots at similar pace with barri er and monoculture treatments. Although the proliferation of root systems can be affected by water and nutrient availability (Livesley et al., 2000) , it is difficult to separate the effects of allelopathy from those of resource competition. Wanvestraut et al . (2004) stated that allelopathy cannot be ruled out as a factor in th e poor development of cotton in non-barrier plants because pecan does produce juglone (Dana and Lerner, 2001), an allelochemical known to have an influence on a variety of tree and agronomic crops. Livesley et al. (2000) stated that the allelopathic effects of Grevillea might have caused the lower root length of maize grown in competition with Grevillea . Implications Knowing morphological characte ristics of the system comp onents is essential in predicting plant performance in agroforestry systems. Prev ious studies on modeling crop processes in agroforestry systems have relied on root data of plants grown in monoculture conditions. Accurate mechanistic modeling in agroforestry systems can only be possible when a spatial, quantitative and temporal understanding of tree crop processes and interactions has been achieved, an important element of whic h is the understanding of the spatial distribution and dynamics of root sy stems. Our study has clearly shown that crop root morphology can vary significantly when grown in association with trees. This

PAGE 79

65 information is critical in developing and refining comprehensive tree-crop interaction models in agroforestry systems.

PAGE 80

66Table 4-1. Growth parameters of cotton gr own in non-barrier, barrier and monoculture treatments in 2002 and 2003 growing season s. 2002 2003 Treatment Treatment Non-Barrier Barrier Monoculture Non-Barrier Barrier Monoculture Whole Plant DW (g plant-1) 29.78b154.04a57.33a 25.13c49.56b72.60a (2.67)2 (3.76)(3.71) (1.88)(3.85)(5.71) Root DW (g plant-1) 3.53c7.60b8.70a 2.83c5.68b12.19a (0.43)(0.23)(0.69) (0.21)(0.41)(0.96) Shoot DW (g plant-1) 26.25b46.44a48.63a 22.29c43.88b60.40a (1.81)(2.54)(3.07) (1.71)(3.56)(4.95) Root: Shoot ratio 0.12c0.16b0.19a 0.13c0.16b0.21a (0.02)(0.01)(0.01) (0.01)(0.01)(0.01) 1 In a given row for every growing season, means followed by a different letters are significantly different at P < 0.05 2 Standard error of the mean is given in parenthesis

PAGE 81

67 Table 4-2. Total root length of cotton in nonbarrier, barrier and monoculture treatments at physiological maturity. Treatment Diameter Class (mm) Non-Barrier Barrier Monoculture < 2.0 337.60b1437.64a 457.82a (18.90)2(17.26) (27.60) > 2.0 < 4.0 12.20b24.85b 55.78a (0.96)(2.10) (5.48) > 4.0 10.10c14.37b 24.98a (0.49)(0.76) (1.90) Total Length (cm) 358.90b476.86a 538.59a (19.73)(19.30) (31.84) 1 In a given row, means followed by different letters are significantly different at P < 0.05 2 Standard error of the mean is given in parenthesis

PAGE 82

68 Table 4-3. Root length density (RLD, cm cm-3) of cotton in barrier, non-barrier and monoculture treatmen ts across soil depth. Growing Season Treatment 2001 2002 2003 Non-Barrier 0.29b10.20c 0.23c (0.04)2(0.02) (0.02) Barrier 0.53a0.42a 0.45a (0.05)(0.04) (0.08) Monoculture 0.27b0.25b 0.29b (0.05)(0.03) (0.02) 1 In a given row, means followed by different letters are significantly different at P < 0.05 2 Standard error of the mean is given in parenthesis

PAGE 83

69Table 4-4. Root length density of cotton under different soil depths. Root Length Density (cm cm-3) 2001 2002 2003 Soil Depth (cm) Soil Depth (cm) Soil Depth (cm) Treatment 0 30 30 60 60 90 0 30 30 60 60 90 0 30 30 60 60 90 Non-Barrier 0.42a1 0.24b0.19b0.21a 0.23a0.17ab 0.41a0.18b 0.11b (0.10)2 (0.04)(0.03)(0.04) (0.04)(0.03) (0.05)(0.04) (0.03) Barrier 0.77a 0.5b0.33b0.5a0.4b0.34a 0.71a0.36b 0.29b (0.10) (0.06)(0.04)(0.05) (0.04)(0.04) (0.14)(0.13) (0.09) Monoculture 0.35a 0.24a0.23a0.37a 0.2b0.18b 0.22b0.43a 0.21b (0.08) (0.04)(0.05)(0.06) (0.03)(0.04) (0.10)(0.05) (0.08) 1 In a given row for every growing season, means followed by different letters are significantly different at P < 0.05. 2 Standard error of the mean is given in parenthesis

PAGE 84

70Table 4-5. Percent amount of root biomass in different layers of soil for the non-barr ier, barrier and monoculture treatments. % amount of roots 2001 2002 2003 Soil depth (cm) Soil depth (cm) Soil depth (cm) Treatment 0 30 30 60 60 90 0 30 30 60 60 90 0 30 30 60 60 90 Non-Barrier 49 28 22 343828 592616 Barrier 48 31 21 403227 522621 Monoculture 43 29 28 492724 265024

PAGE 85

71 Figure 4-1. Root and shoot biomass produc tion of cotton in 2002 and 2003 growing seasons. 40 30 20 10 0 10 20 30 40 50 60 70 20022003Root (g plant 1) Shoot (g plant-1) Non-Barrier Barrier Monoculture0.12*0.160.18 0.13 0.160.21 *means root-shoot ratio 40 30 20 10 0 10 20 30 40 50 60 70 20022003Root (g plant 1 Non-Barrier Barrier Monoculture0.12*0.160.18 0.13 0.160.21 *means root-shoot ratio 40 30 20 10 0 10 20 30 40 50 60 70 20022003Root (g plant 1) Shoot (g plant-1) Non-Barrier Barrier Monoculture0.12*0.160.18 0.13 0.160.21 *means root-shoot ratio 40 30 20 10 0 10 20 30 40 50 60 70 20022003Root (g plant 1 Non-Barrier Barrier Monoculture0.12*0.160.18 0.13 0.160.21 *means root-shoot ratio

PAGE 86

72 Figure 4-2. Root length density of cotton in barrier, non-barrier and monoculture treatments. 90-120 60-90 30-60 0-30 0 Soil Depth (cm) Non-Barrier Barrier Monoculture 90-120 60-90 30-60 0-30 0 0.0 0.2 0.4 0.6 0.8 1.0 Root Length Density (cm cm-3) Non-Barrier Barrier Monoculture 90-120 60-90 30-60 0-30 0 Non-Barrier Barrier Monoculture

PAGE 87

73 Figure 4-3. Temporal changes in root leng th of cotton in barrier, non-barrier and monoculture treatments during 2003 growing season. 0 100 200 300 400 500 600 700 800 900 JuneJulyAugustSeptOctoberTotal Root Length (cm) Non-Barrier Barrier Monoculture

PAGE 88

74 Figure 4-4. Relationship between root dry weight and root length of cotton in 2003 growing season. Barrier y = 77.767Ln(x) + 375.82 R2 = 0.33 P = 0.0001 Non-Barrier y = 71.804Ln(x) + 292.72 R2 = 0.14 P = 0.0065 Monoculture y = 139.37Ln(x) + 342.69 R2 = 0.38 P = 0.00170 200 400 600 800 1000 1200 051015202530 Root Dry weight (g plant-1)Root Length (cm plant-1) Non-Barrier Barrier Monoculture

PAGE 89

75 CHAPTER 5 MODELING COTTON PRODUCTION IN A PECAN ALLEYCROPPING SYSTEM USING CROPGRO Introduction Assessment of crop performance and produc tion in alleycropping and similar agroforestry systems through modeling is beco ming an integral activity in agroforestry research. Modeling is a consid erably useful tool in unders tanding the relationships among soil, plants, trees, and other components in agro forestry systems, particularly in studying the relationships between system component s over time (Young, 1997). In recent years, several agroforestry models have been devel oped with their various foci of interest [e.g., WaNuLCAS (water, nutrient, light capture in agroforestry systems; van Noordwijk and Luciana, 1999, 2000), SCUAF (soil changes under agroforestry; Young and Muraya, 1990), and HyPAR (Mobbs et al., 2003)]. Th e CROPGRO (Simulati on of Crop Growth; Tsuji et al., 1994; Boote et al., 1997, 1998; J ones et al., 2003), model, which has been used predominantly in monoculture agricultura l systems, can also be used to evaluate biophysical interactions in agroforestry sy stems, including intera ctions involving light (Jones et al., 2003). These models were deve loped, not only to understand the processes and interactions involving system compone nts and their effect s upon overall production, but also for their usefulness as decision s upport tools for identifying best management options for attaining optimal production. Because plant components in agroforestry systems draw from the same reserve pool of resources, competition for limited grow th resources (i.e., water, nutrients and

PAGE 90

76 light) between and among them is inevitabl e, making the systems more complex. Under potential production scenarios where water a nd nutrients are availa ble for plant growth, the level of light transmittance to crop canopies determines production (Kropff, 1993, 1994). Several studies in alleycropping and simila r agroforestry systems have indicated that shading affects morphol ogical development in plants, with consequences on photosynthesis and biomass production (Figure 5-1). Shading can also affect carbon allocation patterns in plants. It is in this co ntext that farmers in temperate regions may be hesitant to adopt agroforestry practices, since they may assume that shading may adversely affect crop performance. While this may be true in some circumstances, there exists a need for better understanding the comp lexities of such systems with regard to optimizing light use. In this regard, a pplication of a process-based model (e.g.., CROPGRO) can be useful for understanding th e effects of light on production, which, in turn, can provide a basis for determining wh ether or not to promote adoption of an agroforestry system. CROPGRO, a member of the DSSA T (Decision Support System for Agrotechnology Transfer) model, is a dynamic simulation model that simulates growth and development of most common agronomic crops [e.g., soybean ( Glycine max L.), peanut ( Arachis hypogaea L.), cowpea ( Vigna unguiculata L.)] grown under varying environmental conditions. CROPGRO is widely used in temperate and tropical regions because of its precision in simulating producti on when compared to actual field data, provided that the model is calibrated base d on site-specific conditions. Bannyan et al. (2003) found close agreement between simu lated and actual biomass of wheat ( Triticum

PAGE 91

77 sp.) grown in the United Kingdom, using DSSAT models. A similar observation was noted by Mavromatis et al. (2002) when th ey used CROPGRO to simulate biomass production of soybean grown in North Carolina, USA. The relationships of plant morphological para meters [(i.e., specif ic leaf area (SLA) and leaf area index (LAI)] with light levels have been well established in CROPGRO models (Figure 1). These varia tions in leaf morphological pa rameters due to differences in environmental growth conditions, as can be quantified by the model, influence production. Therefore, this study was conduc ted to simulate production of cotton (Gossypium sp.) under different light levels in a temperate alleycropping system using the CROPGRO-cotton model, where interspecific competition for water and nutrients was assumed to be non-existent. Accordingl y, the following research questions were addressed: 1. How will the CROPGRO-cotton model perform when used to predict leaf morphological changes as affected by sh ading in an alleycropping system? 2. Can the CROPGRO-cotton model quantify the effects of shading on SLA, leaf size and LAI and their effects on biomass production? 3. Can the CROPGRO-cotton model predict biomass production under varying light levels based on differences in leaf morphological parameters? Two hypotheses were therefore tested: 1) the CROPGRO-cotton model, when properly calibrated, can accurately predic t production of cotton grown under different light regimes; and 2) variations in lig ht transmittance will affect morphological development of cotton, with consequences on production that are quantifiable through model simulation.

PAGE 92

78 Materials and Methods Field and Experime nt Description The study was conducted at the West Flor ida Research and Education Center (WFREC) Farm of the University of Florida, located near Jay in northwestern Florida, USA (30 N, 87 W). The climate is te mperate with moderate winters and hot, humid summers. The soil is classified as a Red Bay fine-loamy, sand soil (siliceous, thermic, Rhodic Paleudult) with some increase in clay content with depth. A mature pecan ( Carya illinoensis K. Koch)-cotton ( Gossypium sp.) alleycropping system was established in Spring 2001 from an existing orchard of 50-yr-old pecan trees that had been planted at a uniform spacing of 18.3 m x 18.3 m. For this study, 10 plots were demarcated within the orchard and arra nged into 5 blocks using a randomized block design. Half of the plots were trenched to a depth of 120 cm and width of 20 cm, to prevent belowground interaction; hence these pl ots were referred to as “barrier plots” because of the exclusion of pecan roots th rough trenches covered with polyethylene plastic sheets. For the purposes of simula tion modeling, only the barrier plots were considered in this study. In each plot or al ley, 16 rows of cotton, one meter apart, were sown using conventional tillage in May 10, 2001 and May 16, 2002. Data were collected in rows 1, 4 and 8. Row 8, located at the center of the alley, was approximately 8 m distance from the base of the pecan trees, and hence received greater light transmittance due to less shading from the overhead tree canopy. It was assumed that the trenching us ed in the study eliminated belowground interaction/competition for water and nutrients between pecan and cotton (Wanvestraut et al., 2004; Allen, 2003). Thus, the major drivi ng force in the production of cotton was assumed to be the level of light transm ittance to cotton canopies, which was an

PAGE 93

79 environmental input condition in the CROPGRO -cotton model. Four treatments were set up in the model, as follows: 1) Control (co tton grown in open field, with full amount of light transmittance); 2) Row 1 (cotton grown one meter from the base of the pecan trees, with only about 50% of light transmittance; 3) Row 4 (cotton grown four meters from the base of the pecan trees, with 55% light tran smittance; and 4) Row 8 (cotton grown at the middle of alley, eight meters from the base of the pecan trees, with more exposure to direct sunlight, with approximately 70% light transmittance). The CROPGRO model in DSSAT The CROPGRO model, a component of the DSSAT software package, was developed by the International Benchmark Site Network for Agrotechnology Transfer (IBSNAT) project (Tsuji et al., 1994). CROPGRO is a proc ess-oriented model within DSSAT for general applications ; it is independent of locat ion, season and management systems (Jones et al., 1998; Boote et al., 2003) . CROPGRO simulates e ffects of weather, soil water, and nitrogen content in the soil and plant on crop growth and yield. The model is based on the simulation of carbon, water and nitrogen balances in the soil-plant systems. The simulation is dynamic, with stag es of development, rate of growth and partitioning of biomass each affected by w eather and soil environments. CROPGRO has been well tested to simulate growth and development over a wide range of species. Modeled species include soybean, peanut ( Arachis hypogaea L.), cowpea ( Vigna unguiculata L.), and chickpea ( Cicer arietinum L.), among others (Boote et al., 1998). The CROPGRO-cotton model has only been de veloped recently. Parameter estimates of the CROPGRO-cotton model had only been tested in one experiment (Messina et al., 2004) prior to this study.

PAGE 94

80 The CROPGRO model allows the user to organize and manipulate data and to run the models in various ways and analy ze their outputs (Hoogenboom et al., 1994; Thorthon et al., 1997). DSSAT version 4.0 was used in this study. See Appendices A and B for a sample of the model simulation outputs. CROPGRO Model Inputs Weather data inputs Weather data were obtained from the Florida Automated Weather Network (FAWN) database. The weather station was locat ed at the study site . Input weather data included daily solar radiation (MJ m-2 day-1), maximum and minimum air temperature (oC), and rainfall (mm day-1). Four weather data files we re created and each treatment used a specific weather data file. The weather data files va ried according to the amount of light transmittance that each treatment had received as indi cated earlier. Actual light transmittance was measured using a Cept ometer (Decagon AccuPAR) on a biweekly basis in the 2001 and 2002 growing seasons (Zamora, 2005; Chapter 2). Light transmittance (%) in each row was calculated and then multiplied by the daily solar radiation in each weather f ile for each treatment. Soil parameter inputs CROPGRO requires soil water inputs (low er, upper and satu rated water holding limits), bulk density, total carbon in soil, pH, and nitrate (NO3-N) and ammonium (NH4N) nitrogen. Table 5-1 presen ts soil variables used in the simulation. Data on physical properties (sand, clay and orga nic carbon) of soil in our experiment were used to determine the lower limit of plant water availa bility (LL), and draine d upper limit (DUL), which represent the water availability at field capacity, and wate r at field saturation (SAT), respectively, according to Ritchie ( 1998) and Ritchie et al ., (1998). Other model

PAGE 95

81 management and cultural inputs included pl anting date, planting density, row spacing, and planting depths. Model Execution Model parameterization Parameterization involves identifying parame ters in the model that would fit best based on our study site conditions. The CRO PGRO-cotton model includes various parameters that are stored in the CROPGRO 40.SPE file, and users are not expected to change these parameters. However, for our st udy, some of the fixed parameters (Table 52) were recalculated based on our data so that the effectiveness of the model in predicting production under a shaded environment coul d be evaluated. The defined values and relationships of the parameters listed in Ta ble 5-2, which are shown in Table 5-3, were supplied in the CROPGRO40.SPE file. The CROPGRO-cotton model requires an uppe r and lower limit of SLA to account for the variations that occu r over time based on plant physiological development. Further, the model includes a function of SLW and phot osynthesis to produce observed effects of SLW on photosynthesis (Messina et al., 2004), as this parameter can affect daily canopy assimilation by plants. SLW decreases si gnificantly when light become a limiting resource, thus resulting in thinner leaves and lower net assimilati on per unit leaf area. Hence, modifying of XPGSLW and YPGSLW values in the model was necessary to account for this variation in plant physiol ogical development under shaded conditions. Model calibration A number of cultivar-specific parameters (g enetic coefficients) are used by the crop model to predict cotton daily growth and de velopment in response to weather, soil characteristics, and management actions (B oote et al., 1998). Calibration was focused on

PAGE 96

82 those cultivar parameters (Table 5-4) most likely to be affected under a shaded environment. The genetic coefficients required by the model for these cultivar parameters were estimated (from our data) as follows: 1) candidate coefficient-parameters were selected; 2) the values of the coeffi cient-parameters were changed by running CROPGRO in an optimization shell until the error sum of squares (simulated minus observed) was minimized; and 3) the set of co efficients that produced the lowest RMSE (root mean square error) and higher d-sta tistics value were adopted. Lower RMSE is desirable. Calibration was done by iterative ly running the crop model within an appropriate value of the coefficient concerned that was observed or measured in our field study. Cultivar coefficient values were then changed until the simulated and measured values matched or were within predefined error limits through ev aluation of the RMSE and d-statistics values. The calibration proce ss is an iterative, trial-and-error process described by Hanson (2000) and Hanson et al. (1999). In this particular study, our calibration strategy was to put emphasis on correct simulation of SLA and LAI, as accurate simulation of these parameters shoul d also result in a ccurate simulation of biomass. The model uses a reference maximum area of one leaf originally set at 150 cm2, of which the areas of leaves of co tton were within this value. Ho wever, leaf area of cotton can vary with node position at optimal condi tions between 50 to 250 cm2 (Reddy et al., 1991, 1992, 1993, 1997a, 1997b). The increase in SIZELF value (Table 5-4) was based on the much larger leaf size that cotton can attain in shaded cond itions as was observed for cotton in our experiment; and SLAVAR wa s increased (Table 5-4) to accommodate for the relatively higher area -weight ratio of cotton leav es based on our experiment.

PAGE 97

83 Calibration of these parameters for the cont rol treatment continued until the predicted time series of SLA and LAI were in clos e agreement with th e measured values. Results and Discussion Assessing Model Capability In a comparable study, the CROPGRO-co tton model performed very well in assessing growth performance and producti on of cotton grown in monoculture on a Macopa clay loam soil in New Mexico, as indicated by close agreement between simulated and measured values of the evalua ted parameters (i.e., LAI, SLA and biomass, data not shown) (Messina et al., 2004). However, the model, it its present form and with its original parameters and genetic coefficien t, failed to predict accurately the growth behavior of cotton grown under the shaded c onditions or the monoculture treatment in our study. Higher discrepancies were noted on individual simulated values for SLA, and LAI, and when compared to their respectiv e measured values. Results of simulations indicated little variation (almos t indistinguishable separation) of predicted SLA (contrary to field results), plus, the predicted LAI in all treatments was significantly lower than the actual data. Accurate prediction of phenological developmen t in plants is a vital process in crop models, as this determines model behavior based on actual field conditions. Apparently, some of the parameters of the model were not correct when the model was applied in shaded environments. SLA and LAI are di rectly influenced by and tend to respond dramatically to light, particularly if light becomes a limiting resource for growth.

PAGE 98

84 Model Evaluation Evaluating SLA and LAI After model parameterization and calibrati on, there was significant improvement in the model’s behavior. Use of the new values of the parameters listed in Tables 5-3 and 54, respectively, showed that the model perfor med very well in simulating SLA of cotton grown in each treatment (Figure 5-3). Va riations in light transmittance that each treatment received based on weat her inputs also resulted in differences in simulated SLA among treatments, with the model simulating hi gher SLA for cotton in Row 1. Consistent with actual field data, the model simulated lower SLA for the control treatment. In terms of assessment of simulated LAI, the model im proved its prediction of LAI for the control treatment [(lower RMSE (0.44) and higher dstatistics (0.73) compared to the default model values (Table 5-5, Figure 5-3)], afte r calibration. Assessment of simulated LAI of cotton in the shaded treatment also showed si milar results (Table 55). However, LAI of cotton in Rows 4 and 8 was still under-predicted (Figure 5-4). Assessment of predicted aboveground biomass Although LAI was still underestimated after model parameterization and calibration, results of simulation showed that the model was robust in simulating aboveground biomass production in each of the treatments (Figure 5-4) as shown in 2001. A similar result was observed in simulating biomass of cotton in each treatment in 2002. Analysis of RMSE and d-statistics in each of the treatments showed a close agreement of the measured and simulated values. Regression analysis also confirmed a close agreement between these parameters in both years, with R2 = 0.95, and R2 = 0.92, for 2001 and 2002, respectively (Figure 5-5).

PAGE 99

85 Across treatment in the shaded enviro nment (Rows 1, 4 and 8), the model underpredicted biomass by only less than 5% and 8% in 2001and 2002, respectively, when compared to the measured data (Table 5-10). Evaluating Relationships Effects of leaf size (SIZELF) on LAI Under shaded conditions, plants tend to incr ease their leaf size to capture greater amounts of light, with possible effects on LAI. However, increasing and or decreasing the size of leaves (SIZELF) of cotton in the mode l did not result in any changes in LAI in any of the shaded treatments (Table 5-6). It is thus apparent that the model does not have a defined relationship between SIZELF and LAI as shown in Figure 5-1, because the CROPGRO model is based on source-sink rela tionships (Messina et al., 2004). During the vegetative stage, leaf area expansion a nd photosynthesis can be sink-limited, thus making the model not to take into account the changes in SIZELF in predicting LAI. Other studies have shown that, in cotton, si nk limitation occurs until the plant reaches about seven nodes and with leaf area of 320 cm2, after which th e growth is totally source driven (Reddy et al., 1991, 1992, 1993 1997a, 1997b). This value (320 cm2) was used as a parameter input in the model. Effects of SLA on LAI and biomass prediction SLA, which is generally influenced by shading, affects LAI. SLA and LAI were known to have relationships in the CROPGRO model (Figure 5-1). Results of simulation indicated that LAI prediction appeared to be sensitive to changes in SLA (Table 5-7), which, in turn, influenced biomass prediction for the shaded treatments (Table 5-7). For instance, a 10% increase in SLA provided a be tter estimate for LAI in shaded treatments compared to that of the model’s original cultivar value of SLA (Table 5-7).

PAGE 100

86 Although predicted LAI increased with incr easing SLA, such an increase in SLA lowered the predicted bioma ss of cotton in the shaded treatments (Table 5-7). For instance, increasing SLA of cotton in Row 8 resulted in lower predicted biomass and higher RMSE and lower d-statistics values (Table 5-7). Decreasing SLA, on the other hand, did improve biomass prediction. The associ ated RMSE and d-statis tics were at least close to the values generated using the orig inal model coefficients. Similar observations were also noted for Row 1 and 4 treatments. Plants receiving more light generally have smaller but thicker leaves, resulting in lowe r SLA and greater photos ynthetic apparatuses, and hence higher photosynthesis rates. This mechanism, which was assumed to be present for cotton in row 8, for instance, expl ains greater biomass prediction in row 8 when SLA was reduced compared to that of cotton in rows 1 and 4. Effects of partitioning on LAI and biomass prediction As indicated earlier, plants grown under sh aded environments tend to allocate more carbon to the aboveground component (i.e., leaves ) at the expense of their root systems. This adaptive mechanism in plants, which was observed in our experiment (Chapters 3 and 4), occurs in order to increase leaf ar ea for greater light capture (Smith and Huston, 1989; Sack and Grubb, 2002, Jose et al., 2002 ). Partitioning of assimilates among vegetative plant parts differs among crop sp ecies and also depends on the growth shortage of the crop itself (Boote et al., 1996). In the CROPGRO-cotton model, dry-matter partitioning is described as a fraction of assimilate allocated to leaves (YLEAF). Although the mode l does not have explicitly defined relationships between shading and partitioning, the simulations showed that, under shaded treatments, increased partitioning to leaves resulted in an increase in simulated LAI when compared to that of LAI predicted using cultivar model values.

PAGE 101

87 Apparently, an additional 16% allocation (ove r the current model values) of carbon in the leaves throughout leaf developmental stages for cotton in row 1 improved LAI estimates in that treatment. On the other hand, a 20% increase in carbon alloca tion in rows 4 and 8 was needed to provide close estimates of LAI with the measured values in these treatments (Figure 5-7). Simulated values for LAI were 3.03 and 3.07 for rows 4 and 8, respectively, while their measured values were 3.13 and 3.59 (Table 5-8, Figure 5-3). Effects of LFMAX on biomass prediction When light becomes a limiting factor for growth, plants attempt to increase the efficiency with which they intercept light a nd convert it into bioma ss (Lambers et al., 1998). Thus, plants grown in shaded envir onment had higher radiation or light use efficiency (RUE) than those plants grown under direct exposure to sun. This mechanism was observed in our experiment (Chapter 2) and cotton grown in the shaded treatments had higher RUE than that of cotton in th e monoculture treatment. Although there is no relationship between shading and diffuse radiation in the CROPGRO-cotton model (Figure 5-1), results of s imulations indicated that ch anging LFMAX values, a model parameter that is also influenced by diffuse radiation, resulted in changes in biomass prediction in shaded treatme nts. Increasing LFMAX from 1.1 to 1.3 indicated a 17%, 15% and 13% increase in biomass predictions for cotton in rows 1, 4 and 8, respectively (Table 5-9). Decreasing LFMAX also result ed in a decrease in biomass prediction regardless of treatment conditions. Messina et al. (2004) indicated that changes in model behavior with changes in LFMAX are also probably related to the impact of these parameters on the total available assimilate pool, which appears to be one of the main driving forces for the model.

PAGE 102

88 Spatial Prediction of Biomass As hypothesized, the model also quantified spatial variations in biomass production between and among cotton planted in differe nt rows subjected to varying light transmittance levels. Differences in simulated biomass prediction in each of these treatments were based on the variations in plant morphological development that the model quantified. Consistent with measured va lues, the model predicted greater yield in row 8, both in 2001 and 2002, while row 1 had the lowest (Table 5-10). An increase in light transmittance to cotton in row 8 indicated a 42% incr ease in biomass over that of row 1 in 2001. Overall, biomass was slight ly under-predicted in rows 1, 4 and 8 treatments both in 2001 and 2002 when compared to the measured values (Table 5-10). Although Montieth et al. (1991) did not a pply modeling tools to evaluate spatial differences in sorghum (Sorghum bicolor) pr oduction, they observed that the yield decreased dramatically due to a 50% reducti on in light transmittance in the plants’ canopy, which was also observed in our experiment (Chapter 2) . Chirko et al. (1996) also noted that planting at further distances from tree lines in an intercropping system in China influenced production of wheat. In their study, a higher level of PAR transmittance resulted in an increase in yield of wheat by 45.7 kg ha-1 between rows at 2.0 and 2.5 m from the tree line (Chirko et al., 1996). A similar observation was made by Droppelman et al. (2000), who reported that yield of sorghum in row pos ition closest to the tree row was lower than yields in rows at further di stance from the trees. These authors concluded that shading caused substantial reduction in yield of their experimental plants, although competition for water could have been a factor as well. The extent of shading could alter the e fficiency of conversion of energy to dry matter by affecting light interception and the photosynthetic activity of individual leaves

PAGE 103

89 (Peri et al., 2003). The model was able to a ccount for this mechanism. It was, thus, confirmed that the model could be used to evaluate production under varying light levels caused by shading, provided that belowgr ound resources are not limiting for growth. Conclusions Results of simulation indicated that aboveground mechanisms affecting production, which includes SLA and maximum photos ynthetic capacity of cotton (LFMAX), influence model behavior. Changes in par titioning also influenced model behavior. Apparently, changing the coefficient values of these parameters resulted in significant changes in simulated values of the parame ters of interest (i.e., LAI and biomass) regardless of environmental conditions, but such change s were more pronounced under the shaded treatments. Under shaded conditi ons, plants not only develop thinner and larger leaves but also exhi bit higher light use efficienc y. These crop traits, which the model took into account, mostly explained the observed differences in simulated values among treatments. Results of simulation also indicated that there could be possible interactive effects of partitioning and LFMAX in the course of formulating a desired model output (accurate prediction) when compar ed to that of the measured parameters (i.e., biomass). Results of simulation also showed that the model could be used to predict production under varying light leve ls in an agroforestry system. Future work should focus also on incorporating belowground dynamics (i.e., competition for water and nutrients) into the model to evaluate production as affected by all major competition vectors in agroforestry systems.

PAGE 104

90 Table 5-1. Initial soil conditi ons in Jay, Florida in 2001 growing season used in the CROPGRO-cotton model. Soil Depth (cm) % Sand % Clay % Silt Bulk Density (g cm-3) % OM TKN (ppm) 0-30 66.615.418.01.23.4 1350.0 30-60 64.619.016.41.22.1 1260.0 60-90 62.621.016.41.21.9 1000.0 90-120 63.219.017.81.21.8 830.0 Table 5-2. Definition of parameters used dur ing the parameterization stage of employing the CROPGRO-cotton model to simulate production of cotton under shaded environment. Parameter Definition Unit SLAMAX The maximum specific leaf area (SLA) that can be attained under low light cm2 g-1 SLAMIN The minimum specific leaf area (SLA) that can be attained under high light cm2 g-1 SLAPAR The curvature of the relationship between SLA and Photosynthetically Active Radiation unitless XPGSLW and YPGSLW Relative change in daily canopy assimilation with change in average canopy specific leaf weight (SLW unitless Table 5-3. Computed coefficient paramete r values based on actual data of cotton collected in a pecan-cotton alleyc ropping system in Jay, Florida. Parameters Old Model Values Computed Coefficient Values SLAPAR -0.047 -0.055 XPGSLW 0.003 0.002 YPGSLW 0.162 0.330

PAGE 105

91 Table 5-4. Calibrated genetic coefficien ts of cotton grown in a pecan-cotton alleycropping system in Jay, Florida. Treatments LFMAX SLAVR SIZELF Model Default Value 1.6 200 270 Control 1.1 230 320 Row 1 1.1 275 399 Row 4 1.1 260 386 Row 8 1.1 270 416 LFMAX Maximum leaf photosynthesis rate at 30 C, 350 vpm CO2, and high light (mg CO2 m-2 s-1) SLAVR Specific leaf area of cultivar under standard growth conditions (cm2 g-1) SIZELF Maximum size of full leaf (cm2) Table 5-5. Comparison of RMSE and d-statis tics of simulated and observed LAI based on calibrated and default model values. Default Calibrated Treatment LAI RMSE d-statistics LAI RMSE d-statistics Control 1.53 0.97 0.51 1.61 0.44 0.73 Row 1 1.41 0.85 0.45 1.51 0.17 0.92 Row 4 1.66 1.73 0.43 2.47 0.93 0.63 Row 8 2.10 1.42 0.53 2.38 0.60 0.75 Table 5-6. Effects of varying size of leaf of cotton (SIZELF) on LAI prediction. Treatment SIZELF LAI RMSE d-statistics Control 32011.710.44 0.73 Row 1 39921.590.16 0.92 4351.590.16 0.92 3501.590.16 0.92 Row 4 38621.860.93 0.63 4251.860.93 0.63 3451.860.94 0.63 Row 8 41622.350.60 0.74 4602.350.60 0.74 3752.330.61 0.74 1 SLA value used in for the control 2 SLA value used in the Cultivar file for each treatment

PAGE 106

92 Table 5-7. Effects of varying SLA on LAI and biomass production. Treatment SLA LAI RMSE dstatistics Biomass RMSE dstatistics Control 2301 1.710.440.734331 465 0.98 Row 1 2752 1.590.180.921904 327 0.93 200 1.460.260.852536 802 0.84 305 1.650.140.941746 337 0.94 245 1.530.210.882097 484 0.91 Row 4 2602 1.860.930.632516 235 0.97 200 1.781.010.623210 713 0.91 285 1.920.890.642322 259 0.96 235 1.810.980.622769 346 0.97 Row 8 2752 2.350.600.753304 512 0.96 200 2.180.700.734181 1123 0.88 300 2.410.580.742996 505 0.93 240 2.370.670.753720 720 0.94 1 Calibrated SLA value of cotton 2 Actual SLA of cotton used in the culti var file as indicated in each treatment

PAGE 107

93 Table 5-8. Effects of partitioning on LAI and biomass production. Treatment % Carbon Partitioned to leaves Predicted LAI RMSE dstatistics Biomass RMSE dstatistics Control Model Value1 1.710.440.734331 465 0.98 Row 1 Model Value 1.590.180.921904 327 0.93 162 2.160.150.942039 496 0.95 20 2.630.220.892096 557 0.90 13 2.140.290.811787 325 0.91 Row 4 Model Value 1.860.930.632516 235 0.97 16 2.740.740.702660 347 0.99 20 3.030.640.742721 417 0.96 13 2.511.090.582387 437 0.95 Row 8 Model Value 2.350.600.753304 512 0.97 16 3.290.510.813442 631 0.95 20 2.590.500.813500 697 0.94 13 3.050.730.733167 458 0.97 1 Calibrated partitioning to leaves used in the SPE.File 2 Percent changes used in the sensitivity analysis of partitioning effects on LAI and biomass

PAGE 108

94 Table 5-9. Effects of changing LFMAX in biomass prediction. Treatment LFMAX Biomass RMSE d-statistics Control 1.1* 4331 465 0.98 Row 1 1.1* 1904 327 0.93 1.2 2075 515 0.91 1.3 2222 496 0.93 1.0 1705 421 0.89 Row 4 1.1* 2516 235 0.97 1.2 2721 455 0.96 1.3 2895 403 0.96 1.0 2276 484 0.92 Row 8 1.1* 3304 512 0.97 1.2 3538 554 0.95 1.3 3721 488 0.96 1.0 3027 476 0.94 * model value Table 5-10. Comparison of simulated and observed biomass of cotton in 2001 and 2002 Year Treatment Biomass (kg ha-1) Observed Simulated 2001 Control 5393 5401 Row 1 2455 2498 Row 4 3438 3253 Row 8 4600 4327 2002 Control 4093 4909 Row 1 1808 1731 Row 4 2387 2278 Row 8 3387 2872

PAGE 109

95 Figure 5-1. Conceptual framework showing shading as it affects biomass production under the CROPGRO-cotton model Shading SIZELF LAI SLA Photosynthesis Biomass Light level Diffuse Radiation LFMAX Partitioning Relationships in CROPGRO Relationships not in CROPGRO

PAGE 110

96 0 100 200 300 400 50100150200 Days After PlantingPredicted SLA Simulated Observed 0 100 200 300 400 50100150200 Days After PlantingPredicted SLA Simulated Observed 0 100 200 300 400 50100150200 Days After PlantingPredicted SLA Simulated Observed 0 100 200 300 400 50100150200 Days After PlantingPredicted SLA Simulated ObservedControl Row 8 Row 4 Row 1 0 100 200 300 400 50100150200 Days After PlantingPredicted SLA Simulated Observed 0 100 200 300 400 50100150200 Days After PlantingPredicted SLA Simulated Observed 0 100 200 300 400 50100150200 Days After PlantingPredicted SLA Simulated Observed 0 100 200 300 400 50100150200 Days After PlantingPredicted SLA Simulated ObservedControl Row 8 Row 4 Row 1 Figure 5-2. Simulated SLA of cotton in each treatment after model calibration.

PAGE 111

97 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 50100150200 Days After PlantingPredicted LAI Control (S) Row 1 (S) Row 4 (S) Row 8 (S) Control (O) Row 1 (O) Row 4 (O) Row 8 (O) Figure 5-3. Simulated LAI of cotton in each treatment after model calibration.

PAGE 112

98 0 1000 2000 3000 4000 5000 6000 7000 50100150200 Days After PlantingAboveground biomass (kg ha-1) Control (S) Row 1 (S) Row 4 (S) Row 8 (S) Control (O) Row 1 (O) Row 4 (O) Row 8 (O) 0 1000 2000 3000 4000 5000 6000 7000 50100150200 Days After PlantingAboveground Biomass (kg ha-1) Control (S) Row 1 (S) Row 4 (S) Row 8 (S) Control (O) Row 1 (O) Row 4 (O) Row 8 (O)2001 2002 0 1000 2000 3000 4000 5000 6000 7000 50100150200 Days After PlantingAboveground biomass (kg ha-1) Control (S) Row 1 (S) Row 4 (S) Row 8 (S) Control (O) Row 1 (O) Row 4 (O) Row 8 (O) 0 1000 2000 3000 4000 5000 6000 7000 50100150200 Days After PlantingAboveground Biomass (kg ha-1) Control (S) Row 1 (S) Row 4 (S) Row 8 (S) Control (O) Row 1 (O) Row 4 (O) Row 8 (O)2001 2002 Figure 5-4. Simulated aboveground biomass of cotton in each treatment in 2001 and 2002 growing seasons.

PAGE 113

99 R2 = 0.95 0 1000 2000 3000 4000 5000 6000 7000 01000200030004000500060007000 Simulated Biomass (kg ha-1)Measured Biomass (kg ha-1) R2 = 0.920 1000 2000 3000 4000 5000 6000 7000 01000200030004000500060007000 Simulated Biomass (kg ha-1)Observed Biomass (kg ha-1)2001 2002 R2 = 0.95 0 1000 2000 3000 4000 5000 6000 7000 01000200030004000500060007000 Simulated Biomass (kg ha-1)Measured Biomass (kg ha-1) R2 = 0.920 1000 2000 3000 4000 5000 6000 7000 01000200030004000500060007000 Simulated Biomass (kg ha-1)Observed Biomass (kg ha-1)2001 2002 Figure 5-5. Relationship between simulated and observed aboveground biomass of cotton in 2001 and 2002 growing seasons.

PAGE 114

100 CHAPTER 6 SUMMARY AND CONCLUSION The mechanisms and dynamics of ab oveground competition for light and their effects on production have been only partiall y explored in the context of temperate agroforestry systems. If this information is available, temporal and spatial aspects of agroforestry system design and management can be formulated to mitigate the adverse effects of competitive interactions. This study was conducted with the following major objectives: 1) to determine light distributi on and its effects on co tton biomass and lint yield in a pecan alleycroppi ng system (Chapter 2); 2) to determine the production physiology of cotton as influenced by interspeci fic interaction (Chapter 3); 3) to examine morphological plasticity of cotton roots in res ponse to interspecific competition (Chapter 4); and 4) to apply a proce ss-based model to simulate cotton biomass as affected by varying light levels (Chapter 5). In Chapter 2, we quantified temporal and spatial distribution of light and evaluated its effects on overall performance of cott on based on the hypothesis that growth and productivity of cotton would not be adve rsely affected by shading if belowground competition for water and nutrients was allevi ated. We found that despite lower light transmittance in the barrier compared to the monoculture treatment, aboveground plant biomass was comparable to the monoculture in both years of the study. This indicated that cotton in the barrier treatment tolerated moderate shading and provided at least an acceptable yield when belowground competition for resources was alleviated. However, cotton in the non-barrier treatm ent could not withstand severe competition for both above

PAGE 115

101 and belowground resources, thus exhibiting lo wer production when compared to that in the barrier and monoculture treatments. Result s also indicated that , despite shading, the absence of belowground competition for resour ces increased radiation use efficiency (RUE) by 30% in the barrier compared to th e non-barrier treatment. RUE of cotton in both barrier and non-barrier treatments was high er compared to the monoculture because light was limiting in the former treatments. Light interception and absorption were both affected by LAI. LAI between 3.0 and 4.0 was observed to be the optimum LAI providing the highest light absorbance and yield in our experiment. In Chapter 3, we examined the produ ction physiology of cotton. In general, shading by pecan trees resulted in varying leaf morphology of cotton among treatments, with cotton in the barrier and non-barrier treat ments exhibiting greater leaf area, thus resulting in higher SLA than that in the monoculture plants. However, despite having higher SLA in the non-barrier treatment, co mpared to the monoculture, the competitive presence of trees for belowground resources exhibited a 74% a nd 65% decrease in canopy net photosynthetic index (CNPI) of cotton in the non-barrier (18.3 mol m-2 s-1) compared to monoculture (70.7 mol m-2 s-1 ) and barrier (52.7 mol m-2 s-1) treatments, respectively. Relatively higher CNPI in th e barrier treatment resulted in at least comparable aboveground production to that of the monoculture (Chapter 2). We also observed a non-significant variation in SLN of cotton between the barrier and non-barrier treatments, indicating that competition for nitr ogen was not a factor driving productivity in our system. Lower transpiration rates a nd stomatal conductance observed in the nonbarrier treatment compared to the barrier and monoculture treatments is perhaps an indication of drier soil conditions.

PAGE 116

102 In Chapter 4, we examined morphological plasticity of cotton r oots in response to interspecific competition from pecan. We f ound that competition for light along with belowground competition for water severely a ffected root morphological development of cotton in the non-barrier treatmen t. On a plant basis, average (across month) total root length in the non-barrier plants was 25% and 33% lower than that in the barrier and monoculture treatments, respec tively. This resulted in signi ficantly lower RLD in the non-barrier compared to the other two treatments. Cotton in the barrier treatment exhibited the highest root RLD among trea tments followed by monoculture. Generally, roots were concentrated in the upper 30 cm layer of soil. However, RLD was higher in the 30-60 layer in the non-ba rrier in 2002, indicating that competition for water was severe. Further, as a result of interspecific competition for resources, plants in the barrier and non-barrier treatments allocate d more carbon in the shoot at the expense of their root system in order to enhance light capture. This mechanism resulted in lower root-shoot ratios in these treatments compared to the monoculture. In Chapter 5, we used a process-oriented model, CROPGRO-Cotton, a member of the decision support system for agrotechnology (DSSAT) model, to simulate production of cotton in the barrier treatment where light was assumed to be the only limiting resource for its growth. Parameterizati on and calibration were focused on model parameters that were mostly affected by th e amount of light recei ved. These parameters included SLA, SLW, root-shoot ratio, LAI, a nd leaf area. After model parameterization and calibration, we found that the model was very robust and could be used to simulate cotton production under this condition. Strong si gnificant relationships between measured

PAGE 117

103 and simulated values were observed for biomass (R2 = 0.92 and R2= 0.96, respectively, for 2001 and 2002). In general, our studies have shown that despite having an influence on multiple morphological variables and physiological processe s, light as a competitive vector, is not a major determinant of productivity in p ecan-cotton alleycropping systems. Shading effect of pecan trees in barrier and non-ba rrier treatments signif icantly affected the morphological development of cotton plants. Sh ading negatively affected SLA, LAI and root-shoot partitioning patterns of cotton in our system. This, in turn, contributed to low net photosynthetic capacity and biomass producti on of cotton in the non -barrier treatment despite high RUE (Figure 6-1). Although shading affected the morphological development of cotton plants, overall biom ass production in the barrier treatment was comparable to the monoculture treatment. R oot-root interactions that led to water competition between pecan and cotton (Wanve straut et al., 2004) in the non-barrier treatment resulted to low root biomass production. Such mechanism also negatively influenced leaf area development, SLA and LAI of cotton plants. Low LAI resulted to reduction in light capture and absorption capac ity of cotton plants in the non-barrier treatment (Figure 6-1). Our studies further indicated that temperate alleycropping provides a unique opportunity to attain production comparable to monoculture conditions, provided that belowground comp etition for resources is eliminated or alleviated. Management of temperate alleyc ropping systems, particularly pecan and cotton, should take this into account. Over all, the findings from this study provide a unique contribution to our unde rstanding of competition fo r light and its effects on overall production with and without be lowground competition for resources.

PAGE 118

104 Figure 6-1. Model showing the competitive vectors and their influence on biomass production in a pecan-cotton alleycropping system in Jay, Florida. Shading SIZELF LAI SLA Photosynthesis Biomass Light level Diffuse Radiation RUE Partitioning Nutrient Uptake Water Uptake Root Competition

PAGE 119

105 APPENDIX A MODEL INTERFACE

PAGE 120

106 APPENDIX B SAMPLE GRAPHICAL OUTPUT OF THE CROPGRO COTTON MODEL 0 1000 2000 3000 4000 5000 6000 20406080100120140160180 Days after Planting Tops wt kg/ha (Control) Tops wt kg/ha (Row 1) Tops wt kg/ha (Row 4) Tops wt kg/ha (Row 8) Tops wt kg/ha (UFJY0201 COT) TRT 1 Tops wt kg/ha (UFJY0201 COT) TRT 2 Tops wt kg/ha (UFJY0201 COT) TRT 3 Tops wt kg/ha (UFJY0201 COT) TRT 4

PAGE 121

107 APPENDIX C SAMPLE OVERVIEW OUTPUT OF THE CROPGRO-COTTON MODEL *SIMULATION OVERVIEW FILE *DSSAT Cropping System Model Ver. 4.0.1.000 Feb 21, 2005; 15:42:04 *RUN 1 : Control MODEL : CRGRO040 COTTON EXPERIMENT : UFJY0101 CO JAYFL TREATMENT 1 : Control CROP : COTTON CULTIVAR : TEMPLATE CDM ECOTYPE :CO0001 STARTING DATE : MAY 16 2001 PLANTING DATE : MAY 16 2001 PLANTS/m2 : 12.0 ROW SPACING : 100.cm WEATHER : UFJY 2001 SOIL : UFJA010001 TEXTURE : SL ed Bay sandy loam WATER BALANCE : NOT SIMULATED ; NO H2O-STRESS IRRIGATION : NITROGEN BAL. : NOT SIMULATED ; NO N-STRESS N-FERTILIZER : RESIDUE/MANURE : ENVIRONM. OPT. : DAYL= 0.00 SRAD= 0.00 TMAX= 0.00 TMIN= 0.00 RAIN= 0.00 CO2 = R330.00 DEW =0.00 WIND= 0.00 SIMULATION OPT : WATER :N NITROGEN:N N-FIX:N PHOSPH :N PESTS :N PHOTO :C ET :R INFIL:S HYDROL :R SOM :G MANAGEMENT OPT : PLANTING:R IRRIG :R FERT:A RESIDUE:N HARVEST:M WTH:M *SUMMARY OF SOIL AND GENETIC INPUT PARAMETERS COTTON CULTIVAR :IB0002-TEMPLATE CDM ECOTYPE :CO0001 CSDVAR :23.00 PPSEN : 0.01 EMG-FLW:30.00 FLW-FSD:16.00 FSD-PHM : 63.00 WTPSD :0.150 SDPDVR : 20.0 SDFDUR :22.00 PODDUR :15.00 XFRUIT : 0.55 *SIMULATED CROP AND SOIL STATUS AT MAIN DEVELOPMENT STAGES RUN NO. 1 Control CROP GROWTH BIOMASS LEAF CROP N STRESS DATE AGE STAGE kg/ha LAI NUM kg/ha % H2O N -------------------------------------16 MAY 0 Start Sim 0 0.00 0.0 0 0.0 0.00 0.00 16 MAY 0 Sowing 0 0.00 0.0 0 0.0 0.00 0.00 23 MAY 7 Emergence 3 0.01 0.1 0 4.1 0.00 0.00 23 MAY 7 End Juven. 3 0.01 0.1 0 4.1 0.00 0.00 23 MAY 7 Flower Ind 3 0.01 0.1 0 4.1 0.00 0.00 28 MAY 12 First Leaf 7 0.01 1.2 0 3.7 0.00 0.16

PAGE 122

108 27 JUN 42 Flowering 326 0.41 8.0 12 3.8 0.00 0.00 9 JUL 54 Boll > 6mm 847 0.89 10.2 29 3.4 0.00 0.00 17 JUL 62 First Seed 1296 1.23 11.5 43 3.3 0.00 0.00 11 SEP 118 End Leaf 4868 1.85 17.8 118 2.4 0.00 0.00 2 OCT 139 Bolls>.5sz 5657 1.86 19.4 130 2.3 0.00 0.00 5 OCT 142 End Msnode 5776 1.86 19.7 130 2.2 0.00 0.00 12 OCT 149 Cracked Bl 5869 1.85 19.7 127 2.2 0.00 0.00 22 OCT 159 90%Open Bl 4675 0.20 19.7 91 1.9 0.00 0.00 22 OCT 159 Harvest 4675 0.20 19.7 91 1.9 0.00 0.00 *MAIN GROWTH AND DEVELOPMENT VARIABLES @ VARIABLE SIMULATED MEASURED ----------------------Anthesis day (dap) 42 -99 Pod 1 day (dap) 54 -99 Full pod day (dap) 62 -99 Physiological maturity day (dap) 149 -99 Yield at maturity (kg [dm]/ha) 1830 -99 Pod weight at maturity (kg [dm]/ha) 2481 -99 Number at maturity (no/m2) 1260 -99 Unit wt at maturity (g [dm]/unit) 0.1453 -99 Number at maturity (no/unit) 20.00 -99 Tops weight at maturity (kg [dm]/ha 4675 -99 By-product harvest (kg [dm]/ha) 2059 -99 Leaf area index, maximum 1.86 -99 Harvest index at maturity 0.392 -99 Threshing % at maturity 73.78 -99 Grain N at maturity (kg/ha) 48 -99 Tops N at maturity (kg/ha) 91 -99 Stem N at maturity (kg/ha) 24 -99 Grain N at maturity (%) 2.62 -99 Tops weight at anthesis (kg [dm]/ha 326 -99 Tops N at anthesis (kg/ha) 12 -99 Leaf number per stem, maturity 19.69 -99 Grain oil at maturity (%) 11.73 -99 Canopy height (m) 0.93 -99 Day of Harvest Maturity Stage (dap) 159 -99 Cotton YIELD : 1830 kg/ha [DRY WEIGHT] ***********************************************************************

PAGE 123

109 LIST OF REFERENCES Allen S.C. 2003. Nitrogen dynamics in a pecan ( Carya illinoensis K. Koch)-cotton ( Gossyium hirsutum L.) alleycropping system in th e Southern United States. Ph.D. Dissertation. University of Flor ida, Gainesville, Florida. USA. Azam-ali, S.N., R.B. Matthews, J.H. Willia ms, and J.M. Peacock. 1990. Light use, water uptake and performance of individua l components of a sorghum/groundnut intercrop. Expl. Agric. 26: 413-427. Baker, D.N., and B. Acock. 1986. A conceptual model of stress effects. In: J.R.Mauney and J. M. Stewart (eds), Cotton phys iology. Memphis, Tennessee. Cotton Foundation Press. pp. 245-258. Bange, M.P., and S.P. Milroy. 1998. Growth analysis of short and long season cotton cultivars. Proceedings of the 9th Australian Agronomy Conference. Wagga wagga, Australia. Regional Institut e Limited Press. pp. 145-150. Bannayan, M.N., M.J. Crout, and G. Hooge nboom. 2003. Applicat ion of the CERESwheat model for within-season prediction of winter wheat yield in the United Kingdom. Agron. J. 95: 114-125. Beer, J., R. Muschler, D. Kass, and E. So marriba. 1998. Shade management in coffee and cacao plantations. Agrofor. Syst. 38:139-164. Bellow, J.G., and P.K.R. Nair. 2003. Co mparing common methods for assessing understory light availability in shaded-perennial agrofore stry system. Agric. For. Meteorol. 114: 197-211. Bennett, J.M., T.R. Sinclair, L. Ma, and K. J. Boote. 1993. Single leaf carbon exchange and canopy radiation use effici ency of four peanut cul tivars. Peanut Sci. 20:1-5. Bingham, I.J., J.M. Blackwood, and E.A. Stev enson. 1997. Site, scale and time-course for adjustments in lateral root initiation in wheat following changes in C and N supply. Ann. Bot. 80: 97-106. Bingham, I.J., and A.G. Bengough. 2003. Morphol ogical plasticity of wheat and barley roots in response to spatial variations in soil strength. Plant and Soil. 250: 273-282. Bond, B.J., B.T. Farnsworth, R.A. C oulombe, and W.E. Winner. 1999. Foliage physiology and biochemistry in response to li ght gradients in conifers with varying shade tolerance. O ecologia. 120:183-192.

PAGE 124

110 Boote, K.J., J.W. Jones, and N.B. Pickering. 1996. Potential uses and limitations of crop models. Agron. J. 88: 704-716. Boote, K.J., J.W. Jones, G. Hoogenboom, and G.G. Wilkerson. 1997. Evaluation of the CROPGRO-soybean model over a wide range of experiments. In: M.J. Kropff, P.S. Teng, P.K. Agarwal, J. Bouma, B.A.M. B ourman, J.W. Jones, and H.H. van Laar (eds), Applications of system approach es at the field level. Dordrecht, The Netherland. Kluwer Academic Publishers. pp. 113-133. Boote, K.J., J.W. Jones, and G.H. Hoogenboom. 1998. Simulation of crop growth CROPGRO model. In: R.M. Peart, and R.B. Curry (e ds), Agricultural system modeling and simulation. Copenhagen, Denmark. Marcel Dekker. pp 651-693. Bowen, G.D. 1985. Roots as a component of tr ee productivity. In: M.G.R. Cannell, and J.E. Jackson (eds), Attributes of tees as crop plants. Hun tingdon, United Kingdom. Institute of Terrestrial Ecology Publisher. pp 303-315. Campbell, W.J., L.H. Allen, and G. Bowes. 1990. Response of soybean canopy photosynthesis to CO2 concentration, light and te mperature. J. Exp. Bot. 41:427433. Cannell, M.G.R., M. Van Noordwijk, and C.K. Ong. 1996. The central agroforestry hypothesis: The tree must acq uire resources that the crop would not otherwise acquire. Agofor. Syst. 34: 27-31. Chen, J.L., J.F. Reynolds, P.C. Harley, and J.D.Tenhunen. 1993. Coordination theory of leaf nitrogen distribution in a canopy. Oecologia 93:63-69. Chirko, C.P., A.G. Michael, P.V. Nguyen, and J. P. Jiang 1996. Influence of direction and distance from tree on wheat yield and photos ynthetic photon flux density (Qp) in a Paulownia wheat intercropping system. Fo r. Ecol. and Man. 83:171-180. Chirwa, P.W., P.K.R. Nair, and C.S. Ka mara. 1994. Soil moisture changes and maize productivity under alley cropping with Leucaena and Flemingia hedgrows at Chalimbana near Lusaka, Zambia . For. Ecol.and Man. 64: 231-243. Corlette, J.E., C.R. Black, C.K. Ong, and J.L. Monteith. 1992. Aboveground and belowground interactions in a leucae na/millet alleycropping system. II. Light interception and dry matter production. Agric. For. Metreol. 60:73-91. Cregg, B.M., R.O. Teskey, and P.M. Dought erty. 1993. Effects of shade on stress on growth morphology, and carbon dynamics of loblolly pine branches. Trees 7:208213. Dana, M.N., and B.R. Lerner. 2001. Black Walnut toxicity. Purdue University Cooperative Extension Service Paper General Horticulture HO – 193 –W.

PAGE 125

111 Davis, M.A., K.J. Wrage, P.B. Reich, M.G. Tjoelker, T. Schaeffer and C. Muermann. 1999. Survival, growth, photosynthesis of tree seedlings competing with herbaceous vegetation along a water-light-n itrogen gradient. Plant Ecol. 145: 341350. Droppelmann, K.J., J.E. Ephrath, and P.R. Be rlinger. 2000. Tree/crop complementarity in an arid zone runoff agroforestry system in northern Kenya. Agrofor. Sys. 50:1-15. Egli, D.B., and L.H. Bruening. 1992. Planting date and soybean yield: Evaluation of environmental effects with a crop simulation model: SOYGRO. Agric. For. Meteorol. 63: 19-29. Egli, P., and B. Schmid. 1999. Relationships between leaf nitrogen and limitations of photosynthesis in canopies of Solidago altissima . Acta Oecologia. 20:559-570. Ellsworth, D.S, and P.B Reich. 1993. Leaf mass per area, nitrogen content and photosynthetic carbon gain in Acer saccharum seedlings in contrasting forest light environments. Func. Ecol. 6:432-435. Evans, J.R. 1989. Photosynthesis and nitr ogen relationships in leaves of C3 plants. Oecologia. 78:9-19. Field, C. 1983. Allocating leaf nitrogen for maximization of carbon gain: leaf age as a control of the a llocation program. Oecologia. 56: 341-347. Field, C.B., and H.A. Mooney. 1986. The photos ynthesis-nitrogen relationship in wild plants. In: T. J. Givnish (ed), On th e economy of plant form and function. San Diego, California. Cambridge University Press. pp 25-65. Garett, G.H.E., and L. Buck. 1997. Agroforestry practice and policy in the United States of America. For. Ecol. and Man. 91:5-15 Gersani, M., and T. Sachs. 1992. Developmental correlations between root in heterogeneous environments. Plant Cell Environ. 15: 463-469. Gillespie, A.R., S. Jose, D.B. Mengel, W.L. Hoover, P.E. Pope, J.R. Seifert, D.J. Biehle, T. Stall, and T.J. Benjamin. 2000. Defini ng competition vectors in a temperate alleycropping system in the Midweste rn USA 1. Production Physiology. Agrofor. Syst. 48:25-40. Gordon, A. M., S. M. Newman, and P. A Williams. 1997. Temperate Agroforestry System: An overview. Wallingford, E ngland. CAB International. pp 1-6. Govindajaran, M., M. R. Rao, M. N. Mat huva, and P.K.R. Nair. 1996. Soil–water and root dynamics under hedgerow intercroppi ng in semi-arid Kenya. Agron J. 88: 513520.

PAGE 126

112 Grassi, G., and G. Minotta. 2000. Influence of nutrient supply on shade-sun acclimation of Picea abies seedlings: effects on foliar morphology, photosynthetic performance ad growth. Tree Physiol. 20: 645-652. Gregory, P. J. 1994. Root growth and activity. In: Physiology and determination of crop yield. Madison, Wisconsin. American Soci ety of Agronomy Crop Science Society of America, Soil Science Society of America. pp 65-93. Gregory, P. J. 1996. Approaches to modeli ng the uptake of water and nutrients in agroforestry systems. Agrofor. Syst. 34: 51-65. Grongier. J. W., J. R. Seiler, J.A. Peterson, and R.E. Kreh. 1996. Growth and photosynthetic responses of four Virginia piedmont tree spec ies to shade. Tree Physiology 16:773-778. Gulmon, S. L., and C. C. Chu. 1981. The effect s of light and nitrog en on photosynthesis, leaf characteristics, and dry ma tter allocation in chaparral shrub, Diplacus ausantiacus . Oecologia 49: 207-212. Hanson, J. D. 2000. Generic crop production m odel for the root zone water quality model. In L. R. Ahuja (ed), The root zone water quality model. Highland Ranch, Colorado. Water Resource Publishing. pp.145-156. Hanson, J. D., K. W. Rojas, and M. J. Sh affer. 1999. Calibrating the root zone water quality model Agron. J. 91: 171-177. Harrington, R.A., B.J. Brown, and P.B. Reic h. 1989. Ecophysiology of exotic and native shrubs in southern Wisconsin. I. Relati onship of leaf characteristics, resource availability, and phenology to seasonal pa tterns of carbon gain. Oecologia. 80: 356367. Heitholt J.J., W. T. Pettigrew, and W. R. Meredith, Jr. 1992. Light interception and lint yield of narrow-row cotton. Crop Sci. 32:728-733. Hollinger, D.Y. 1996. Optimality and nitrogen allocation in a tree canopy. Tree Physiol. 16: 627-634. Hoogenboom, G., J.W. Jones, P.W. Wilkens, W.D. Batchelor, W.T. Bowen, L.A. Hunt, N.B Pickering, U. Singh, D.C. Godwin, B. B aer. K.J. Boote, J.T. Ritchie and J.W. White. 1994. Crop models. In: G. Y. Tsugi, G. Uehara, and S. Balas (eds), DSSAT v3. Honolulu, Hawaii. University of Hawaii Press. pp. 95-244. Immo, M., and V.R. Timmer. 2000. Vect or competition analysis of a Leucaena -maize alleycropping system in western Ke nya. For. Ecol. and Man.126: 255-268.

PAGE 127

113 Itimu O. A., and Giller K.E., 1997. Distribution of Senna spectabilis , Gliricidia sepium and Maize ( Zea mays L.) roots in an alley croppi ng trial on the Lilongwe Plain, Central Malawi. Ph.D Thesis. Wye Colle ge, University of London, Kent. United Kingdom. Jagtap, S., and C.K. Ong. 1997. Perspectiv e on issues, needs and opportunities for agroforestry models. In: Agroforestry modeling and Research Coordination. Annual Report to ODA. Edinburg, United Kingdom. Institute of Terrestrial Ecology. pp. 110-112. Jones, J.W., G. Hoogenboom, C. H. Porter, K.J. Boote, W.D. Bachelor, L.A. Hunt, P.W. Wilkens, U. Singh, A.J. Gijsman, J.T R itchie. 2003. The DSSAT cropping system model. Europ. J. Agronomy 18: 235-265. Jones, J.W., G.Y. Tsuji, G. Hoogenboom, L. A. Hunt, P.K. Thornton, P.W. Wilkens, D.T. Imamura, W.T. Bowen, U. Singh. 1998. Decision support system for agrotechnology transfer: DSSAT v3. In: G. T. Tsuji, G. Hoogenboom, and P. K. Thornton (eds), Understanding options for agricultural production. Dordrecht, The Netherlands. Kluwer Academic Publishers. pp. 157-177. Jose, S., and A.R. Gillespie. 1997. Leaf area productivity relationship among mixed species hardwood forest communities of th e central hardwood region. For. Sci. 43: 56-64. Jose, S., A.R. Gillespie, J.R. Seifer, and D.J. Biehle. 2000a. Defining competition vectors in a temperate alley cropping system in the Midwestern USA 2. Competition for water. Agrofor. Syst. 48:41-59. Jose, S., A.R. Gillespie, J.R. Seifert, D.J. Biehle. 2000b. Defining competition vectors in a temperate alley cropping system in th e midwestern USA: 3. Competition for nitrogen and litter decomposition dynamics. Agrofor. Sys.48: 61-77. Jose, S., A.R. Gillespie, J.R. Seifert and P.E. Pope. 2001. Comparison of minirhizotron and soil core methods for quantifying root biomass in a temperate alley cropping system. Agrofor. Syst. 52: 161-181. Jose, S., S. Merritt, and C.L. Rams ey. 2002. Growth, nutrition, photosynthesis and transpiration responses of longleaf pine s eedlings to light, water and nitrogen. For. Ecol. and Man. 2003: 335-344. Jose. S., S. Merrit, and C. L. Rams ey. 2003. Growth, nutrition, photosynthesis and transpiration responses of longleaf pine seed lings to light, water, and nitrogen. For. Ecol. and Man. 180:335-344. Jose S., A. R. Gilliespie and S. G. Pallardy. 2004. Interspecific interactions in temperate agroforestry. Agrofor. Syst. 61:237-255.

PAGE 128

114 Karim, A.B., P.S. Savill, and E.R. Rhoade s. 1993. The effects of between-row (alley width) and with-in row spacing of Gliricidia sepium on alley cropped maize in Sierra Leone: growth and yield of maize. Agrofor. Syst. 24:81-93. Kasperbauer, M.J. and W.J. Busscher. 1991. Genotypic differences in cotton root penetration of a compacted subsoil layer. Crop Sci. 31: 1376-1378. Keino, J. 1998. Characterization of cotton ro ots in relation to plant growth, water management and potassium kinetic upta ke parameters. Ph. D. Dissertation. University of Arkansas, Fa yetteville, Arkansas, USA. Kiniry, J.R., C.A. Jones, J.C. O’Toole, R. Blanchet, M. Caberguene, and D.A. Spanel. 1989. Radiation use efficiency in biomass accumulation prior to grain filling for five grain-crops species. Field Crop Res. 20:51-64. Klepper, B., H.M. Taylor, M.G. Huck, and F. L. Ficus. 1973. Water relations and growth of cotton in drying soils. Agron. J. 65: 307-310. Klepper, B. 1991. Root-shoot relationships. In: Y. Waisel et al. (ed). Plant roots: The hidden half. 1st ed. New York City, Marcel Decker Publisher. pp: 15-26. Kozlowski, T.T. and S.G. Pallar dy. 1997. Physiology of Woody Plants, 2nd ed. San Diego, California. Academic Press. p. 411. Kropff, M.J. 1993. Mechanisms of competition fo r light. In: M. J. Kropff, and H. H. van Laar (eds), Modeling crop-week inte ractions. Wallingford, England. CAB International. pp. 33-60. Kropff, M. J. 1994. Competition for resource cap ture in agricultural crops. In: Montieth, J. L., R. K. Scott, M. H. Unswor th (eds), Resource capture by crops. Loughborough, Leicestershire. Nottingha m University Press. pp. 233-255. Kull, O., and P.G. Jarvis. 1995. The role of nitrogen in a simple scheme to scale-up photosynthesis from leaf to ca nopy. Plant Cell Environ. 18:1174-1182. Lai, C.T., G. Katul, J. Butnor, M. Siqueira, D. Ellsworth, C. Maier, K. Johnsen, S. MCkeand, and R. Oren. 2002. Modeling the limits on the response of net carbon exchange to fertilization in a south-ea stern pine forest. Plant, Cell Environ. 20:1095-1119. Lambers, H., F S. Chapin III and T. L. Pons. 1998. Plant Physiological Ecology. New York City. Springer-Verlag. p. 540. Lawson, T. L. and B. T. Kang. 1990. Yield of maize and cowpea in an alleycropping system in relation to available lig ht. Agric. For. Meterol. 52:347-357.

PAGE 129

115 Lehmann, J., I. Peter, C. Steglich, G. Gebauer, B. Huwe, and W. Zech. 1998. Belowground interactions in dryland agro forestry. For. Ecol. and Manage. 111: 157-169. Leuning, R., F. M. Keliher, D.G.G. De Pur y, and E. D. Schulze. 1995. Leaf nitrogen, photosynthesis, conductance and transpiratio n: scaling from leaves to canopies. Plant Cell Environ. 18:1183-1200. Livesley, S.J., P.J. Gregory, and R.J. Bure sh. 2000. Competition in tree-row agroforestry systems. 1. Distribution and dynamics of fi ne roots length and biomass. Plant and Soil. 227: 149-161. Lynch, J. 1995. Root architecture and plant productivity. Plant P hysiology. 109: 7-13. Matthews, R.B. and G. J. Lawson. 1997. Struct ure and applications of the HyCAS model Summary of Proceedings of 3rd DFID (ODA) Agroforestry Modelling Workshop. Edinburgh, United Kingdom. Newba ttle Abbey Publishers. pp. 45-64. Mavromatis, T., K.J. Boote, J.W. Jones, G.G. Wilkerson, and G. Hoogenboom. 2002. Repeatability of model genetic coeffici ents derived from soybean performance trials across different States. Crop. Sci. 42:76-89. Mayus, M., H. Van Keulen, and L. Stroosn ijder. 1999. A model of tree-crop competition for windbreak systems in the Sahel: desc ription and evaluation. Agrofor. Syst. 43:183-201. McMichael, B.L., and J.E. Quisenberr y. 1991. Genetic variation for root-shoot relationships among cotton germplasm. Environ. Exp. Bot. 31, 464-470. McMichael, B.L., and J.E. Quisenberry. 1993. The impacts of the soil environment on the growth of root systems. Environ. Exp. Bot. 33, 53-61. McMichael, B.L., D.R. Upchurch, and J. J. Burke. 1996. Soil temperature derived prediction of root density in cotton. . Environ. Exp. Bot. 36, 303-312. Messina, C.D., J.W. Jones, K. J. Boote, G. Hoogenboom, and J. T. Ritchie. 2004. A model for cotton cropping system design. Bu lletin-Department of Agricultural and Biological Engineering. University of Florida. Gainesville, FL. USA. Miller, A.W., and S.G. Pallardy. 2001. Re source competition across the crop-tree interface in a maize-silver maple temp erate alleycropping stand in Missouri. Agrofor. Syst. 53:247-259. Milroy, P.S., and M.P. Bange. 2003. Nitr ogen and light responses of cotton photosynthesis and implications fo r crop growth. Crop Sci. 43:904-913.

PAGE 130

116 Mitchell, A.K., and T. M. Hinckley. 1993. Effe cts of foliar nitrogen concentration on photosynthesis and water use efficiency in Douglas-fir. Tree Physiology 12:403410. Mobbs, D.C., M.G.R. Cannell, G.J. Lawson, and N.M.J. Crout. 2003. HyPAR: A mechanistically-based agroforestry m odel. Edinburgh, United Kingdom. Institute of Terrestrial Ecology. p.150. Monteith, J. 1978. Reassessment of maximum growth rates for C3 and C4 crops. Exp. Agric. 14: 1-5. Monteith, J.L., C.K. Ong, and J.E. Corl ett. 1991. Microclimatic interactions in agroforestry system. For. Ecol. and Man. 45: 31-44. Monteith, J. L. 1994. Principles of resource ca pture by crops. In: Montie th, J. L., Scott R. K. and Unsworth, M. H. (eds), Recource Capture by Crops. Loughborough, Leicestershire. Nottingham Un iversity Press. pp. 1-16. Mooney, H.A., C. Field, S.L. Gulmon, and F. A. Bazzaz. 1981. Photosynthetic capacity in relation to leaf position in desert versus old-field a nnuals. Oecologia. 50:109-112. Muchow, R.C., and T.R. Sinclair, 1993. Nitr ogen response of leaf photosynthesis and canopy radiation use efficiency in fiel d-grown maize and sorghum. Crop Sci. 34:721-727. Muetzelfeldt R.B., and J. Taylor. 1997. Th e agroforestry modeling environment. In: Agroforestry modeling and research coordination. Annual Report to ODA. Edinburg, United Kingdom. Institute of Terrestrial Ecology. pp. 10-20. Naidu, C.V., and P.M. Swamy. 1993. Effects of shade on growth, biomass production and associated physiological parameters in Pongamia pinnata (Linn.) Pierre. Indian J. Plant Physiology 36: 212-214. Nair, P.K.R. 1993. An introduction to agroforestry. Dordrecht, The Netherlands. Kluwer Academic Publishers. p. 499. Nelson, R.A., J.P. Dimes, E.P. Paningbata n, D.M. Silburn. 1998. Erosion/productivity modeling of maize farming in the Ph ilippines uplands: Parameterizing the agricultural production system si mulator. Agric. Sys. 58:129-146. Newman, E.I. 1966. A method of estimating the to tal length of roots in a sample. J. Appl. Ecol. 3: 139-145. Nippert, J.B. and J.D. Marshall. 2003. S ources of variation in ecophysiological parameters in Douglas-fir a nd grand fir canopies. 23:591-601.

PAGE 131

117 Nissen, T.M., D.J. Midmore, and M.L. Cabrera. 1999. Aboveground and belowground competition between intercropped cabbage and young Eucalyptus torelliana . Agrofor. Sys. 46:83-93. Ong, C.K., J.E. Corlett, R.P. Singh, a nd C.R. Black. 1991. Above and belowground interactions in agroforestry syst ems. For. Ecol. Manage. 45: 45-57. Ong, C.K., and C.R. Black. 1995. Complementarity in resource use in agroforestry systems. In: B.T. Kang, A.O. Osiname, and A. Larbi (eds), Alley farming research and development. Ibadan, Nigeria. Intern ational Institute of Tropical Agriculture (ITTA) Publisher. pp 73-89. Ong, C.K., R.C. Black, F. Marshall, and J. E. Corlette. 1996. Principles of resource capture and utilization of light and water. In: Ong C.K. and P. A. Huxley (eds), Tree-crop interactions–A physiologica l approach. Wallingford, England. CAB International. pp.150-175. Paul, M.J., and C.H. Foyer. 2001. Sink regul ation of photosynthesis. J. Exp. Bot. 52:1383-1400. Pereira, J.S., M.M. Chaves, F. Fonseca, M.C. Araujo and F. Torres. 1992. Photosynthetic capacity of leaves of Eucalyptus globules (Labill) growing in the field with different nutrient and water supplies. Tree Phys. 11: 381-389. Peri, P.L., D.J. Moot, D.L. McNeil, and R.J. Lucas. 2003. Modeling net photosynthetic rate of field grown cocksf oot leaves to account for re growth duration. New Zealand J. Agric. Res. 46:105-115. Pettigrew, W.T., J.C. McCarty Jr., a nd K.C. Vaugh. 2000. Leaf senescence-like characteristics contribute to cotton’ s premature photosynthetic decline. Photosynthesis Res.. 65:187-195. Pettigrew, W.T., J.J. Heitholt, and K. C Vaghn. 1993. Gas exchange differences and comparative anatomy among cotton leaf-t ype isolines. Crop Sci. 33: 1295-1299. Plaut, Z., A. Carmi, and A. Grava. 1996. Cott on root and shoot res ponses to sub-surface drip irrigation and partial wetting of th e upper soil profile. Irrig. Sci. 16: 107-113. Rao, M. R., P. K. R. Nair, and C. K. O ng. 1998. Biophysical interact ions in tropical agroforestry systems. Agrofor. Sys. 38: 3-50. Ramakrishna, A., and C.K. Ong. 1994. Productiv ity and light interception in upland ricelegume intercropping systems. Tropical Agric. 71: 5-11. Ramsey, C.L., and S. Jose. 2002. Management ch allenges of pecan and pine based alley cropping system of the Southern United States . In: W. Schroder, and J. Kort (eds), Temperate agroforestry: Adaptive and mitigative roles. Regina, Canada. Plaines and Prairie Forestry Association. pp. 158-163.

PAGE 132

118 Reddy, V.R., Z. Wang, and K.R. Reddy. 1997a. Gr owth responses of cotton to aldicarb and temperature. . Environ. Exp. Bot. 38, 39-48. Reddy, K.R., H.F. Hodges, and J.M. McKi nion. 1997b. Crop modeling and applications: a cotton example. Adv. Agron. 59: 225-290. Reddy, K.R., H.F. Hodges, and J.M. McKi nion. 1993. Temperature effects on Pima cotton lead growth. Agron. J. 85:681-686. Reddy, K.R., D.N. Baker, and H.F. Hodges. 1991. Temperature effects on cotton canopy growth, photosynthesis, and resp iration. Agron. J. 83: 699-704. Reddy, K.R., V.R. Reddy, and H.F. Hodges. 1992. Temperature effects on early season cotton growth and development. Agron. J. 84:229-237. Reich, P.B., M.B. Walters, M.G. Tjoelker , D. Vanderklein and C. Buschena. 1998a. Photosynthesis and respiration rates depend on leaf and root morphology and nitrogen concentrati on in nine boreal trees species di ffering in relative growth rate. Func. Ecol. 12:395-405. Reich, P.B., D.S. Elsworth, and M.B. Walters . 1998b. Leaf structure (specific leaf area) modulates photosynthesis-ni trogen relationship: evidence from within and across species and functional gr oups. Func. Ecol. 12:948-958. Reich, P.B., D.S. Ellsworth, M.B. Walters, J.M. Vose, C. Greshman, J.C. Volin, and W.D. Bowman. 1999. Generality of leaf trai t relationship: a test across six biomes. Ecol. 80:1955-1969. Robinson, D. 1994. The response of plants to non-uniform supplies of nutrients. New Phytol. 127: 635-374. Rosenthal, W.D., and T.J. Gerik. 1991. Radi ation use efficiency among cotton cultivars. Agron. J. 83:655-658. Rosolem, C.A., P.T. Joao, S. Vanzolini, a nd V.J. Ramos. 1999. The significance of root growth on cotton nutrition in an acidic low-P soil. Plant and Soil. 212: 185-190. Ritchie, J.T. 1998. Soil water balance and plan t stress. In: G. Y. Tsuji, G. Hoogenboom and P.K. Thornton (eds), Understandi ng options for agricultural production. Dordrecht, The Netherlands. Kluwer Academic Publisher. pp. 41-54. Ritchie, J.T., U. Singh, D.C. Godwin, and W.T. Bowen. 199 8. Cereal growth, development and yield. In: G. Y. Tsuji, G. Hoogenboom, and P.K. Thornton (eds), Understanding options for agricultural production. Dordrecht, The Netherlands. Kluwer Academic Publisher. pp. 79-98.

PAGE 133

119 Royce, F.S., J.W. Jones, and J.W. Ha nsen. 2001. Model-based optimization of crop management for climate forecast application. Amer. Soc. Agric. Eng. 44: 13191327. Sack, L., and P.J. Grubb. 2002. The combined im pacts of deep shade and drought on the growth and biomass allocation of shadetolerant woody seedlings. Oecologia. 131: 175-185. Samuelson, L.J. 2000. Effects of nitrogen on leaf physiology and growth of different families of loblolly and sl ash pine. New For. 19:95-107. Schlegel, P., B. Huwe, and W.G. Teixeira . 2004. Modeling species and spacing effects on root zone water dynamics using Hydrus-2D in an Amazonian agroforestry system. Agrofor. Syst. 60:277-289. Schroth, G. 1999. A review of belowground in teractions in agroforestry: focusing on mechanisms and management options. Agrofor. Syst. 43:5-34. Schroth, G. and W. Zech. 1995. Root leng th dynamics in agroforestry with Gliricidia sepium as compared to sole cropping in the semi-deciduous rainforest zone of West Africa. Plant and Soil. 170: 297-306. Sinclair, T.R., and T. Horie. 1983. Leaf n itrogen, photosynthesis and crop radiation use efficiency: A review. Crop Sci. 29:90-98. Sinclair, T.R., J.M. Bennett, and K.J. B oote. 1993. Leaf nitrogen, photosynthesis and radiation use efficiency in peanut. Peanut Sci. 20: 40-43. Singh, R.P., C.K. Ong, and N. Saharan. 1989. Above and below ground interactions in alleycropping in semi-arid India. Agrofor. Sys. 9: 259-274. Shelton, M.G., and G.L. Switzer. 1984. Variat ion in the surface ar ea relationships of loblolly pine fascicles. For. Sci. 30:355-363. Smith, T., M. Huston. 1989. A theory of the spatial and temporal dynamics of plantcommunities. Vegetatio. 83:49-69. Stenberg, P., T. Kuuluvainen, S. Kellomaki, J.C. Grace, E.J. Jokela, and H.L. Gholz. 1994. Crown structure, light interception and pro ductivity of pine trees and stands. Ecological Bulletins 43:20-34. Stenberg, P., S. Palmroth, B.J. Bond, D.G. Sprugel, and H. Smolander. 2001. Shoot structure and photosynthetic e fficiency along the light gradient in a Scots pine canopy. Tree Physiol. 21: 805-814. Taiz, L., E. Zeiger. 2000. Plant Physio logy. Sunderland, Massachusetts. Sinauer Associates. p.425.

PAGE 134

120 Taylor, H.M., and B. Klepper. 1974. Water rela tions of cotton. I. Root growth and water use as related to top growth and so il water content. Agron. J. 66:584-588. Teklehaimanot, J.B., and S.J. Quedraogo. 2004. Fi ne root distribution of pruned trees and associated crops in parkland system in Burkina Faso. Agrofor. Syst. 60: 13-26. Tennant, D. 1975. A test of a modified line inte rsect method of estimating root length. J. Ecol. 63: 995-1001. Traw, B.M., and D. Ackerly. 1995. Leaf position, light levels, and nitrogen allocation in five species on rain forest pioneer species. Amer. J. Bot. 82:1137-1143. Tsuji, G.Y., G. Uehera, and S. Bala s (eds). 1994. DSSAT v3. Honolulu, Hawaii. University of Hawaii Press. p. 400. United States Department of Agriculture (U SDA). 1999. Agroforestry Notes: AF-Notes # 12. Lincoln, Nebraska, USA. Van Noordwijk M., L.Y. Spek, and P. de Willegen. 1994. Proximal root diameter as predictor of total root size for fractal branching mode ls. Plant and Soil. 161: 107117. Van Noordwijk, and M., P. Purnomosidhi. 1995 . Root Architecture in relation to treesoil-crop interactions and s hoot pruning in agroforestry. Agrofor. Syst. 30: 161173. Van Noordwijk, G. Lawson, A. Soumare, F.R. Groot, and K. Hairiah. 1996. Root Distribution of Tree and Crop: Competition and/or complementarity. In. P. Huxley, and C. K. Ong (eds), Tree Crop Inte raction: A Physiological Approach. Wallingford, England. CAB International. p 386. Van Noordwijk, M. and B. Luciana. 1999. WANULCAS. Modeling global change impacts on the soil environment. SEAMEO BIOTROP-GCTE IMPACTS CENTRE. GCTE Working Document #28. Ca nbera, Australia. Southeast Asia Regional Center for Tropical Biology Press. pp. 46-65. Van Noordwijk M., and B. Luciana. 2000. WANULCAS version 2, Background on a model of water, nutrient and light captur e in agroforestry system. Bogor, Indonesia. International Center for Res earch in Agroforestry. p. 186. Wanvestraut, R.H., S. Jose, P.K.R. Nair, a nd B.J. Brecke. 2004. Competition for water in a pecan ( Carya illinoensis K. Koch) – cotton ( Gossypium hirsutum L.) alleycropping system in southern Un ited States. Agrofor. Syst. 60:167-179. Werger, M.J.A., and T. Hirose. 1991. Leaf nitrogen distribution and whole canopy photosynthetic carbon gain in herbace ous stands. Vegetation. 97:11-20.

PAGE 135

121 Young, A. 1997. Agroforestry for soil management. 2nd ed. Wallingford, England. CAB International. p. 320. Young, A., and P. Muraya. 1990. SCUAF: Soil changes under agroforesrtry. Nairobi, Kenya. International Center for Research in Agroforestry. p. 124. Zamora, D. S. 2005. Competition in a pecan -cotton alleycropping systems in the Southern, United States: Is light the limiti ng factor? Agrofor. Syst. J (in press). Zhao, D., and D. Oosterhuis. 1998. Cotton respons es to shade at different growth stages: Nonstructural carbohydrate compos ition. Crop Science 38: 1196-1203. Zhang, H.M., and B.G. Forde. 1998. An arabidopis MADS box genes that controls nutrient-induced changes in root architecture. Science. 279: 407-409. Zhenmin, L., R.G. Percy, C.O. Qualset, and E. Zeiger. 1998. Stomatal conductance predicts yield in irrigated Pima cotton and bread wheat grown at high temperatures. J. Exper. Bot. 49:453-460. Zinkhan, F.C. and D.E Mercer. 1997. An assessme nt of agroforestry systems in southern USA. Agrofor. Syst. 35: 303-321.

PAGE 136

122 BIOGRAPHICAL SKETCH Diomides Santos Zamora was born in August 16, 1971, in the province of Tarlac, Philippines. He attended the University of th e Philippines at Los Banos, College Laguna, where he obtained both his B achelor of Science in forestry and Master of Science in forestry in 1994 and 1999, respectively. He worked as a Researcher at the Philippine Council for Agriculture, Forestry and Natu ral Resources Research and Development (PCARRD) from 1994 to 2001 for the USAID -financed Sustainable Agriculture and Natural Resource Management-Collaborativ e Research Suppor t Program (SANREMCRSP). In May 2001, he was granted a graduate research assistantship to pursue his doctoral degree in the School of Forest Res ources and Conservation at University of Florida, Gainesville, FL.