Agronomy Journa l • Volume 104 , I s sue 3 • 2012 675

Bio

met

ry, M

odel

ing

& S

tati

stic

s

Physical Modeling of U.S. Cotton Yields and Climate Stresses during 1979 to 2005

Xin-Zhong Liang,* Min Xu, Wei Gao, K. Raja Reddy, Kenneth Kunkel, Daniel L. Schmoldt, and Arthur N. Samel

Published in Agron. J. 104:675–683 (2012)Posted online 5 Mar. 2012doi:10.2134/agronj2011.0251Available freely online through the author-supported open access option. Copyright © 2012 by the American Society of Agronomy, 5585 Guilford Road, Madison, WI 53711. All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.

Global warming impacts on regional climate varia-tions are expected to modify our ability to produce food,

fi ber, feed, and fuels on our working lands. Climate-driven regional changes in crop production could dramatically alter management practices and may, in the future, preclude certain crops from being cultivated in their current production zones (Intergovernmental Panel on Climate Change, 2007; Climate Change Science Program, 2008; Karl et al., 2009). Wide-spread alterations in rural economies would aff ect many U.S. agriculture-dependent residents and, ultimately, downstream consumers nationally. Crop growth models, which are tightly coupled with climate models to project future agricultural production scenarios, can aid in the development of eff ective technologies and policies that will allow us to adapt to future

climate trajectories. Th e ability of these models to reproduce the observed relationships between crop yields and climatic stresses must be evaluated to establish credibility for future projections.

Climate variability and changes aff ect crop yields by causing climatic stresses on components of the agricultural ecosystem (Bazzaz and Sombroek, 1996). Climate impacts the production of crops by modifying the biophysical environment in which they grow (Monteith, 1981). Temperature is one of the most important variables controlling crop growth and phenology (Hodges et al., 1993), while rainfall and atmosphere–soil interactions play determining roles in crop productivity (Reddy et al., 2002). Weather conditions interact with soil water availability and aff ect water status in plants, which, in turn, infl uences how plant processes respond to other environmental factors. Photosynthesis, photorespiration, and transpiration are the main plant processes directly aff ected by atmospheric CO2 changes. Elevated CO2 generally enhances photosynthesis due to increased concentrations of the substrate and suppression of photorespiration in C3 plants such as cotton (Reddy et al., 2000) that in turn causes partial stomatal closure (Morison, 1987) and reduces transpiration (Rosenzweig and Hillel, 1998; Kimball and Idso, 1983). Th is increases water use effi ciency (Reddy et al., 2000) and plant yields (Kimball, 1983). It is imperative that any crop growth model fi rst be evaluated for its ability to reproduce these observed relationships between crop yields and climate stresses before its general use to study crop–climate interactions under present and future conditions.

Th erefore, the objective of this study was to evaluate a geo-graphically distributed cotton growth model newly redeveloped from GOSSYM for its ability to simulate historical cotton yield variations and their responses to climatic stresses under actual

ABSTRACTClimate variability and changes aff ect crop yields by causing climatic stresses during various stages of the plant life cycle. A crop growth model must be able to capture the observed relationships between crop yields and climate stresses before its credible use as a prediction tool. Th is study evaluated the ability of the geographically distributed cotton growth model redeveloped from GOSSYM in simulating U.S. cotton (Gossypium hirsutum L.) yields and their responses to climate stresses during 1979 to 2005. Driven by realistic climate conditions, the model reproduced long-term mean cotton yields within ±10% of observations at the 30-km model resolution across virtually the entire U.S. Cotton Belt and correctly captured the critical dependence of their geo-graphic distributions on regional climate characteristics. Signifi cant correlations between simulated and observed interannual variations were found across 87% of the total harvest grids. Th e model also faithfully represented the predictive role of July to August air temperature and August to September soil temperature anomalies on interannual cotton yield changes on unirrigated lands, with a similar but weaker predictive signal for irrigated lands as observed. Th e modeled cotton yields exhibited large, positive correlations with July to August leaf area index. Th ese results indicate the model’s ability to depict the regional impact of climate stresses on cotton yields and suggest the potential predictive value of satellite retrievals. Th ey also provide a baseline reference for further model improvements and applications in the future study of climate–cotton interactions.

X.-Z. Liang, Dep. of Atmospheric and Oceanic Science, Univ. of Maryland, College Park, MD 20742, and Dep. of Atmospheric Sciences, Univ. of Illinois, Urbana, IL 61801; X.-Z. Liang and M. Xu, Earth System Science Interdisciplinary Center, Univ. of Maryland, College Park, MD 20740, and Division of Illinois State Water Survey, Institute of Natural Resource Sustainability, Univ. of Illinois, Champaign, IL 61820; W. Gao, USDA UV-B Monitoring and Research Program, Natural Resource Ecology Lab., and Dep. of Ecosystem Science and Sustainability, Colorado State Univ., Fort Collins, CO 80523; K.R. Reddy, Dep. of Plant and Soil Sciences, Mississippi State Univ., Mississippi State, MS 39762; K. Kunkel, Cooperative Institute for Climate and Satellites, North Carolina State Univ., Asheville, NC 28801, and National Oceanic and Atmospheric Administration, National Climatic Data Center, Asheville, NC 28801; D.L. Schmoldt, USDA National Institute of Food and Agriculture, Washington, DC 20024; and A.N. Samel, Dep. of Geography, Bowling Green State Univ., Bowling Green, OH 43403. Received 10 Aug. 2011 *Corresponding author (xliang@umd.edu).

Abbreviations: CWRF, Climate–Weather Research and Forecasting; LAI, leaf area index; NARR, North American Regional Reanalysis.

676 Agronomy Journa l • Volume 104, Issue 3 • 2012

climatic conditions across the U.S. Cotton Belt. In a continuous eff ort to study cotton–climate interactions, GOSSYM has been totally reengineered to the FORTRAN 90 soft ware standard (Xu et al., 2005), and its representation of physical processes and specifi cation of adjustable parameters (Liang et al., 2012a) have been greatly improved to facilitate full coupling with the mesoscale regional Climate–Weather Research and Forecasting (CWRF) model (Liang et al., 2012b). Th e present stand-alone evaluation, as driven by the best available observational analysis of climate, eliminates complications from any CWRF climate biases via nonlinear feedbacks and thus enables eff ective identi-fi cation of physical and numerical defi ciencies specifi c to GOS-SYM. Th is eff ort is a prerequisite for application of the fully coupled CWRF–GOSSYM across the same computational domain and grid resolution.

MATERIALS AND METHODS

Model Simulations and Observational Data

Th e redeveloped, geographically distributed GOSSYM, at the CWRF 30-km grid spacing, is integrated continuously for the entire growing season (with the longest span from 30 April to 12 November in New Mexico) during 1979 to 2005 under realistic climate conditions and agricultural practices. Th e original GOSSYM has 55 parameters for users to calibrate information specifi c to the cotton cultivar at a farm site (Boone et al., 1993; Reddy et al., 2003) and an additional 30 parame-ters to be specifi ed as input for soil conditions and management practices. To facilitate the spatially distributed modeling of climate–crop interactions, Liang et al. (2012a) fi rst minimized the input parameter list by replacing most of these quantities with the best available physical representations and observa-tional estimates. Th e remaining two new adjustable parameters with regional dependence (initial NO3 amount in the top 2 m of soil and the ratio of irrigated water amount to potential evapotranspiration) were then determined through inverse modeling to minimize local root mean square errors (RMSEs) of annual cotton yields under realistic climate conditions. Th is new set of parameter specifi cations was adopted in the present study to evaluate GOSSYM’s representation of cotton yields and climate stresses across the U.S. Cotton Belt.

Th e climatic conditions driving the stand-alone GOSSYM integration were taken from the North American Regional Reanalysis (NARR) (Mesinger et al., 2006). It is a long-term, consistent, high-resolution climate data set, represent-ing the best available proxy of observations. Th e NARR adopts a 32-km grid, close to that of CWRF, and provides 3-h atmospheric and land surface data, including precipita-tion, evapotranspiration, runoff , net radiation fl ux, surface air temperature, humidity and wind, and soil temperature and moisture. Given large biases in the NARR model-based product, especially for the southeastern United States, we replaced daily precipitation with an objective analysis based on gauge measurements from 7235 U.S. National Weather Service cooperative stations (see Liang et al. [2004] for the data source and analysis procedure).

Figure 1 illustrates the geographic distributions of rain-fall, daily maximum surface air temperature, incident solar radiation, evapotranspiration, and wind speed at 10-m height

Fig. 1. Geographic distributions of observed (a) rainfall (mm/d), (b) daily maximum surface air temperature (°C), (c) incident solar radiation (W/m2), (d) evapotranspiration (mm/d), and (e) wind speed at 10-m height (m/s) averaged across the growing season during 1979 to 2005.

Agronomy Journa l • Volume 104, Issue 3 • 2012 677

averaged across the growing season during 1979 to 2005. Th ese mean climatic conditions determined the overall planting and yield distributions across the U.S. Cotton Belt, discussed below.

For this study, the main verifi cation data were observed cot-ton yields. Annual yield data for each county are available from the National Agricultural Statistics Service (NASS) archive and were interpolated into the CWRF grids using ArcGIS. (Th e interpolation included six steps using ArcGIS commands: “joinfi eld” to incorporate the plant area, harvest area, and yield data from NASS into the county shape fi le; “shapegrid” to transform the data in ArcGIS shape fi le format to ArcGIS GRID format at 0.01° resolution by nearest neighbor assign-ment approach; “project” to project the data onto the CWRF Lambert Conformal Conic Projection; “resample” to resample the projected data to 1-km resolution by the nearest neigh-bor assignment method; “latticeclip” to clip the above 1-km data into the 30-km grid; and “zonalmean” to get the mean of the 1-km data within the 30-km grid. All of the steps were programmed into one ArcGIS script to be run in batch man-ner.) Th ey provided the reference against which the GOSSYM simulated cotton yields were evaluated. Th e same observational data described above were used to assess the model’s ability to represent cotton growth dependence on climate.

Cross-Validation for Parameter Optimization

Th e model optimization required a long time series of his-torical observations and retrospective predictions to train the redeveloped GOSSYM and derive the appropriate geographic distribution for the two new adjustable parameters. Many previous studies have adopted the cross-validation approach (Efron and Gong, 1983; Michaelsen, 1987) to assess perfor-mance for climate models (Peng et al., 2002; Kharin and Zwi-ers, 2002; Liang et al., 2007) and crop growth models (Irmak et al., 2000; Wallach et al., 2001; Th orp et al., 2007; Xiong et al., 2008) because of constraints imposed by short data records. For example, Irmak et al. (2000) evaluated the ability of a soybean [Glycine max (L.) Merr.] model to predict anthesis, maturity, and yields and found that the errors using cross-vali-dation were similar to those from a variety of trial data. Th orp et al. (2007) evaluated CERES-Maize yield simulations using 5 yr of observed yields and found that cross-validation was an appropriate method of model assessment. Th ey also pointed out that a more stable optimized parameter can be obtained when additional growing seasons are used in the cross-validation.

We focused on annual cotton yield variations during 1979 to 2005 using the leave-one-out cross-validation approach, where the two new adjustable parameters are trained with data from all years except one and then used for GOSSYM to make a prediction of the cotton yield for that excluded year. Specifi -cally, 26 yr of data were used for training in each optimization solution and 27 yr of the resulting prediction were applied for verifi cation. Th us, the total number of annual data samples for training was 26, and 27 for verifi cation, which is suffi cient to obtain robust statistics. Figure 2 illustrates the relative mean biases [

p o o( )/Y Y Y− ] of the cotton yields predicted by GOS-SYM using the cross-validation approach, where Y denotes the mean of the cotton yields from 1979 to 2005, while the subscripts p and o represent predictions and observations, respectively.

Th e redeveloped GOSSYM prediction performed well, where the long-term mean modeled yields were within ±10% of observations in the 30-km grids across major areas of the U.S. Cotton Belt, and –4% overall (Fig. 2). Th e largest underestima-tions (less than –20%) of the predictions occurred along the Arizona–California border and in much of central and eastern Texas. In the Southeast, predictions were generally within ±10% of observations, with larger overestimations (>20%) in some areas. In contrast, the original GOSSYM overestimates yields by 27 to 135% on the state level and 92% overall (Liang et al., 2012a). Clearly, the redeveloped GOSSYM signifi cantly improved the prediction of mean cotton yields. In addition, the mean bias distribution based on the cross-validation shown here is very close to that using a single optimization with all 27 yr of data (Liang et al., 2012a). Th ese results indicate that the method for optimizing the two new parameters through inverse model-ing is appropriate and robust. Th is establishes the credibility of using the redeveloped model to study the relationships between cotton yields and climate stresses.

RESULTS

Mean Cotton Yields and Climate Dependences

Figure 3 depicts the observed and GOSSYM-simulated geographic distributions of 1979 to 2005 average cotton yields. In observations, the yields were greatest (>1200 kg ha–1) in California and Arizona, least (<600 kg ha–1) in the Texas High Plains and Oklahoma, medium (about 800 kg ha–1) in New Mexico and the Mississippi Delta, and low (about 600 kg ha–1) in the remaining states. Th e GOSSYM-simulated values agreed well with observations, where the long-term mean modeled yields were within ±10% of observations across most areas of the U.S. Cotton Belt. Th is provides a credible baseline to study crop yield interannual variability and its dependence on climate variations.

Comparison of Fig. 1 and 3 provides a general picture of how the geographic distribution of mean cotton yields depends on the distinct regional climate. Most critical to crop growth are temperature, incident solar radiation, water, and nutrient

Fig. 2. Geographic distribution of cotton yield relative mean biases [ p o o( ) /Y Y Y− ] for the cross-validation, where Y denotes the mean of cotton yields during 1979 to 2005 while the subscripts p and o represent predictions and observations, respectively.

678 Agronomy Journa l • Volume 104, Issue 3 • 2012

supply. Rainfall is abundant in the southeastern states, exceed-ing 3 mm d–1 on average, and is accompanied by a tempera-ture of approximately 30°C, which is close to the 26 to 28°C threshold for optimum growth (Reddy et al., 2002). Mean-while, increased cloud cover causes incident solar radiation to be reduced, and both moisture and nutrients are depleted by rapid drainage in sandy soil (which requires, in addition to rich rainfall, light irrigation <100 mm yr–1 in certain areas, like Georgia), each of which is less than optimal for cotton growth. As a result of the competition between the two oppos-ing environmental conditions, the actual yields in these states fall in the low range of the U.S. Cotton Belt. On the other hand, the southwestern states are identifi ed with abundant daily solar radiation, in excess of 300 W m–2, that promotes cotton growth but excessively dry and hot weather, which causes large water and heat stresses. High cotton yields are achieved in this region through heavy irrigation (?500–900 mm yr–1). Across Texas and Oklahoma, the prevailing low-level southerly jet stream facilitates large evapotranspiration and produces additional water stress. Th e limited availability of water for irrigation in this region contributes to the lowest yields in the U.S. Cotton Belt. Along the Lower Mississippi River, moderate rainfall and large irrigation potential as well as warm temperatures, adequate solar radiation, and deep soils allow large planting acreages and average yields. Th e GOSSYM model realistically captured these physical cotton–climate correspondences.

Figure 4 illustrates the simulated geographic distributions of water and carbohydrate stresses on plant growth, N stress on the cotton boll, as well as annual maximum leaf area index (LAI) and plant height averaged during the growing season.

Fig. 3. Geographic distributions of (a) observed and (b) modeled 1979 to 2005 mean cotton yields.

Fig. 4. Geographic distributions of GOSSYM simulated mean (a) water stress factor, (b) carbohydrate stress factor, (c) N stress factor on the cotton boll, (d) maximum leaf area index (LAI), and (e) maximum plant height as averaged across the growing season during 1979 to 2005.

Agronomy Journa l • Volume 104, Issue 3 • 2012 679

Each stress factor increased as the actual plant stress decreased from the most severe (0.0) to none (1.0). Th e water stress factor (Fig. 4a) was generally larger under irrigated than unirrigated conditions, with mean values of 0.53 and 0.37, respectively. Th is indicates that water availability or precipitation timing for cotton growth was less optimal in the rainfed regions during 1979 to 2005. Higher cotton yields in these regions can be produced as water becomes more abundant. Across the irri-gated regions, the larger water stress factor resulted from model optimization for GOSSYM to produce realistic cotton yields. Uncertainties exist, however, due to unknown actual irrigation practices. Th e water stress factor was largest (?0.58) across Arizona and California, where cotton growth is maintained by heavy irrigation presuming an abundant water supply; it was smallest (?0.25) in the dryland areas of the Texas High Plains, where both irrigation and rainfall are insuffi cient; and it was medium (?0.33–0.54) in the Lower Mississippi River basin and the southeastern states, where abundant rainfall or irrigation occurs.

Across most areas, regardless of irrigation practice, the simu-lated carbohydrate stress factor (Fig. 4b) varied between 0.37 and 0.75. Th us, the CO2 fertilization eff ect on cotton growth is not saturated at present, and higher yields can be projected in the future as atmospheric CO2 concentrations increase. Th e lowest carbohydrate stress factor existed across the dryland areas of the Texas High Plains where the water stress factor also reached its minimum. Across this region, greater cotton productivity can be achieved when water availability and CO2 concentration are enhanced. On the other hand, the model simulated weak N stress across the U.S. Cotton Belt, where the stress factor was between 0.86 and 1.0 everywhere (Fig. 4c). Th is may result from the tendency of farmers to apply adequate fertilizers to maximize cotton yields. Th e impact of this factor on cotton growth shall be addressed in the future, when robust observational data become available.

Note that the annual yield (Fig. 3b), maximum LAI (Fig. 4d), and plant height (Fig. 4e) were all higher across areas with larger water stress factor values. Th ese variables were maximized in Arizona and southern California, where irrigation is unlimited, and were somewhat smaller in northern California and along the Lower Mississippi River, where irrigation is also suffi cient. Under unirrigated conditions, mostly in the southeastern states, the water stress on cotton growth may be substantial, depend-ing on regional rainfall abundance and timing. Across Texas and Oklahoma, severe water stress greatly reduced LAI, plant height, boll weight, and number of open bolls, and consequently led to very low cotton yields (<500 kg ha–1).

Interannual Cotton Yield Variations and Climate Stresses

Figure 5 depicts the observed and GOSSYM-simulated cotton yield interannual standard deviations. Th ere is reason-able agreement between the modeled and observed variabili-ties. Given that the yearly records are independent and have 26 degrees of freedom, the modeled and observed deviations are not statistically diff erent at the 95% confi dence level when their ratios are within 0.67 to 1.50. By this measure, GOSSYM simulated cotton yield interannual variability close to observa-tions across 57% of the harvested areas.

A more important capability in modeling variability, how-ever, is the temporal correspondence between GOSSYM and observed anomalies from 1 yr to another. Th is can be measured by correlating the time series. Figure 5c shows that GOSSYM and observed crop yields are positively correlated across 99% of the grids of the U.S. Cotton Belt. Assuming yearly indepen-dence, correlation coeffi cients >0.32 are statistically signifi cant at the 95% confi dence level. By this measure, signifi cant cor-relations were found across 79% (irrigated), 83% (unirrigated), and 87% (all) of the harvest grids. Note that a grid was defi ned as irrigated when the irrigated area percentage is >80%. while unirrigated grids are those in which the unirrigated area percentage is >80%. Th is result demonstrates the important contribution of climate variations to cotton yields in these areas and explains >50% of the observed interannual variance in 53% of the harvest grids. Th ere are, however, about 3% of the harvest grids having correlations <0.1. Th ese areas may be identifi ed with minor climate control, large model defi ciencies, or observational inaccuracies due to the mapping of USDA county-level data to the CWRF 30-km grid. In particular, correlations were small in southern California and western Arizona. Th is model failure probably resulted from simulation of the incorrect cotton cultivar. In reality, California produces a special upland cotton cultivar known as San Joaquin Valley

Fig. 5. Geographic distributions of cotton yield interannual standard deviations for (a) observations and (b) GOSSYM simulations, and (c) their interannual correlation during 1979 to 2005.

680 Agronomy Journa l • Volume 104, Issue 3 • 2012

Acala, variations of which are also grown in Arizona. Its yield is higher than those of the upland cotton cultivars planted in the southeastern, Mississippi Delta, and Plains states and results from the longer growing season, greater number of hot days, and close control of irrigation. Unfor-tunately, the current GOSSYM does not possess the ability to simulate the growth of this Acala variant.

Previous empirical studies have strived to establish statistical relation-ships between crop yields and both major climate variables and stresses for prediction purposes; however, the relationships, if any, are very complex because it is the integration across many factors (climate stresses, soil characteris-tics, cultural practices) during the entire crop life cycle that determines the fi nal yield. A physically based model, such as GOSSYM, provides a unique tool to discover the underlying climatic controls on crop yields. For example, we sought to identify which climate variables in particular growth periods play a major role in fi nal cotton yield. Th is can be depicted by the time evolu-tion during the life cycle of interannual correlation coeffi cients R(s, t ) between yearly cotton yields Y(s,y) and 30-d running means V(s, t ,y) of the climate variable of interest, where (s,y,t) denotes the specifi c grid, individual year, and Day of the Year, while ( )t indicates the running mean of the variable between days t and t + 29.

Figure 6 illustrates the percent-age of the total harvest areas across the U.S. Cotton Belt where 30-d running correlations of observed and GOSSYM-simulated yields with rainfall, incident solar radiation, daily maximum surface air (2-m height) temperature and root-zone (top 1-m depth) soil temperature are statistically signifi cant at the 95% confi dence level during the growing season, averaged from 1979 to 2005. Note that the correlations may be positive or negative, depending on actual heat, water, light, and nutrient stresses. For example, positive temperature anomalies may foster higher (lower) yields in a cold (hot) region, where the average tempera-ture is below (above) the optimum for cotton growth. Th us, correlation absolute values were considered here for both rainfed and irrigated lands.

Across the rainfed lands, cotton growth in GOSSYM is determined predominantly by climate variations. For the four climate variables examined, the simulated percentage of areas with signifi cant correlations exceeded observations,

especially during the summer peak dependence periods. Th is overestimation probably occurred because the model does not consider plant mortality caused by weather, weeds, and insects or cultural practices. Th e model, however, faithfully captured the overall observed climate control on cotton yields. Th e most important climatic factor that aff ected cotton yields in both observations and GOSSYM is the daily maximum air tempera-ture, which had a distinct plateau of signifi cant correlations in >68% of the harvest areas during late June to early August. Given the use of a 30-d sliding window, this result suggests that the July to August maximum air temperature is a good predic-tor of the annual cotton yield. Th e second most important factor is root-zone soil temperature, where the observed peak,

Fig. 6. Evolution during the growing season (1 April–27 October) of the percentage of area with significant interannual correlations between cotton yield and 30-d running windows of rainfall (Rain), incident solar radiation (Rswi), daily maximum temperature at 2-m height (Tmax), and root-zone soil temperature at 1-m depth (ST1m) as observed (solid lines) and simulated by GOSSYM (dashed lines) during 1979 to 2005. Assuming yearly independence, statistical significance occurs at the 95% level when the correlations are >0.32. The result is shown for rainfed (left) and irrigated (right) lands.

Agronomy Journa l • Volume 104, Issue 3 • 2012 681

relative to maximum temperature, occurred approximately 30 d later and signifi cant correlations occurred across a smaller percentage of the harvest area (51 vs. 68% for maximum temperature). Th e model also correctly simulated this delayed eff ect. Note that soil temperature was not directly measured but produced by the NARR model with certain biases. Th is may partially explain the larger discrepancy that existed between the modeled and observed yield dependence on soil temperature when compared with air temperature. Th e third most important factor is incident solar radiation, for which the observed overall yield dependence peaked in late May to early June across about 35% of the harvest areas. Th e model correctly depicted this peak but maintained an elevated percentage of signifi cant correlations through early July. Th e fourth factor is rainfall, with which signifi cant cotton yield correlations were observed during May to August across about 20% of the harvest areas. Th e model produced a much higher percentage of signifi cant correlations, in excess of 60% of the harvest areas, during late May to early July. Th is large overestimation may have occurred, in part, because the model does not represent the impact of rainfall damage on cotton growth.

For irrigated lands, the percentage of areas having signifi cant cotton yield correlations with the climate variables shows both the magnitude and phase agreement between GOSSYM and observations. Th e predictive signals, however, are dramatically diff erent than those for rainfed lands. In particular, a relatively large percentage of areas (?40%) had signifi cant correlations with rainfall, incident solar radiation, and daily maximum air temperature during mid-April to May and with root-zone soil temperature (40–60%) in mid-April to June. Th ese peaks occurred approximately 2 mo earlier than they did for rainfed lands. On the other hand, there was a rapid decrease in the percentage of signifi cant areas during early June for the fi rst three variables and late June for the fourth variable. Th is transition is coincident with the development of cotton squares and fl owers. Model output showed that irriga-tion was essentially not applied during the early growing season because the surface temperature and evapotranspiration were relatively low. As such, initial crop growth strongly depended on local climate conditions. As the growing season progressed, more frequent and heavier irrigation regulated crop growth, making climate control less evident. A rapid increase in the percentage of area with signifi cant correlations with daily maximum air temperature occurred in July and early August. Th e model correctly simulated this increase, although it overestimated its magnitude.

Th e above correlation results are not intended to indicate that

rainfall and solar radiation are any less important than tempera-ture at specifi c locations. Th ey only suggest a general tendency under the prevailing regional climate conditions. Across the U.S. Cotton Belt, water stress is generally not severe because rainfall and irrigation are suffi ciently plentiful. Given that nutrient stress is also weak and solar radiation is abundant, heat stress becomes the dominant factor that controls cotton yields.

Plant physical conditions at certain growth stages in the cotton life cycle are anticipated to be related to fi nal yields (Li et al., 2001; Doraiswamy et al., 2004, 2005; Zhao et al., 2007). One representative variable is the LAI, which can be retrieved through remote sensing (Myneni et al., 2002). Th e LAI pre-dicts photosynthetic primary production and is oft en used as a reference for crop growth (Hay and Porter, 2006). In general, increased LAI implies a more vigorous vegetative canopy during the plant growing period but may reduce boll production in the mature stage as vegetative and fruiting structures compete for the available energy and nutrients. To depict the general yield–LAI relationship across the entire U.S. Cotton Belt, four statistical measures were calculated: gross correlation R( t ), mean correla-tion ( )R t , and frequency (across space) of negative and positive R(s, t ) correlations signifi cant at the 95% confi dence level, where R ( )t is the same as R(s, t ) except that all values at individual grids (s) and years (y) are used as data samples, while ( )R t is the average of R(s, t ) across all grids.

Figure 7 shows the time evolution of the four statistical measures of cotton yield correlations with preceding LAI as simulated by the redeveloped GOSSYM. Across the entire U.S. Cotton Belt, annual yields were identifi ed with a distinct LAI correlation maximum during July to August, reaching 0.6 to 0.7 for ( )R t and 0.4 to 0.7 for ( )R t . Th e number of grids with signifi cant positive interannual correlations increased from

Fig. 7. Evolution from 1 May to 30 October of interannual relationships between annual cotton yields and preceding 30-d running means of leaf area index as measured by the gross correlation (solid thick line) and mean correlation (dashed thick line) as well as the frequency of positive (solid thin line) and negative (dashed thin line) correlations significant at the 95% confidence level across the harvest areas of the entire U.S. Cotton Belt.

682 Agronomy Journa l • Volume 104, Issue 3 • 2012

25% in early July to 60% in late August. Aft er that, as cotton plants were in their mature stage with fully developed fruit-ing structures, the correlations rapidly decreased and became highly negative in late September to October, reaching –0.4 to –0.5 for ( )R t . Th e number of grids with signifi cant negative interannual correlations increased from 18% in mid-September to 38% in mid-October. Th ese modeled relationships are physi-cally intuitive from the perspective of cotton growth charac-teristics and, when validated by observations, have important implications. In particular, the summer LAI information can be an eff ective predictor of fi nal yields at a lead time of 2 to 3 mo (Zhao et al., 2007).

DISCUSSION AND CONCLUSIONSTh e geographic distribution of cotton yields is critically deter-

mined by distinct regional climate characteristics. In the south-western states, plentiful solar radiation and heavy irrigation counteract dry and hot weather to achieve high yields. Along the Lower Mississippi River, moderate rainfall plus convenient irrigation, warm temperatures, and abundant solar radiation produce average yields. In the southeastern states, adequate rainfall and warm temperatures compete against cloud-cover-attenuated solar radiation and drainage-induced moisture and nutrient depletion to maintain lower yields. Finally, across Texas and Oklahoma, large evapotranspiration and insuffi cient water availability for irrigation result in the lowest yields in the U.S. Cotton Belt. Th ese physical cotton–climate correspondences were realistically captured by GOSSYM.

Th e GOSSYM-simulated water stress was generally large without irrigation. Th is indicates that rainfed regions experi-ence less optimal water availability conditions for cotton growth, although higher yields can be produced when water becomes more abundant. Across the irrigated regions, water stress is relatively low when water supplies allow heavy irrigation. Lower yields may be anticipated in the future with a drier climate and less water available for irrigation. Across most irrigated and unirrigated areas, modeled carbohydrate stress ranged from medium to high. Th is suggests that the CO2 fertilization eff ect on cotton growth is not currently saturated, and higher yields can be projected in a future with enhanced CO2 concentrations. Th e GOSSYM model, however, simulated weak N stress, the credibility of which warrants tests against detailed observational measurements when they become available.

Th e GOSSYM model simulated interannual cotton yield variability that was in substantial agreement with observa-tions, with signifi cant correlations in 87% of all harvest grids. Th is indicates the model’s ability to simulate regional climate impacts on cotton yields that explains >50% of the observed interannual variance in 53% of the harvest grids. Lag correla-tion analysis revealed that the July to August maximum air (August to September soil) temperature anomalies may predict annual cotton yield diff erences in unirrigated lands. Th ese time-specifi c air temperature and delayed soil temperature sig-nals are more distinguishable than other climate variables for both observed and GOSSYM yields. A similar, but somewhat weaker, predictive signal was found for irrigated lands in both modeled and observed yields. In addition, simulated cotton yields were highly positively correlated with July to August LAI. Th ese relationships, if confi rmed by observations, have

important implications for predicting annual cotton yields from summer LAI as retrieved from satellite measurements.

Th is study demonstrates that GOSSYM simulations are very promising for the geographically distributed modeling of climate-driven cotton growth. Th ere remain, however, important model output discrepancies with observations. In particular, the model generally overestimated cotton yield interannual variabil-ity, where standard deviations across the entire U.S. Cotton Belt were, on average, about 45% larger than observations. Th e per-centage of areas containing signifi cant cotton yield correlations with climate variables was also greater in GOSSYM than obser-vations, especially during the summer peak dependence periods. Th ese overestimations may have occurred because the model does not currently consider plant mortality caused by catastrophic weather events, such as hurricanes or heavy rainfall (especially in the southeastern states) and strong wind gusts (especially in Texas) or yield loss due to weeds, diseases, and insects. Large diff erences in technology and management, including irrigation, fertilizer and pesticide application, tillage practice, hybrid, and perhaps nutrient stresses such as K defi ciency (Reddy and Zhao, 2005), also exist across individual farm fi elds and time periods, none of which are explicitly treated in the model. Th ese missing factors are presumed to account for a large portion of the dis-agreement between modeled and observed variance and, hence, imperfect temporal correspondence.

In conclusion, the redeveloped GOSSYM model, with signifi -cant physics improvements and the best available soil and cultural specifi cations, realistically reproduced the geographic distribution of mean annual yields across the U.S. Cotton Belt and captured key interannual signals of climatic stresses observed during 1979 to 2005. Th is provides a baseline reference from which further model improvements and applications will be made. In particular, outcomes from the present study form a solid foundation for our next research phase. Th at work will examine the utility of the fully coupled CWRF–GOSSYM model to predict cotton–climate interactions and project future yield changes.

ACKNOWLEDGMENTS

This study benefited from constructive discussions with Carl J. Bernacchi on crop modeling. We thank David Kristovich for valuable comments on the manuscript. This research was mainly supported by the USDA UV-B Monitoring and Research Program, Colorado State University, USDA-CSREES-2009-34263-19774 (subawards to the University of Illinois, G-1449-1), USDA-NIFA-2010-34263-21075 (subawards to the University of Maryland, G1470-3), and USDA-NRI-2008-35615-04666 (subawards to the University of Maryland, G-1469-3). The modeling was conducted at the National Center for Supercomputing Applications facility. The views expressed are those of the authors and do not necessarily reflect those of the sponsoring agencies or the affiliating institutions of the authors.

REFERENCES

Bazzaz, F., and W. Sombroek, editors. 1996. Global climate change and agri-cultural production. John Wiley & Sons, Chichester, UK.

Boone M.Y.L., D.O. Porter, J M. McKinion. 1993. Calibration of GOS-SYM: Th eory and practice. Comput. Electron. Agric. 9:193–203. doi:10.1016/0168-1699(93)90038-3

Agronomy Journa l • Volume 104, Issue 3 • 2012 683

Climate Change Science Program. 2008. Th e eff ects of climate change on agri-culture, land resources, water resources, and biodiversity in the United States. Synthesis and Assessment Product 4.3. U.S. Climate Change Sci-ence Program, Washington, DC.

Doraiswamy, P.C., J.L. Hatfi eld, T.J. Jackson, B. Akhmedov, J. Prueger, and A. Stern. 2004. Crop condition and yield simulations using Land-sat and MODIS. Remote Sens. Environ. 92:548–559. doi:10.1016/j.rse.2004.05.017

Doraiswamy, P.C., T.R. Sinclair, S. Hollinger, B. Akhmedov, A. Stern, and J. Prueger. 2005. Application of MODIS derived parameters for regional crop yield assessment. Remote Sens. Environ. 97:192–202. doi:10.1016/j.rse.2005.03.015

Efron, B., and G. Gong. 1983. A leisurely look at the bootstrap, the jackknife, and cross‐validation. Am. Stat. 37:36–48. doi:10.2307/2685844

Hay, R., and J.H. Porter. 2006. Th e physiology of crop yield. Blackwell Publ., Oxford, UK.

Hodges, H.F., K.R. Reddy, J.M. McKinion, V.R. Reddy. 1993. Temperature eff ects on cotton. Bull. 990. Miss. Agric. For. Exp. Stn., Mississippi State, MS.

Intergovernmental Panel on Climate Change. 2007. Climate change 2007: Impacts, adaptation, and vulnerability. Contribution of Working Group II to the Th ird Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge Univ. Press, Cambridge, UK.

Irmak, A., J.W. Jones, T. Mavromatis, S.M. Welch, K.J. Boote, and G.G. Wilkerson. 2000. Evaluating methods for simulating soybean cultivar responses using cross validation. Agron. J. 92:1140–1149. doi:10.2134/agronj2000.9261140x

Karl, T.R., J.M. Melillo, and T.C. Peterson. 2009. Global climate change impacts in the United States. Cambridge Univ. Press, New York.

Kimball, B.A. 1983. Carbon dioxide and agricultural yield: An assemblage and analysis of 430 prior observations. Agron. J. 75:779–788. doi:10.2134/agronj1983.00021962007500050014x

Kimball, B.A., and S.B. Idso. 1983. Increasing atmospheric CO2: Eff ects on crop yield, water use and climate. Agric. Water Manage. 7:55–72. doi:10.1016/0378-3774(83)90075-6

Kharin, V.V., and F.W. Zwiers. 2002. Climate predictions with multimodel ensem-bles. J. Clim. 15:793–799. doi:10.1175/1520-0442(2002)0152.0.CO;2

Li, H., R.J. Lascano, E.M. Barnes, J. Booker, L.T. Wilson, K.F. Bronson, and E. Segarra. 2001. Multispectral refl ectance of cotton related to plant growth, soil water and texture, and site elevation. Agron. J. 93:1327. doi:10.2134/agronj2001.1327

Liang, X.-Z., L. Li, K.E. Kunkel, M. Ting, and J.X.L. Wang. 2004. Regional climate model simulation of U.S. precipitation dur-ing 1982–2002, Part 1: Annual cycle. J. Clim. 17:3510–3528. doi:10.1175/1520-0442(2004)0172.0.CO;2

Liang, X.-Z., M. Xu, W. Gao, K.R. Reddy, K.E. Kunkel, D.L. Schmoldt, and A.N. Samel. 2012a. A distributed cotton growth model developed from GOSSYM and its parameter determination. Agron. J. 661–674 (this issue).

Liang, X.-Z., M. Xu, K.E. Kunkel, G.A. Grell, and J.S. Kain. 2007. Regional climate model simulation of U.S.–Mexico summer precipitation using the optimal ensemble of two cumulus parameterizations. J. Clim. 20:5201–5207. doi:10.1175/JCLI4306.1

Liang, X.-Z., M. Xu, X. Yuan, T. Ling, H.I. Choi, F. Zhang, et al. 2012b. Cli-mate–Weather Research and Forecasting model (CWRF). Bull. Am. Meteorol. Soc. (in press).

Mesinger, F., G. DiMego, E. Kalnay, K. Mitchell, P. C. Shafran, W. Ebisuzaki, et al. 2006. North American regional reanalysis. Bull. Am. Meteorol. Soc. 87:343–360. doi:10.1175/BAMS-87-3-343

Michaelsen, J. 1987. Cross-validation in statistical climate forecast models. J. Appl. Meteorol. 26:1589–1600. doi:10.1175/1520-0450(1987)0262.0.CO;2

Monteith, J.L. 1981. Climatic variation and the growth of crops. Q. J. R. Mete-orol. Soc. 107:749–774. doi:10.1256/smsqj.45401

Morison, J.I.L. 1987. Intercellular CO2 concentration and stomatal response to CO2. In: Z. Zeiger and G.D. Farquhar, editors, Stomatal function. Stanford Univ. Press, Stanford, CA. p. 229–251.

Myneni, R., S. Hoff man, Y. Knyazikhin, J. Privette, J. Glassy, Y. Tian, et al. 2002. Global products of vegetation leaf area and fraction absorbed PAR from year one of MODIS data. Remote Sens. Environ. 83:214–231. doi:10.1016/S0034-4257(02)00074-3

Peng, P., A. Kumar, H. van den Dool, and A.G. Barnston. 2002. An analysis of multimodel ensemble predictions for seasonal climate anomalies. J. Geo-phys. Res. 107:4710. doi:10.1029/2002JD002712

Reddy, K.R., P.R. Doma, L.O. Mearns, M.Y.L. Boone, H.F. Hodges, A.G. Richardson, and V.G. Kakani. 2002. Simulating the impacts of cli-mate change on cotton production in the Mississippi Delta. Clim. Res. 22:271–281. doi:10.3354/cr022271

Reddy, K.R., H.F. Hodges, and B.A. Kimball. 2000. Crop ecosystem responses to climatic change: Cotton. In: K.R. Reddy and H.F. Hodges, editors, Climate change and global crop productivity. CAB Int., Wallingford, UK. p. 161–187.

Reddy, K.R., V.G. Kakani, D. Zhao, A. Mohammed, and W. Gao. 2003. Cot-ton responses to ultraviolet-B radiation: Experimentation and algo-rithm development. Agric. For. Meteorol. 120:249–265. doi:10.1016/j.agrformet.2003.08.029

Reddy, K.R., and D. Zhao. 2005. Interactive eff ects of elevated CO2 and potas-sium defi ciency on photosynthesis, growth, and biomass partitioning of cotton. Field Crops Res. 94:201–213. doi:10.1016/j.fcr.2005.01.004

Rosenzweig, C., and D. Hillel. 1998. Climate change and the global harvest. Oxford Univ. Press, Oxford, UK.

Th orp, K., W. Batchelor, J. Paz, A. Kaleita, and K. DeJonge. 2007. Using cross-validation to evaluate CERES-Maize yield simulations within a decision support system for precision agriculture. Trans. ASABE 50:1467–1479.

Wallach, D., B. Goffi net, J.E. Bergez, P. Debaeke, D. Leenhardt, and J.N. Aubertot. 2001. Parameter estimation for crop models: A new approach and application to a corn model. Agron. J. 93:757–766. doi:10.2134/agronj2001.934757x

Xiong, W., I. Holman, D. Conway, E. Lin, and Y. Li. 2008. A crop model cross calibration for use in regional climate impacts studies. Ecol. Modell. 213:365–380. doi:10.1016/j.ecolmodel.2008.01.005

Xu, M., X.-Z. Liang, W. Gao, K.R. Reddy, J. Slusser, and K.E. Kunkel. 2005. Preliminary results of the coupled CWRF–GOSSYM system. Proc. SPIE 5884:68–74.

Zhao, D., K.R. Reddy, V.G. Kakani, J.J. Read, and S. Koti. 2007. Canopy refl ectance in cotton for growth assessment and prediction of lint yield. Eur. J. Agron. 26:335–344. doi:10.1016/j.eja.2006.12.001