HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (4) Citing Articles via Google Scholar Google Scholar Articles by Hu, C. Articles by Ahuja, L. R. Search for Related Content PubMed Articles by Hu, C. Articles by Ahuja, L. R. GeoRef GeoRef Citation Agricola Articles by Hu, C. Articles by Ahuja, L. R. Related Collections Sustainable Agriculture Crop Rotation Systems Nutrient Management

ORIGINAL RESEARCH

Evaluating Nitrogen and Water Management in a Double-Cropping System Using RZWQM

^a Center for Agricultural Resources Research, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences
^b USDA-ARS Great Plains Systems Research Unit, Fort Collins, CO 80526
^* Corresponding author (tim.green{at}ars.usda.gov)

Received 11 January 2005.

	ABSTRACT

TOP EXECUTIVE SUMMARY ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
Simulation of water and nutrient processes can enhance intensive agriculture to help feed the world's population in a sustainable manner. Due to excessive N application, environmental protection and agricultural sustainability have become major issues in agriculture. In this study, we calibrated and tested the RZWQM model to assess N management in a double-cropping system comprised of winter wheat (Triticum aestivum L.) and corn (Zea mays L.) at Luancheng, in the North China Plain. Data, including biomass, grain yield, soil water, and soil and crop N, were used from 2001–2003 field trials applying 200 to 800 kg N ha^–1 yr^–1 for five cropping seasons. In general, soil water, biomass, and grain yields were predicted better than plant N uptake or soil residual N. Once it had been tested and used to improve the understanding of N processes in this cropping system, the model was further used to evaluate the effects of alternative water and N management scenarios on N leaching. Typical application rates of both water and N could be reduced by about half based on these results, which would have high economic, social, and environmental impacts in China. The results also demonstrate the potential of RZWQM for evaluating N and water management practices in other regions and climates of the world with intensive agriculture.

Abbreviations: MRE, mean relative error • NCP, North China Plain • NUE, nitrogen use efficiency • OM, organic matter • RE, relative error • RMSE, root mean squared error • SD, standard deviation

	INTRODUCTION

TOP EXECUTIVE SUMMARY ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
NITROGEN MANAGEMENT remains one of the most challenging tasks for sustainable agriculture, as it affects not only productivity and profit, but also other issues of importance to the environment such as NO₃ contamination of ground and surface waters, greenhouse gas emissions and global warming. Elevated NO₃ in groundwater has often been associated with overapplication of N fertilizers and irrigation water by producers anxious to achieve optimum yields (Schepers et al., 1991). Sexton et al. (1996) reported that NO₃ leaching from corn grown on a sandy loam soil in central Minnesota increased rapidly as the N application rate exceeded 100 kg N ha^–1 yr^–1. As N rates increased to about 250 kg N ha^–1 yr^–1, NO₃ leaching increased exponentially.

Population and economic pressures worldwide have driven a general intensification of crop production. Sustained agricultural production is a particularly important issue for China, with its huge population of 1.3 billion. Winter wheat and summer corn are the staple grain crops. The deficit of grain (demand minus supply) has been about 27 million Mg yr^–1 recently in China (Zhou, 2004). In some high-yielding areas in the North China Plain (NCP), with great expansion of irrigated lands and large input of fertilizers (N application rates up to 450 kg N ha^–1 yr^–1) since the late 1970s, total grain production has increased three times and annual yield is up to 14 Mg ha^–1. The NCP is one of the major agricultural areas in China (3 x 10⁵ km²) with a population of about 130 million people. It provides one-fifth of the nation's food supply. In the future, increased total grain production will be mainly dependent on the increase of grain yield per hectare, because urbanization is decreasing the area of arable land. Fertilizer application is one of the key factors contributing to increased grain yield. Using 1 Mg of fertilizer can cause an increase grain production by 8.84 Mg (Peng, 2000). Nitrogen fertilizer use is expected to rise continuously in China in the 21st century: 25 million Mg in 2005, 27 million Mg in 2010, and 30 million Mg in 2020 (Peng, 2000).

Water management issues are coupled with fertilizer management. The mean annual rainfall of the NCP is about 480 to 500 mm. About 70% of the total rainfall occurs from July to September, the growing season of corn. Rainfall during the winter wheat growing season (October–June) is about 60 to 150 mm. Supplemental irrigation is required to support wheat production, as the consumptive water use of winter wheat is about 430 to 470 mm. Farmers in this region generally irrigate winter wheat four to five times each season, and corn is irrigated once or twice per year.

Excess fertilizer application and overdraft of aquifers for irrigation have caused contamination and depletion of groundwater in the NCP (Zhang et al., 1995, 2003; Hu et al., 2005). Thus, for sustainable agriculture, it is necessary to develop comprehensive and improved N management recommendation tools to advise farmers. To maintain high yields while keeping adverse environmental impacts to a minimum, we need to effectively synthesize the knowledge accrued in various disciplines of agricultural research for timely solutions to practical problems in agriculture. Integration of information from field studies and processes through simulation modeling is recommended to identify solutions to such problems (Elliott and Cole, 1989; Peterson et al., 1993). Agricultural systems simulation models can help synthesize much of the information accumulated from various experiments at various locations (Mathews and Blackmore, 1997). In general, all the reported studies for determining fertilizer recommendations for a locality made use of field experiments that have been done periodically with limited multiyear, multilocation replications, and conclusions are extrapolated statistically or otherwise. Fertilizer responses, depending on the soil and climate (especially rainfall) at the location, vary a great deal among years and locations. Field experiments to capture all the multiyear, multilocation variability are nearly impossible. In this context, dynamic cropping system simulation models, well calibrated and validated against field experimental data, hold the promise of extrapolating the short-duration field experimental results to other years and other locations (universal), making use of long-term weather and soil information (Mathews et al., 2002; Knisel and Turtola, 2000; Mathews and Blackmore, 1997; Godwin and Jones, 1991; Paz et al., 1998, 1999). Concepts from the present study using such simulation tools may be adapted to other locations, climates, and crops, such as other double-cropping systems in Australia (Bresson 1995), Japan (Nakamura et al., 2004), and Portugal (Trindade et al., 1997) where NO₃ leaching is a common issue.

The RZWQM model is a comprehensive, process-based, agricultural system model that can simulate the complex interactions in the soil–plant–atmosphere system as affected by irrigation, fertilization, tillage, and other management practices, and predict their effect on crop yield and water quality (Hanson et al., 1998; Ahuja et al., 2000). The generic crop model included in RZWQM can be parameterized to simulate specific crops (Hanson, 2000). The generic crop model of RZWQM has already been parameterized to simulate corn, wheat, and soybean [Glycine max (L.) Merr.], including validation with measured data (Farahani et al., 1999; Ghidey et al., 1999; Hanson et al., 1999; Jaynes and Miller, 1999; Landa et al., 1999; Martin and Watts, 1999; Ma et al., 2001, 2003; Saseendran et al., 2004; Wu et al., 1999). Furthermore, various N management strategies for rainfed winter wheat in eastern Colorado have been developed using RZWQM (Saseendran et al., 2004). The RZWQM model has not been tested for N management in double-cropping systems involving corn and winter wheat (two crops harvested in any single year). The objectives of this study were to: (i) calibrate and test the RZWQM for simulation of an irrigated wheat–corn double-cropping system in the NCP under a range of N application rates, and (ii) demonstrate the use of an agricultural simulation model for developing general nutrient and irrigation management recommendations that can balance negative environmental impacts with better agronomic yields. The model and methodology have potential for similar applications in other regions of the world.

	MATERIALS AND METHODS

TOP EXECUTIVE SUMMARY ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
Site Description
Field experiments were conducted at Luancheng Agroecosystem Experimental Station (37°53'N, 114°41'E, elevation 50 m), Chinese Academy of Sciences, which is located at the piedmont of the Taihang Mountains, in the NCP. Daily meteorological data of air temperature, solar radiation, wind speed, humidity, and precipitation were collected from a weather station located about 300 m from the experimental site. Mean annual precipitation at the station is about 536 mm, 70% of which occurs during July to September. Annual average air temperature is 12.2°C. The soil type of the area is predominately silt loam (agric, rusty Ustic Cambisols; Zitong, 1999). The predominant cropping system in the region is a winter wheat–corn double-cropping system (two crops harvested in any single year) without a fallow period between the crops. Crops are generally flood irrigated with pumped groundwater.

The field experiment was conducted during a 5-yr period (1998–2003), but as the first 2 yr were used as "start up" to remove system memory effects, detailed data for the simulation study were available only from 2001 to 2003. A randomized block design was used, and there were four N treatments (200, 400, 600, and 800 kg N ha^–1 yr^–1) with three replications each. These rates reflect feasible inputs (below average, average, high, and very high) currently used in the NCP. Each plot was 2.5 by 2.5 m, and was bordered by a concrete wall to the 2-m depth to prevent lateral flow of water and nutrients. A summary of the crop management details is shown in Table 1. A moldboard plow was used for tillage operations. Nitrogen was applied to the ground surface in the form of urea in four equal split doses at (i) preplanting of winter wheat, (ii) the jointing stage of winter wheat, (iii) the seedling stage of corn, and (iv) the tassel stage of corn. The crops were flood irrigated at a rate equivalent to a 70-mm depth of water after each N fertilizer application, i.e., four times a year. The actual irrigation amount applied to each plot every time (Table 1) was recorded with flow meters. Winter wheat was planted in late September or early October, and summer corn was planted in early June the following year after wheat harvest, with little or no fallow period in between the crops. The crop cultivars were Gaoyou 503 for winter wheat and Yedan24 for corn.

The model needs initial conditions for soil water, soil temperature, soil chemistry, soil organic matter, and microorganism pools. Model default values of soil chemistry were used. Unit gradient flow was assumed as the soil bottom boundary condition at 220 cm; however, comparisons between simulated and measured variables were made only to 180 cm.

Model Description
Detailed descriptions of the different components of the RZWQM are available elsewhere (Ahuja et al., 2000; Hanson et al., 1998). In the model, water infiltration is calculated with the Green–Ampt equation (Green and Ampt, 1911) and vertical water redistribution is calculated by solving the Richards equation. Soil hydraulic properties are estimated using the Brooks–Corey equation (Brooks and Corey, 1964) with modifications (Ahuja et al., 2000). The generic crop model included in RZWQM can be parameterized to simulate specific crops (Hanson, 2000). The plant model simulates both plant population development and plant growth. Crop phenology, while not explicitly simulated, is handled through seven growth stages: (i) dormant seeds, (ii) germinating seeds, (iii) emerged plants, (iv) established plants, (v) plants in vegetative growth, (vi) reproductive plants, and (vii) senescent plants. Simulated management practices include crop plantings at different dates, fertilizer (inorganic and organic) and irrigation applications, and tillage operations. Organic manure and crop residue (C and N) dynamics on the soil surface and subsoil are also simulated.

Organic Matter–Nitrogen Cycling
The OMNI submodel simulates Organic Matter–NItrogen cycling in RZWQM (Shaffer et al., 2000). Residues (crop stover, manure, and other organics) are partitioned between fast and slow pools based on their C/N ratios. There are three soil organic matter (OM) pools with C/N ratios of 8 (fast pools), 10 (medium pool), and 12 (slow pool). All of the above pools are dynamically linked. In addition, there are three living microorganism pools for aerobic heterotrophs, autotrophs, and facultative heterotrophs. Residue and OM pools are subject to first-order decay with respect to each C concentration, as affected by the soil water O₂ concentration, soil pH, ion strength, heterotrophic microbial population, soil temperature, and the degree of soil water saturation (Shaffer et al., 1992).

A fraction of the decayed residue and OM is transferred to other pools and the rest of the C is converted to CO₂. The model also uses CO₂ as a C source. Nitrogen is released as NH₄ from the residue and OM pools during the decaying process. Ammonium is nitrified to NO₃ following a zero-order equation as a function of the soil temperature, soil O₂ concentration, soil pH, ion strength, autotrophic microbial population, and the degree of water saturation (Shaffer et al., 1992). Nitrate from nitrification and applied commercial fertilizers is subject to denitrification under anaerobic conditions as a first-order equation. Ammonia volatilization is estimated from the partial pressure gradient of NH₃ in the soil and air, with due consideration for wind speed and soil depth. The denitrification rate coefficient is a function of the soil temperature, soil pH, ion strength, total soil C, anaerobic microbial population, and the degree of soil water saturation. The RZWQM model default values of OM decay, nitrification, denitrification, and volatilization rates were used (Ahuja et al., 2000).

Model Parameterization and Calibration
Field-measured values of soil profile water content, profile soil temperature, crop residue mass, soil profile NO₃, bulk density, and soil OM content of different N experiments at planting in the year 2001 were used as initial conditions for model simulations (Table 2). Crop and soil data of the experiment with N dosage at 200 kg N ha^–1 yr^–1 (WC200N) from 2001 to 2003 were used for model calibration, and data from 400 kg N ha^–1 yr^–1 (WC400N), 600 kg N ha^–1 yr^–1 (WC600N), and 800 kg N ha^–1 yr^–1 (WC800N) dosages were used in the model testing. The RZWQM model needs to be calibrated for soil hydraulic properties, nutrient properties, and plant growth parameters (Hanson et al., 1999).

The crop parameters were calibrated after a sensitivity analysis to identify input parameters that needed calibration. Sensitivity analyses of RZWQM to changes in input parameters have been reported (Ma et al., 2000, Walker et al., 2000). The biomass-to-leaf-area conversion coefficient (SLW) in the regional crop parameter set has the strongest influence on crop yield (Walker et al., 2000; Ma et al., 2000; Table 4); however, changes in SLW alone to achieve correct yield results can lead to unrealistically high leaf area index (LAI) in simulations (Ma et al., 2000). In this study, we optimized all nine regional crop parameters recommended for calibration (Table 4; Hanson et al., 1999). Twelve species-specific parameters also were calibrated to correctly simulate the crop cultivars used in the experiment (Table 5). Procedures and methods for calibrating the crop parameters are described elsewhere (Hanson et al., 1999; Ahuja and Ma, 2002). Farahani et al. (1999) developed plant-growth parameters for corn to test the model in Colorado, USA. These parameters were modified by Ma et al. (2003). Saseendran et al. (2004) parameterized the generic crop model of RZWQM for simulation of winter wheat in eastern Colorado. Calibration of parameters for the corn cultivar Yedan24 and for winter wheat cultivar Gaoyou 503 used in the study for the NCP climate are based on Ma et al. (2003) and Saseendran et al. (2004), respectively. These parameters were calibrated further for the location through optimization using an objective function based on field-measured values of grain yield, biomass, soil water content at different layers, N uptake by grain, N uptake by aboveground biomass, and residual NO₃–N in the soil profile (180 cm) at harvest.

[1]

[2]

[3]

[4]

where P_i is the ith predicted value, O_i is the ith observed value, O_avg is the average of observed values, and n is the number of data pairs. Values of E are equivalent to the coefficient of determination (R²) if the values fall around a 1:1 line of predicted vs. observed data, but E is generally lower than R² when the predictions are biased, and E can be negative.

The objective function is calculated as the average of the normalized prediction mean errors of the above variables. For normalization of the prediction mean errors, their respective average measured values were used. We used a direct search for optimization of the crop parameters, constraining their values to increments of 5% at a time (the number of model runs at this increment level exceeded 100000) between an upper limit of +50% and lower limit of –50% from Saseendran et al. (2004) for the winter wheat crop. A similar procedure was conducted for optimization of the crop parameters for corn from the reported parameters of Ma et al. (2003). The calibrated crop parameters are listed in Tables 4 and 5.

Measured as well as simulated data of soil water, plant N uptake, residual soil N, and grain and biomass yield were analyzed for treatment differences (P < 0.05) by one-way analysis of variance (Dowdy and Wearden, 1991).

Model Applications for Nitrogen and Water Management
Once an agricultural system model is adequately calibrated and tested, it has the potential for use in evaluation of alternative crop–soil management practices for the location in terms of their production potential and impact on the environment. Conventional methods for evaluation of the alternative management practices are based on site- and season-specific experiments that need to be repeated when soils and climates are changed, at immense costs of revenue and time. The model-based method is less expensive and can be used for evaluating alternative crop production scenarios. To demonstrate the potential use of the RZWQM model for testing various management alternatives in corn–winter wheat double-cropping systems, we ran the model with different irrigation and N application rates for the location during the experimental period from 2001 to 2003. We did not have long-term data available for running the model for longer periods in order to analyze the risks related to climate variability of the location on different N and water management practices. Therefore, the results from these investigations should be considered preliminary.

First, irrigation rates of 0 (rainfed), 25, 50, and 100% of the current irrigation rates used in the field experiments (Table 1) were evaluated under a single N rate of 200 kg N ha^–1 yr^–1. This rate was chosen for the analysis based on the fact that, in this study, in both measured and model predicted data, there is no statistically significant difference (as discussed below) in grain yield, biomass yield, N uptake of aboveground biomass, or N uptake of grain between field-based N treatments at 200, 400, 600, and 800 kg N ha^–1 yr^–1. Hence the lowest N rate of 200 kg N ha^–1 yr^–1 was selected for further evaluating the different irrigation rates. In a second step, the optimum irrigation rate was further evaluated for different N rates at 0, 50, 100, 200, 400, 600, and 800 kg N ha^–1 yr^–1. In this way, the model was used to extend field experimental results to more optimal application rates, and to evaluate N leaching in conjunction with grain production at these new rates.

	RESULTS AND DISCUSSION

TOP EXECUTIVE SUMMARY ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
Soil Water Content Simulations
Grain yield, biomass, grain N uptake, biomass N uptake, and residual NO₃–N in the soil at harvest for the WC200N experiment during all five cropping seasons of the experimental period, 2001 to 2003, were used to calculate the objective function (calibration targets) for model calibration. Evaluation of the calibration data set shows that the model can be calibrated to achieve a satisfactory match between simulated and field-measured crop growth (Table 6), soil water contents (Fig. 1 and 2a ), and crop and soil N parameters (Table 6). Simulated water contents in different soil layers showed reasonably good agreement with field-measured values, with an RMSE of 0.053 m³ m^–3 (Fig. 1). In contrast to deeper soil layers with RMSE values of 0.049, 0.050, 0.053, 0.048, 0.052, 0.040, 0.029, and 0.062 m³ m^–3, respectively, for the 40-, 60-, 80-, 100-, 120-, 140-, 160-, and 180-cm depths, simulated water contents in the top 20-cm soil layer (RMSE of 0.060 m³ m^–3) showed less correspondence with the field-measured values. Soil water dynamics in the surface soil layers are more complex than deeper layers due to high spatial and temporal variations in OM contents, macropores, and other properties. In addition, a neutron probe was used for measurement of soil water for all soil layers in this experiment. At the 20-cm depth, it is possible that the neutron probe measurements were not very reliable (Wu et al., 1999; Jaynes and Miller, 1999). Although macropores can be simulated in the model (Ahuja et al., 2000), we did not have field measurements and quantifications for simulating macropores in this study. Nonetheless, simulated soil profile water storage (0–180 cm) showed reasonable agreement with the measured values (Fig. 2a), with an RMSE of 5.9 cm (0.04 m³ m^–3 for depth-averaged volumetric water content).

View this table: [in this window] [in a new window]   Table 6. Comparison of observed and model simulated results during the calibration period (exp.WC200N)

View larger version (36K): [in this window] [in a new window]   Fig. 1. Comparison between model-predicted and field-measured soil water contents at various soil depths for the 200 kg N ha–1 yr–1 treatment (model calibration). Values in parentheses are root mean squared errors of soil water predictions (m3 m–3).

View larger version (21K): [in this window] [in a new window]   Fig. 2. Comparison between the model-predicted and field-measured soil profile (180-cm) water storage: (a) calibration results for the 200 kg N ha–1 yr–1 treatment (WC200N) during 2001 to 2003; (b) validation results for 400, 600, and 800 kg N ha–1 yr–1 (WC400N, WC600N, and WC800N, respectively) during 2001 to 2003. Numbers in parentheses are corresponding root mean squared errors.

View larger version (22K): [in this window] [in a new window]   Fig. 3. Comparison between model-predicted and field-measured soil profile (180-cm) water contents for the 400 kg N ha–1 yr–1 treatment (other N rates have similar trends, not shown).

View larger version (40K): [in this window] [in a new window]   Fig. 4. Comparison between model-predicted and field-measured soil water contents at various soil depths for the 400 kg N ha–1 yr–1 treatment (other N rates have similar trends, not shown). Values in parentheses are root mean squared errors.

During model validation runs, predictions of grain yield of both corn and wheat for the five cropping seasons were in good overall agreement with the field-measured values (Fig. 5 ), and accuracies were comparable to those during calibration. Measured grain yields for all the seasons and N treatments ranged from 3267 to 5502 kg ha^–1 for wheat, and from 5481 to 6919 kg ha^–1 for corn. Corresponding ranges of simulated grain yields were from 3394 to 6145 kg ha^–1 for wheat, and from 5736 to 7317 kg ha^–1 for corn. Relative errors of model predictions of grain yields during individual crop seasons across different N treatments were between –14 and 12% for wheat and 3 and 19% for corn (data not shown). The MRE and RMSE values of grain yield predictions (corn and wheat predictions combined) ranged from 3 to 8% and 524 to 891 kg ha^–1 (compared with a calibration RMSE of 334 kg ha^–1), respectively. Out of 20 grain yield predictions (this includes the calibration treatment), eight were within ±1 standard deviation (SD) of the measured values from the mean (Fig. 5). Of the remaining 12 predictions, nine were within ±2 SD, and the rest within ±3 SD.

View larger version (21K): [in this window] [in a new window]   Fig. 5. Comparison between model-predicted and measured grain yields for the 200, 400, 600, and 800 kg N ha–1 yr–1 (WC200N, WC400N, WC600N, and WC800N, respectively) treatments during 2001 to 2003. Error bars represent ± one standard deviation of the measured grain yield from the mean of replications.

View larger version (24K): [in this window] [in a new window]   Fig. 6. Comparison between model-predicted and measured aboveground biomass for the 200, 400, 600, and 800 kg N ha–1 yr–1 (WC200N, WC400N, WC600N, and WC800N, respectively) treatments during 2001 to 2003. Error bars represent ± one standard deviation of the measured aboveground biomass from the mean of replications.

Field-measured total N uptake by biomass varied from 95 to 201 kg N ha^–1 across application rates and crop seasons, whereas the model predictions during the validation runs produced a comparable range in prediction values from 112 to 220 kg N ha^–1 (Fig. 7 ). In general, both biomass and grain N uptake predictions during validation runs were less accurate than calibration. During validation, the REs of biomass N uptake predictions across crop seasons and N rates ranged from –20 to 35%, and the RMSE values ranged from 29 to 37 kg N ha^–1.

View larger version (21K): [in this window] [in a new window]   Fig. 7. Comparison between the model-predicted and measured aboveground biomass N uptake for the 200, 400, 600, and 800 kg N ha–1 yr–1 (WC200N, WC400N, WC600N, and WC800N, respectively) treatments during 2001 to 2003. Error bars represent ± one standard deviation of the measured aboveground biomass N uptake from the mean of replications.

View larger version (21K): [in this window] [in a new window]   Fig. 8. Comparison between the model-predicted and measured grain N uptake for the 200, 400, 600, and 800 kg N ha–1 yr–1 (WC200N, WC400N, WC600N, and WC800N, respectively) treatments during 2001 to 2003. Error bars represent ± one standard deviation of the measured grain N uptake from the mean of replications.

View larger version (22K): [in this window] [in a new window]   Fig. 9. Comparison between the model-predicted and measured soil residual NO3–N (0–180 cm) at harvest of the crops for the 200, 400, 600, and 800 kg N ha–1 yr–1 (WC200N, WC400N, WC600N, and WC800N, respectively) treatments during 2001 to 2003. Error bars represent ± one standard deviation of the measured soil residual NO3–N from the mean of replications.

Simulated Nitrogen Losses
No field-measured values of volatilization, seepage, and denitrification losses of N from the crop–soil environments are available for this study. We compared the model predictions of these variables with those reported in the literature. Li et al. (2001), Wang et al. (2002), and Cai et al. (2002) reported volatilization losses of 10 to 37, 18, and 1 to 18% of applied N in the field, respectively. The predicted N volatilization losses increased with N application rate from 2 to 13% (Table 7). Cai et al. (2002) reported a 1.8% denitrification loss of applied N in the field, whereas the simulated denitrification losses ranged from 0.2 to 0.4% for various N application rates. In general, model predictions were reasonably close to the values reported for the region. Numerical simulations showed consistent increases in N losses due to volatilization, seepage, and denitrification with increased N application rates (Table 7, Fig. 10 ). Simulation results also show that NH₃ volatilization and N seepage are the main sources of loss of fertilizer-applied N. Nitrogen lost through deep seepage is a potential source for contamination of groundwater in these areas with increased application rates in agriculture (Zhang et al., 1995, 2003), requiring more attention for its control through proper N fertilizer and water management.

View larger version (34K): [in this window] [in a new window]   Fig. 10. The RZWQM-predicted N losses due to denitrification, volatilization, and deep seepage per crop season in the wheat–corn double cropping system under 200, 400, 600, and 800 kg N ha–1 yr–1 (WC200N, WC400N, WC600N, and WC800N, respectively) treatments.

Accumulation of inorganic N in the soil profile accompanies increased N leaching (seepage) as fertilizer amounts are increased under full irrigation rates (Table 7). This indicates excess amounts of both N and water in the soil profile, which are not taken up by the crop. Table 7 also shows increased accumulation of N during the simulation period with increasing application rate, but the N dynamics are not shown here. Due to the relatively short experimental period (1998–2003), full dynamic equilibrium in soil profile N has not been obtained, particularly at the highest application rate. Because crops are planted soon after the harvest of the previous crop in the double-cropping system, the soil residual N generally will be available for use by the subsequent crop. As the rate of N application increases, the inorganic N left in the soil at harvest may not be taken up fully by the subsequent crop. This residual N will be subject to leaching to groundwater when rainfall occurs between crop seasons, potentially causing serious environmental degradation in the region.

Nitrogen Use Efficiency
Nitrogen use efficiency is defined as

The NUE of different N treatments were computed for both measured and model-predicted scenarios (Table 7). To make comparisons of NUE for the different N rates without a control experiment with a zero-N treatment, the calibrated model was run with a zero-N rate and the results were used in the NUE computations of both measured and model-predicted cases. Predicted NUE was in close agreement with the measured values (Table 7). The highest NUE was observed for the 200 kg N ha^–1 yr^–1 treatment in both model-predicted and measured cases.

Nitrogen and Water Management
Differences in predicted grain yield, biomass, and plant N uptake between irrigation treatments at the full rate (Table 1) and 50% of the full rate were not found to be statistically significant in a paired t-test (p = 0.17); however, predicted values of these variables at 0 and 25% of the full rate were significantly different (p = 0.02) from those at the full rate.

Model simulations showed that about 10% of the water applied at the full rate was lost to deep seepage out of the root zone of the crop vs. 1% in the 50% irrigation. In Table 8, predicted deep seepage loss of N to groundwater in the full irrigation treatment (41 kg ha^–1) was about 10 times more than that predicted for the 50% irrigation experiment (4 kg ha^–1). This low seepage rate explains the insensitivity of N seepage to further reductions in irrigation or increased N application rates.

The 50% irrigation rate was further evaluated for optimizing the N application rate. The RZWQM model was run for the five cropping seasons, with N rates of 0, 50, 100, 200, 400, 600, and 800 kg N ha^–1 yr^–1 under irrigation at 50%. For N rates of 200 kg ha^–1 yr^–1, average predicted grain yields were 6276 and 3613 kg ha^–1 for corn and wheat, respectively, and the predicted total plant N uptake was 850 kg N ha^–1 (Table 8). Nitrogen use efficiencies were >0.98 for N rates of 200 kg ha^–1 yr^–1 and below. Deep N seepage losses remained more or less constant, with an average value of 4 kg ha^–1 yr^–1 for N rates of 50, 100, and 200 kg ha^–1 yr^–1. Nitrogen seepage losses did not increase with increased fertilizer rates at the 50% irrigation rate. Predicted N volatilization losses were 8, 6, 6, 8, 12, and 33% of the total applied N for N rates of 50, 100, 200, 400, 600, and 800 kg ha^–1 yr^–1, respectively. Residual soil N at harvest at the end of five cropping seasons for the 400, 600, and 800 kg N ha^–1 yr^–1 treatments were higher by 220, 768, and 1168 kg ha^–1 compared with the 200 kg N ha^–1 yr^–1 treatment (142 kg N ha^–1). Again, N accumulated in the soil profile due to limitations in plant uptake and seepage under reduced irrigation. During longer periods, some of the accumulated N would be leached from the soil profile in very wet years. Extrapolation to longer time periods is possible with the model, but was not pursued in this study. Taking into account the plant N uptake, NUE, residual N accumulation in the soil, and volatilization, the 200 kg N ha^–1 yr^–1 N treatment with 50% of the existing irrigation level is recommended for water and N management in the corn–winter wheat double-cropping system in the NCP. These values could be further refined with higher resolution optimization, but the main effects of water and N have been identified here.

	CONCLUSIONS

TOP EXECUTIVE SUMMARY ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
The RZWQM model was evaluated for modeling a double-cropping system comprised of winter wheat and corn (two crops per year) in the NCP. This is the first evaluation of RZWQM for modeling a double-cropping system. Grain yield, biomass, N uptake, soil water content (total profile and at soil depths), and residual NO₃–N were selected as indicators of model performance, because both production and environmental quality should be considered. Testing of the model against observed data for five cropping seasons during the 3-yr period under four different N treatments indicated that the model is able to simulate many variables of the winter wheat–corn double-cropping system reasonably well. As such, the model has potential for developing sustainable N and water management practices. The best predictions were obtained for grain yield, biomass, and soil water content. The model was able to simulate the wet and dry periods of water in the soil profile. Plant N uptake and residual soil NO₃–N at harvest were predicted with a lesser degree of accuracy in comparison to biomass, grain yield, and soil water. Attempts to rectify this problem through better calibration of the model did not lead to improved results. The generic plant model needs to be improved further to get better predictions of crop N uptake.

Taking into account the variability in the measured values of soil residual NO₃–N, the model predictions were accurate enough to provide guidance for decision making. The simulated N losses indicated that increased N application rates will result in increased N losses by volatilization, seepage, and denitrification. Although we do not have field data for verification of these results in this case, the results fall within the range of values reported in the literature. The simulation results showed that as much as half of the applied N at 800 kg N ha^–1 yr^–1 can be lost through the combined effects of volatilization, seepage, and denitrification. At the present irrigation rate, seepage losses (N leaching) increased by more than 200 kg ha^–1 across the range of application rates. In an initial attempt to optimize the N and water application rates at the location using the validated model, we found that an N rate of 200 kg N ha^–1 yr^–1 with irrigation at 50% of the current application rate will give near-optimum grain yields with reduced contamination of the environment.

High levels of agricultural production in the NCP depend largely on irrigation. Like many other locations worldwide, irrigation is causing rapid declines in the groundwater table, at local rates of 0.8 to 1.0 m yr^–1 (Hu and Yin, 1999). This is a serious concern for agricultural sustainability. Water-saving agricultural practices for reducing irrigation and a net reduction in the crop evapotranspiration losses are necessary to slow down the groundwater table decline. Reducing irrigation by 50% in this region could stabilize the groundwater table (Hu and Yin, 1999), because most of the excess irrigation water is evaporated. Thus, the findings of this study have important implications for groundwater management in the NCP. These conclusions are preliminary, but the resource, environmental, and socioeconomic implications of reducing water and N inputs by approximately 50% are dramatic. To make more concrete and precise recommendations, the model needs to be run with many years of weather data to incorporate the climate variability impacts on crop productivity in this area.

Finally, N application in excess of crop requirements is a potential source for contamination of groundwater, and the N volatilization flux from the applied fertilizer on agricultural lands is a potential source of greenhouse gases in the atmosphere that can contribute to global warming. The calibrated and tested RZWQM model can be a useful tool for developing management practices to alleviate these problems.

	ACKNOWLEDGMENTS

 
This work was supported in part by the National Natural Science Foundation of China (30570335). Comments from Hamid Farahani, Newell Kitchen, and Thomas Harter helped us improve the paper.

	REFERENCES

TOP EXECUTIVE SUMMARY ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 

• Ahuja, L.R., K.W. Rojas, J.D. Hanson, M.J. Shafer, and L. Ma. (ed.). 2000. Root Zone Water Quality Model. Modeling management effects on water quality and crop production. Water Resources Publ., Highlands Ranch, CO.
• Ahuja, L.R., and L. Ma. 2002. Parameterization of agricultural system models: Current approaches and future needs. p. 273–316. In L.R. Ahuja, L. Ma, and T.A. Howell (ed.) Agricultural system models in field research and technology transfer. Lewis Publ., Boca Raton, FL.
• Bresson, L.M. 1995. A review of physical management for crusting control in Australian cropping systems. Research opportunities. Aust. J. Soil Res. 33:195–209.
• Brooks, R.H., and A.T. Corey. 1964. Hydraulic properties of porous media. Hydrol. Pap. 3. Colorado State Univ., Fort Collins.
• Cai, G.X., D.L. Chen, and H. Ding. 2002. Nitrogen losses form fertilizers applied to maize, wheat and rice in the North China Plain. Nutr. Cycling Agroecosyst. 63:187–195.[CrossRef]
• Dowdy, S., and S. Wearden. 1991. Statistics for research. 2nd ed. John Wiley & Sons, New York.
• Elliott, E.T., and C.V. Cole. 1989. A perspective on agroecosystem science. Ecology 70:1597–1602.
• Farahani, H.J., G.W. Buchleiter, L.R. Ahuja, G.A. Peterson, and L.A. Sherrod. 1999. Seasonal evaluation of the Root Zone Water Quality Model in Colorado. Agron. J. 91:212–219.[Abstract/Free Full Text]
• Ghidey, F., E.E. Alberts, and N.R. Kitchen. 1999. Evaluation of the Root Zone Water Quality Model using field measured data from the Missouri MSEA. Agron. J. 91:183–192.[Abstract/Free Full Text]
• Godwin, D.C., and C.A. Jones. 1991. Nitrogen dynamics in soil–plant systems. p. 287–321. In R.J. Hanks and J.T. Ritchie (ed.) Modeling plant and soil systems. Agron. Monogr. 31. ASA, CSSA, and SSSA, Madison, WI.
• Green, W.H., and G.A. Ampt. 1911. Studies on soil physics. 1. Flow of air and water through soils. J. Agric. Sci. (Cambridge) 4:1–24.
• Hanson, J.D. 2000. Generic plant production. p. 81–118. In L.R. Ahuja. K.W. Rojas, J.D. Hanson, M.J. Shaffer, and L. Ma (ed.) Root Zone Water Quality Model. Modeling management effects on water quality and crop production. Water Resources Publ., Highlands Ranch, CO.
• Hanson, J.D., K.W. Rojas, and M.J. Shaffer. 1999. Calibrating the Root Zone Water Quality Model. Agron. J. 91:171–177.[Abstract/Free Full Text]
• Hanson, J.D., L.R. Ahuja, M.D. Shaffer, K.W. Rojas, D.G. DeCoursey, H. Farahani, and K. Johnson. 1998. RZWQM: Simulating the effects of management on water quality and crop production. Agric. Syst. 57(2):161–195.
• Hu, C., and Y. Yin. 1999. Modeling dynamics of groundwater in Luancheng County. (In Chinese.) Progr. Geogr. 17(suppl.):26–31.
• Hu, C., J.A. Delgado, X. Zhang, and L. Ma. 2005. Assessment of groundwater use by wheat (Triticum aestivum L.) in the Luancheng Xian region and potential implications for water conservation in the northwestern North China Plain. J. Soil Water Conserv. 60:80–88.
• Jaynes, D.B., and J.G. Miller. 1999. Evaluation of the Root Zone Water Quality Model using data from the Iowa MSEA. Agron. J. 91:192–200.[Abstract/Free Full Text]
• Knisel, W.G., and E. Turtola. 2000. GLEAMS model application on a heavy clay soil in Finland. Agric. Water Manage. 43:285–309.[CrossRef]
• Landa, F.M., N.R. Fausey, S.E. Nokes, and J.D. Hanson. 1999. Plant production model evaluation for the Root Zone Water Quality Model (RZWQM). Agron. J. 91:220–227.[Abstract/Free Full Text]
• Li, G.T., B.G. Li, and D.L. Chen. 2001. Method for measurement of ammonia volatilization from large area field by Bowen ratio system. (In Chinese.) J. China Agric. Univ. 6(5):56–62.
• Paz, J.O., W.D. Batchelor, B.A. Babcock, T.S. Colvin, S.D. Logsdon, T.C. Kaspar, and D.L. Karlen. 1999. Model-based technique to determine variable rate nitrogen for corn. Agric. Syst. 61:69–75.
• Paz, J.O., W.D. Batchelor, T.S. Colvin, S.D. Logsdon, T.C. Kaspar, and D.L. Karlen. 1998. Analysis of water stress effects causing spatial yield variability. Trans. ASAE 41:1527–1534.
• Peng, L. 2000. Progressive process, prospects and distribution of fertilizer use and grain production in China. Res. Agric. Modernization. 21:14–18.
• Ma, L., J.C. Ascough II, L.R. Ahuja, M.J. Shaffer, J.D. Hanson, and K.W. Rojas. 2000. Root Zone Water Quality Model sensitivity analysis using Monte Carlo simulation. Trans. ASAE 43:883–895.
• Ma, L., D.C. Nielsen, L.R. Ahuja, R.W. Malone, S.A. Saseendran, K.W. Rojas, J.D. Hanson, and J.G. Benjamin. 2003. Evaluation of RZWQM under varying irrigation levels in eastern Colorado. Trans. ASAE 46:39–49.
• Ma, L., M.J. Shaffer, and L.R. Ahuja. 2001. Application of RZWQM for soil nitrogen management. p. 265–301. In M.J. Shaffer, L. Ma, and S. Hansen (ed.) Modeling carbon and nitrogen dynamics for soil management. Lewis Publ., Boca Raton, FL.
• Martin, D.L., and D.G. Watts. 1999. Evaluation of the Root Zone Water Quality Model for conditions in central Nebraska. Agron. J. 91:201–211.[Abstract/Free Full Text]
• Mathews, R., and B.S. Blackmore. 1997. Using crop simulation models to determine optimum management practices in precision agriculture. p. 413–420. In John V. Stafford (ed.) Precision agriculture '97: European Conference on Precision Agriculture, 1st, Univ. of Warwick, Coventry, UK, 7–10 Sept. 1997. BIOS Scientific Publ., Abingdon, UK.
• Mathews, R., W. Stephens, T. Hess, T. Middleton, and A. Graves. 2002. Applications of crop/soil simulation models in tropical agricultural systems. Adv. Agron. 76:31–124.
• Nakamura, K., T. Harter, Y. Hirono, H. Horino, and T. Mitsuno. 2004. Assessment of root zone nitrogen leaching as affected by irrigation and nutrient management practices. Vadose Zone J. 3:1353–1366.[Abstract/Free Full Text]
• Peterson, G.A., D.G. Westfall, and C.V. Cole. 1993. Agroecosystem approach to soil and crop management research. Soil Sci. Soc. Am. J. 57:1354–1360.[Abstract/Free Full Text]
• Saseendran, S.A., D.C. Nielsen, L. Ma, L.R. Ahuja, and A.D. Halvorson. 2004. Modeling nitrogen management effects on winter wheat production using RZWQM and CERES-wheat. Agron. J. 96:615–630.[Abstract/Free Full Text]
• Schepers, J.S., M.G. Moravek, E.E. Alberts, and K.D. Frank. 1991. Maize production impacts on groundwater quality. J. Environ. Qual. 20:12–16.[Abstract/Free Full Text]
• Sexton, B.T., J.F. Mocrief, C.J. Rosen, S.C. Gupta, and H.H. Cheng. 1996. Optimizing nitrogen and irrigation inputs for corn based on nitrate leaching and yield on a coarse-textured soil. J. Environ. Qual. 25:983–992.
• Shaffer, M.J., K.W. Rojas, D.G. DeCoursey, and C.S. Hebson. 1992. Nutrient chemistry process. p. 131–162. In RZWQM Technical documentation. GPSR Tech. Rep. 2. USDA-ARS, Great Plains Systems Res. Unit, Fort Collins, CO.
• Shaffer, M.J., K.W. Rojas, D.G. DeCoursey, and C.S. Hebson. 2000. Nutrient chemistry process—OMNI. p. 119–144. In L.R. Ahuja, K.W. Rojas, J.D. Hanson, M.J. Shaffer, and L. Ma (ed.) Root Zone Water Quality Model. Modeling management effects on water quality and crop production. Water Resources Publ, Highlands Ranch, CO.
• Trindade, H., J. Coutinho, M.L. Van Beusichem, D. Scholefield, and N. Moreira. 1997. Nitrate leaching from sandy loam soils under a double-cropping forage system estimated from suction-probe measurements. Plant Soil 195:247–256.[CrossRef]
• Walker, S.E., J.K. Mitchell, M.C. Hirschi, and K.E. Johnsen. 2000. Sensitivity analysis of the Root Zone Water Quality Model. Trans. ASAE 43:841–846.
• Wang, Z.H., X.J. Liu, and X.T. Jiu. 2002. Measurement of soil volatilization. (In Chinese.) Plant Nutr. Fert. Sci. 8:205–209.
• Wu, L., W. Chen, J. M. Baker, and J. A. Lamb. 1999. Evaluation of the Root Zone Water Quality Model using field-measured data from a sandy soil. Agron. J. 91:177–182.[Abstract/Free Full Text]
• Zhang, W., Z. Tian, and X. Li. 1995. Survey of nitrate contamination in groundwater in relation to N fertilizer use in North China. (In Chinese.) Plant Nutr. Fert. Sci. 1:80–86.
• Zhang, X., D. Pei, and C. Hu. 2003. Conserving groundwater for irrigation in the North China Plain. Irrig. Sci. 21:159–166.
• Zhou, J. 2004. Problems of future grain security in China and suggestions of increasing ensuring capacity of grain production. Bull. Chin. Acad. Sci. 19:40–44.
• Zitong, G. (ed.). 2001. Chinese soil taxonomy. Science Press, Beijing.

This article has been cited by other articles:

				J. A. Kozak, L. Ma, L. R. Ahuja, G. Flerchinger, and D. C. Nielsen Evaluating Various Water Stress Calculations in RZWQM and RZ-SHAW for Corn and Soybean Production Agron. J., June 27, 2006; 98(4): 1146 - 1155. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cite