Elsevier

Geoderma

Volume 140, Issue 3, 15 July 2007, Pages 247-259
Geoderma

RZWQM simulation of long-term crop production, water and nitrogen balances in Northeast Iowa

https://doi.org/10.1016/j.geoderma.2007.04.009Get rights and content

Abstract

Agricultural system models are tools to represent and understand major processes and their interactions in agricultural systems. We used the Root Zone Water Quality Model (RZWQM) with 26 years of data from a study near Nashua, IA to evaluate year to year crop yield, water, and N balances. The model was calibrated using data from one 0.4 ha plot and evaluated by comparing simulated values with data from 29 of the 36 plots at the same research site (six were excluded). The dataset contains measured tile flow that varied considerably from plot to plot so we calibrated total tile flow amount by adjusting a lateral hydraulic gradient term for subsurface lateral flow below tiles for each plot. Keeping all other soil and plant parameters constant, RZWQM correctly simulated year to year variations in tile flow (r2 = 0.74) and N loading in tile flow (r2 = 0.71). Yearly crop yield variation was simulated with less satisfaction (r2 = 0.52 for corn and r2 = 0.37 for soybean) although the average yields were reasonably simulated. Root mean square errors (RMSE) for simulated soil water storage, water table, and annual tile flow were 3.0, 22.1, and 5.6 cm, respectively. These values were close to the average RMSE for the measured data between replicates (3.0, 22.4, and 5.7 cm, respectively). RMSE values for simulated annual N loading and residual soil N were 16.8 and 47.0 kg N ha−1, respectively, which were much higher than the average RMSE for measurements among replicates (7.8 and 38.8 kg N ha−1, respectively). The high RMSE for N simulation might be caused by high simulation errors in plant N uptake. Simulated corn (Zea mays L.) and soybean [Glycine max (L.) Merr.] yields had high RMSE (1386 and 674 kg ha−1) with coefficient of variations (CV) of 0.19 and 0.25, respectively. Further improvements were needed for better simulating plant N uptake and yield, but overall, results for annual tile flow and annual N loading in tile flow were acceptable.

Introduction

Agricultural system models have been improved tremendously in the last two decades because of advancements in computer technology, integration with field research, and real-time data collection. The Root Zone Water Quality Model (RZWQM) has evolved steadily since its debut in 1992 (Ahuja et al., 2000). By partnering with scientists at MSEA (Management Systems Evaluation Areas) sites, team members developing RZWQM have been able to integrate field research results (Watts et al., 1999) and test the model for: 1) water and pesticide movement in Minnesota (Wu et al. 1999); 2) surface runoff, nitrate and pesticide losses to seepage and runoff, and crop yield in Missouri (Ghidey et al., 1999); 3) crop yield, and water, nitrate and pesticide movement in Iowa (Jaynes and Miller, 1999); 4) crop yield, N uptake, plant biomass, leaf area index, soil water content, and soil N in Nebraska (Martin and Watts, 1999); and 5) leaf, stem, and seed biomass of corn in Ohio (Landa et al., 1999).

At another MSEA site near Nashua, IA the RZWQM has been evaluated extensively under tile-drained conditions to estimate pesticide leaching, N loss, and crop growth (Bakhsh et al., 2004a, Bakhsh et al., 2004b, Singh et al., 1996, Kumar et al., 1998a, Kumar et al., 1998b). This site is composed of 36 0.4-ha plots and has gone through three phases of study (1978–1992, 1993–1998, 1999–2003). In previous model applications, RZWQM was used only for selected plots (management systems) and years. Kumar et al. (1998a) tested RZWQM on three plots (Plot #13, #22, and #35) and for 1993 and 1995 growing seasons to study swine manure application on nitrate leaching in tile drains. In a subsequent study, they evaluated RZWQM on six plots (#14, #25, #31 for no till and #13, #22, and #35 for moldboard plow) from 1990 to 1992 to investigate macroporosity effects on pesticide transport to tile drains (Kumar et al., 1998b) and pesticide distributions in soils (Azevedo et al., 1997). Singh and Kanwar, 1995a, Singh and Kanwar, 1995b) and Singh et al. (1996) simulated tile drainage and nitrate-N losses to tile drains using RZWQM for 12 of the plots (continuous corn) from 1990 to 1992. The same set of data was later used by Kumar et al. (1999) to study tillage effects on water and nitrate movement with RZWQM. Bakhsh et al. (1999) applied RZWQM to three manure plots (#11, #23, and #27) for 1993–1996 data to study swine manure application on nitrate-N transport to tile drains. In order to fully integrate RZWQM with the long-term field study near Nashua, we need to simulate all the management practices (crop rotation, tillage, and N and manure management) for all the years on all the plots. Doing so will: 1) quantify weather variability; 2) identify possible experimental errors due to equipment failure or mismanagement; and 3) fill in data gaps that will allow analysis of water and N balances in the soil–plant–atmosphere continuum.

Since most RZWQM applications at other locations used only 2 to 4 years of field data as shown above (Ma et al., 2000, Ma et al., 2003, Bakhsh et al., 2004a, Bakhsh et al., 2004b, Jaynes and Miller, 1999, Kumar et al., 1998a, Kumar et al., 1998b), in this study, we recalibrated an improved version of RZWQM for one plot using the 26 years of data collected in Nashua, Iowa of the USA and then applied the calibrated soil and plant parameters to other plots with various management practices. Therefore, our objectives were to (1) demonstrate the calibration of RZWQM for plant and soil parameters using long-term field experiments and (2) analyze the overall goodness-of-simulation of RZWQM for crop yields as well as long-term water and N balances under different management practices. However, detailed analyses of simulation results for a specific management effect (rotation, tillage, and N management) were discussed by Ma et al., 2007a-this issue, Malone et al., 2007-this issue.

Section snippets

Materials and methods

The experiment was conducted at Iowa State University's Northeast Research Center near Nashua, IA. The three dominant soils at this site are Floyd loam (fine-loamy, mixed, superactive, mesic Aquic Hapludolls), Kenyon silty-clay loam (fine-loamy, mixed, superactive, mesic Aquic Hapludolls), and Readlyn loam (fine-loamy, mixed, superactive, mesic Aquic Hapludolls). These soils are moderately well to poorly drained, lie over loamy glacial till, and belong to the Kenyon–Clyde–Floyd soil

Model calibration

Model calibration was conducted following the procedure outlined by Ma et al. (2003). Measured soil hydraulic properties were used for the 10 soil layers (Table 2). Since the lateral saturated soil hydraulic conductivity (Lksat) was not sensitive except for the layers immediately above (#6) and between (#7) the tile drains, Lksat was therefore assumed to be the same as the vertical saturated soil hydraulic conductivity (Ksat) for all the layers except the two (#6 and #7) layers. Lksat for these

Summary and conclusions

We evaluated RZWQM using long-term (1978 to 2003) field data for 30 out of 36 plots. RZWQM was calibrated using data for Plot No. 25 and then used with the same soil and plant parameters, except for lateral hydraulic gradient, to simulate responses for the other 29 plots. Simulated water and N balances and crop production were compared to experimental measurements whenever possible. In general, the model correctly simulated year to year variations in tile flow and N loading in the tile flow.

References (37)

  • A. Bakhsh et al.

    Simulating the effect of swine manure application on NO3–N transport to subsurface drainage water

    Trans. ASAE

    (1999)
  • A. Bakhsh et al.

    Simulating nitrate losses from Walnut Creek Watershed

    J. Environ. Qual.

    (2004)
  • A. Bakhsh et al.

    Using RZWQM to predict herbicide leaching losses in subsurface drainage water

    Trans. ASAE

    (2004)
  • F. Ghidey et al.

    Evaluation of RZWQM using field measured data from the Missouri MSEA

    Agron. J.

    (1999)
  • D.B. Jaynes et al.

    Evaluation of RZWQM using field measured data from Iowa MSEA

    Agron. J.

    (1999)
  • B.P. Jones et al.

    Double-crop soybean leaf area and yield responses to Mid-Atlantic soils and cropping systems

    Agron. J.

    (2003)
  • D.L. Karlen et al.

    Twelve-year tillage and crop rotation effects on yields and soil chemical properties in northeast Iowa

    Commun. Soil Sci. Plant Anal.

    (1991)
  • D.L. Karlen et al.

    Challenges of managing liquid swine manure

    Appl. Eng. Agric.

    (2004)
  • Cited by (77)

    • Comparison of simulated nitrogen management strategies using DRAINMOD-DSSAT and RZWQM2

      2022, Agricultural Water Management
      Citation Excerpt :

      The two commonly used models for evaluating a wide range of management practices to curtail NO3-N losses from artificially drained agricultural lands across the Midwestern corn-belt of the United States are RZWQM2 (Ahuja et al., 2000) and DRAINMOD (Skaggs, 1978) /DRAINMOD-NII (Youssef et al., 2005). RZWQM2 (Root Zone Water Quality Model) is a one-dimension field-scale model that simulates hydrology, nutrient and pesticide fate and transport, crop growth, and agricultural management practice components within and below the crop root zone of a unit area (Ahuja et al., 2000; Ma et al., 2005, 2007). DRAINMOD/DRAINMOD-NII is a field-scale process-based model that simulates hydrology and C and N dynamics, respectively, in drained agricultural lands under a wide range of soil, climatic conditions, and management practices (Skaggs and Chescheir, 2003; Skaggs, 1978; Youssef and Skaggs, 2006; Youssef et al., 2005).

    • Evaluating RZ-SHAW model for simulating surface runoff and subsurface tile drainage under regular and controlled drainage with subirrigation in southern Ontario

      2020, Agricultural Water Management
      Citation Excerpt :

      Process-based models have been widely evaluated as reliable tools to simulate crop growth, nutrient dynamics and transport as well as water movement under various conditions. Consequently, they were extensively applied to predict the long-term effects of different climate and management practices on agricultural production and water quality (Ma et al., 2007a, b; Thorp et al., 2007, 2008; Qi et al., 2011), thereby to provide mitigation and adaptation BMPs to maximize sustainability goals for farmers and policymakers (Plaza-Bonilla et al., 2015; Uzoma et al., 2015; Wang et al., 2016; Jiang et al., 2019a). Numbers of hydrological models are capable of simulating the subsurface drainage and surface runoff in the field scale, for example, the ADAPT (Sands et al., 2003), MACRO (Kuzmanovski et al., 2015), CREAMS/GLEAMS (Chinkuyu et al., 2005; Knisel and Douglas-Mankin, 2012), RZWQM2 (Jiang et al., 2019b; Kuzmanovski et al., 2015), WEPP (Savabi, 1993), and DRAINMOD (Sands et al., 2003; Luo et al., 2000; Skaggs et al., 2012).

    • Simulating the effects of conventional versus conservation tillage on soil water, nitrogen dynamics, and yield of winter wheat with RZWQM2

      2020, Agricultural Water Management
      Citation Excerpt :

      We adjusted the model parameters such as the soil hydraulic parameters, soil organic and soil microbial pools, and winter wheat cultivar parameters using the automatic parameter optimization program PEST (model independent parameter estimation) linked with RZWQM2. Final adjustment of the parameters was conducted by trial and error (Ma et al., 2007a). The calibrated plant cultivar parameter values are shown in Table 2.

    View all citing articles on Scopus
    View full text