RZWQM Simulation Of Long-Term Crop Production, Water And Nitrogen Balances In Northeast Iowa

Authors: Ma, Liwang; Malone, Robert - Rob; Heilman, Philip - Phil; Karlen, Douglas; KANWAR, R.S. - IOWA STATE UNIVERSITY; Cambardella, Cynthia; Ahuja, Lajpat

Submitted to: Geoderma
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/15/2006
Publication Date: 5/17/2007
Citation: Ma, L., Malone, R.W., Heilman, P., Karlen, D.L., Kanwar, R., Cambardella, C.A., Ahuja, L.R. 2007. RZWQM simulation of long-term crop production, water and nitrogen balances in northeast Iowa. Geoderma 140(3):247-259. doi: 10.1016/j.geoderma.2007.04.009.

Interpretive Summary: Agricultural system models provide the best possible simulation or representation of major processes and their interactions so that they can be better understood. 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. 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 each plot. Keeping all other soil and plant parameters constant, RZWQM correctly simulated year to year variations in tile flow and N load in tile flow. Yearly crop yield variation was simulated with less satisfaction although the average yields were reasonably simulated. Root Mean Square Errors (RMSE) for simulated soil water storage, water table, and yearly tile flow were 3.0, 22.1, and 5.6 cm, respectively. These values were close to the average RMSE between replicates (3.0, 22.4, and 5.7 cm, respectively). RMSE values for simulated yearly N load and residual soil N were 14 and 50 kg N ha-1, respectively, which were much higher than the average RMSE among replicates (8 and 39 kg N ha-1, respectively). The high RMSE for N balance might be caused by high simulation errors in plant N uptake. Simulated corn and soybean yields had high RMSE (1465 and 647 kg ha-1) even though average simulated yields were within 10% for corn and 3% off for soybean. Further improvements are needed for simulating plant N uptake and yield, but overall, RZWQM has the capability of simulating the major soil and plant processes with reasonable accuracy, and results for tile flow and N loading in tile flow are acceptable.

Technical Abstract: Agricultural system models provide the best possible simulation or representation of major processes and their interactions so that they can be better understood. 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. 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 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 load in tile flow (r2=0.81). Yearly crop yield variation was simulated with less satisfaction (r2=0.50 for corn and r2=0.35 for soybean) although the average yields were reasonably simulated. Root Mean Square Errors (RMSE) for simulated soil water storage, water table, and yearly tile flow were 3.0, 22.1, and 5.6 cm, respectively. These values were close to the average RMSE for the measured data (3.0, 22.4, and 5.7 cm, respectively). RMSE values for simulated yearly N load and residual soil N were 14 and 50 kg N ha-1, respectively, which were much higher than the average RMSE among replicates (8 and 39 kg N ha-1, respectively). The high RMSE for N balance 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 (1465 and 647 kg ha-1) even though average simulated yields were within 10% for corn and 3% off for soybean. The high simulation errors for corn yields were primarily caused by under-prediction for manure plots. Further improvements are needed for simulating plant N uptake and yield, but overall, RZWQM has the capability of simulating the major soil and plant processes with reasonable accuracy, and results for tile flow and N loading in tile flow are acceptable.