Runoff, Sediment and Total Phosphorus Loss Predictions Using The Apex Model.

Agricultural fields are an important source of phosphorus (P) enrichment to water bodies. High soil erosion and runoff, especially when coincident with high soil test P or high P applications, result in excessive P loss to surface water. It is in the best interest of agriculture and environmental protection to reduce P loss from agricultural fields. Reduction of P loss from agricultural production requires accurate estimates of how management practices impact P loss relative to soils, landscape position, slope, and hydrology. Computer models can be used as a tool to assess the impact of management practices on P loss. This study determines the accuracy of out of box (OOB) model in predicting runoff, sediment and total P loss from agricultural fields using Agricultural Policy/Environmental eXtender (APEX) model. The OOB APEX model was run using field-scale runoff data collected from P loss studies in Eastern Kansas, U.S.A with 9 different treatments from 2001-2007 in two different studies. The APEX model’s performance indicators coefficient of determination (r²), Percent Bias (PBIAS) and Nash-Sutcliffe Efficiency (NSE) were calculated. The PBIAS for corn and soybean yield was -4.46 and -95.52 respectively. For both study the overall PBIAS, r² and NSE values for runoff ranges from -24.41 to -183.5, 0.21 to 0.74, and 0.59 to -9.32 respectively. Likewise, the PBIAS, r² and NSE values for sediment loss were 1.95 to -204.45, 0.001 to 0.30 and 0.02 to -5.45. And the PBIAS, r² and NSE values for total P loss were and 58.21 to -546.06, 0.21 to -139.60, and 0.06 to 0.77, respectively. Results indicated that the statistics for runoff for some treatments were within the acceptable range but the statistics for sediment and total P loss were poor. The model predicted P loss form agricultural fields amended with turkey litter was much higher than measured data.  We conclude that using the out of box model without calibration and validation would not predict the sediment and phosphorus loss accurately even though the statistics for runoff in some of the treatments were within the acceptable range. Moreover, results also suggest P sub-routines may need to be updated to better estimate P loss with the APEX model.