Proceedings of the 2020 Mississippi Water Resources Conference

Year: 2020 Authors: Mendez-Monroy J.F., Ramirez-Avila J.J.

The use of computer programs to understand the dynamics and interactions of soil, hydrologic and crop production processes, have allowed scientists, researchers and farmers to facilitate and improve decision-making procedures. The Agricultural Policy/Environmental Extender Model (APEX) is a tool capable of performing long term simulations (1-4000 years) on a daily time step, for the managing of whole farms or small watersheds. APEX allows users developing analyses related to water balance, nutrient balance, and sediment transport. To highlight the potential of APEX in predicting soil, hydrologic and crop production processes of agricultural scenarios in the Mississippi Delta, a study was developed for two fields with soybean and cotton under reduced tillage and winter cover crop practices. Available information about runoff, sediment transport, soils, operation schedules and crop yield for a period of four years was used for the simulation. A sensitivity analysis (SA) and a calibration (CA) and validation (VA) process were carried out using the APEX Auto-Calibration and Uncertainty Estimator tool (APEX-CUTE) to compare observed and simulated runoff depths from the study fields. Four parameters associated to the prediction of soil evaporation (PARM17), runoff (PARM20 and PARM42), and potential evapotranspiration (PARM 34) were recognized as sensitive. To perform the CA-VA process, APEX-CUTE uses an objective function designed to maximize the model performance efficiency (NSE) and reduce the tendency of the simulated values to be larger or smaller than their observed values (PBIAS). The objective function was evaluated 1000 times resulting in optimized values of NSE>0.70 and PBIAS≌0.070, and NSE>0.90 and PBIAS≌0.449 for CA and VA, respectively. The best parameter set was used to simulate runoff that was compared with observed data. In general terms, APEX was successful in predicting daily runoff. For the monthly and annual runoff, some variations were noticed (under and overestimations). Despite this, the model had an optimal performance, and in future studies, it could potentially be used as an assistant tool for agricultural, research and engineering purposes.