Interfacing FMBM with PRZM-2 for modeling agricultural water quality
Simulations with the PRZM model developed by USEPA show that the timing of the application of chemical relative to the occurrence of significant infiltration events has major impact on the amount of chemical reaching the shallow ground water. Simulations of surface runoff from the Field Moisture Balance Model (FMBM) compared favorably with the measured storm runoff at Gingles and Four Mile Creek watersheds in Iowa as compared to those with the PRZM-2 (PRZM, version 2.0) model;The FMBM model capability was enhanced to include water quality predictions using the chemical algorithms of the PRZM-2 model. The new model is referred to as the Field Moisture Balance Water Quality (FMBWQ) model. Simulations were conducted (with FMBWQ and PRZM-2) during the four growing seasons (1967-70) for the Gingles watershed and five growing seasons (1976-80) for the Four Mile Creek watershed with the available weather data, and assessments on model evaluations made. For the Gingles watershed, the FMBWQ model predicted higher R-squares for the years 1967, 1968, and 1970 with correspondingly lower squared sum of residuals (SSR) and root mean square errors (RMSE), and for the Four Mile Creek watershed, the FMBWQ model predicted higher R-squares and low RMSE values for the years 1977, 1978, 1979, and 1980. The FMBWQ model generally predicted runoff on days when some surface runoff was measured, whereas PRZM-2 predicted runoff on days when no measurements were recorded. Soil moisture predictions by the FMBWQ model compared more favorably with the measured data than by PRZM-2;Four agricultural pesticides (atrazine, alachlor, metribuzin, and chloramben) were selected for simulation. Due to the lack of measured chemical fate data for the experimental sites during the study period, a quantitative assessment was made between the two models for the components of pesticide mass balance. The FMBWQ model predicted less pesticide mass in the profile with higher surface concentrations, in general. A sensitivity analysis of the FMBWQ model was performed in relation to the pesticide organic carbon partitioning coefficient (K[subscript] oc) and half-lives (t[subscript]1/2) to determine the trend and magnitude of their fate and transport.