Estimating nitrogen and phosphorus contents using model integrated in small agricultural watersheds.

The Agricultural Nonpoint Source Pollution Model (AGNPS) was applied to estimate the loads of nitrogen and phosphorus according to antecedent soil moisture conditions (AMC) using probable rainfall in a stream draining small agricultural watersheds. Calibration and verification of the model were performed using observed data collected from rainfall events in the Imgo watersheds during 1997-1998. Evaluation of model outputs was based on graphical displays contrasting observed and simulated values for each rainfall-runoff event, and standard statistics such as coefficient of efficiency. The coefficient of efficiency of peak flow, total N, and total P in the estimation were 0.97, 0.93, and 0.74, respectively. Probable rainfall of Imgo watersheds was computed for several return periods using the frequency analysis of Gumbel's extreme value distribution. As the probable flood increase, concentrations of total N and total P generally decrease. Loss of N and P was proportional to the preceding soil water content in the same probable rainfall. The excessive fertilization under AMC-III condition can cause relatively higher outflow of N and P from the soil. It can be concluded that surface runoff and sediment transport influence soil quality and the quality of water absorption and AGNPS can provide realistic estimates of nonpoint source nutrient yields.