Post‐Audit of Lake Model Used for NPS Management

Abstract

Ability of a continuous simulation model to predict nonpoint pollution loads and reservoir water quality was reevaluated for conditions significantly different from those used for calibration/verification. Lower pollution loadings than calibration/verification period resulted from start‐up of regional advanced wastewater treatment (AWT) plant and significant switchover from high tillage cropland practices within the reservoir's 580 sq mile $\left(1\text{,}500{\mathrm{km}}^2\right)$ drainage area. Use of both parametric and nonparametric statistical tests to assess model goodness‐of‐fit is demonstrated. Nonpoint pollution loading factors from small testing sites are reverified with gaging data at the mouths of two watersheds representing more than 90% of reservoir drainage area. Simulated water quality statistics for two‐layer reservoir are compared with daily average of recorded data from 7 monitoring stations. There was typically no significant difference between simulated and recorded statistics on nonpoint pollution loadings and concentrations of critical indicators of lake eutrophication. Results are further evidence of the model's reliability for water quality studies of land use plans or control measures that also produce major changes in pollution loadings.