Parameter uncertainties exert a significant effect on nonpoint source
pollution (NPS) modeling results. A decision made on the basis of such
results may thereby be inappropriate. In this work, the parameter unc
ertainty is analyzed to explore an improved modeling procedure. Draina
ge patterns generated from digital elevation data and rainfall are the
major parameters examined. A case study for the watershed of the Posa
n off-stream reservoir is implemented. A significant spatial variation
of NPS distribution simulated with a drainage pattern generated from
varied methods is observed. The effects of rainfall randomness on the
spatial loading distribution are assessed and computed based on a Mont
e Carlo simulation. The proposed procedure is capable of improving the
quality of modeling results and the decision for an appropriate contr
ol strategy.