Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Watershed models are used to evaluate and develop Total Maximum Daily Loads (TMDLs). When analyzing model outputs, modelers are faced with various uncertainties in input and output data, model parameters, and model structure. These uncertainties negatively affect the usefulness of watershed models. Error analysis propagation throughout different components of a watershed model is one of the major challenges when a model is evaluated. Probabilistic point estimation methods propagate the parameter uncertainty by performing point estimations of the parameter space instead of calculating the entire probability density function (PDF). The objective of this study is to evaluate precipitation and parameter uncertainty propagation on simulated flows using the Monte Carlo simulation and Harrs probabilistic point estimate method. Daily certainty bound flows are modeled using the Hydrological Simulation Program - FORTRAN (HSPF) for a watershed in Alabama and Mississippi. Twelve HSPF parameters and from +/- 5% to 50% observed precipitation data are evaluated. Observed daily streamflow data from 01/01/2000 to 11/30/2004 are used to evaluate the uncertainty methods. The comparison shows that Harrs method could be a suitable replacement for Monte Carlo simulations in watershed models with several parameters. Additionally, computational efficiency is reached using 24 runs in Harrs method to estimate HSPF certainty bounds versus 10,000 runs using Monte Carlo methods. Rainfall errors lower than 70% observed precipitations become more important than parameter uncertainty.