Uncertainty Estimation of HEC-HMS Flood Simulation Model using Markov Chain Monte Carlo Algorithm

There are some parameters in hydrologic models that cannot be measured directly. Estimation of hydrologic model parameters by various approaches and different optimization algorithms are generally error-prone, and therefore, uncertainty analysis is necessary. In this study we used DREAM-ZS, Differential Evolution Adaptive Metropolis, to investigate uncertainties of hydrologic model (HEC-HMS) parameters in Tamar watershed (1530 km2) in Golestan province. In order to assess the uncertainty of 24 parameters used in HMS, three flood events were used to calibrate and one flood event was used to validate the model. The results showed that the 95% total prediction uncertainty bounds bracketed most of the observed data especially peak discharge values but the uncertainty due to other sources than parameter uncertainty (e.g. forcing data (rainfall) and model structure error) are significant. Coefficient of variation for curve number (CN) was small for all flood events, therefore this parameters is more sensitive than the others. Histograms of the posterior probability density functions (pdfs) show that most of the individual parameters are well-defined and occupy only a relatively small region of the uniform prior distributions. Best simulation under DREAM-ZS was obviously better than simulation results of Nelder and Mead search algorithm.