Parameters Uncertainty of Hydrological Models and Its Influence on Probabilistic Flood Forecasting

In order to study the parameters uncertainty of Nash model,the Markov chain Monte Carlo based on adaptive Metropolis method coupled with Bayesian forecasting system and genetic algorithm,a new adaptive metropolis method based on Bayesian forecasting system(AM-MCMC based on BFS) is built.And the probabilistic flood forecasting is implemented by the AM-MCMC based on BFS.Then the effect of parameters uncertainty of Nash model on probabilistic flood forecast is analyzed.The case results show that the posterior distributions of parameters of Nash model,which are caught by the AM-MCMC based on BFS to describe the characteristics of all parameters of hydrologic model,are useful for improving the accuracy of probabilistic flood forecasting and reducing the uncertainty of flood forecasting.