Uncertainty analysis of bias from satellite rainfall estimates using copula method

摘要: Abstract The aim of this study is to develop a copula-based ensemble simulation method for analyzing the uncertainty and adjusting bias two high resolution satellite precipitation products (PERSIANN TMPA-3B42). First, set sixty daily rainfall events that each them occurs concurrently over twenty 0.25° × 0.25° pixels (corresponding both PERSIANN TMPA spatial resolution) determined perform simulations validations. Next, number fifty-four out (90%) selected events, differences between rain gauge measurements as reference surface data estimates (SREs) are considered termed observed biases. Then, multivariate Gaussian copula constructed from normal distribution fitted Afterward, employed generate multiple fields randomly based on In fact, invariant monotonic transformations random variables thus generated have same dependence structure Finally, simulated biases imposed original in order obtain an bias-adjusted realizations estimates. area implementation proposed methodology region southwestern part Iran. reliability performance developed model regard correction SREs examined six those (10%) events. Note these not participated steps generation. addition, three statistical indices including bias, root mean square error (RMSE), correlation coefficient (CC) used evaluate model. results indicate RMSE improved by 35.42% 36.66%, CC 17.24% 14.89%, 88.41% 64.10% TMPA-3B42 estimates, respectively.