Design flood estimation is a fundamental task in hydrology. We propose a machine- learning-based approach to estimate design floods anywhere on the global river network. This approach shows considerable improvement over the index-flood-based method, and the average bias in estimation is less than 18 % for 10-, 20-, 50- and 100-year design floods. This approach is a valid method to estimate design floods globally, improving our prediction of flood hazard, especially in ungauged areas.

Multi-model ensemble hydrological simulation has been an effective method for improving simulation accuracy. This study explored the feasibility of applying a multi-model ensemble simulation to the upper Yalongjiang River Basin. The results of the BPNN multi-model ensemble simulation are better than that of a single model. Multi-model ensemble simulation should become an important direction in hydrological simulation research.

Sediment and runoff changes of seven hydrological stations along the Lower Yellow River from 1980 to 2003 were analyzed at multiple time scale. Research results indicate that (1) The runoff and sediment of the Lower Yellow River during 20 years show the different trend.The runoff reduced, but sediment were not same.（2）The time scale has big effect on the analysis results. (3) The changes of runoff and sediment show high persistence without additional measures, which proved the necessary of WSRS.