The research into rainfall-runoff plays a very important role in water resource management. However, runoff simulation is a challenging task due to its complex formation mechanism, time-varying characteristics and nonlinear hydrological dynamic process. In this study, a nonlinear autoregressive model with exogenous input (NARX) is used to simulate the runoff in the Linyi watershed located in the northeastern part of the Huaihe river basin. In order to better evaluate the performance of NARX, a distributed hydrological model, TOPX, is used to simulate the discharge as a reference, and runoff classification by cluster analysis is used to further improve the accuracy of runoff simulation. Based on the four statistics indexes of the Nash–Sutcliffe efficiency (NSE), correlation coefficient (CC), root mean square error (RMSE) and mean relative bias (Bias), the NARX model is capable of simulating the rainfall-runoff dynamic process satisfactorily, although there is a little underestimation of the peak flow. After runoff classification, underestimation has been improved, and discharge simulation driven by NARX based on runoff classification (C-NARX) is well consistent with the observation. It is feasible to take it as a promising method, which also can be seen as a good reference and replacement for the current rainfall-runoff simulation.

Frequency estimates of extreme precipitation are revised using a regional L-moments method based on the annual maximum series and Chow’s equation at lower return periods for the Jiangsu area in China. First, the study area is divided into five homogeneous regions, and the optimum distribution for each region is determined by an integrative assessment. Second, underestimation of quantiles and the applicability of Chow’s equation are verified. The results show that quantiles are underestimated based on the annual maximum series, and that Chow’s formula is applicable for the study area. Next, two methods are used to correct the underestimation of frequency estimation. A set of rational and reliable frequency estimations is obtained using the regional L-moments method and the two revised methods, which can indirectly provide a robust basis for flood control and water resource management. This study extends previous works by verifying underestimation of the quantiles and the provision of two improved methods for obtaining reliable quantile estimations of extreme precipitation at lower recurrence intervals, especially in solving reliable estimates for a 1-year return period from the integral lower limit of the frequency distribution.

This study aims to gain insights into turbidity currents in approach channels with a closed end. A vertical two-dimensional numerical model was introduced, which considers the viscosity of sediment and assumes rheological behavior of the very near-bed layer. The dynamic propagation of the turbidity current, both before and after it reaches the closed end, was studied. It is found that before the turbidity current reaches the closed end, the velocity of the current head is unaffected by the length of the channel, while the velocity and height of the turbidity head are decreased along the channel. When the head reaches the wall, a surge and reverse flow occur. After a quasi-stable state has formed, results show that the speed of current along the channel decreases gradually to zero speed at the closed end. The current speed increases asymptotically with the initial sediment concentration toward a limiting value at high concentrations. The depth of the turbidity current is about 0.3–0.6 times the total depth. The sediment concentration is reasonably constant along the channel. The study deepens the understanding of turbidity currents in these special channels and provides theoretical guidance for engineering practice.