新安江模型在资料匮乏的长江中下游山区中小流域洪水预报应用

水文资料匮乏流域的洪水预报(PUBs)是水文科学与工程中一个尚未解决的重大挑战.中国湿润山区中小流域大多是水文资料匮乏的流域,在此地区进行洪水预报的重要手段之一就是水文模型参数的估计.对基于参数物理意义的估算方法(以下简称物理估算法)及两种区域化方法进行了研究,将其用于新安江模型参数的估算及移植.皖南山区的29个中小流域被选作水文资料丰富的测量流域,鄂西山区的3个中小流域被视为水文资料匮乏的目标流域,目的是研究目标流域与测量流域空间位置较远但物理条件相似时,区域化等方法是否可以有效估计模型参数.结果表明,即使目标流域与测量流域空间距离较远,区域化及物理估算法也能一定程度上减少参数估计导致的模型效率损失,且在研究区的最优参数估计方案为单流域物理相似法结合回归法及物理估算法.为长江中下游资料匮乏的山区中小流域提出了可行的新安江模型参数估计方案,为该地区的洪水预报提供指导.

Application of Xin'anjiang Model in the flow prediction of ungauged small- and medium-sized catchments in the middle and lower reaches of the Yangtze River Basin

Flow prediction in ungauged basins (PUB) is as important as it is difficult. PUB is a major unsolved challenge in scientific and engineering hydrology. So far, empirical and stochastic methods have mainly been used for this purpose. Many small catchments in humid mountainous areas of China are with little or no hydrological data. The principal method of flood forecasting in this area is the parameter estimation of the hydrological model. Xin''anjiang Model was used to predict flow in this study, and Xin''anjiang Model parameters were estimated based on the physical meaning of them and regionalization. Regression and physical similarity approach were included in regionalization methods in this study. The 29 small basins in mountainous areas in southern Anhui Province were selected for gauged catchments, and three small basins in mountainous regions in western Hubei Province were treated as ungauged catchments. The main goal of this approach was to research whether regionalization and other methods could effectively estimate model parameters when the spatial location of the gauged catchments and the ungauged catchments were far away, but the physical conditions were similar. It is shown that:(1) Xin''anjiang Model has high accuracy in flow prediction of study areas. (2) Even if the space distance of gauged catchments and ungauged catchments is far, regionalization and physical-based estimation method can also reduce the efficiency loss caused by parameter estimation to some extent. (3) With the increase in the number of donor catchments, the efficiency of the physical similarity approach is reduced. (4) The optimal parameter estimation scheme in the study area is the ingenious combination of physical similarity approach (only one donor catchment), regression approach and physical-based estimation method. A feasible parameter estimation scheme of the Xin''anjiang Model is proposed in this study, which provides guidance for flow forecasting in small ungauged catchments in humid mountainous areas among the middle and lower reaches of the Yangtze River.