摘要: 水文模型的不确定性评价已经成为水文学者研究的热点问题之一。本文介绍了一种基于水文模型模拟偏差修正的改进GLUE方法,并采用3个评价指标(即覆盖率、平均带宽和平均不对称性)评价了分别由该方法求得的4个水文模型(即新安江模型、SMAR模型、SIMHYD模型和Tank模型)依次在汉江流域中的5个子流域上的预测区间的不确定性。结果表明,改进的GLUE方法可以显著的提高预测区间的覆盖率以及降低预测区间的不对称性。
Abstract:
The uncertainty analysis of hydrological models has received increasing attention of hydrological researchers. This paper applies a modified GLUE (generalized likelihood uncertainty estimation) methodology to estimate the prediction bounds of four hydrological models (namely, Xinanjiang model, SMAR model, SIMHYD model, and Tank model) on five sub-catchments in the Hanjiang river basin, respectively. Three indices (namely, containing ratio, average band width, and average asymmetry degree) are adopted to assess the prediction bounds obtained by the modified GLUE. The results show that the modified GLUE methodology can increase containing ratio and reduce average asymmetry degree of prediction bounds significantly.