期刊
HYDROLOGY AND EARTH SYSTEM SCIENCES
卷 22, 期 9, 页码 5021-5039出版社
COPERNICUS GESELLSCHAFT MBH
DOI: 10.5194/hess-22-5021-2018
关键词
-
资金
- Norwegian Research Council's Enhancing Snow Competency of Models and Operators (ESCYMO) project (NFR) [244024]
- NOTUR/NORSTORE [NS9333K, NN9333K]
Parameter uncertainty estimation is one of the major challenges in hydrological modeling. Here we present parameter uncertainty analysis of a recently released distributed conceptual hydrological model applied in the Nea catchment, Norway. Two variants of the generalized likelihood uncertainty estimation (GLUE) methodologies, one based on the residuals and the other on the limits of acceptability, were employed. Streamflow and remote sensing snow cover data were used in conditioning model parameters and in model validation. When using the GLUE limit of acceptability (GLUE LOA) approach, a streamflow observation error of 25% was assumed. Neither the original limits nor relaxing the limits up to a physically meaningful value yielded a behavioral model capable of predicting streamflow within the limits in 100% of the observations. As an alternative to relaxing the limits, the requirement for the percentage of model predictions falling within the original limits was relaxed. An empirical approach was introduced to define the degree of relaxation. The result shows that snow- and water-balance-related parameters induce relatively higher streamflow uncertainty than catchment response parameters. Comparable results were obtained from behavioral models selected using the two GLUE methodologies.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据