Evaluation of Rain-gauge Network in Upper Teesta Basin Using Entropy

Abstract

The Upper Teesta River basin has been severely affected by excessive flash flood events due to vast rainfall in recent years. Evaluating existing networks is crucial for planning, designing, and adequately managing water resource projects. The network design helps lower project failure risk and minimize economic losses. The network should be reviewed and updated promptly until it reaches the optimal network so that it gives adequate information like rainfall. This paper uses the Shannon entropy theory to assess the existing precipitation network in the upper Teesta River basin using marginal entropy and trans-information index. The statistical parameters of nine rain gauge stations are determined. Daily rainfall records of TRMM, APHRODIRE and CFSR have been compared with ground-based observation records for the four different stations. Results show that satellite-based rainfall records are underestimated regarding ground-based observations at all rain gauge stations. The maximum information minimum redundancy concept is applied to the existing rain gauge network, and the result reveals that only nine rain gauge stations in the network are insufficient to capture the precipitation variability in the mountainous region as per WMO standards. These findings support the need to improve the existing network to understand flood modelling, prediction and water resource management.