Evaluation of different precipitation inputs on streamflow simulation in Himalayan River basin

Abstract

This study evaluates four high-resolution Satellite Precipitation Products (SPPs) and gauge-based precipitation (Gauge) in a well-performed J2000 hydrological model across the Himalayan River basin originated from the Tibetan Plateau. The four high-resolution SPPs include Tropical Rainfall Measuring Mission 3B-42 (TRMM), Climate Hazards Group Infrared Precipitation version 2 (CHIRPS), Multi-Source Weighted-Ensemble Precipitation version 1 (MSWEP), and Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks- Climate Data Record (PERSIANN-CDR). Two methods were deployed to evaluate hydrological response on a daily and monthly scale. The first method (Method I) consists of calibration based on gauge and re-running the model for SPPs with gauge-based calibration parameters whereas input-specific calibration of SPPs was adopted in the second method (Method II). The model-driven by TRMM demonstrated superior result followed by CHIRPS, PERSIANN-CDR, and MSWEP respectively both on the daily and monthly scale under method I. Comparing the performance metric with the method I, the substantial improvement of Nash Sutcliffe efficiency (NSE) and Percentage bias error (Pbias) were observed in MSWEP whereas improvement of Pbias was found in PERSIANN-CDR for method II. In the context of method I, MSWEP could not predict hydrological behavior properly in the Himalayan River basin while it performed best in method II on both a daily and monthly scale. SPPs could be used in the future to predict hydrological responses in the Himalayan River basin with the application of the hydrological model but it contains error and bias which should be concisely examined and corrected before practical application for planning and management of water resources project.