Assessing Suitability of Satellite Rainfall Data for Estimation of Daily Streamflows of a Small Tropical Catchment in India

Reliable estimation of streamflow is crucial for developing effective water resources management strategies. However, there are several watersheds in India which are ungauged or contain inconsistent data archives for rainfall and discharge products, particularly for small watersheds at daily scale. This paper investigates the efficacy of the remote sensing rainfall products, precisely Tropical Rainfall Measuring Mission (TRMM) on daily scale over upper Tungabhadra sub-basin, a small tropical catchment in India. This precipitation dataset was corrected with in-situ rainfall data and used as a input to the physically based Variable Infiltration Capacity (VIC) hydrological model for estimation of streamflows. Streamflows generated with original TRMM rainfall data has resulted in larger difference in both the high and low streamflows when compared with observed discharge values, and resulted in high positive bias and low Nash-Sutcliffe efficiency (NSE) at the daily timescale. The corrected TRMM rainfall data enhanced this daily hydrological simulation with significant improvement in different performance indicators (i.e., NSE, bias and `goodness of fit'). Thus this study finds that the TRMM data products with appropriate correction can be used for estimation of daily streamflows over small watersheds.