Impact of GPM Precipitation Error Characteristics on Hydrological Applications

Precipitation-measuring satellites constitute a constellation of microwave and infrared sensors in geosynchronous earth orbit. The limited sampling of passive microwave constellations continues to be a problem, affecting applications such as hydrological modeling. Recent constellations have contributed in the construction of the next generation of earth and space science missions by allowing measurement settings to be customized to meet changing scientific understanding. Our study focuses on examining the Global Precipitation Measurement (GPM) constellation mission. The aim of the study is to examine the impact of different uncertainties carried by the GPM constellation on hydrological applications. Firstly we investigated the evaluation and comparison of spatial sampling error for the Global Precipitation Measurement (GPM) mission orbital data products. The region over India with high seasonal rainfall appears to have lower sampling uncertainty, and vice versa, with some exceptions due to differences in precipitation variability and space-time correlation length.  Second, we investigated how intermittency produced by low temporal sampling propagates through a hydrological model and contributes to stream flow uncertainty. We also examined the effect of grid resolution and how it relates to Clausius-Clapeyron scaling. This paper proposes and discusses techniques for quantifying the influence of grid resolution as a function of spatial–temporal characteristics of heavy precipitation based on these findings. Thirdly, we have quantified the influence of two different algorithms i.e top down and bottom up approach utilizing precipitation products that includes the Global Precipitation Measurement mission's (GPM) integrated Multi-satellite Retrievals (IMERG) late run, the SM2RAIN-Climate Change Initiative (SM2RAIN-CCI), and the SM2RAIN-Advanced SCATerometer (SM2RAIN-ASCAT) on hydrological simulations. The results from our study indicate that precipitation forcing at 6-hourly integration outperforms the stream flow simulations as compared to 3-hourly and 12-hourly forcing integration times. IMERG based precipitation also contains significant bias which is propagated into hydrological models when used as precipitation forcing.