The planned NASA/JAXA Global Precipitation Measurement (GPM) mission is expected to enhance existing satellite-based rainfall estimates via improved spatial and temporal coverage from the GPM constellation of satellites. Precipitation is one of the primary forcing in hydrological and land surface models, often evolving over short space and time scales where properly instrumented surface networks may not be available. Yet, satellite data often provides the only source of timely precipitation data over many of the world's remote watersheds. The planned GPM constellation is likely to include a number of satellites and sensor types which, owing to launch schedules, will likely change before and during the mission. Hence, it is essential to understand the sensitivity of the GPM multi-satellite precipitation estimation algorithms (and the subsequent effect upon GPM-initialized land surface models) to the availability of types of satellites and sensors over the mission lifetime.

On such means to evaluate the potential effectiveness of GPM-era constellation precipitation products is by conducting satellite omission experiments. We have adopted the “NRL-Blend” precipitation algorithm to generate an ensemble of precipitation products by systematically omitting categories of satellites and/or instruments. Since June 2007, we have been generating an ensemble of GPM proxy data, using data from existing satellites and based on the NRL-blend algorithm, over the continental United States and surrounding areas (0N-50N, 130W-50W). The set of ensemble members are configured (a) to omit the morning or afternoon and/or all cross-track sounders; (b) to omit the TRMM TMI and/or precipitation radar; and (c) to omit all morning or afternoon satellites. The different satellite configurations (ensembles) are compared against ground truth data. Our preliminary analyses indicate that compared to the “all satellites” configuration, the omission of the morning satellites (specifically the across-track sounders) showed the largest performance degradation compared to the all-satellites configuration.

GPM is currently planned to be active during a companion NASA mission, the Soil Moisture Active Passive (SMAP) program. There exists significant GPM-SMAP overlap in terms of science goals and measurement requirements. Therefore, validation efforts also include the use of land surface models (LSM) and other types of hydrological observations (other than raingauge) to examine the impact of these GPM proxy data upon streamflow, discharge, soil moisture and other runoff measurements which will be directly or indirectly inferred from SMAP. We are employing the Noah LSM, incorporated with the NASA Land Information System (LIS), to simulate land surface and hydrological states and partition the energy and moisture fluxes that are relevant for water resources management applications. Besides the routine evaluation techniques and metrics, such as root mean squared error, false alarm ratio and other skill scores, used in the research community, we are also adopting novel fuzzy-based methodologies to characterize the uncertainties in the satellite derived rainfall data and predictions.