Emmanouil Anagnostou
University of Connecticut

E-mail:
Phone:
Fax:

This proposal aims at bridging the gap between satellite precipitation data that is available at coarser meteorological scales and its effective use overland at finer hydrological scales by developing an integrative data-modeling uncertainty framework to provide feedback to the GPM algorithm community. Specifically, we shall investigate the following research questions:

1. What is the current state of flood/water cycling predictability from existing (pre-GPM) precipitation satellite constellation observations and what is the potential improvement anticipated in the microwave-rich GPM era? These questions will be addressed using hydrologic error simulations for a variety of watershed/storm characteristics, scales, and climate regimes. The study of improvements from current to GPM era will consider the combined effect of increased microwave sampling and realistic projections of anticipated improvements in passive microwave rain estimation over land.
2. What are the optimal satellite precipitation products and resolutions in terms of prediction uncertainty of various hydrological variables of interest (runoff, soil moisture, energy fluxes, etc.)? The GPM goal is to produce precipitation products at 5-km square grid and 3-hour time resolutions, but, the selection of this criterion has not been verified on the basis of actual investigation of the hydrologic prediction error. In this project we intend to identify the complete range of responses of hydrologic variables associated with different satellite precipitation products and resolutions, and then devise a merging scheme to produce optimal ensembles of hydrologic parameters from existing satellite retrievals. The proposed research will provide critical quantitative information regarding the impact of satellite rain estimation error in water cycling/hydrologic prediction uncertainty. It will identify the optimal scales, resolutions and merging approaches where satellite precipitation data can be most useful in the prediction of hydrological variables. Our project is expected to enhance the ability to extract hydrologically useful information from satellite precipitation estimates than what is possible today with far-reaching societal implications. The results anticipated will therefore constitute an interim milestone to our on-going agenda to provide better focus to the development of next generation multi-sensor algorithms for satellite rainfall estimation. In the longer term perspective, the developed error propagation tools and findings from the proposed research would contribute to the establishment of metrics and criteria for assessing the impact of progress made in satellite rainfall remote sensing for advancing hydrologic predictability overland. Finally, the work accomplished through this proposal will form the stepping stone to a larger scope investigation of a number of open issues that we plan to address, such as, for example, the combined role played by precipitation remote sensing uncertainty and hydrologic modeling complexity in error propagation studies.