In order to satisfy the National Weather Service's (NWS) hydrologic forecasting needs, the Office of Hydrologic Development (OHD) developed and implemented a precipitation estimation algorithm called the Multisensor Precipitation Estimator (MPE). This algorithm optimally merges radar rainfall estimates with rain gauge observations. However, this algorithm has limited applicability in mountainous regions of the conterminous United States, because of radar range limitations and beam blockage, and variable rain gauge network density. Satellite Precipitation Estimates (SPE's) provide spatially-continuous precipitation estimates in such regions. Of all the existing SPE's available operationally to date, the Hydroestimator product produced by the National Environmental Satellite Data and Information Service (NESDIS) is the one best suited for hydrologic forecasting needs in terms of resolution. At present, Hydroestimator fields can be inserted into the MPE analysis only through a manual interactive process.

In this study, we made an attempt to seamlessly integrate the hourly Hydroestimator product within the operational MPE algorithm. This process involved several steps. The first step was to correct the Hydroestimator product for local statistical biases using rain gauge data. The second step was to fill the gaps in the corresponding mosaicked radar estimate field with the bias-corrected Hydroestimator field, thereby creating a spatially-continuous field of radar and satellite estimates. In the last step, we optimally merged this remote-sensor field with rain gauge observations to create the final multisensor analysis.

We are currently testing this technique in the California Nevada River Forecast Center region for the period of October 2002 to September 2003. If the rain gauge network is dense enough to catch all the rainfall, then gauge only analysis may be adequate. Where the rain gauge density is not high enough to capture all the rainfall details, adding radar and satellite precipitation estimates may improve QPE. Initial validation results based on limited sample cases showed promising results. We will present samples of the multisensor precipitation fields and results of more robust independent validation at the conference.