Uses of Satellite Data for Hurricane Modeling and Coastal Precipitation Forecasts

In the past two decades, significant progresses have been made in numerical weather forecast skill, attributing mainly to improvements in satellite instrumentation, numerical models, data assimilation techniques and computer technological advances. Direct assimilation of satellite radiance observations into NWP models has resulted in steady increases in the global medium range forecast skills at all major NWP centers. Today, satellites provide more than 90% of the data ingested by NWP models. However, the utilized satellite data in NWP is only a small fraction of available data. Many satellite data are excluded to make the numbers more manageable for timely processing by a thinning process, to remove horizontally-correlated observation errors. and to avoid cloud- or rain-contamination that renders assimilation more prone to error. These latter observations naturally contain some information about the clouds and precipitation present. Better use of their content in general requires estimating some properties of any clouds and precipitation affecting the local radiative transfer, charactering errors in radiance observations, formulating covariances of hydrometeor variables. However, the properties of clouds and precipitation are difficult to resolve in NWP models that are run at 10 km or even coarser. The near-term priorities for satellite data assimilation shall focus on better uses of those data that are less/not affected by clouds and precipitation. In hurricane and severe storm conditions, the data obtained from satellite microwave temperature and water vapor sounders provide three dimensional warm core features that could be directly used for hurricane model initialization. The Advanced Technology of Microwave Sounder (ATMS) and the Cross-track Infrared Sounder (CrIS) on board the recently launched Suomi National Polar-Orbiting Partnership (NPP) satellite, are providing data for profiling atmospheric temperature and moisture under all weather conditions and supporting continuing advances in data assimilation and NWP modeling. It has been demonstrated that the ATMS describes better warm core structures than the predecessor instruments such as AMSU-A/MHS and is more valuable for hurricane initialization. It was also demonstrated that GOES imager radiances significantly improves coastal quantitative prediction forecasts (QPFs). Further improvements in short- to medium-range weather forecasts are expected when the full potential of these new instruments as well as instruments on board other polar-orbiting satellite (NOAA-18, MetOp-A, MetOp-B etc.) and GOES satellites are explored. Revisiting and optimizing the quality control and bias correction schemes in the current GSI for all satellite data types are shown to be necessary and urgently needed.