Improving Quality Control of JPSS and GOES-R Radiances for NWP Applications

Satellite passive microwave and infrared radiance observations are of high quality in both accuracy and precision and are available globally with unprecedented amount. They have routinely been assimilated into global NWP systems for about two decades. Consistent and significant improvements were obtained for medium-range weather forecasts from direct assimilations of satellite radiance observations at the National Centers for Environmental Prediction (NCEP) and European Centre of Medium-Range Weather Forecasts (ECMWF). Direct assimilation of radiance observations requires forward radiative transfer model simulations for any given input model state. Observations for which the forward radiative transfer model simulations are of poor precision must be identified and eliminated from assimilation through a proper QC procedure. With a rapid advancement in computer technology, global NWP models have reached a very high resolution that can resolve cloud-populated mesoscale systems. This imposes a much more stringent requirement on QC of satellite data than that for coarser resolution global forecast models. The Advanced Technology Microwave Sounder (ATMS) on board the Suomi National Polar-orbiting Partnership (SNPP) satellite combines the capability of its two predecessors, AMSU-A and MHS, into a single instrument. The two ATMS lowest-frequency oxygen channels are available at the same locations as the six high-frequency water vapor channels. This unique feature of ATMS makes the cloud detection part of QC much more effective for ATMS applications in NWP through data assimilation than for MHS data. When MHS data is assimilated as a single data stream as was done traditionally, an improved QC for MHS data in the GSI system leads to an improved instead of a degraded 24-hour alongshore and offshore precipitation forecasts in Gulf of Mexico. Atmospheric state variables have large spatial and temporal variability within and around a mesoscale system. Instead of a climatologically valid static scheme, a local, regime-dependent and objective cloud mask (CM) algorithm is required for effectively isolating cloud-free pixels from cloudy pixels for Geostationary Operational Environmental Satellite (GOES) imager radiance assimilation in high resolution forecast models. Through various cases, it is shown that the importance of the quality assurance of satellite observations for NWP applications can be emphasized.