Microwave Integrated Retrieval System (MIRS) - Background

The Microwave Integrated Retrieval System (MIRS) is an upgrade to the regression-based microwave retrieval system known as MSPPS (Microwave Surface and Precipitation Products System) which lacked the profiling capability and was specific to a single instrument. Some of the objectives of MIRS are (1) to perform the retrieval in all-weather conditions and (2) over all-surface types, with a major benefit being the extension of the spatial coverage to critically important active regions.

MIRS is based on an assimilation-type scheme (1DVAR) capable of optimally retrieving atmospheric and surface state parameters simultaneously. More details could be found under the 'Algorithm' section. Thanks to its dynamic memory design, it is capable of performing retrievals using different instrumental configurations. It is also envisioned that it will serve as a tool for performing sensor design studies and performances estimation for future satellites and concepts like the future MIS sensor onboard NPOESS or a potential Geostationary-based microwave sensor.

Currently at NOAA/NESDIS/STAR, MIRS is applied routinely to NOAA-18 and METOP-A AMSU/MHS sensors pair and is being extended to run routinely with DMSP-F16 SSMI/S data. It is also expected that MIRS will be the retrieval algorithm for the future sensors: NPP/ATMS and NPOESS microwave sensors (MIS, ATMS).

MIRS uses the Community Radiative Transfer Model (CRTM) as its forward operator, leveraging therefore a large amount of effort being undertaken at the Joint Center for Satellite Data Assimilation (JCSDA).

The direct outputs from MIRS include temperature, moisture and several hydrometeors atmospheric profiles, land surface temperature and emissivity (at all channels). From these core products are derived a set of secondary products using the Vertical Integration and Post-Processing (VIPP) process. These include: Total Precipitable Water (TPW), vertically integrated Cloud Liquid Water (CLW), Ice Water Path (IWP), Graupel-size ice Water Path (GWP), Rain Water Path (RWP). In addition, surface properties are also derived from the retrieved emissivities and associated skin temperature. These include Snow Water Equivalent (SWE), Sea Ice Concentration (SIC), Soil Wetness Index (SWI) or Soil Moisture, etc.

MIRS algorithm generates new products; vertical profiles of cloud amount, precipitation, ice, snow and graupel and surface emissivity spectra. Although the accuracy of the hydrometeors vertical distribution is not great because of the lack of information content in the measured radiances used, it nevertheless sets the stage for a future capability to retrieve accurately liquid and frozen precipitations profiles (either by using more channels or by getting help from external data such forecast outputs).

Because the same algorithm will eventually be used consistently across the platforms (different channels spectra, polar and geostationary orbits), the time series of these retrievals will thus be self consistent, making the resulting climate data records free of jumps due to changes in the algorithms.