Assimilation impacts of AMSU snow/ice emissivity model improvements

Radiance data from AMSU-A/B onboard NOAA satellites are one of the most important data for operational data assimilation systems. However, in the polar region, only a few radiances sensitive to the surface or lower troposphere are used because of difficulty in estimating of snow/ice surface emissivity.

New snow/ice statistical emissivity models (Yan, et al., 2003), developed at the JCSDA , have been incorporated in NCEP's Spectral Statistical Interpolation (SSI) assimilation system. After quality control procedures and observation errors were adjusted, analysis and forecast impacts were evaluated. Many more lower sensing channels were assimilated in the polar regions and the lower tropospheric temperature was increased. This near surface temperature change may be correcting a low-temperature bias of the forecast model. The impacts were localized to latitudes poleward of 60-degrees and did not substantially propagate to lower latitudes during the assimilation or forecast. Further impacts on the forecasts are currently under investigation.