• cloudy radiance;
  • satellite observation;
  • data assimilation;


Infrared radiances from the Multi-functional Transport Satellite (MTSAT)-1R satellite were assimilated in cloudy conditions where effective cloud fractions were greater than 0.8. These cloudy radiances provide new information that currently assimilated clear-sky radiances from geostationary satellites and microwave sounders do not have.

The radiance data to be assimilated were created by averaging pixels from the original radiances. We investigated how the characteristics of spatially averaged observations (super-observations or super-obs) vary with the super-ob size. The cloudy super-ob radiances were simulated by using a simple radiative transfer model with cloud-top pressure (Pc) and effective cloud fraction (Ne). The model assumed a single-layer cloud and that cloud emissivity does not depend on spectral wavelength. These two parameters were estimated by a minimum residual method with the infrared channel 1 (10.8 µm) and channel 2 (12.0 µm). To further ensure the validity of these assumptions, we limited use of super-ob radiances to almost completely overcast (Ne ≥ 0.8) conditions with homogeneous, middle to relatively high clouds (160 ≤ Pc ≤ 650 hPa).

These overcast super-ob radiances (OSRs) from geostationary satellites have the advantages of (i) providing temperature information that is highly vertically resolved at the cloud top, (ii) providing frequent measurements, and (iii) being available in cloudy areas where clear radiances are often rejected by cloud quality-control procedures. Data assimilation experiments revealed that OSRs generally had a small impact on analyses and forecasts but provided a slightly improved forecast of temperatures in the upper troposphere and winds in the low troposphere. Copyright © 2012 Royal Meteorological Society