Detection and Correction of Diurnal Sampling Bias in HIRS/2 Brightness Temperatures

Abstract

Diurnal sampling biases arise in the High-Resolution Infrared Radiation Sounder (HIRS) satellite observations because some of the NOAA polar-orbiting satellites drift significantly from their original local observation time. Such bias adversely affects interpretation of these data for climate studies. Twenty-six years of HIRS/2 radiance satellite data (1979?2004) were examined by creating monthly mean gridded data that categorize the observations by local observing time through averaging ascending and descending orbits separately. Corresponding HIRS/2 simulated radiance data from the Geophysical Fluid Dynamics Laboratory (GFDL) climate model were constructed using HIRS/2 satellite sampling and were found to accurately represent the diurnal sampling bias. Correction of the HIRS/2 observations from the observed diurnal sampling bias was using the model simulations of HIRS brightness temperatures to adjust the observed brightness temperatures to the model daily mean. The diurnal bias was found to vary with channel, surface type, latitude, satellite, and cloud cover, but showed little dependence on satellite scan angle. Diurnal bias is most pronounced for ascending orbit observations of the afternoon [1400 local solar time (LST)] satellites with 60°N to 60°S domain averaged brightness temperatures variations up to 0.78 K yr?1. Lower tropospheric temperature and water vapor channels contained the largest bias, and biases over land were more than twice as large as those over the ocean. Brightness temperature adjustments of up to 10 K were needed in the most extreme situations.