A correction to the diurnal sampling bias in HIRS observations using GFDL climate model simulations

The 26-year HIRS data record (1979-2004) has become important for monitoring and assessing climatic change of temperature, water vapor, and clouds. The NOAA polar-orbiting satellites, from which HIRS instruments have flown, are nominally sun-synchronous orbits; however, some of these satellites have drifted several hours away from their initial local time of observation. The changing local sampling times on these satellites mainly impacts HIRS observations of lower tropospheric temperature since surface temperature is most directly affected by the diurnal cycle. This study draws from previous work done to identify the diurnal sampling bias and devises a diurnal sampling bias correction to HIRS channels residing in the 6.7 - 15 μm spectral range. The correction is based on the difference between GFDL model simulations of HIRS brightness temperatures using HIRS-sampled model data and the full 3-hour model data. The correction adjusts all HIRS observations to a diurnal mean brightness temperature thus removing both diurnal sampling bias caused by drifting satellite orbits and local time differences between ascending and descending observations. The implications of cloud and limb effects on this correction will be discussed. Removal of the diurnal sampling bias makes the HIRS data more suitable for climate studies particularly those studies comparing HIRS observations to climate model simulations.