Diurnal‐seasonal and weather‐related variations of land surface temperature observed from geostationary satellites

[1] The time series of clear-sky Land Surface Temperatures (LST) for one year, 2001, obtained from pyrgeometric observations at five selected US surface radiation (SURFRAD) stations and independently retrieved for the locations of these stations from Infrared Imager hourly observations of two geostationary satellites, GOES-8 and GOES-10, are presented as a sum of time-dependent expected value (diurnal and seasonal cycles), and weather-related anomalies. The availability of three independent observations is used to assess random and systematic errors in LST data. Temporal variation of the expected value is approximated as a superposition of the first two annual and diurnal Fourier harmonics. This component of temporal variations of LST absorbs all systematic errors; which themselves are often a subject of diurnal and seasonal variations. The results revealed that the weather-related temporal variation of LST is much smaller than the temporal variations of the expected value, but much larger than the random errors of observation. Scale of temporal autocorrelation of weather-related component of clear-sky LST variations is about 3 days.