A Examination of the Time/space Averaging Process Used in the Earth Radiation Budget Experiment

This study examines the Time/Space Averaging (TSA) process applied to scanner measurements of the Earth Radiation Budget Experiment. The TSA process produces monthly averages from the temporally sparse flux measurements on a regional scale. This diurnal distribution of the ERBS flux measurements for a given month and region shows considerable bias in high latitudes due to the inclination of the ERBS satellite orbit. The ERBS measurements for 72 days (approximately the precessing period of the ERBS satellite orbit) have a nearly complete diurnal distribution for each region between 60^circN- 60 ^circS. To examine the TSA process, simple averages obtained from two 72-day measurement sets in 1985 and 1986 are assumed as the expected true averages. For clear-sky, simple averaging methods are employed to infer the true averages where one can expect a diurnal sampling bias due to clouds. The TSA process is examined in detail for different data sets of ERBS and NOAA 9. This report shows that the TSA process is adequate to account for the diurnal variation of radiation for both ERBS and NOAA 9 clear-sky measurements. The monthly clear-sky longwave flux differences between ERBS and NOAA 9 are explained mostly by the ERBS diurnal sampling bias and the different scene identification procedures for daytime and nighttime. When applied to NOAA 9 cloudy-sky measurements, considerable errors result in the summer hemisphere. This is due to the substantial diurnal variations of cloud properties and the NOAA 9 sampling at only around 2:30 and 14:30 local time in most latitudes. For ERBS cloudy-sky measurements, the above errors are seen to be reduced significantly because of better diurnal sampling. A similar examination is suggested for the case of NOAA 10, the third ERBS satellite, which samples at around 7:30 and 19:30 local time in most latitudes.