Global monthly precipitation estimates from satellite-observed outgoing longwave radiation

Abstract

The relationship between the flux of outgoing longwave radiation (OLR) estimated from satellite observations and precipitation is investigated using monthly OLR data from the NOAA polar-orbiting satellites and the merged analysis of precipitation of Xie and Arkin for the 8-yr period from July 1987 to June 1995. The mean annual cycle of OLR in the Tropics is dominated by changes in cloudiness and exhibits a strong negative correlation with precipitation, while in the extratropics the strongest influence on the annual cycle of OLR is surface temperature and a positive correlation with precipitation is found. However, the anomaly of OLR exhibits a negative correlation with precipitation over most of the globe. The regression coefficient relating the anomaly of precipitation to that of OLR is spatially inhomogeneous and seasonally dependent but can be expressed with high accuracy as a globally uniform linear function of the local mean precipitation. Based on these results, a new technique is developed to estimate monthly precipitation over the globe from OLR data. First, the mean annual cycle of precipitation is calculated from the merged analysis of precipitation for the 8-yr period. The precipitation anomaly is then estimated from the OLR anomaly field using the coefficient value appropriate for the mean annual cycle of precipitation at each location. Finally, the total precipitation is estimated as the sum of the mean annual cycle and the anomaly. Verification tests showed that this estimate, which is referred to here as the OLR-based precipitation index (OPI), is able to represent large-scale precipitation with globally uniform and temporally stable high quality, similar to geostationary satellite IR-based estimates over the Tropics and to estimates based on microwave scattering observations over extratropical areas. The OPI estimates are then produced for the 22-yr period from 1974 to 1995 and are used to investigate the annual and interannual variability of global precipitation. The mean distribution and seasonal variations as observed in the 22-yr set of OPI estimates agree well with those of several published long-term means of precipitation estimated from station observations, and the interannual variability in precipitation associated with the El Niño-Southern Oscillation phenomenon resemble those found in previous studies but with additional details, particularly over ocean areas.