An Intercomparison of Gauge Observations and Satellite Estimates of Monthly Precipitation

In order to further our quantitative understanding of the advantages and the shortcomings of the various sources of data used to represent climatic-scale precipitation, monthly gauge observations and satellite estimates are intercompared for global grid areas of 2.5° latitude/longitude for a period from July 1987 to June 1990. The results show that 1) at least five gauges are necessary to construct an areal-averaged monthly mean for the grids with accuracy of 10%, and 10% of the global land grids satisfy the requirement; 2) both microwave- and IR-based satellite estimates give similar spatial distributions of precipitation with good agreement with gauge observations for the warm seasons and over the tropical Pacific Ocean; and 3) the satellite estimates, especially those from the IR-based algorithm, exhibit poorer correspondence with gauge observations over land areas for the cold seasons. These results show that, for many applications, no single type of data can be used as the source for a monthly precipitation dataset with full global coverage, suggesting the need to improve the algorithms and to develop methods of combining the individual data sources, particularly in estimating extratropical precipitation.