|Coverage||Tropics (40N-40S, 0-360E)|
|Time Period||January 1986 - Present|
|Resolution||2.5 x 2.5 degree, monthly|
Satellite infrared window channel measurements around the 11 to 12 micron band have long been used to estimate precipitation over the tropics. The availability of IR window channels on board geostationary satellites, which see through to the surface except when clouds are present, provides high temporal sampling (3-hourly or better) from space. Since clouds are opaque at IR frequencies, the resulting brightness temperatures over cloudy regions corresponds to the temperature of the cloud top, providing a measure of the cloud-top height. In the tropics where deep convection is prevalent, high-level cloudiness is reasonably well correlated with precipitation, at least when averaged over large space and time scales. Unfortunately the relationship between cloud-top height and precipitation changes with latitude and over land as well as over regions with significant amounts of non-precipitating cirrus or other clouds. As a result, IR only techniques suffer from significant regional and time dependent biases, which is why the merged techniques use passive microwave rain estimates to "adjust" the IR estimates.
This dataset is one of the intermediate products from version 2 of the Global Precipitation Climatology Project (GPCP) merged rainfall dataset.
The GPI dataset is computed from geostationary IR window channel data using the technique originally developed by Arkin and Meisner . The GPI dataset is one of the component rainfall estimates used in the GPCP merged dataset [Huffman et al., 1997].
This dataset and other GPCP intermediate datasets are available through either of the following sources.