A Merged Algorithm for Quantitative Rainfall Estimates Using Thermal Infrared Bands

We present a new algorithm to generate rainfall estimates with the thermal infrared (IR) band. The algorithm merges ancillary passive microwave (PMW) data from SSM/I with IR images to generate rainfall rates at the spatial and temporal resolutions of the IR sensor. Comparison against independent land-GPCC data shows how the algorithm slightly improves the GOES precipitation index (GPI) estimates, being also very easy to implement. The algorithm combines a maximum entropy estimate of the probabilistic distribution function (PDF) of the rainfall rates with the IR brightness temperature of the clouds, generating a fused product. This combines the potentialities of the more direct rainfall estimation of the PMW sensors with the timely IR data, increasing also the spatial resolution. The algorithm can be applied both with a bulk parameterisation derived with ancillary data, or be fine-tuned with timely PMW measurements. We show how to apply the algorithm for the direct generation of estimates, and the possible applications of the method.