High-frequency estimation of rainfall from thunderstorms via satellite infrared and a long-range lightning network in Europe

A rain retrieval technique called Omvrios, that combines geostationary satellite infrared (IR) observations and cloud-to-ground (CG) lightning information retrieved from a long-range lightning detection network (Zeus) in Europe, is presented. Cloud systems are defined in the IR temperature array by the 255 K isotherm. Bulk parametrizations that relate cloud top IR temperature, morphological characteristics and CG fightning information (location and flash rate) are used to discriminate rainy from non-rainy cloud systems, and to evaluate the convective and stratiform rain areas and associated area-averaged rain rates of the rainy systems. The technique's parameters were calibrated using collocated and instantaneous rain fields derived from Special Sensor Microwave/Imager passive microwave data. Retrieved rain estimates are aggregated to 0.1degrees grid resolution and 6 h temporal accumulation. The technique is validated during the warm season from May to August 2002, based on independent 6 h rain accumulation measurements from a network of 700 rain-gauges located across Europe. Statistical analysis shows high correlation with gauge rainfall data (0.88) and low overall systematic difference (similar to5%). Besides direct validation against rain-gauge data, Omvrios has been compared to existing passive microwave-calibrated IR rain-retrieval techniques. It is shown that lightning information can lead to a significant (25-40%) reduction in the random error of IR retrievals and to an increase of nearly 0.3 in correlation with gauges. In terms of systematic differences (retrieval bias) Omvrios is shown to be consistent, as conditional and unconditional biases are nearly equal (within 5%), while for the other IR rain retrievals the variation between conditional and unconditional biases was significant (34-75%).