Opportunities in Probabilistic Quantitative Precipitation Forecasting (PQPF) Research

Steve Mullen
University of Arizona

Abstract:

Precipitation (amount and type) is arguably the surface sensible weather element of greatest importance to society. Unfortunately, precipitation remains the most challenging weather element to forecast skillfully. There are three primary reasons for this behavior, all of which can not be ignored: 1) large model errors, 2) high levels of analysis-observational uncertainty and inherently short predictability limits. The talk will touch upon all three aspects of the challenge. We argue that inclusion of stochastic forcing in our NWP constructs may offer a viable, cost-effective approach to improve certain aspects of the performance of ensemble forecast systems that avoids the overhead associated with an operational center supporting numerous versions of models and parameterizations. Incremental improvements in forecast skill and value will remain difficult to prove unequivocally, however, in the face of the observational uncertainty and the short predictability limits that characterize precipitation, especially heavy precipitation.