Bayesian Processor of Output for Probabilistic Quantitative Precipitation Forecasting

Roman Krzysztofowicz  

University of Virginia  

Abstract:

The Bayesian Processor of Output (BPO) is a new, theoretically-based technique for probabilistic forecasting of weather variates.  It processes output from a numerical weather prediction model and optimally fuses it with climatic data in order to quantify uncertainty about a predictand.  There are three versions, for binary predictands, multi-category predictands, and continuous predictands. The first version of the BPO was developed and tested for the probability of precipitation (PoP) occurrence forecast.

This will be a tutorial introduction to the third version of the BPO that produces the probabilistic
quantitative precipitation forecast (PQPF), conditional on precipitation occurrence. The conditional PQPF is in the form of a continuous distribution function of the precipitation amount. The talk will explain the forecasting equation, present examples of forecasts, and report preliminary performance comparisons of the BPO with the Model Output Statistics (MOS) technique;  then it will point to future research on the Bayesian Processor of Ensemble (BPE).