An Overview of the 2012 NOAA Hazardous Weather Testbed Spring Forecasting Experiment

Each spring, the Experimental Forecast Program of the NOAA/Hazardous Weather Testbed conducts a collaborative forecasting experiment. Organized by the Storm Prediction Center (SPC) and National Severe Storms Laboratory (NSSL), these annual forecasting experiments test emerging concepts and technologies designed to improve the prediction of hazardous mesoscale weather. The fundamental motivation for these experiments is to accelerate the transfer of promising new tools from research to operations and inspire new initiatives for operationally relevant research. The 2012 Spring Forecasting Experiment (SFE2012) was conducted 7 May – 8 June. It had 2 experimental forecasting components, one focusing on convection initiation (CI) and coverage and another aimed specifically at severe convection. In addition, there was a third element of SFE2012 involving experimental observing systems. An overarching theme of the forecast components was prediction of temporal trends in deep convection over regional domains. For severe convection, this emphasis involved breaking down 20-h probabilistic Convective Outlooks into 4- and 8-h sub-intervals. This breakdown was done independently 1) through a traditional decision-making process by the forecast teams, and 2) using an automated temporal disaggregation algorithm that ingested the longer period human forecast as a constraint and calibrated ensemble model guidance from the SPC Storm Scale Ensemble of Opportunity as input. Manual and automated sub-interval forecasts were compared to observations the next day to assess whether the automated forecasts might be useful as a first-guess in SPC operations where there is a desire to provide higher temporal resolution outlooks that are consistent with the current Day 1 Convective Outlooks. Separate forecasts for day-1 convective coverage also focused on 4-h periods over regional domains. Meanwhile, forecasts for initiation of specific convective episodes/events were designed to provide temporal resolution as fine as 1 h and a quantitative estimate of the uncertainty in the timing forecasts. The evaluation of experimental observations focused on multiple vertical sounding systems, with an emphasis on retrieved thermodynamic data. Contributing instruments included a passive microwave radiometer (model MP-3000, on loan from Radiometrics Corporation) and radiosonde sounding systems from Intermet, Vaisala, and Sippican. Emphasis was on evaluating the utility of each of these systems for both forecasting operations and experimental field programs. In addition, data gathered using these instruments were compared directly to model-forecast soundings to assess and compare the performance of different PBL parameterizations in the WRF model. Finally, SFE2012 also featured evaluations of experimental model output, guidance products, and forecasts. These evaluations included examinations of output from a 24-member 4-km, WRF- model ensemble generated by the University of Oklahoma Center for Analysis and Prediction of Storms (CAPS). WRF-model sensitivities to 1) microphysical parameterizations and 2) data-assimilation strategies were examined. In addition, a systematic comparison of objective and subjective assessments of human and model forecasts was performed. An overview of SFE2012 and each of these topics would be provided at the conference. An oral presentation is preferred.