Investigation of an Automated Temporal Disaggregation Technique for Convective Outlooks during the 2012 Hazardous Weather Testbed Spring Forecasting Experiment

Storm Prediction Center (SPC) forecasters have demonstrated skill over the years in creating severe weather outlooks for the convective day (i.e., 12Z-12Z). Although increasing the temporal resolution of convective outlooks is desirable, the question arises about how to accomplish this task without requiring a significant increase in workload for the forecaster. An automated approach to begin addressing this topic was developed for the 2012 NOAA Hazardous Weather Testbed Spring Forecasting Experiment (SFE). The temporal disaggregation technique takes a forecaster-generated, longer-period total severe storm probability outlook and breaks it into shorter individual time periods using ensemble model guidance as the input. The end result is an automated higher temporal resolution forecast of total severe weather probability consistent with the forecaster outlook.

During the 2012 SFE, calibrated total severe storm probability guidance from the SPC storm-scale ensemble of opportunity (SSEO) was used to temporally disaggregate an experimental longer-period (i.e., 16Z-12Z) human forecast of total severe storm probability (i.e., unlike SPC Day 1 Convective Outlooks, which include separate probabilities for severe hail, wind, and tornadoes) over a regional domain. The general procedure involved scaling the longer-period SSEO guidance to match the human forecast, applying the scaling factor to the SSEO guidance for each individual period (i.e., 16-20Z, 20-00Z, 00-04Z, and 04-12Z), and utilizing checks to ensure consistency with the human forecast. Subjective and objective comparisons of automated individual period outlooks to forecaster-generated outlooks for identical time periods during the 2012 SFE demonstrate the effectiveness and skill of this technique in producing higher temporal resolution forecasts of total severe storm probability. The results also reveal the importance of incorporating the human longer-period forecast as input to this technique rather than depending solely on numerical model guidance.