Probabilistic medium range snow and ice forecasting at the Weather Prediction Center

During the past two winter seasons, the Development and Training Branch (DTB) and Hydrometeorological Testbed (HMT) have developed and tested a probabilistic winter weather outlook for the medium range forecast period (Day 4-7). Initially, forecasts issued during the 2013-2014 winter season were made by the Weather Prediction Center (WPC) forecasters on the Winter Weather Desk. These forecasts were not disseminated to the public but were evaluated internally. Additional testing was conducted during the 2014 HMT-WPC Winter Weather Experiment (WWE) and both subjective and objective verification results from the inaugural 2013-2014 season lead to several product development changes which included shifting the responsibility to the medium range forecasters. Further, during the 2014-2015 season, the forecasts are now being disseminated to the National Weather Service (NWS) Weather Forecast Office (WFO) forecasters.

The Day 4-7 probabilistic winter weather outlook was developed by WPC DTB meteorologist Michael Bodner. The quantitative precipitation forecast (QPF) component of the outlook leverages the expertise of the WPC medium range forecasters by utilizing their deterministic QPF prepared for the 48-hour periods (Day 4-5 and Day 6-7). This deterministic QPF is then disaggregated into 24-hour amounts utilizing a blended ensemble mean of the 6-hour and 48-hour QPF from the Global Ensemble Forecast System (GEFS), European Centre Ensemble System (ECENS), and the Canadian Meteorological Center Ensemble (CMCE). The disaggregated 24-hour WPC QPF is then used as a mean, and the 24-hour QPF from each of the members from the GEFS (20 members), ECENS (50 members), and CMCE (20 members) are used to compute a cumulative distribution function (CDF) for amounts equal to or greater than 0.10”. The result is used to create a probability of WPC QPF being greater than or equal to 0.10” in the 24-hour period. Meanwhile, the thermal component of the outlook is addressed by computing an ensemble probability of frozen precipitation using derived precipitation type fields from each of the 90 members of the GEFS, ECENS, and CMCE. These derived precipitation type fields are computed for each ensemble member using the 2-meter temperatures and temperatures at several mandatory levels (925-mb, 850-mb, and 700-mb). A mosaic of snow, sleet, and freezing rain are used in computing the ensemble probability of frozen precipitation. The ensemble probability of frozen precipitation is then combined with the probability of WPC QPF greater than 0.10” to create the resultant probability of winter precipitation greater than 0.10”. Several cases will be presented to illustrate the utility of the winter weather outlook, as well as verification from the past two winter seasons. During the 2014-2015 season, probabilistic forecasts of freezing rain exceeding 0.01” and 0.05” thresholds, as well as snow and sleet exceeding 0.5” were tested in the 2015 HMT-WPC WWE. Findings of these test thresholds will also be presented.