A Blizzard Warning and Winter Storm Warning has been issued parts of the Southern and Central Plains, the second such blizzard in a week for some residents. However a concept in disaster preparedness can be readily displayed with the forecast for this event. Currently, the National Weather Service is forecasting a foot as the upper limit to the snowfall values in Kansas and 15″ as the upper limit to snowfall values in extreme Northeastern part of the Texas Panhandle. But the highest snowfall totals for this storm could be much much higher….
Recently, the NOAA Hydrometeorological Prediction Center (HPC) began issuing probabilistic snowfall graphics (shown below) that show snowfall forecasts where snowfall values are exceeded 90%, 75%, 50%, 25% and 10% of the time.
HPC 50th Percentile Snow Forecast
(Click for larger image)
HPC 90th Percentile Snow Forecast
(Click for larger image)
Essentially a 50th percentile snow forecast is the forecast that 50% of the time the amounts will be higher and 50% of the time, the amounts will be lower. For planning purposes, this is the likely amount forecasted if you’re going to play the middle of the road.
However the 90th percentile forecast is quite different. It shows the amounts that will be exceeded only 10% of the time. While many people want to forecast snow amounts accurately, the 10% probability event is a great resource to “Plan for the worst” and the 50% probability event is “hoping for the best”.
The wild card in events like this is thunder snow. Essentially, areas where thunder snow occurs can receive locally higher snowfall amounts. The snow probabilities point to this potentiality, although the bands of intense snowfall will not cover the entire area. Depending on where the most intense bands set up, it will dramatically impact the amounts of snow received in those areas. This again points to the importance of the 10% exceedance threshold. Most people will receive snow amounts closer to the 50th percentile amount, but there will be pockets where people receive substantially higher amounts.
Comparison between NWS forecast (left) and HPC 90th percentile / 10% exceedance forecast (right) Current as of 0800CT on 2/24/2013
You may be saying right now, “that’s great but I hate math and hate probabilities”. Communicating potential risk, especially in low probability, high impact events is critical for anticipating the worst and taking protective action while hoping for the best. If there was a 10% chance of an intruder in your house, going after you and your family, would you take protective action? If there was a 1 in 10 chance that you could lose your job, would you start developing a backup plan?
We’ll see how this specific event unfolds, but the current forecast (left in map above) isn’t even at the levels depicted in the 50th percentile event (likely underestimating snow amounts). Between that and the incredible disparity between the forecast and 10% potential snowfall amounts, this is a classic example where people can and likely will be caught surprised by the event.