Wednesday, March 02, 2005
The highly anticipated "post mortem" discussion...
Jason Samenow @ 2:00 PM
In yesterday's editorial, I stressed that we would be accountable for our forecasts here at CapitalWeather.com. So here you go, a 'no-spin' assessment of our forecast for Monday's winter storm...
What we forecasted
What we forecasted
- On Thursday night we indicated a major storm could move up the eastern seaboard bringing anything from heavy rain to historic snow.
- On Friday night we said "the most likely outcome is snow to rain to snow, with a light to moderate net snowfall" but that the uncertainties were "gargantuan."
- By Saturday night, we had become increasingly confident about a coastal storm track (which verified) and forecasted moderate to heavy snow accumulations north and west of a line from Vienna, Va to Bethesda, Md. Inside the beltway, we forecasted moderate snow accumulations and indicated the snow would mix with and change to sleet and/or rain before changing back to snow. (Interestingly, our forecast 36 hours prior to the event was superior to our forecast 12 hours before the first flakes fell.)
- On Sunday afternoon, we reiterated uncertainties and indicated 1-18" was still within the realm of possibility, but, in a subsequent post, noted that the most likely snowfall was in the range of 6 or 7 inches and graphed other probabilities.
- In our final pre-storm forecast Sunday night, we forecasted 6-10" for most of the area and indicated the snow would start late at night and become heavy by Monday morning through the afternoon, with light snow Monday night from the upper low.
- On Monday morning, as the snow was late arriving and guidance suggested a possible mix for some areas, we lowered accumulations inside the beltway to 4-8".
- On Monday afternoon, as the snow was not sticking in many areas, we indicated accumulations would likely be near the lower end of our revised projections.
- The snow began about five hours later than forecast.
- Snow fell but did not stick in many areas given barely freezing temperatures
- Total accumulations ranged from 2-6 inches. Most areas in metropolitan Washington got four or five inches except downtown areas, and areas adjacent to the river. Here is a nice map of snowfall totals from the National Weather Service.
- Saw the possibility of a storm 96 hours in advance
- Correctly characterized some of the appropriate uncertainties
- Did not hit the panic button or waffle during the storm. Some forecasters kept changing their forecast each hour. We eased our totals downward and appropriately emphasized the lower end of our forecasted accumulation range.
- Correctly forecasted accumulating snow from the upper low Monday night (which bailed us out a bit by giving us the additional 1-2" of accumulation)
- We miscalculated on the onset of the snow. I still can't figure out how it took the snow 8 hours to travel from Charlottesville to DC (110 miles) with the storm racing northeastward. Yeah, it was dry overhead, but not THAT dry. In 9/10 cases, snow in central Virginia takes just a few hours to get here. The later onset was important since the snow was starting during the day when it was more difficult for it to stick.
- We did not take into enough consideration the relatively warm temperatures and overestimated the storm's moisture. When temperatures are 32 or 33 degrees during a late February storm, unless it's snowing very hard, it will not stick well during the day time in the urban zone. And because the storm was less moisture laden than we anticipated, it didn't snow hard consistently. Hence, less snow and less snow sticking.
- We did "scoop" the competition by posting a credible accumulation forecast before any of the news media or National Weather Service. (This forecast, posted Saturday night, was actually better than our subsequent *updated* maps and would've verified quite well had we stuck with it.)
- Our Sunday night forecast of 6-10" tied with the National Weather Service as the least accurate forecast of those we documented. The best was WUSA's (Channel 9), which forecasted 4-6" -- so props to Tony Pann and Topper Shutt.
- For heavy snow accumulations in Washington in late February and March, either very cold air (below 30) or very heavy and persistent snowfall is required for snow to stick during the day light hours. We did not have a strong cold air source for this storm, which we should've recognized as the kiss of death for significant daytime accumulations -- especially in and around the city.
- We should really rely a little less on models, and look at what the weather maps are telling us. We did not pay enough attention to the fact there was this low in the Ohio Valley that was, in a sense, "competing" with the coastal low. This likely limited the intensification of the coastal low and resulted in less precipitation over the area. I observed this, and it bothered me, but I did not account for it in my forecast.
Comments are closed for this archived entry | Link | 
Email this post
Email this post
