Martinsville Weather - Winter Storm Forecast Review for Sun 01/25/04
The main thing I learned while forecasting this winter storm is that I really need to improve my nowcasting skills. I will explain that more in a moment. I will break down this review into several sections to address each aspect of forecasting this storm.
Long Range Forecast: My long range forecast for this storm was my strong point. I predicted this storm 10 days in advance (with the exception of 8 days out when I briefly changed my forecast to Sunny). As a result, my 10 day forecast accuracy for Sunday was 89.26%. (Keep in mind that this percentage does not account for accumulation forecasts, just the general weather as well as high and low temperatures.) Even when others were downplaying the storm or calling for warmer temperatures and rain, I stuck to my forecast and it payed off in the end. The only bad part of my long range forecast is that I did not catch this lingering freezing drizzle until just a couple of days in advance because I didn't expect the wedge of cold air to stay in place this long.
Short Range Timing Forecast: My timing forecast for this storm was pretty good. In Saturday morning's update, I said the precip would start around 9 AM and end around 7 PM. Sunday morning's update was similar. The precip actually started around 8 AM, so an hour off is pretty good. It ended around 10 PM, but it lasted a little longer in this area than in most of the region, so this prediction wasn't too bad either.
Precipitation Type Forecast: Precip type is always very tricky to determine, and this storm was no exception. I originally predicted (in Saturday morning's update) a changeover to sleet around noon Sunday, then changed that to around 4 PM in Sunday morning's update. The 4 PM call ended up being pretty good, as the precip changed back and forth between snow and sleet between about 3 and 7 PM Sunday before finally changing over to all sleet. No one expected the cold air to push as far south as it did (all the way to coastal South Carolina and central Georgia), which is why the snow lasted longer than most people predicted.
Temperature Forecast: This was a very unique storm with regard to temperatures. It was a nearly perfect setup for very cold air to pour in east of the mountains. A strong high pressure system in the Northeast caused the incredibly strong wedge of cold air to form, with temperatures plunging into the lower 20s as far south as the North Carolina/South Carolina border. Locally, my temperature forecast was very good overall. The lowest temperature on Sunday was 16 and I had forecasted 19, and I expected the temperature to start rising a little Sunday night but it didn't start to rise until Monday morning. Still, my temperature forecast overall was very good.
Accumulation Forecast: Now we come to the worst part of my forecast. Saturday morning, I predicted 2 to 3 inches of snow and sleet before a changeover to freezing rain. By Saturday evening, it became apparent that the precipitation was not going to change to freezing rain as quickly as I had thought, so I increased my forecast accumulation to 3 to 5 inches. When I updated my forecast Sunday morning, I made the worst mistake. The snow was already falling when I updated my forecast at 8 AM, several hours before the computer models had predicted. Because of that, I upped my forecast significantly to 6 to 10 inches, which turned out to be a very bad decision. For some reason (I am still not sure why), the snow just didn't accumulate very quickly Sunday morning, and by the time my 2 PM update rolled around only about 2.5 to 3 inches had fallen. I then lowered my forecast to 6 to 8 inches, which was better, but still too high. Based on reports I have received, most locations picked up between 5 and 7 inches, after accounting for a little bit of compression of the snow caused by the sleet that fell Sunday evening. The National Weather Service predicted 3 to 6 inches, so they were a little on the low side. The best forecast would have been 4 to 8 inches, which is what the computer-generated forecast from The Weather Channel predicted. I actually considered predicting 4-8 or 5-8 when I was updating Sunday morning, but I decided on 6-10 inches when I found out that it had already started snowing. This is why I said at the beginning of this review that I need to work on my nowcasting skills. (Nowcasting refers to forecasts for the immediate future, such as the next few hours.) If I had only gone up to 4-8" instead of making such a drastic increase by going with 6-10", I would be much happier with my forecast based on the way everything turned out. Also, I saw a forecast (from a private-sector meteorologist who did very well with Sunday's storm) predicting a major ice storm on Monday after my Sunday evening update. I glanced at a couple of computer model forecasts and did a quick update, saying that 3/4 to 1 inch of ice accumulation was possible for Monday afternoon, then I had to drop that back down to around 1/2 inch in Monday morning's update. Both forecasts ended up being way too high as very little freezing rain fell Monday afternoon. This is another example of why I need to improve my nowcasting. When I see something that indicates the possibility of a significant snow or ice accumulation, I tend to jump on it before considering all the possible negative things that could cause that forecast to be wrong. Overall this winter I have been much more conservative with my snowfall forecasts (after I did so horribly with predicting way too much snow accumulation with practically every storm last winter), but I didn't put as much thought into just how much snow 6-10 inches is and how unlikely it was that 10 inches of snow would fall.
Conclusion: I give myself an A- for this forecast. I did a good job with most of the aspects of my forecast, but my accumulation forecast should have been a little bit lower. If I had predicted 4-8 inches or so instead, I would have earned an A instead. Still, as difficult a forecast as this was, I did a very good job overall.
New Record: My website set a new record for daily visitors on Sunday. I received a total of 759 hits, shattering the old record of 585 which occurred on February 26, 2003. Thanks to everyone who made this possible! I still hope that, if we get another major winter storm, I can reach my goal of 1,000 daily hits at least once this winter. Please continue to spread the word about my website so it can continue to grow!