Well… that didn’t go according to plan. The low pressure that was supposed to drop roughly 6 inches of snow on average, only provided roughly 1 or 2 inches on average. So, what happened? What went wrong? Why did the region see less than half of what was forecasted?
The primary cause for the forecast bust… was very dry air, the magnitude of which was dramatically understated within the forecast data. An area of dry air in the mid levels (700mb layer) of the atmosphere pushed into the back end of the storm system. The result was a rapid “drying out” of the storm from within. We watched on radar as a storm with a “healthy” looking radar, quickly was eroded away from within. But the computer guidance just could not see how quickly the dry air was eroding the storm… even after it had already begun.
Here’s an example…
On the 9pm run of the HRRR model, it produced a futurecast radar image for 10:30pm that depicted a storm that looked very “healthy”. A large swath of moderate to heavy snow extending back into central PA, before it mixed with sleet and freezing rain further west. This image looked exactly like what we had been expecting all along. This data is what our forecast… and most other forecasts… were based on.
However… in reality… the storm was already being invaded by dry air. And even at the time the 9pm HRRR was running, the computer model couldn’t see it. It produced the futurecast for 10:30pm, only 1.5 hours into the future, that depicted a ‘healthy’ storm. The actual radar for 10:20pm looked nothing like what the computer model data was showing. Dry air had already shut off snow into the western half of the Hudson Valley.
What all this shows, is that even with the amazing technology that we have… that there are limitations. We usually are able to forecast the weather with accuracy… but sometimes key pieces of data are missing, and a forecast goes up in smoke at the very last minute. Alex was able to see the warning signs of trouble with the forecast, and tried to post about it… https://www.facebook.com/share/p/1FWa3tGtd8/ but by that point, it was too late… it was 7pm, and the storm was underway… there was no saving the forecast. All we could do was hope that what we were seeing as a false alarm… and once it became clear it was not a false alarm, try to adjust expectations in real time.
We do our best to provide you the most accurate forecast in the region. But sometimes we’re presented with inaccurate data, and that is not possible. Even at 11pm and 12:30am… as the Saturday night runs of the GFS and European computer data were coming out… those computer models still had Poughkeepsie seeing 7″ of snow. By the time the computer model data was available, it was already invalid. In the rare cases where something like this happens, the best we can do is to try and explain WHY we got it wrong. So that we can all understand why things didn’t go as expected… and try to avoid it happening again in the future. Thanks for all your continued support! Have a great Superbowl Sunday!