Wednesday, February 25, 2009

Whoops

Okay, I admit to screwing up the DeKalb forecast. I put overcast, and it was clear, all day. I can only blame myself for trusting model data, instead of myself. Initially, MOS (model output statistics) guidance indicated overcast conditions all day in DeKalb. I was not fully convinced, but just to make sure I did not miss anything, I checked the area forecast discussions from 3 National Weather Service offices: Chicago, St. Louis, and Davenport. All three indicated cloudiness, but all three had different reasons. One suggested clouds were moving in from western Illinois, which after looking at infrared satellite, showed no clouds in or around Illinois. Another suggested a 500 mb shortwave trough, which after looking at a 500 mb map (and practically squinting at it), I could barely find evidence of a shortwave. The third office suggested moisture flowing in with the low level jet, which is what I went with after seeing a nice low level jet off the Gulf of Mexico at that time. My initial thoughts made me question whether the discussions were trying to explain what was happening or if they were trying to verify what the model suggested.

Needless to say, instead of forecasting with my instinct, which told me little to no clouds in DeKalb, I went with the "meteorological cancer" (which is going with what the model suggests). I regret my decision, especially since I had read a few hours before forecasting that the models were not handling the weather very well. So, I admit my mistake. To avoid this next time, I am going to "challenge" the models to a forecast contest. Next forecast, I will NOT use any models, only data (which includes observed conditions, satellite, radar, and hand analysis maps) and my meteorological knowledge and instinct. After this, I will post a forecast using model guidance, without my input (meaning I will just put down what the models suggest, none of my own thoughts). My goal is to show that meteorologists need to use models only for guidance, and not base a forecast on them alone. Stayed tuned for the contest of man vs. machine!

No comments:

Post a Comment