Forecasting beyond 7 days: What's our Christmas forecast?
I get questions like these quite often before important holidays. The last several days I've received a lot about the Christmas forecast, much more than usual.
Earlier in the week the GFS, the American model, showed Christmas week with temperatures in the teens and 20s. Obviously, that's record-breaking cold!
On Friday, that same model shows temperatures in the 30s in the morning with highs in the 50s, similar to what we had Wednesday and Thursday this week. So what is the weather going to be like?
"Any idea what the weather will be like on Christmas day?" -Brent Barclay
"I am seeing on the internet about a Siberian Express coming down around Christmas day or little before...ruling out and understanding that it is two weeks out with predictions tough, what are your thoughts on this?.” - Mike Beauchamp
“We hear that Houston may have an ice storm Christmas from another news station here at the office. What's your prediction?.” - Sharon Stephens
The question is: How do I make plans for a forecast beyond seven days? There isn't a comforting answer.
Weather models are pretty good three to five days out. However, there is a drastic drop in accuracy once you go beyond six days. Also the GFS (American Long Range Model) is notorious for being highly inaccurate beyond seven days. All of these questions are asking for a forecast more than two weeks away. A lot can change with the weather in that time.
Low pressure disturbances that create rain usually only show up on the weather models 12 hours to three days out. Cold fronts, especially for southeast Texas, can stall north of us or even on top of us, and that will completely change a forecast two to three days out. This explains why our seven-day forecast will change for the weekend Monday through Thursday of the same week.
I could keep giving examples, but the point is the weather is constantly in motion and always changing.
With Christmas still a week and a half away, we'll see more changes to the forecast. Anyone want to make a friendly wager and predict the low and high temperature for Christmas Day within two degrees?
Studies have been done on long-range forecasting past seven days and found no skill in prediction. In fact, climatology, or the seasonal average, will beat a computer-generated long-range forecast.
AccuWeather is marketing a 25-day forecast. I would love a middle school school student to do their science fair project comparing the forecast to what actually happens.