I looked at the GFS and ECMWF model forecasts for temperature for Sandusky, Ohio for the heat of the day (high temperature) for next Monday. Now, what the models forecasted was for the high temperature to be anywhere from 77 to 82 degrees. Then I looked at the National Weather Service's forecast for that day, and it said that the high temperature would be 87 degrees. Is there some other factor that they use that I am missing, or did they just mess up or what? By the way, I'm pretty sure that there is no
NWS office in Sandusky, Ohio, if that matters. Any help is appreciated!
NWS office in Sandusky, Ohio, if that matters. Any help is appreciated!
-
A few possibilities...
1) The model run you saw was a different run than the one used for the NWS forecast.
2) 'Raw' model output from the GooFuS and EC is known as a 'perfect prog'...which of course the 'prog' is seldom if ever perfect.
3) To compensate for known errors in the numerical forecast...a technique called 'model output statistics' (MOS) is applied to the model's forecast values. MOS forecasts are also 'shaded' toward climatology toward the end of the forecast period (Day 5 though Day 7)
The NWS forecast is quite often very close to MOS. 'Forecasters' can tweak the numbers up or down but generally use MOS as the starting point.
1) The model run you saw was a different run than the one used for the NWS forecast.
2) 'Raw' model output from the GooFuS and EC is known as a 'perfect prog'...which of course the 'prog' is seldom if ever perfect.
3) To compensate for known errors in the numerical forecast...a technique called 'model output statistics' (MOS) is applied to the model's forecast values. MOS forecasts are also 'shaded' toward climatology toward the end of the forecast period (Day 5 though Day 7)
The NWS forecast is quite often very close to MOS. 'Forecasters' can tweak the numbers up or down but generally use MOS as the starting point.
-
They are using different models and came up with different answers?