Noticed two examples of numerical model flip-flops this morning for consecutive runs of both the WRF-NAM (short term forecast) and the ECMWF (long term global forecast).
First the WRF-NAM forecasts for this afternoon (Friday, November 21st). Above graphic shows accumulated precipitation through midnight tonight from the forecast run initialized at 12 UTC yesterday morning - with the model squeezing out a precipitation event for eastern Pima County as a moisture-starved short wave passes by. However, the current early run of same model (below), initialized at midnight this morning, forecasts only a bit of rain by midnight tonight over the far southeast corner of the state.
The changes in the forecasts reflect changes within the NWS NAM forecasts, since the high resolution WRF is driven by the lower resolution NAM model. Current forecasts for this afternoon, if they verify, indicate an unsettled situation with fits of wind, virga, mammatus, and some sprinkles around eastern Pima County.
At the global scale, the ECMWF has been forecasting a nasty winter storm along the Front Range of the Rockies for Thanksgiving day. The two panels shown above are from the ECMWF forecasts initialized at 12 UTC yesterday morning (500 mb at left and surface with 1000-500 mb thickness at right). The forecast is valid at 5 am MST on Thanksgiving day, and indicates strong, cold upslope flow and a nasty snow storm from Wyoming down to northern New Mexico.
However, the forecast that was initialized at 00 UTC last evening (below, same panels valid at 5 pm on Thanksgiving day). Has a completely different forecast (this flip-flop taking the ECMWF into sync with the GFS forecasts of past few days). So the current forecast from the ECMWF would bring a mild Thanksgiving day for the Front Range with some downslope, Chinook winds. This is one of the largest run-to-run flip-flops by the ECMWF that I can recall.
So many uncertainties still exist in weather forecasting for some situations - both for short and long terms.
No comments:
Post a Comment