Examples of “good” and “bad” model runs at 144 hours from this morning

First, let us examine the National Center for Environmental Prediction (NCEP) model that came out based on this morning’s data from around the globe.  This first panel is for the winds around 500 mb (millibars, or “hectopascals”) and is at about 16,ooo to 20,ooo feet in the atmosphere, depending on how high or low the point at which a pressure of 500 is reached.  The winds blow along the lines, and the strength is noted by colors; red  and purple are real strong.  The colors tell you where the jet stream is, and storms are steered  by and develop/dissipate along it.

This panel below is for 144 hours from now or is valid Tuesday, October 4th in normal speak.  Note the area off southern California, where the winds are light, come on to the coast and then turn to the NE over Arizona.  Where the winds turn like that is called a “trough.”   The air tends to rise on east side of a trough, and descend on the west side.   This particular “trough” over southern California would be considered rather weak.  However, some tropical moisture is still around and some spotty showers are forecast for eastern Arizona (not shown) with this pattern.

This is not a particular “good” model output; perhaps it could be considered even quite “bad” since we need a lot of rain, and this trough, according to THIS model, is quite weak, and can’t do the job.

So, we start looking around to see if there is a “good” model run that for this same forecast hour, 144 h from now, on Tuesday morning, October 4th, which shows that there will be a LOT of rain.

Sure enough, Environment Canada has supplied what we are looking for based on the SAME data taken around the globe, but using a version of the European Center for Medium Range Forecasting model.  The next figure shows the “classic” four panel projection for that time from the Canadians.  Here, the jet stream part of the forecast is the upper left hand corner, and LOOK at the difference along the West Coast!  Instead of a jet stream ramming into Oregony and northern Califiornia as it does in the US model, its going into Juneau, AK!  And more importantly, that Canadian jet stream map shows a whole low pressure center along the California coast!  And with a low center there, and tropical air to the east of that low, it acts like a pully system to bring a gush of that tropical air across Arizona with substantial rain, as predicted as noted by the green areas in the panel in the lower right.  Not shown is the rain even more substantial rain that PRECEDES this map, also substantial for the 24 h prior to this map.  The heavier rain is due to the low center along the California coast being a stronger, better organizer of the storms in the tropical air over us.

Thus, the Environment Canada model output for 144 h,  would be called a “good” model prediction since we need rain and this model run “delivers.”

So, this is what forecasters deal with a lot of the time, though the differences shown here would be considered a little out of the ordinary as great as they are in this “model divergence.”

In real life the models are perturbed slightly at the beginning of the prediction period to assess how different the results could be, and a whole group of predictions is obtained from the SAME model having slightly different starting conditions.   The resulting outputs are called “ensembles”.   The more likely a situation will be observed in the future is in how LITTLE the ensembles vary.  Here, to be serious for a second, this :model divergence” for the situation in the SW means that neither model output can be relied upon heavily.


The End.