Fog!                                                           11/12

     1.  Why is it that some weather forecasting models don't agree?
              Because the atmosphere is much more complex than any model could ever be, models are based on simplifying
                 assumptions.  Those assumptions differ between models, thus so do the calculated predictions.  There are also
                 a number of other differences between models.  These will be discussed in class and are summarized in the
                 graphic Atmospheric Model Variations.

    2.  What forecasting model is best to use when forecasting just everyday weather?  Is there one model that
         is better than the others?
              If there were an easily identifiable "best" model, probably everyone would use it and all the others would
                 disappear.  Many older models have fallen into disuse because they have been replaced by newer, better
                 models.  However of the models currently widely used, no single model stands out as best.  Some models work
                 best at certain times of year, others for forecasting the weather in certain locations, others work best in certain
                 weather patterns.  Current local forecasters have recently been favoring the GFS (Global Forecast System), the
                 ECMWF (European Centre for Meteorology Forecast Model), the NAM (North American Model) and the MRF
                 (Medium Range Forecast Model).

    3.  What models do local forecasters use on television?  Is it a specific one or a combination?
              To my knowledge, no TV forecaster has indicated which models are his favorites.  That's probably regarded as
                 information too technical to be appreciated by the average viewer.  Almost certainly the most favored models
                 are among those mentioned in the answer to the last question.  Often forecasters use an "ensemble" forecast
                 which is an average based upon more than one model.  Even ensemble models disagree, depending upon
                 which models make up the ensemble.

    4.  How accurate are weather forecasts?
              That's a difficult question to answer because it is difficult to define "accuracy."  Generally forecasters are very good
                 at predicting general patterns, say a stormy period, a heat wave or a period of fair weather.  However they are not
                 so good at forecasting the timing of changes.  A storm may often arrive hours (or occasionally even a day or two)
                 ahead of or behind schedule.  Also they are not so good at forecasting local differences in conditions.  For example
                 a forecast general rainstorm may bring an inch of rain to Ogden, a tenth of an inch to Provo, and only a sprinkle to
                 Salt Lake.  Unforecast variations that occur with location are especially common in mountainous terrain, e.g., in
                 Utah.  Also forecasts are seldom highly precise.  It is rare that the actual minimum and maximum temperatures in a
                 location are exactly the forecast values.  More commonly they are off by 1-5 degrees.  It is also rare that they are off
                 by large amounts.  Clearly forecasts are much better than they were several years ago and are still improving in
                 accuracy.

     5.  How far into the past do we look while analyzing weather using the analogue method?
              That's a question with at least two interpretations, over how long a period do we try to match conditions, and how far
                 back into the past do we look for similar patterns?  The answer to the first interpretation of the question is a few
                 days, the answer to the second is several years, even a few decades in some cases.  (Because the extended western
                 drought of the first several years of the 21st century has been unmatched since the 1930s, forecasters have looked back
                 that far for clues that would indicate whether we have now emerged from that drought or whether it will yet persist.
                 However they do not look to the 1930s for analogous situations in day-to-day forecasting.)

    6.  In what cases would someone use the persistence method?
                 The persistence method is not really used by professional forecasters.  However everyday people use it all the time.
                 Usually when we make plans without checking forecasts, it is with the supposition that tomorrow will be much like
                 today, the basic assumption of the persistence method.  At certain times of the year, especially in certain locations, the
                 persistence method is very accurate.