Caution with Using Model Guidance

[LSU2001]

I have read a number of posts(mostly from newer members) discussing model guidance like it was some type sure thing. I thought I would post simply to explain my thinking about models. So here goes.

Models are one very important aspect of tropical weather forecasting. However, they are by far not the only aspect. Simply because a model track shows a certain track does not mean that any given tropical cyclone will follow that track. Models are simply computer generated mathematical estimates of TC behavior given current data. Many other things affect the actual behavior of cyclones. I have witnessed models shift back and forth many times as new data is incorporated so to say that this place or that place is “safe” based on model data can be very problematic. It is always best for, amateurs like me, to refrain from interpreting model data. If tropical weather forecasting was so easy there would be no need for the myriad of experts employed by various countries to try and predict where a TC will go and how strong it will be. I am not bashing anyone nor am I trying to stop debate. I am only trying to inform some of the newcomers about the limitations of computer models.
Thanks for reading my truly amateur opinion,
Tim

[JenBayles]

You got it Tim - garbage in, garbage out.

[Derecho]

Actually, on the contrary, there isn't enough faith in models, IMHO.

I see far more people with silly little homilies about how you have to "use your gut" and "don't just look at the models" than people that are promoting the blind usage of models.

People seem to have a visceral, emotional dislike of computer modeling; I almost think it's beginning to bother people that the track forecasting of well-formed TCs is now so accurate with the use of models.

A serious problem is a lack of understanding of the relative accuracy of the various models; all too often people decide where they want the storm to go, and then select the model that matches that path, to support it; but often it's the "wishcaster's friend" Canadian, or the ETA, both proven horrible tropical models.

Consensus models like GUNA and CONU are very, very, very, very accurate and people should really think twice before going with a track that isn't very close to the above two models. The overwhelming majority of the time NHC tracks now are essentially identical to one of the consensus models above (or the FSU Superensemble.)

[LSU2001]

I agree with much of what you say but I am only saying that caution should be used when using model guidance. I for one am not knowledable enough to intrepret model data and form a forecast. I am not knocking models,only the lack of understanding that seems to be prevalent. I don't know how many posts I saw with Cindy and Dennis talking about the MM5 or other models clearly without the level of understanding that you are talking about. It is this kind of reliance that can be dangerous.
TIm

[Wthrman13]

In my opinion, there will always be room for human interpretation of model output, no matter how good they get, and I am one of the biggest advocates of numerical modeling of the atmosphere out there (it's essentially my entire graduate research). Without question numerical models are the backbone of modern weather forecasting. The problem is that peoples expectations of what models can do has advanced at a far more rapid rate in recent years than the actual increase in model accuracy and physical reasonability (which has been considerable, even so). Thus, you hear statements like "We were better off 30 years ago when we didn't have the models", or general feelings that computer models have actually made forecasting worse since they often disagree, particularly on hard to forecast phenomena like tropical cyclones. The reality is, models have advanced our ability to forecast numerous weather phenomena far beyond any human's ability alone, and plenty of statistics prove it.

That said, without a human forecaster who is well versed in both meteorology as a science and in a decent understanding of the basic workings, strengths, and limitations of the various models, the model output is, while not quite useless, at the very least much less valuable. As models get better and better, they will start pushing into such frontiers as forecasts of the development and evolution of individual or clusters of severe thunderstorms (something I'm working on in my current research), or even perhaps the forecast of tornado formation in a storm before it even happens (although we have a VERY long way to go before that becomes even a blip on the radar screen), and thus there will always be a need for humans studying these frontiers to be ready to understand the meteorology behind the phenomena that is being predicted by the models, to be able to interpret, make connections, and draw conclusions based on what the models are saying, something that arguably computers may never be able to do, even in principle (although this is debatable and touches on philosophy). I do believe that "gut feelings" and such, MAY have some value, as long as the person saying so has had lots of experience in forecasting. The "gut feeling" may actually be a reflection of that forecaster's understanding and experience at a subconcious level. The phrase is WAY overused however. The person saying so should be willing and able to back it up with sound scientific reasoning and/or appeals to past experience, otherwise such appeals to "gut feelings" are meaningless. Finally, it's just as naive, in my opinion, to think that computer models will replace human forecasters as it is to think that humans can forecast as well or better without the models.

[Hurricaneman]

The models have some value, it gives forecasters a guesstimate of where storms might go or how strong the may be

[Javlin]

Dittos WT13!

[wayoutfront]

which model do you guys think will be most accurate with the ridge ?

newbie question...

[weatherSnoop]

Quite eloquent Wthrman, I was thinking the models provide guidence of the storm itself and the influencial systems that may or may not form. As a very interested observer of the tropics, I think the models serve to bring out the creativity (Gut feelings) of the forcasters. I do not follow them as gospel, but listen intently to those who fully understand their usefulness.