Too much reliance on computer models?

[TSmith274]

What happened to the days when a meterologist put pencil to paper and figured these things out? It seems to me that the NHC is a slave to these computer models. If all they are going to do is replace all the model tracks with a yellow cone, or circle in Jeanne's case, then why do we even need the NHC? In recent years(since computer models), have they ever gone against them? I'm not saying that these models are bad, but we rely on computers so much these days that the art of forecasting these things is being lost. At least that's what it seems like to me. Anyone agree?

[Stormchaser16]

NOPE NOT AT ALL

what the heck kind of argument is this? Sorry we are having advancements in technology.......These models were designed to aid in the art of forecasting........

[x-y-no]

I don't think that's a fair criticism at all, at least in general. The discussions will frequently mention synoptic reasoning, and they will dismiss individual model runs if there is cause.

It would be irresponsible, IMHO, to simply disregard the models. This may lead to an appearance of over-reliance since they will generally get substantial mention.

[Ixolib]

The problem is not the models. It's relying on the validity of the models beyond three days. It's when the NHC started issuing the five day forecast - based largely on models - that all of this pre-storm attention reached its all-time high. Couple that with the "line", and the perception is that the models aren't reliable. They are - just not beyond three days out.

[Stormchaser16]

They are there to help GUIDE after 3 days... the NHC makes that clear, its the MEDIA that cannot understand that, thus the original post saying the NHC puts to much reliance on the models is TRASH

[TSmith274]

Wow I must have touched a nerve. Just wanted to clarify that I really didn't have an opinion either way. Just thought it was an interesting question. Guess it made me devil's advocate. But that's ok. Just as long as it's used, as John Gumm put it, as a tool and not a crutch.

[Ixolib]

Wow I must have touched a nerve. Just wanted to clarify that I really didn't have an opinion either way. Just thought it was an interesting question. Guess it made me devil's advocate. But that's ok. Just as long as it's used, as John Gumm put it, as a tool and not a crutch.

It is an interesting question, and one that will probably have hugely different perspectives in the answer. I believe the technology has both empowering and limiting impacts. Just like when I take my car to get it fixed. It's hard to find mechanics who can "troubleshoot" anything unless the "computer" shows them what the root cause of the problems is. Hopefully, in WX forecasting, the same isn't true and forecasters will always consider synoptic patterns and conditions - looking beyond what the "computer" is telling them. I just wonder what good ole Nash Roberts has to say about the advent of weather forecasting via computer modeling??

[Lockhart]

What happened to the days when a meterologist put pencil to paper and figured these things out? It seems to me that the NHC is a slave to these computer models. If all they are going to do is replace all the model tracks with a yellow cone, or circle in Jeanne's case, then why do we even need the NHC? In recent years(since computer models), have they ever gone against them? I'm not saying that these models are bad, but we rely on computers so much these days that the art of forecasting these things is being lost. At least that's what it seems like to me. Anyone agree?

They do weighted averages of the models based upon past perfomance and their personal beliefs. Sure, if they were going to just take the average of the models, you wouldn't need a meteorologist. But that's not all they do. Also, what's your alternative? Take a pencil and paper and do what, exactly? Work out what you think the storm is going to do? Uh, don't you think a dozen supercomputers might manage that a *little* better? Not only that, but they have numerous different views of the situation, so they're able to "guess" based on multiple models. Very strange question.

[wxwatcher2]

Models are good but they are only as good as the information that is fed into them.
I think a combination of models and good ole fashioned forecasting is a good combination.



GO GATORS

[Stormsfury]

Model guidance is just that ... guidance ...

It's up to the forecaster in charge to decipher a code within what the overall pattern is, and the key to success is to find the right combination to unlock the secrets of the upcoming weather ...

SF

[rbaker]

i believe the models are pretty good within 72 hrs, after that they shift every 6 or 12 hrs depending on the model. The TPC made a mistake coming out with a 5 day models, it puts too much hype, panic , hysteria, and overall LARGE errors, that gives the average joe blow on the street too much anxiety. The TPC will even tell you their own forecast beyond the 72 hrs frame is subj to large errors, which we have seen numerous times this year.
Then we have the infamous black line, which people who are not weather enthesusis like us, pay too much attention to. Prime example Charley. Get rid of the black line that changes like a dog wagging its tail. Keep the 3 day forecast.

[Sean in New Orleans]

The models will never be THE forecast and I think that they could possibly, one day, come into about 10% consideration of the actual forecast. It's good to view them and see what they think, but, that's about all they are and should ever be. They are a small tool in the big picture of making a forecast. There are many other simple tools to making any forecast like a personal, individual analysis of barometric pressure, winds, satellite photos, history, what's happening in China and the Indian Ocean, etc. Granted they are programmed into the computer models, but, these figures need the human analysis seperately all of the time.