Hurricane Irene Model and Forecast Discussion

[BethesdaWX]

It's not going to go any further west and will likely trend even further east before all is said and done. Do you really think it's going to do anything other than brush the NC outer banks and Cape Cod?

Do you have reasoning for this assertion? Given how far out we currently are, I can't see how one could assert "this won't happen". It may be a fish in relevance to your backyard, but to those in highly populated areas on the Eastern Seaboard it could be a major headache. The NOGAPS apparently has a fish, but until we get closer to the event it is hard to write anything off the table.

[HurricaneJosh]

This isn't hard. A storm that spends most of its time offshore and only landfalls on barrier islands is pretty much a fish. The trend is also pretty clear. It's not going to go any further west and will likely trend even further east before all is said and done. Do you really think it's going to do anything other than brush the NC outer banks and Cape Cod?

Again, please post the credible model that shows a fish. No more word spinning-- just produce it.

[Ian]

i would not have massive faith in the ridge pushing it straight north tho the gfs solution is not crazy by any means. too many variables.. if that lead trough gets any stronger it could help push down the edges of the ridge a bit, etc.. you don't nee much of a route to make these things want to go east that's for sure. the ukmet track is not bad but it doesnt seem to get that inland through those two panels at least. over the last 2 days we've come way east so i dont think it's crazy to worry there could be more despite a "great setup" or whatnot.

[yoda]

Does the UKIE go beyond 144? Or does it stop there? Cause it would be interesting to see where Irene goes after 144 sitting about 50-100 miles west of OBX over inland NC. It appears as if its going NNE after it makes landfall IMO

[ORH_wxman]

Does the UKIE go beyond 144? Or does it stop there? Cause it would be interesting to see where Irene goes after 144 sitting about 50-100 miles west of OBX over inland NC. It appears as if its going NNE after it makes landfall IMO

It stops at 144h.

[HurricaneJosh]

For the record, I'm not saying it won't fish. I have no clue. I don't get in the game of forecasting-- it's not my cup of tea. I'm a chaser/historian.

It's just amazing the model-hugging here: people see one run of one model, and it's now this big, hardwired trend in their minds. The number-one rule that we all learn here-- at least, I thought we did-- is that you don't flip out and predict the outcome based on one model run: you wait and see if it pans out over subsequent runs, and in other models. That's what the NHC does, at least.

I'm not even a forecaster and I know these basic rules Re: how to use and apply models.

[Riptide]

More people are jumping onto these model runs because of the newly injected G-IV dropsonde data which supposedly makes the models more accurate.

[yoda]

GGEM at 120 brushes OBX -- http://www.weatheroffice.gc.ca/data/model_forecast/136_100.gif

[HurricaneJosh]

More people are jumping onto these model runs because of the newly injected G-IV dropsonde data which supposedly makes the models more accurate.

Yes, but that doesn't make them magic-- it simply makes them more informed than they were before.

Again, the darn thing might fish-- I certainly don't know.

[ORH_wxman]

Fore the record, I'm not saying it won't fish. I have no clue. I don't get in the game of forecasting-- it's not my cup of tea. I'm a chaser/historian.

It's just amazing the model-hugging here: people see one run of one model, and it's now this big, hardwired trend in their minds. The number-one rule that we all learn here-- at least, I thought we did-- is that you don't flip out and predict the outcome based on one model run: you wait and see if it pans out over subsequent runs, and in other models. That's what the NHC does, at least.

I'm not even a forecaster and I know these basic rules Re: how to use and apply models.

Don't worry, people do it in winter too. The power of ensemble forecasting shows itself during this time frame. Probabilistic forecasting as well. But there will always be people who model hug regardless.

Pattern recognition is huge too. We can sometimes gauge how realistic model solutions are based on synoptic patterns. TC forecasting is definitely more fickle, but these rules still apply.

[yoda]

144 GGEM -- http://www.weatherof...ast/495_100.gif

Just inside the benchmark

So it went NNE/NE after scraping OBX at 120

[Bozart]

It's just amazing the model-hugging here: people see one run of one model, and it's now this big, hardwired trend in their minds. The number-one rule that we all learn here-- at least, I thought we did-- is that you don't flip out and predict the outcome based on one model run: you wait and see if it pans out over subsequent runs, and in other models. That's what the NHC does, at least.

But it's more than one model run, Josh. In a fairly short time span, we've gone from entertaining a threat to Mobile, to Pensacola, to the FL peninsula, to Savannah, to Charleston, to Wilmington, to barely eeking out any hit at all. You don't see a trend here?

[Ian]

Don't worry, people do it in winter too. The power of ensemble forecasting shows itself during this time frame. Probabilistic forecasting as well. But there will always be people who model hug regardless.

Pattern recognition is huge too. We can sometimes gauge how realistic model solutions are based on synoptic patterns. TC forecasting is definitely more fickle, but these rules still apply.

it's not all model hugging either. some of us have been watching lows be they tropical or otherwise on the east coast for decade/s. there are some general ideas that tend to work out more often than not. sure that can dig you into a hole if you let it but at d5-7 how often do you usually want a perfect track or one on the east of the envelope? if you look back through these threads you'll see numerous cases where this storm alone did things most people thought it would not yet there is still some semblance of certainty to what lies ahead.

[ORH_wxman]

it's not all model hugging either. some of us have been watching lows be they tropical or otherwise on the east coast for decade/s. there are some general ideas that tend to work out more often than not. sure that can dig you into a hole if you let it but at d5-7 how often do you usually want a perfect track or one on the east of the envelope? if you look back through these threads you'll see numerous cases where this storm alone did things most people thought it would not yet there is still some semblance of certainty to what lies ahead.

I'm not sure how this contradicts anything I said. I was talking about some people model hugging, and I said some experience with synoptic patterns and a good feel for ensemble forecasting will produce better results than the former....I think you are kind of saying the same thing.

[QVectorman]

But it's more than one model run, Josh. In a fairly short time span, we've gone from entertaining a threat to Mobile, to Pensacola, to the FL peninsula, to Savannah, to Charleston, to Wilmington, to barely eeking out any hit at all. You don't see a trend here?

Thank you!! Glad someone finally pointed this out after 3 pages

[HurricaneJosh]

But it's more than one model run, Josh. In a fairly short time span, we've gone from entertaining a threat to Mobile, to Pensacola, to the FL peninsula, to Savannah, to Charleston, to Wilmington, to barely eeking out any hit at all. You don't see a trend here?

We're getting lost here. I asked which credible model shows a fish. Joel said "all of them". I asked him to show me one. None was shown. Now the discussion has become something else.

I already said that I'm not making a prediction of whether this will or won't be a fish-- I just wanted Joel to produce what he claims exists.

This is becoming a very boring tangent. Sorry, everyone.

[mattmfm]

48 hour errors now out, consistent bias of nearly 200 miles too far SW. That's just terrible. And we're still seeing even a 12 hr bias of 50 miles or so too far SW across the board.

[Cheeznado]

Thank you!! Glad someone finally pointed this out after 3 pages

I agree, the overall trend with a few wobbles has been east east east. A total fish is certainly an increasing possibility, though not quite in the likely category yet.

[Ian]

I'm not sure how this contradicts anything I said. I was talking about some people model hugging, and I said some experience with synoptic patterns and a good feel for ensemble forecasting will produce better results than the former....I think you are kind of saying the same thing.

I thought your first sentence kind of validated that it's model hugging. Someone like widre might get his posts removed but I think he's actually sort of coming from the right place. There have been plenty of red tags saying this or that is highly unlikely and that it more or less has to hit the coast as more than a swipe. I'm not saying it won't but there is some support for the 'unthinkable' happening.

[ArtRosen]

But it's more than one model run, Josh. In a fairly short time span, we've gone from entertaining a threat to Mobile, to Pensacola, to the FL peninsula, to Savannah, to Charleston, to Wilmington, to barely eeking out any hit at all. You don't see a trend here?

Actually, the reliable computer models have been showing a South/North Carolina hit since last Wednesday. I even posted on my Facebook page for people in the Carolinas to be on the lookout. The only trend is that the hurricane re-emerges into the Gulf Stream and heads up the coast to hit Eastern LI/Rhode Island/MA after it hits the Carolinas.

The EURO nor GFS have shown a hurricane hitting anywhere other than the Carolinas.

[NortheastPAWx]

We're getting lost here. I asked which credible model shows a fish. Joel said "all of them". I asked him to show me one. None was shown. Now the discussion has become something else.

I already said that I'm not making a prediction of whether this will or won't be a fish-- I just wanted Joel to produce what he claims exists.

This is becoming a very boring tangent. Sorry everyone.

Josh, he was responding to you claiming not to freak out about it going OTS because of one model run. He's correctly pointing out that the whole trend has been east. You're not concerned at all?

[atownwxwatcher]

But it's more than one model run, Josh. In a fairly short time span, we've gone from entertaining a threat to Mobile, to Pensacola, to the FL peninsula, to Savannah, to Charleston, to Wilmington, to barely eeking out any hit at all. You don't see a trend here?

I am going to use these images instead of the Atlantic ones..

12 Z GFS 156 hrs..

Now compare this to 00z GFS

Same time frame. Either this run is slower or it is slightly further southwest. When you open the images in side by side windows it becomes apparent. So just in comparing this to the 12 Z run because i was always taught to compare 00z to 12 Z and 6 Z to 18 Z to see the differences, if anything its edging towards the ECM.

Keep in mind that the ECM was much further west then this in it took the system straight over Albany..

Before I left for work tonight NHC mentioned that the rest of guidance was essentially following the leader and they named the ECM...

Lets see what the ECM does as it should be coming out shortly....

[HurricaneJosh]

I agree, the overall trend with a few wobbles has been east east east. A total fish is certainly an increasing possibility, though not quite in the likely category yet.

I agree with this.

The whole discussion started when I challenged a poster explain why he insists this hurricane has a 10% chance of hitting the USA.

[mattmfm]

I thought your first sentence kind of validated that it's model hugging. Someone like widre might get his posts removed but I think he's actually sort of coming from the right place. There have been plenty of red tags saying this or that is highly unlikely and that it more or less has to hit the coast as more than a swipe. I'm not saying it won't but there is some support for the 'unthinkable' happening.

It's a fair sentiment to have, and I think it's wholly reasonable to consider an OTS solution. That said, I think this it's also fairly unreasonable to consider a landfall HSE-SNE hurricane as merely a fish.

[ORH_wxman]

Josh, he was responding to you claiming not to freak out about it going OTS because of one model run. He's correctly pointing out that the whole trend has been east. You're not concerned at all?

I'd be concerned about a trend east...but I certainly wouldn't use the 200+ hour progs as evidence of it. That time frame is so far off it doesn't matter. We'd be unaware of it a few years ago. I'd probably use the last 3-5 runs. But since is the first realistic landfall threat in a long time, I think its extremely easy to forget at the time frame we are working with. We are talking 4-5 days before potential landfall. These aren't 48 hours forecasts. The track errors at this time range are enormous.

[Amped]

For the record, I'm not saying it won't fish. I have no clue. I don't get in the game of forecasting-- it's not my cup of tea. I'm a chaser/historian.

It's just amazing the model-hugging here: people see one run of one model, and it's now this big, hardwired trend in their minds. The number-one rule that we all learn here-- at least, I thought we did-- is that you don't flip out and predict the outcome based on one model run: you wait and see if it pans out over subsequent runs, and in other models. That's what the NHC does, at least.

I'm not even a forecaster and I know these basic rules Re: how to use and apply models.

Okay as long as you havn't forgeton about 1995 and 1996, when just about every storm except Fran, went way further east than expected.The lesson I go out of those years was , when there's lack of a well defined steering flow it will go east, this applies mostly for cyclones north of florida.

[HurricaneJosh]

Josh, he was responding to you claiming not to freak out about it going OTS because of one model run. He's correctly pointing out that the whole trend has been east. You're not concerned at all?

Of course I don't like the E trend. However, I disagree with these claims:

* The cyclone has a 10% chance of hitting the USA now.

* All of the models show a fish.

These statements are extreme and don't raise the discussion here. The reality is that the setup is complex and we don't know what's going to happen. Of course it may fish.

[Ian]

It's a fair sentiment to have, and I think it's wholly reasonable to consider an OTS solution. That said, I think this it's also fairly unreasonable to consider a landfall HSE-SNE hurricane as merely a fish.

I'm not.. But I do remember that's still d4-7 or thereabouts. The GFS showed a solid us hit for 197 runs before swiping. What makes you confident at range?

[QVectorman]

We're getting lost here. I asked which credible model shows a fish. Joel said "all of them". I asked him to show me one. None was shown. Now the discussion has become something else.

I already said that I'm not making a prediction of whether this will or won't be a fish-- I just wanted Joel to produce what he claims exists.

This is becoming a very boring tangent. Sorry everyone.

Live by the models die by the models....excellent forecasters pay attention to model trends, bias, etc and their goal is to out perform the models. Not pick the best model in the current run. All these guys are forecasting what the next run will do based on the trend trying to out perform the models.

[mattmfm]

What makes you confident at range?

Absolutely nothing It's 150 hours out.