ATL: MATTHEW - Post-Tropical - Discussion

[TimeZone]

It's only a small tip, but there are some pretty large mountains there. I'd expect weakening at a very decent rate to begin now.

[stormreader]

Your post is maybe the strongest I've ever seen on this site. To just endlessly run models with little or no commentary, or worse, have model runs based on info that pro mets perhaps can't decipher leaves one feeing under nourished. That's what happened with this storm. I knew we were in trouble with the original out to sea forecast when no one could really explain why. The only sure thing was the consistent strength of the Atlantic Ridge. That's why I felt a Fl landfall (even Key West and affect West Coast) was likely. The public has been reduced to running endless models with not an inkling about what's really going on. Good post.

I don't think the "public" is watching model runs but rather listening to official sources. Enthusiasts may be but I'm still not sure what point you are trying to make here. The models are automated pen and paper done with equations, it's just been moved to something that can process things faster. Guessing about what a ridge will do or why leads to a bad forecast every time, that's why planes (and balloons) are out making upper air measurements all around Matthews path, so the data being fed into the models is improved.

Romantic notions of old pen and paper days negates just how bad some of the predictions were and the consequences of a bad forecast now that millions and millions more people live in harms way.

In my opinion, of course.

I do see your point. Perhaps I was a little frustrated with the course of model runs and analysis dealing with this storm in particular. The models were projecting a course that could not be justified by mets looking at future weather trends. They were left guessing, even baffled, by the projected weakness in the Atlantic Ridge forecast by the models. Now you have this projected loop off the East Coast, and someone said that it may be a "bad" model run, and that for right now the NHC is not even considering that part of the run. I don't know, this storm has left a bad taste in my mouth as far as models go. I apologize if I've been too negative.

[MaineWeatherNut]

It's only a small tip, but there are some pretty large mountains there. I'd expect weakening at a very decent rate to begin now.

The entire eye will never be over land at once. So it will always have water to feed into it. Some weakening is likely but I don't think it will be at a decent rate...

[SouthDadeFish]

Wow you gotta be on point not to includ dade in hurricane watch wouldn't be prudent Incase wind field expands or jog. Broward 20 min from Miami.

You gotta draw the line somewhere. At this time it doesn't seem as likely that Miami-Dade county will experience hurricane conditions. That can change, however, so remain vigilant.

[Michele B]

Oh, okay. I wasn't clear. To me this was a really bad run and it looks like a malfunction. It (and the other British system UKMET) made what we call an outlier so far off the mark that you discard it, throw it out. What NHC apparently did was use it in blends up until the time when it goes crazy.

Thank God for humans! Still think that mets from 50 years ago, with their pens and charts, would have done better with this storm than our super sensitive-super complex Euro and GFS.

Super sensitive is right. Only concern I do have is the fact that Euro and especially GFS always kind of overestimate troughs and underestimate ridging which means they are always off in their track errors to the right. (East and North)

I had said this days ago. It also goes along with a lot of stuff Joe Bastardi believes, too. That the 'models' are based on data from just a few years ago, for instance, and don't take into account what storms were doing 40, 50 years ago, who "moved like" these. Of course, it isn't like we have a lot of reliable data about the "ridging" and exact positions of things like Atlantic highs, etc.

Again: "Garbage in, garbage out."

[tolakram]

I do see your point. Perhaps I was a little frustrated with the course of model runs and analysis dealing with this storm in particular. The models were projecting a course that could not be justified by mets looking at future weather trends. They were left guessing, even baffled, by the projected weakness in the Atlantic Ridge forecast by the models. Now you have this projected loop off the East Coast, and someone said that it may be a "bad" model run, and that for right now the NHC is not even considering that part of the run. I don't know, this storm has left a bad taste in my mouth as far as models go. I apologize if I've been too negative.

Actually I had a similar rant a few pages back. I'm mad at both the out to sea and go west crowds ... which I think might encompass everyone.

[Exalt]

News stations up here in Virginia are making a big stink out of it yet aren't really telling people to prepare for anything despite almost half the state being within the cone. I'd expect at least a "stock up on food", or a "possible tropical storm/hurricane conditions could be coming our way".

[SapphireSea]

Thank God for humans! Still think that mets from 50 years ago, with their pens and charts, would have done better with this storm than our super sensitive-super complex Euro and GFS.

Super sensitive is right. Only concern I do have is the fact that Euro and especially GFS always kind of overestimate troughs and underestimate ridging which means they are always off in their track errors to the right. (East and North)

I had said this days ago. It also goes along with a lot of stuff Joe Bastardi believes, too. That the 'models' are based on data from just a few years ago, for instance, and don't take into account what storms were doing 40, 50 years ago, who "moved like" these. Of course, it isn't like we have a lot of reliable data about the "ridging" and exact positions of things like Atlantic highs, etc.

Again: "Garbage in, garbage out."

The complaint from JB on the models does not really hold that much water. Dynamical fluid modeling systems barely and generally do not directly use statistical data to plot courses. That's like complaining about the CLP5 performance. Ludicrous.

Think about the scale of what needs to be modeled. You know like a butterfly effect but with poor resolution of air data. The fact that some models can tell you today with a 200 mile error or so (in some cases less) what a small 16nm low center will do is astounding and a great sign of human progress on forecasting earth weather. Do they still error of course and what worries is that generally to the right of the envelope in many cases.

[SeaBrz_FL]

Actually I had a similar rant a few pages back. I'm mad at both the out to sea and go west crowds ... which I think might encompass everyone.

Not wishing this on anyone to my west, but right now at my location, I'm cheering for both teams.

[TheStormExpert]

18Z has a loop from Hatteras back down to Florida???

Nope, not even close.

But is trending to the right after the Carolinas.
You cannot say that the whole 12z Euro model was a bad run because it does that loop after 120h forecast, that range is always questionable. IMO.

What about in terms of intensity?

[Exalt]

I do see your point. Perhaps I was a little frustrated with the course of model runs and analysis dealing with this storm in particular. The models were projecting a course that could not be justified by mets looking at future weather trends. They were left guessing, even baffled, by the projected weakness in the Atlantic Ridge forecast by the models. Now you have this projected loop off the East Coast, and someone said that it may be a "bad" model run, and that for right now the NHC is not even considering that part of the run. I don't know, this storm has left a bad taste in my mouth as far as models go. I apologize if I've been too negative.

Actually I had a similar rant a few pages back. I'm mad at both the out to sea and go west crowds ... which I think might encompass everyone.

I had a conversation with one of our local meteorologists in DC while Matthew was bombing out, and he refused to accept anything other than what was on the models, even though I provided him with recon evidence that it had further strengthened to a Cat 4 at 150 mph, and kept telling me it was a Cat 3 at 120 mph. He also doubted me on the idea that Matthew could make a run for Cat 5 by last Saturday or potentially that Friday, and continued harping on the models saying it would weaken immediately despite the conditions.

This is what is so dangerous about models, is that meteorologists, especially the ones on TV whose jobs it is to advise the public to take caution and get the message out there about potentially catastrophic situations, seem to only lock onto what they say until the NHC spouts something completely contradictory. They don't look at the conditions in the area, which models often get wrong, or sometimes EXTREMELY wrong as with Matthew. Models never took into account the severe drop in shear and high SSTs in their forecasts while Matthew was bombing out, they constantly said he would weaken prior to primary landfall, yet it never happened.

I wish people would just open their eyes and look past the models, and look at the actual conditions within the area, because of situations like these. Thankfully we've gotten enough warning here in the US, but the situation in Haiti/Cuba could've easily been handled better and communication relating to intensity could've been much better if people had just taken a look at conditions in the particular area for the present and the future.

[RL3AO]

What? The UKMET had a very similar solution. It's a weird solution, but please don't call it a "bad run from a system malfunction" when that is not the case.

Oh, okay. I wasn't clear. To me this was a really bad run and it looks like a malfunction. It (and the other British system UKMET) made what we call an outlier so far off the mark that you discard it, throw it out. What NHC apparently did was use it in blends up until the time when it goes crazy.

Thank God for humans! Still think that mets from 50 years ago, with their pens and charts, would have done better with this storm than our super sensitive-super complex Euro and GFS.

Well you're wrong but you can keep believing it. Meanwhile the brilliant mathematicians and meteorologists that have saved billions (trillions?) in property and costs and countless lives creating these "super sensitive" models will keep doing their thing.

[windnrain]

Latest MIMIC

[SeGaBob]

When is recon?

[Exalt]

Latest MIMIC

Wow what a bounce back after that Haiti hit, convection blew up almost immediately.

[ozonepete]

Nope, not even close.

But is trending to the right after the Carolinas.
You cannot say that the whole 12z Euro model was a bad run because it does that loop after 120h forecast, that range is always questionable. IMO.

What about in terms of intensity?

this is devolving into a discussion of inches and minutes, lol. There will be no changes from the NHC based on the 18Z GFS model run. They cannot do much until the major model runs come in at 00Z and the results don't come in until after 11PM. So don't expect major changes from the MHC before late overnight.

[SouthDadeFish]

But is trending to the right after the Carolinas.
You cannot say that the whole 12z Euro model was a bad run because it does that loop after 120h forecast, that range is always questionable. IMO.

What about in terms of intensity?

this is devolving into a discussion of inches and minutes, lol. There will be no changes from the NHC based on the 18Z GFS model run. They cannot do much until the major model runs come in at 00Z and the results don't come in until after 11PM. So don't expect major changes from the MHC before late overnight.

This assumes the HWRF doesn't change much either, which instead looks to be a bit SW of the previous run, so far. While I don't think it will change much, it could slightly shift a bit more left if the HWRF pulls a 12Z GFDL.

[Soonercane]

Super sensitive is right. Only concern I do have is the fact that Euro and especially GFS always kind of overestimate troughs and underestimate ridging which means they are always off in their track errors to the right. (East and North)

I had said this days ago. It also goes along with a lot of stuff Joe Bastardi believes, too. That the 'models' are based on data from just a few years ago, for instance, and don't take into account what storms were doing 40, 50 years ago, who "moved like" these. Of course, it isn't like we have a lot of reliable data about the "ridging" and exact positions of things like Atlantic highs, etc.

Again: "Garbage in, garbage out."

The complaint from JB on the models does not really hold that much water. Dynamical fluid modeling systems barely and generally do not directly use statistical data to plot courses. That's like complaining about the CLP5 performance. Ludicrous.

Think about the scale of what needs to be modeled. You know like a butterfly effect but with poor resolution of air data. The fact that some models can tell you today with a 200 mile error or so (in some cases less) what a small 16nm low center will do is astounding and a great sign of human progress on forecasting earth weather. Do they still error of course and what worries is that generally to the right of the envelope in many cases.

Well said... we are not talking about MOS or SHIPS here, while some empirical tuning is done on things like parameterization schemes dynamical models generally are not "trained" on historical data in the literal sense.

[NC George]

Just thought I would share this little blurb from our local news media.

http://www.wect.com/story/33315748/uncw ... -emergency

I would point out the UNC-Wilmington was going to be on fall break this Thursday -Sunday anyway. Basically they are kicking out the few remaining stragglers, not cancelling class for 2 days. They are probably cancelling Wed night classes to encourage everyone to leave Wed, and not wait until Thursday.

[JaxGator]

People in the Jax and St. Augustine area are taking Matthew seriously. A lot of the roads are busy with folks getting off work getting supplies. Home Depot, Lowes and other stores have full parking lots and are very busy.