Tropical Depression Ernesto - 35 mph - 1004 mb - W 14 mph

[dtk]

ALMOST always is the correct answer!

You are pretty much making a fool of yourself with statements like this.

[LkHopatGuy]

Why is there a fight going on here in this thread. How are the weenies like myself supposed to understand or learn anything when this is bordering on Banter? Please stop the fighting!

[HurricaneJosh]

Why is there a fight going on here in this thread. How are the weenies like myself supposed to understand or learn anything when this is bordering on Banter? Please stop the fighting!

Actually, the debate is interesting. Read it and learn about the models.

[dtk]

I think he was questioning how the forecasts of the empirical parameters in the new GFS have changed the SHIPS (if at all).

Honestly, we don't have a single side-by-side comparison of before/after the DA upgrade since we (NCEP/EMC) don't run the statistical guidance ourselves.

However, given the substantial improvements we see in the general tropical forecasts (particularly wind forecasts), I would say that all things being equal the current statistical guidance (based on GFS forecasts) > guidance before the upgrade. This would be consistent with other findings, but frankly I have no real scientically sound basis to make this argument. It's also entirely possibly that parameters/equations were changed in the statistical guidance confusing things even further.

[am19psu]

Honestly, we don't have a single side-by-side comparison of before/after the DA upgrade since we (NCEP/EMC) don't run the statistical guidance ourselves.

However, given the substantial improvements we see in the general tropical forecasts (particularly wind forecasts), I would say that all things being equal the current statistical guidance (based on GFS forecasts) > guidance before the upgrade. This would be consistent with other findings, but frankly I have no real scientically sound basis to make this argument. It's also entirely possibly that parameters/equations were changed in the statistical guidance confusing things even further.

Since it's just MLR, if the previous iteration of the GFS was biased and the new hybrid EnKF is unbiased (or less biased) in the parameters, it could bias the new forecasts. But that is just a general statistical observation, I have no clue how the SHIPS is performing based on the new GFS

[CUmet]

Does the SHIPS guidance use mid-lvl shear as a predictor? I feel as if those situations are when globals most strongly outperform the statistical intensity guidance.

Kind of, it's the "shear adj" parameter in the SHIPS file. I believe it represents the "generalized shear" that's mentioned in Mark Demaria's conference presentation in 2010 (extended abstract below). It's basically a vertical integral of shear, such that it includes information from the whole layer instead of just 850 and 200. To quote from the abstract: "Positive values of the GS deviation indicate that environmental wind profile has a nonlinear structure so that levels besides 850 and 200 hPa are contributing to the shear."

https://ams.confex.c...pers/167916.pdf

[wxsmwhrms]

Agreed. The modeling for Dean more than a week out was spectacular. In Oz, Cyclone Yasi was similarly well-modeled by the Euro way, way in advance.

Sorry, y'all are right, it was Felix I was thinking of.

Why is there a fight going on here in this thread. How are the weenies like myself supposed to understand or learn anything when this is bordering on Banter? Please stop the fighting!

Uh, this IS how you learn. Would you prefer that we allow people to post things that we believe are not correct, and let you believe them?

[bluewave]

How so? If anything, we have some evidence that the DA upgrade in May of this year yielded (marginally) improved intensity forecasts. This is likely a secondary impact through the improved initialization of the storm environment (and not necessary the storm itself), which leads to improved track guidance, which then subsequently leads to improved intensity guidance (they are not independent, obviously).

Thanks for the answer. I was thinking more about the rainfall issues that you said that they were working on a fix for

after the upgrade. Could that have an impact on the tropical cyclone forecasts?

[David Reimer]

You know a tropical storm is lame when the mets start degenerating into conversation and discussion on the technical aspects of a weather model.

[dtk]

Since it's just MLR, if the previous iteration of the GFS was biased and the new hybrid EnKF is unbiased (or less biased) in the parameters, it could bias the new forecasts. But that is just a general statistical observation, I have no clue how the SHIPS is performing based on the new GFS

Fair point, and SHIPS does use some form of multiple regression (and I clearly have no clue how the statistical guidance based on the post-ugrade GFS is doing either). Though, since the upgrade was almost exclusively DA (the only model changes were related to a code restructuring, non-science), the impact on the systematic (bias) errors should have been small-ish. For example, the MDL folks didn't have to do a retraining of the MOS equations with the upgrade (and were still able to benefit from the improved forecasts......ignoring the model problems that have been documented this summer with the too cool/moist during the heat waves over Central CONUS, 100% unrelated to the May 2012 upgrade).

[mappy]

You know a tropical storm is lame when the mets start degenerating into conversation and discussion on the technical aspects of a weather model.

Poor Joshypoo

[am19psu]

Fair point, and SHIPS does use some form of multiple regression (and I clearly have no clue how the statistical guidance based on the post-ugrade GFS is doing either). Though, since the upgrade was almost exclusively DA (the only model changes were related to a code restructuring, non-science), the impact on the systematic (bias) errors should have been small-ish. For example, the MDL folks didn't have to do a retraining of the MOS equations with the upgrade (and were still able to benefit from the improved forecasts......ignoring the model problems that have been documented this summer with the too cool/moist during the heat waves over Central CONUS, 100% unrelated to the May 2012 upgrade).

That;s a good point. If MOS didn't need to be adjusted, SHIPS is probably still reasonable as well.

[dtk]

Thanks for the answer. I was thinking more about the rainfall issues that you said that they were working on a fix for

after the upgrade. Could that have an impact on the tropical cyclone forecasts?

No, because it turns out that particular issue is predominatly an "over-land" issue. We've isolated some of the problem to a change that was made to the land model (in 2010, I think)....which results in a big increase in soil moisture (which is clearly not an issue over the open ocean). The only impact would be for landfalling storms, where the model may not degenerate things quickly enough....

[bluewave]

No, because it turns out that particular issue is predominatly an "over-land" issue. We've isolated some of the problem to a change that was made to the land model (in 2010, I think)....which results in a big increase in soil moisture (which is clearly not an issue over the open ocean). The only impact would be for landfalling storms, where the model may not degenerate things quickly enough....

Thanks. That's great to know.

[CUmet]

To be honest, Ernesto looks better right now than it ever has IMO (although that might not be saying much). The low-level circulation looks better defined, convection continues to pop up near to the center (although restricted downshear at outer radii), and the upper-level outflow has expanded on the northern side.

Often times in strongly sheared TCs, you'll get a huge convective burst that results in temporary intensification, but the downdrafts from that burst flood the inner core with cool, dry air that shuts everything off, resulting in a naked swirl. The convective trends in Ernesto over the past 6 hours or so seem contrary to this, as the convection appears more "bubbly" or "popcorny", and this is consistent with a reduction of vertical wind shear. The lack of global model support for intensification still gives me a lot of pause though.

[HurricaneJosh]

To be honest, Ernesto looks better right now than it ever has IMO (although that might not be saying much). The low-level circulation looks better defined, convection continues to pop up near to the center (although restricted downshear at outer radii), and the upper-level outflow has expanded on the northern side.

Often times in strongly sheared TCs, you'll get a huge convective burst that results in temporary intensification, but the downdrafts from that burst flood the inner core with cool, dry air that shuts everything off, resulting in a naked swirl. The convective trends in Ernesto over the past 6 hours or so seem contrary to this, as the convection appears more "bubbly" or "popcorny", and this is consistent with a reduction of vertical wind shear. The lack of global model support for intensification still gives me a lot of pause though.

Nice to see one of the mets deciding to get off the Hate Ernesto Express Train.

[andyhb]

To be honest, Ernesto looks better right now than it ever has IMO (although that might not be saying much). The low-level circulation looks better defined, convection continues to pop up near to the center (although restricted downshear at outer radii), and the upper-level outflow has expanded on the northern side.

Often times in strongly sheared TCs, you'll get a huge convective burst that results in temporary intensification, but the downdrafts from that burst flood the inner core with cool, dry air that shuts everything off, resulting in a naked swirl. The convective trends in Ernesto over the past 6 hours or so seem contrary to this, as the convection appears more "bubbly" or "popcorny", and this is consistent with a reduction of vertical wind shear. The lack of global model support for intensification still gives me a lot of pause though.

Finally some positivity for Josh.

[Cary]

I will outforecast you any day, time, and place. Name the event and I'll do it, it won't even be close!

One of the funniest/saddest things I have ever read on here.

[gulfcane]

"unfavorable conditions ahead - will ernesto survive"

huh?

[Srain]

HPC Final Afternoon Update:

...ERNESTO...

NHC FORECASTS A HURRICANE ON DY 5 NEARING THE YUCATAN. ENSEMBLE

GUIDANCE IS REASONABLY WELL CLUSTERED...ALLOWING US TO MOVE THE

CENTER OF CIRCULATION SWIFTLY NORTHWESTWARD INTO THE GULF OF

MEXICO ON DY 6/7. STAY TUNED TO THE LATEST ADVISORIES FROM NHC

CONCERNING ERNESTO.

[PSUBlizzicane2007]

HPC Final Afternoon Update:

...ERNESTO...

NHC FORECASTS A HURRICANE ON DY 5 NEARING THE YUCATAN. ENSEMBLE

GUIDANCE IS REASONABLY WELL CLUSTERED...ALLOWING US TO MOVE THE

CENTER OF CIRCULATION SWIFTLY NORTHWESTWARD INTO THE GULF OF

MEXICO ON DY 6/7. STAY TUNED TO THE LATEST ADVISORIES FROM NHC

CONCERNING ERNESTO.

Hmmm.

[boulderrr]

TROPICAL STORM ERNESTO DISCUSSION NUMBER 9

NWS NATIONAL HURRICANE CENTER MIAMI FL AL052012

500 PM AST FRI AUG 03 2012

THE OVERALL STRUCTURE OF ERNESTO HAS BECOME A LITTLE BETTER

ORGANIZED DURING THE PAST SEVERAL HOURS...WITH AN INCREASE IN

BANDING FEATURES AND CENTRAL CONVECTION. HOWEVER...WESTERLY SHEAR

IS STILL APPARENT WITH MOST OF THE SHOWER AND THUNDERSTORM ACTIVITY

EAST OF THE CENTER. THE INITIAL INTENSITY IS HELD AT 45 KT PENDING

THE ARRIVAL OF AN AIR FORCE RESERVE PLANE THIS EVENING. WESTERLY

SHEAR IS FORECAST TO CONTINUE FOR THE NEXT COUPLE OF DAYS...WHICH

COULD IMPEDE STRENGTHENING. THERE REMAINS A LARGE DIVERGENCE IN

THE INTENSITY GUIDANCE...WITH THE GFS AND ECMWF BASICALLY

DISSIPATING THE SYSTEM AND THE STATISTICAL MODELS SHOWING ERNESTO

BECOMING A HURRICANE IN A COUPLE DAYS. THE NHC INTENSITY FORECAST

REMAINS NEAR THE INTENSITY CONSENSUS...AND IS CLOSE TO THE PREVIOUS

PREDICTION. NEEDLESS TO SAY...THIS IS AN UNCERTAIN FORECAST.

ERNESTO CONTINUES TO MOVE TOWARD THE WEST OR 275/18. THERE IS NO

CHANGE TO THE SYNOPTIC REASONING FROM THE PREVIOUS FORECAST AS THE

CYCLONE REMAINS EMBEDDED WITHIN A FAST EASTERLY FLOW SOUTH OF THE

SUBTROPICAL RIDGE. THIS STEERING PATTERN SHOULD KEEP THE CYCLONE ON

A GENERAL WESTWARD TRACK FOR THE NEXT COUPLE OF DAYS. AFTER

THAT...THE RIDGE WILL PROBABLY WEAKEN SOMEWHAT...CAUSING ERNESTO TO

SLOW DOWN AND GAIN A LITTLE LATITUDE. HOWEVER...TRACK MODELS HAVE

A LARGE SPREAD IN THE LONG TERM WITH LARGE DIFFERENCES IN THE

STRENGTH OF THE RIDGE. THE OFFICIAL FORECAST IS NUDGED TO THE

WEST...AND LIES ON THE SOUTH SIDE OF THE MODEL CONSENSUS. ALTHOUGH

THE NHC FORECAST IS WELL TO THE NORTH OF THE NORMALLY RELIABLE GFS

AND ECMWF MODELS...THESE MODELS ARE SHOWING A RATHER WEAK SYSTEM IN

A FEW DAYS...AND ARE NOT THOUGHT TO BE THE MOST REPRESENTATIVE IN

THIS CASE.

FORECAST POSITIONS AND MAX WINDS

INIT 03/2100Z 13.9N 64.1W 45 KT 50 MPH

12H 04/0600Z 14.2N 66.7W 45 KT 50 MPH

24H 04/1800Z 14.7N 69.9W 50 KT 60 MPH

36H 05/0600Z 15.3N 73.2W 50 KT 60 MPH

48H 05/1800Z 15.9N 76.5W 55 KT 65 MPH

72H 06/1800Z 17.0N 81.0W 65 KT 75 MPH

96H 07/1800Z 19.0N 84.5W 75 KT 85 MPH

120H 08/1800Z 21.0N 88.0W 65 KT 75 MPH...INLAND

$$

FORECASTER BLAKE

[Analog96]

HPC Final Afternoon Update:

...ERNESTO...

NHC FORECASTS A HURRICANE ON DY 5 NEARING THE YUCATAN. ENSEMBLE

GUIDANCE IS REASONABLY WELL CLUSTERED...ALLOWING US TO MOVE THE

CENTER OF CIRCULATION SWIFTLY NORTHWESTWARD INTO THE GULF OF

MEXICO ON DY 6/7. STAY TUNED TO THE LATEST ADVISORIES FROM NHC

CONCERNING ERNESTO.

Guidance is in pretty good agreement in that period.

[phil882]

To be honest, Ernesto looks better right now than it ever has IMO (although that might not be saying much). The low-level circulation looks better defined, convection continues to pop up near to the center (although restricted downshear at outer radii), and the upper-level outflow has expanded on the northern side.

Often times in strongly sheared TCs, you'll get a huge convective burst that results in temporary intensification, but the downdrafts from that burst flood the inner core with cool, dry air that shuts everything off, resulting in a naked swirl. The convective trends in Ernesto over the past 6 hours or so seem contrary to this, as the convection appears more "bubbly" or "popcorny", and this is consistent with a reduction of vertical wind shear. The lack of global model support for intensification still gives me a lot of pause though.

Good observation... this is certainly not a typical TC shear pattern where there is a huge convective burst on the down shear side of the tropical cyclone. The global models are what give me pause as well, although a more thorough investigation shows some interesting things. Lets try to break down whats going on in each model.

GFS:

The GFS is rather strange, as the system starts off vertically coupled (500 hPa vort max on top of the 925 hPa center). However, for some reason this vort max slides southward as a new vort max develops northward. Traditionally, deep stratiform convection results in cyclonic vorticity maxima in the middle levels of the atmosphere through diabetic heating. Thus the GFS seems to indicate the convection will have difficulty staying over the center in order to maintain vertical coherence. A traditionally westerly shear pattern results in a eastward displaced 500 hPa vorticity maxima due to convection forming on the downshear side of a low-level vortex. However, the GFS seems to suggest that the 500 hPa vorticity maxima over the system will shift southward (perhaps a result of northerly shear) while a new vorticity maxima develops east and merges with this feature by 48 hours. Meanwhile once the 925 hPa moves out from under the 500 hPa vorticity maxima, it races westward and the system becomes completely decoupled from the mid-level circulation. Thats why we see the GFS weaken Ernesto in the short term. The key apparently is this northerly shear that originally causes the vortex to become tilted which allows the low-level flow to capture the vortex more fully.

I'll work on the ECMWF next!

[lakeeffectkid383]

Looks pretty nice considering.

http://www.ssd.noaa.gov/PS/TROP/floaters/05L/imagery/rgb0-lalo.jpg

[CUmet]

Good observation... this is certainly not a typical TC shear pattern where there is a huge convective burst on the down shear side of the tropical cyclone. The global models are what give me pause as well, although a more thorough investigation shows some interesting things. Lets try to break down whats going on in each model.

GFS:

The GFS is rather strange, as the system starts off vertically coupled (500 hPa vort max on top of the 925 hPa center). However, for some reason this vort max slides southward as a new vort max develops northward. Traditionally, deep stratiform convection results in cyclonic vorticity maxima in the middle levels of the atmosphere through diabetic heating. Thus the GFS seems to indicate the convection will have difficulty staying over the center in order to maintain vertical coherence. A traditionally westerly shear pattern results in a eastward displaced 500 hPa vorticity maxima due to convection forming on the downshear side of a low-level vortex. However, the GFS seems to suggest that the 500 hPa vorticity maxima over the system will shift southward (perhaps a result of northerly shear) while a new vorticity maxima develops east and merges with this feature by 48 hours. Meanwhile once the 925 hPa moves out from under the 500 hPa vorticity maxima, it races westward and the system becomes completely decoupled from the mid-level circulation. Thats why we see the GFS weaken Ernesto in the short term. The key apparently is this northerly shear that originally causes the vortex to become tilted which allows the low-level flow to capture the vortex more fully.

I'll work on the ECMWF next!

One thought I just had was that perhaps the TC 500 mb vort max and the developing 500 mb vort max in the model are corotating to some extent, such that the TC 500 mb vort max gets dragged back artificially while the LLC outruns it.

[Amped]

Yeah the sytem looks better than I thought. However I am still keeping my blind faith in the most pessimistic model.

[phil882]

One thought I just had was that perhaps the TC 500 mb vort max and the developing 500 mb vort max in the model are corotating to some extent, such that the TC 500 mb vort max gets dragged back artificially while the LLC outruns it.

That's another possible theory, and you can see that to some extent in the animation. In that hypothetical scenario, for the GFS forecast to be correct, there has to be a significant blow up on convection on the east side of Ernesto over the next 18-24 hours. Given the way the convection is co-locating over the center currently, I'm skeptical of that solution.

To lend some argument towards the northerly shear theory, there is some weak northerly that was observed in Puerto Rico and Hispaniola on the 12z soundings. Will be interesting to see if this has persisted in the 00z soundings. I know this might be under-representative for the Caribbean, but just trying to grasp a physical reason why the GFS is depicting the vorticity the way it does. The GFS does show some northerly flow in the 12-18 hour periods, but beyond that it actually develops a nice 200 hPa anticyclone over the 500 hPa vorticity. The other possible solution of course is that the GFS is poorly handling the convective features and is just wrong. (convective feedback issues were a much bigger issue prior to the 2010 model physics modification).

[CUmet]

The latest satellite scans are very impressive, with convective towers blowing up close to the center, and more importantly in all quadrants relative to the center. It'll be interesting to see what happens tomorrow as the strengthening ridge to the north should accelerate Ernesto, but right now the convective organization is inconsistent with strongly sheared TC/tilted vortex behavior.

[phil882]

The latest satellite scans are very impressive, with convective towers blowing up close to the center, and more importantly in all quadrants relative to the center. It'll be interesting to see what happens tomorrow as the strengthening ridge to the north should accelerate Ernesto, but right now the convective organization is inconsistent with strongly sheared TC/tilted vortex behavior.

No kidding... -80C tops going up over the center currently. Recon is in the storm currently, and looks like they missed the center on the first pass... or the center is less defined??