Katrina
Gary Gray has invented and sells his own Tropical Cyclone model software that's widely used - been pretty accurate in the past. Sharing the address for his forecasts - they're interesting.
Tropical Weather Discussion
by Gary Gray
--------------------------------------------------------------------------------
...Katrina Reaches Hurricane Strength [8/25 5PM EDT]...
Things are coming together for Katrina, both in terms of her becoming stronger and in terms of her forecast. She developed an eye-like feature very early today, then lost it. But, since mid to late morning, she redeveloped her eye and maintained it. It has been quite ragged most of the time, which is to be expected for a mere tropical storm (which is what she was up until a couple of hours ago). But it has been persistent. Her pressure also dropped quite significantly. From a model perspective, MOST of the models have come somewhat in line, showing a track initially favoring the further west tracks more off the Florida Peninsula west coast towards a landfall between Pensacola and Apalachicola. Post-landfall (second landfall, that is) some large differences remain... like the 12Z GFS hooking Katrina east off the NC coast, while the 12Z ECMWF hauls her north up the Ohio Valley! So, I'm certainly not saying everything's neat and tidy. But, in the overall big picture and in the nearer term for dealing with the landfalls, things are coming together...
Current Conditions: As I write this, early evening Thursday, Katrina is looking quite healthy... not tremendous - she is, after all, barely a hurricane, but healthy. Convection is deep, persistent and has become more symmetric today. As noted above, Katrina has also maintained an eye for many hours now. It has become small and rather filled in at times, but has remained largely identifiable. The latest recon aircraft reports have measured maximum flight level winds of 64kts, corresponding to surface winds of just about 55kts. And surface pressures from recon have been reported at about 990mb. That's about a 10mb drop from yesterday, and it is low enough to support SLIGHTLY higher winds... indicating that winds may had not, at that time, quite spun up yet. That recon report was from a few hours ago, and since then, the recon's stepped frequency microwave radiometer (SFMR) has been used to measure Katrina's winds. They appear to have responded to the pressure falls and Katrina is now a hurricane. As I write this, JUST before the official 5PM EDT update, I don't know what Katrina's "official" intensity is (TPC/NHC did not provide an intensity when they upgraded her). But, given that her intensification has been a pretty steady pace, albeit accelerating somewhat, I'd estimate Katrina to be at about 65kts now... meaning she's at minimal hurricane status and should gain a bit more intensity just prior to landfall. As for her motion, attempting to track her on radar has given indications of a mainly west motion with occasional nudges southward. However, this is likely misleading due to Katrina's lesser quality appearance earlier in the day. In fact, recon fixes show an ever so slightly north of due west motion... primarily due west. The forward speed is at about 5kts.
Model Discussion: In terms of landfall number one, that now seems to be a foregone conclusion, near or a short distance north of Fort Lauderdale late tonight (Thu PM) and the very wee hours Friday AM. So, with the models let's concentrate on the 2nd landfall and, less so (since it's less important but not wholly irrelevant), on the post-landfall track. Here's how the models are stacking up. Let me first hammer on the 12Z GFDL, since it continues to, erroneously, show a significant initial southward motion. In fact, at 17:33Z recon fixed Katrina's eye at 26.3N/79.9W. The 12Z GFDL, in its mere 06hr forecast (by 18Z Thu) had Katrina down to 25.8N/80.0W. That's a massive error in a short term forecast... 30nm, plus it severely alters the motion at 18Z... for example, the motion at 18Z was about 275-degrees at 6kts, while the GFDL projected it to be 251-degrees at 8kts. So, the GFDL is off considerably. But, if you care... it has Katrina cutting across extreme south Florida as a major hurricane, then making a wider look than almost every other model, to a landfall just west of Mobile Monday evening. The 12Z NAM also continues to be suspect, as it quickly loses Katrina in the near term, then grasps back onto her further SW. This positions her such that she'd be more inclined to drive straight west; that's what the NAM does, but it's difficult to accept given its short term confusion.
We can pretty much run through the other models, as they're all more trustworthy in the near term, albeit still varied in the long range. The UKMET has the same issue as the GFDL, which results in a far west landfall. But, we can't lump it in with the poor short-term performers, because the UKMET is nowhere near as severe with the immediate southward motion (the GFDL has Katrina down to 25.5N by 00Z Friday; the UKMET has her to 26.1N by then). So, it may not end up RIGHT, but it's a "legitimate" forecast. So, with that early displacement, it takes Katrina longer to make the bend around. The result is a landfall identical to the GFDL... Monday night just west of Mobile. After landfall the UKMET flips Katrina NEward; by D+6 (Wed AM) it has her in SW VA/SE WV. The 12Z ECMWF has also shifted west, but not as far. It shows Katrina coming ashore near Pensacola late Mon afternoon or very early Mon evening. Thereafter, though, as compared to the UKMET, it has a more northward component and pulls west of the UKMET, to NW WV/SE OH by Wed AM. The EC shows healthy intensification for Katrina prior to landfall. The 12Z NOGAPS shows less change than the other models, though it was already on the west side of the guidance, so while some other models shifted west, it just needed to hold the line. It shows Katrina making landfall Monday morning near Fort Walton Beach, FL. With a flatter Atlantic ridge than in past runs, by D+6 (Wed AM) the NOGAPS is well east of the EC/UK, putting Katrina just offshore of Cape May, NJ. The 12Z GFS maintains its more eastern solution. It shows Katrina making landfall in the Big Bend of Florida, a short distance east of Apalachicola early Monday. Then, with a MUCH flatter Atlantic ridge, shoves her hard to the east once she reaches North Carolina; by D+6 the 12Z GFS has her just SE of Cape Hatteras. The 12Z GGEM agrees with the NOGAPS/GFS flat Atlantic ridge and splits the difference on the specifics... It brings Katrina ashore near Panama City, FL midday Monday, then rides her up inland to near Norfolk by D+6 (Wed AM).
My thinking: First off, I don't want to ignore the first landfall, but it is pretty much a done deal, so I'm not going to dwell on it. Simply put, folks in southeastern Florida obviously should have completed their preparations for a full-fledged hurricane. And as a mere Category 1 hurricane, assuming Katrina doesn't rapidly intensify to a Category 2 by landfall, if you're properly prepared and don't take foolish risks, you should be able to ride out Katrina with only the typical problems (power outages, fresh water flooding in the prone areas due to heavy rains, etc). I specifically expect landfall near Fort Lauderdale in the overnight hours with Katrina as a Category One hurricane, at about 70kts, perhaps 75kts.
The real questions arise with the second landfall. The models have converged a bit, but with at least one "reasonable" solution out past Mobile (the UKMET), that actually tries to increase the spread. So, we've not converged a whole lot. Plus we see a discrepancy in the overall way in which Katrina gets caught in the mid-latitude system. The ECMWF, UKMET and *apparently* the GFDL (the GFDL only goes to 126hrs and it's not 100% clear by then) tend toward more of a merging with the main mid-latitude trough, sending Katrina more northnortheastward up the east side of the Mississippi/Ohio Valleys. The opposing trio, the GFS, NOGAPS and GGEM, keep her separate, throwing her out ahead of the trough, with a track more to the NE, and even eventually ENE and E on the GFS. Of course, the complexity of those bends is largely after landfall (though the curve begins just before landfall, so it has some impact there as well, but it's minimal). What's more important is what happens before that. Well, I'm opting to shoot down the middle. That is not only the "path of least regret" as it's likely to be "close" since it's a consensus forecast, but there are also physical reasons why I prefer the "down the middle" solution. For one thing, I'm very doubtful of any significant southward component of motion. It's certainly POSSIBLE, as there is some upper level flow that would favor this. But, it's weak and retreating flow and we've not seen such a motion develop yet. So, I'm skeptical. Such a motion would allow Katrina to get further west, like the UKMET and GFDL show. I'm not buying it, so I'm holding further east. On the flip side, we've seen a pretty solid westward shift in the models and there is, in fact, going to be some ridging in place, albeit extremely weak, to the north of Katrina for a good solid three days. So, I'm highly doubtful of the consistently east GFS/GGEM solutions. I'm favoring something along the lines of the NOGAPS and ECMWF... with landfall Monday somewhere between Pensacola and Panama City.
As for intensity at landfall... I'd anticipate slow, steady strengthening once Katrina is over the Gulf. With very warm water temps, her intensification could occur at a pretty good rate. However, upper level conditions are too chaotic, in my opinion, to support any sustained, rapid intensification. So, Katrina will likely emerge into the Gulf from Florida very late Friday as a weakened system, down to about 40kts or so. She'll have about two and a half days to intensify. She should manage about 15kts per day... ROUGHLY. That'll put her at about 75kts by landfall... a Category 1 hurricane. Category 2 strength is certainly within reach (even Category 3 isn't impossible if she can defy my expectations and undergo a rapid intensification cycle or two), but it's not my forecast right now.
The post-landfall track/intensity I think will also be something of a NOGAPS/ECMWF solution. Those two diverge after landfall, actually. But, my point is, I think there'll be more absorption aloft than the NOGAPS shows (meaning a further west track embedded in the main trough, rather than running out ahead). But I also think that Katrina will remain strong enough to hold some of her own identity, even aloft, thereby keeping her from simply melting into the trough and, therefore, running a little east of the ECMWF. Going in between the EC and NOGAPS would, arguably, leave us close to the UKMET (almost up the Appalachians). But the UK's westward bias early (it's a "bias" in my opinion, since I'm not siding with its significant southward component early on) means that it's not an ideal compromise, as it's still a bit too far west. Basically, I suspect a track up the Piedmont may be best. But all of this is somewhat "fluff" (not entirely, it's not inconsequential, as some flooding rains could come out of it); the actual hurricane impact/landfall is not only more critical, but a day or two more near-term than the post-landfall track. So, reiterating that... My expectation right now is for Katrina to make landfall Monday in the approximate vicinity of Fort Walton Beach, FL as about a 70kt hurricane. I use the "approximate" and "about" qualifiers because, yes, some question remains. The entire northeastern Gulf of Mexico should be paying close attention to the forecast for Katrina once she comes across the Peninsula on Friday.
Gary Gray has invented and sells his own Tropical Cyclone model software that's widely used - been pretty accurate in the past. Sharing the address for his forecasts - they're interesting.
Tropical Weather Discussion
by Gary Gray
--------------------------------------------------------------------------------
...Katrina Reaches Hurricane Strength [8/25 5PM EDT]...
Things are coming together for Katrina, both in terms of her becoming stronger and in terms of her forecast. She developed an eye-like feature very early today, then lost it. But, since mid to late morning, she redeveloped her eye and maintained it. It has been quite ragged most of the time, which is to be expected for a mere tropical storm (which is what she was up until a couple of hours ago). But it has been persistent. Her pressure also dropped quite significantly. From a model perspective, MOST of the models have come somewhat in line, showing a track initially favoring the further west tracks more off the Florida Peninsula west coast towards a landfall between Pensacola and Apalachicola. Post-landfall (second landfall, that is) some large differences remain... like the 12Z GFS hooking Katrina east off the NC coast, while the 12Z ECMWF hauls her north up the Ohio Valley! So, I'm certainly not saying everything's neat and tidy. But, in the overall big picture and in the nearer term for dealing with the landfalls, things are coming together...
Current Conditions: As I write this, early evening Thursday, Katrina is looking quite healthy... not tremendous - she is, after all, barely a hurricane, but healthy. Convection is deep, persistent and has become more symmetric today. As noted above, Katrina has also maintained an eye for many hours now. It has become small and rather filled in at times, but has remained largely identifiable. The latest recon aircraft reports have measured maximum flight level winds of 64kts, corresponding to surface winds of just about 55kts. And surface pressures from recon have been reported at about 990mb. That's about a 10mb drop from yesterday, and it is low enough to support SLIGHTLY higher winds... indicating that winds may had not, at that time, quite spun up yet. That recon report was from a few hours ago, and since then, the recon's stepped frequency microwave radiometer (SFMR) has been used to measure Katrina's winds. They appear to have responded to the pressure falls and Katrina is now a hurricane. As I write this, JUST before the official 5PM EDT update, I don't know what Katrina's "official" intensity is (TPC/NHC did not provide an intensity when they upgraded her). But, given that her intensification has been a pretty steady pace, albeit accelerating somewhat, I'd estimate Katrina to be at about 65kts now... meaning she's at minimal hurricane status and should gain a bit more intensity just prior to landfall. As for her motion, attempting to track her on radar has given indications of a mainly west motion with occasional nudges southward. However, this is likely misleading due to Katrina's lesser quality appearance earlier in the day. In fact, recon fixes show an ever so slightly north of due west motion... primarily due west. The forward speed is at about 5kts.
Model Discussion: In terms of landfall number one, that now seems to be a foregone conclusion, near or a short distance north of Fort Lauderdale late tonight (Thu PM) and the very wee hours Friday AM. So, with the models let's concentrate on the 2nd landfall and, less so (since it's less important but not wholly irrelevant), on the post-landfall track. Here's how the models are stacking up. Let me first hammer on the 12Z GFDL, since it continues to, erroneously, show a significant initial southward motion. In fact, at 17:33Z recon fixed Katrina's eye at 26.3N/79.9W. The 12Z GFDL, in its mere 06hr forecast (by 18Z Thu) had Katrina down to 25.8N/80.0W. That's a massive error in a short term forecast... 30nm, plus it severely alters the motion at 18Z... for example, the motion at 18Z was about 275-degrees at 6kts, while the GFDL projected it to be 251-degrees at 8kts. So, the GFDL is off considerably. But, if you care... it has Katrina cutting across extreme south Florida as a major hurricane, then making a wider look than almost every other model, to a landfall just west of Mobile Monday evening. The 12Z NAM also continues to be suspect, as it quickly loses Katrina in the near term, then grasps back onto her further SW. This positions her such that she'd be more inclined to drive straight west; that's what the NAM does, but it's difficult to accept given its short term confusion.
We can pretty much run through the other models, as they're all more trustworthy in the near term, albeit still varied in the long range. The UKMET has the same issue as the GFDL, which results in a far west landfall. But, we can't lump it in with the poor short-term performers, because the UKMET is nowhere near as severe with the immediate southward motion (the GFDL has Katrina down to 25.5N by 00Z Friday; the UKMET has her to 26.1N by then). So, it may not end up RIGHT, but it's a "legitimate" forecast. So, with that early displacement, it takes Katrina longer to make the bend around. The result is a landfall identical to the GFDL... Monday night just west of Mobile. After landfall the UKMET flips Katrina NEward; by D+6 (Wed AM) it has her in SW VA/SE WV. The 12Z ECMWF has also shifted west, but not as far. It shows Katrina coming ashore near Pensacola late Mon afternoon or very early Mon evening. Thereafter, though, as compared to the UKMET, it has a more northward component and pulls west of the UKMET, to NW WV/SE OH by Wed AM. The EC shows healthy intensification for Katrina prior to landfall. The 12Z NOGAPS shows less change than the other models, though it was already on the west side of the guidance, so while some other models shifted west, it just needed to hold the line. It shows Katrina making landfall Monday morning near Fort Walton Beach, FL. With a flatter Atlantic ridge than in past runs, by D+6 (Wed AM) the NOGAPS is well east of the EC/UK, putting Katrina just offshore of Cape May, NJ. The 12Z GFS maintains its more eastern solution. It shows Katrina making landfall in the Big Bend of Florida, a short distance east of Apalachicola early Monday. Then, with a MUCH flatter Atlantic ridge, shoves her hard to the east once she reaches North Carolina; by D+6 the 12Z GFS has her just SE of Cape Hatteras. The 12Z GGEM agrees with the NOGAPS/GFS flat Atlantic ridge and splits the difference on the specifics... It brings Katrina ashore near Panama City, FL midday Monday, then rides her up inland to near Norfolk by D+6 (Wed AM).
My thinking: First off, I don't want to ignore the first landfall, but it is pretty much a done deal, so I'm not going to dwell on it. Simply put, folks in southeastern Florida obviously should have completed their preparations for a full-fledged hurricane. And as a mere Category 1 hurricane, assuming Katrina doesn't rapidly intensify to a Category 2 by landfall, if you're properly prepared and don't take foolish risks, you should be able to ride out Katrina with only the typical problems (power outages, fresh water flooding in the prone areas due to heavy rains, etc). I specifically expect landfall near Fort Lauderdale in the overnight hours with Katrina as a Category One hurricane, at about 70kts, perhaps 75kts.
The real questions arise with the second landfall. The models have converged a bit, but with at least one "reasonable" solution out past Mobile (the UKMET), that actually tries to increase the spread. So, we've not converged a whole lot. Plus we see a discrepancy in the overall way in which Katrina gets caught in the mid-latitude system. The ECMWF, UKMET and *apparently* the GFDL (the GFDL only goes to 126hrs and it's not 100% clear by then) tend toward more of a merging with the main mid-latitude trough, sending Katrina more northnortheastward up the east side of the Mississippi/Ohio Valleys. The opposing trio, the GFS, NOGAPS and GGEM, keep her separate, throwing her out ahead of the trough, with a track more to the NE, and even eventually ENE and E on the GFS. Of course, the complexity of those bends is largely after landfall (though the curve begins just before landfall, so it has some impact there as well, but it's minimal). What's more important is what happens before that. Well, I'm opting to shoot down the middle. That is not only the "path of least regret" as it's likely to be "close" since it's a consensus forecast, but there are also physical reasons why I prefer the "down the middle" solution. For one thing, I'm very doubtful of any significant southward component of motion. It's certainly POSSIBLE, as there is some upper level flow that would favor this. But, it's weak and retreating flow and we've not seen such a motion develop yet. So, I'm skeptical. Such a motion would allow Katrina to get further west, like the UKMET and GFDL show. I'm not buying it, so I'm holding further east. On the flip side, we've seen a pretty solid westward shift in the models and there is, in fact, going to be some ridging in place, albeit extremely weak, to the north of Katrina for a good solid three days. So, I'm highly doubtful of the consistently east GFS/GGEM solutions. I'm favoring something along the lines of the NOGAPS and ECMWF... with landfall Monday somewhere between Pensacola and Panama City.
As for intensity at landfall... I'd anticipate slow, steady strengthening once Katrina is over the Gulf. With very warm water temps, her intensification could occur at a pretty good rate. However, upper level conditions are too chaotic, in my opinion, to support any sustained, rapid intensification. So, Katrina will likely emerge into the Gulf from Florida very late Friday as a weakened system, down to about 40kts or so. She'll have about two and a half days to intensify. She should manage about 15kts per day... ROUGHLY. That'll put her at about 75kts by landfall... a Category 1 hurricane. Category 2 strength is certainly within reach (even Category 3 isn't impossible if she can defy my expectations and undergo a rapid intensification cycle or two), but it's not my forecast right now.
The post-landfall track/intensity I think will also be something of a NOGAPS/ECMWF solution. Those two diverge after landfall, actually. But, my point is, I think there'll be more absorption aloft than the NOGAPS shows (meaning a further west track embedded in the main trough, rather than running out ahead). But I also think that Katrina will remain strong enough to hold some of her own identity, even aloft, thereby keeping her from simply melting into the trough and, therefore, running a little east of the ECMWF. Going in between the EC and NOGAPS would, arguably, leave us close to the UKMET (almost up the Appalachians). But the UK's westward bias early (it's a "bias" in my opinion, since I'm not siding with its significant southward component early on) means that it's not an ideal compromise, as it's still a bit too far west. Basically, I suspect a track up the Piedmont may be best. But all of this is somewhat "fluff" (not entirely, it's not inconsequential, as some flooding rains could come out of it); the actual hurricane impact/landfall is not only more critical, but a day or two more near-term than the post-landfall track. So, reiterating that... My expectation right now is for Katrina to make landfall Monday in the approximate vicinity of Fort Walton Beach, FL as about a 70kt hurricane. I use the "approximate" and "about" qualifiers because, yes, some question remains. The entire northeastern Gulf of Mexico should be paying close attention to the forecast for Katrina once she comes across the Peninsula on Friday.
