| Tropical Update |
[Sep. 30th, 2010|12:21 pm] |
Tropical Depression 16 crossed Cuba as expected and aimed at South Florida became Tropical Storm Nicole as was barely possible. Nicole lasted for all of 6 hours. Unfortunately, once Nicole was no longer a tropical storm, my detailed data was cut! This is really rather aggravating, because it should have lasted awhile longer while Nicole was still a tropical depression but instead the weather service downgraded it immediately. I'm not sure why they have been so reluctant to consider Nicole a depression during times when it clearly had a closed circulation. In any case, that means I can't run a storm model and have to drop back to the much less reliable general weather model and look at some of the official models for comparison, primarily NAM and GFS.
Based on the most likely path up the east coast, which is now first over water and then over land, rather than following the coastline much more closely, then reintensification up to hurricane status is very unlikely. Nicole is certainly pursuing a faster course up the east coast than I predicted two days ago.
Jamaica got pounded by Nicole, but Florida seems to have gotten off lightly. The remnants of Nicole south of Cuba still look like they will redevelop. The new low east of Florida on Friday still looks to be a viable prediction. It won't approach shore any closer and won't have a chance to develop before heading into waters north of the US on Oct. 3.
This is about the same time as the new low should be crossing Cuba aimed East of Florida. But I show it dawdling, not developing much and on Oct 5, due east of the south tip of Florida, turning and heading back south recrossing Cuba the next day. It then makes several more tries at developing in fantasy territory before dieing out. In short, it's not showing as much of a Gulf refire this run, but leaves lots of possibilities open with a low loose in the Gulf. Also an independent low forming east in the Caribbean should head north east and in fantasy territory loop back in the direction of the Gulf.
I'm showing nothing significant coming off of Africa until way into fantasy territory on Oct. 11.
The weather service is showing a new invest, Invest 97, moving into the Caribbean. Because it's an invest, I have detailed data to run a storm model. The national hurricane system is giving it a 30% chance of developing in the next 5 days, but I'm showing far less than that. My storm model shows it following a bit south of the BAMS model which is the only reliable official model steering it into the area south of Cuba. If Invest 97 speeds up a bit, it could be part of the Gulf refire, and if it slows down a lot it could be part of the system that forms in the Caribbean to the east of that system on its way back south.
But bottom line is, I'm not showing Invest 97 develop in the next 5 days and certainly not in the next three. Having said that I should note that while my Storm model does an excellent job at predicting path and the general weather model does an excellent job at showing a general picture of things that occur (and don't occur) some time in advance, none of my models are very good with intensity. Most of the reliable models do a much better job at intensity prediction than I do, and the weather service does a better job still. Both the SHIPS and IVCN models show increasing intensity over the next 5 days and SHIPs has it as a tropical storm in 12 hours and IVCN in 24 hours.
We'll see.
Posted Sept 30 at 12:19 PM |
|
|
| Disappearing posts and the coming storm. |
[Sep. 28th, 2010|12:30 am] |
I was getting a bit irritated that my last two tropical weather posts on Live Journal hadn't automatically shown up on Facebook, and come to find out, one of them wasn't actually on Live Journal. Don't know what happened but it clearly wasn't Facebook's fault..
The weather associated with Matthew is out in the Pacific now and has about a10% chance of redeveloping in the next couple of days. The low associated with Matthew almost certainly returned to the base of the peninsula and has been slowly growing ever since. The weather service seemed reluctant to consider it a tropical depression in spite of the fact that it has obviously had closed circulation for a while now. In the missing post from yesterday afternoon, I pointed out that there were WNW winds at Roatan in Honduras, south winds at Grand Cayman, north winds at Chetuma, and so forth proving a very broad based closed circulation as well. By that time the weather service had finally recognized the system as Invest 96 and were just holding off on recognizing it as a tropical Depression thought the detailed data I need for a storm model run was still being held up. Today it is officially Tropical Depression 16 and I have good data.
Yesterday, my general weather model, and the NOGAPS and GFL models were the only ones showing the disturbance coming up over Cuba, crossing Florida and heading up the east coast. Today all of the reliable models do. The two official models were showing the crossover starting at south Florida, whereas my general weather model was showing the crossing starting further north. With good data, my storm model is also showing a south Florida crossing starting in about 36 hours. This is likely not enough time to reach past tropical storm strength, if that. Once in the Atlantic, I show it heading back up the coast, following it very closely. My prediction of path is right on top of the GFDL model at this point. The BAMM model shows it tracking the coast off-shore and the rest of the official models show it heading over land. I show it continuing to strengthen as it follows the coast and it should make hurricane strength around Thursday.
My model and the GFL are the only ones showing that the weather left behind south of Cuba will once again reform, growing into a much stronger system that will also cross Florida and head up the coast. By Friday we should have an established low south of Cuba again, and a new low east of Florida from remnants left their by the current storm. The one east of Florida will develop as it heads north staying off shore, while the Caribbean will be crossing Cuba on Saturday and then swing back west as we enter complete fantasy territory on the following Monday. By Nov. 7 the low is back west of Cuba and two days later is heading to central Floriday as a strong tropical storm becoming a hurricane just before it hits. Crossing Florida nocks it back down to a tropical storm but almost immediately its a hurricane again travelling up just off shore.
It's unusual this late in the season to see the tropical weather being dominated by the Gulf originated storms. Usually by now it's all coming off Africa, but that has been strangely quiet.
Posted Sept 28 at 12:30pm |
|
|
| Where's Matthew??? |
[Sep. 25th, 2010|02:31 pm] |
Lisa did a brief stint as a hurricane just to spite me. It was really a rather remarkable recovery, but by this morning Lisa was back to a tropical storm and is now back to steadily dying.
Matthew took a large byte out of Honduras and Nicaragua as my model and some of the official models predicted, Then today Matthew started across the Yucatan towards the Gulf. I didn't have time to post this morning before leaving for a meeting, but I did get a run in, and it showed that Matthew would get halfway across and then do a speedy return. When I look at the data sets that have come in since then, that's exactly what I see with tropical depression Matthew again sitting at the south base of the Yucatan Peninsula,
Except that's not what the weather service is saying, The weather service still has Matthew dying halfway across the Yucatan. And this makes no sense to me. In calculating the position of a tropical disturbance, the weather service never looks at the weather associated with a low, they always base the position on where the low itself is, And looking at the data I have, there's no sign of a low at the point where the weather service says Matthew is and by their own data there the biggest low in the area right where my model says Matthew should be.
Now granted, the weather service has data that's a lot more detailed and fine grained than what I can rip off the Internet every 6 hours, but this is still not making sense. If their version of Matthew is still a tropical depression then why isn't the area at the south base of the Yucatan shown as a depression as well. That only makes any sense if the major low doesn't have closed circulation which seems pretty unlikely. But even if you grant that it doesn't, then it damn well should be considered an invest because the point where my model was the only one showing this thing spinning back up is long long gone. At this point all of the official models are showing the new development, though where they take it varies considerably.
The weather service probably has good reason for what they are saying, and if not, they certainly aren't going to retract. So at this point, what I think is still Matthew will probably not go by that name. It may in fact not get a name at all. Let's look into the future as the general weather model shows it. (I can't use the storm model because I'd need the more detailed data on the current low that I would only get if it's declared at least an invest,) For the next 24 hours, it takes it easy, with some back and forth east west movement that will briefly take it back over land again. But the second 24 hours we have a steady slow movement eastward. The third day we start getting a bit of northward movement in our eastward movement and it appears we are headed for Cuba. On the fourth day the eastward movement stops and we actually graze the west end of Cuba as we build a little strength. Day 5 we are pointed at the middle of Florida and almost reaching tropical storm status while a new low has formed from the remains south of Cuba. Day 6 we cross Florida to the northeast corner, losing most of our strength, and the new low is crossing Cuba about where it's predecessor did but is already stronger. Day 7, the last we have any loose connection with reality, what I still think of as Matthew is dead and its remains are moving away at a fast clip. The new storm is a tropical system that stalls and spends the next week building into a significant hurricane that finally crosses Florida and heads up the east coast of the US.
Attempting to predict tropical weather is certainly entertaining.
:-/ |
|
|
| Tropical Update |
[Sep. 23rd, 2010|10:05 pm] |
Lisa is back up to a Tropical Storm, which doesn't sound like it's dying, but it is. This is just a blip and the weather service concurs in their latest discussion.
Matthew is with us now and headed towards Central America, Honduras and Nicaragua, all models agree. But from there things get complicated. The majority of the reliable models are showing Matthew grazing land, which is what my model was showing three days ago. But now two of the official models and mine are showing a deeper byte into land. The NOGAPs model shows an eventual sharp turn to the west and the Pacific. But my model and the GFS curve back out to water just in time to start crossing the Yucatan peninsula. From there the GFS and my model are showing similar paths but different timing. The GFS gets almost across the peninsula and starts curving west towards the Pacific at the end of the 5th day. By that time, my model has completed the turn and is almost through with retracing it's steps to the south of the Yucatan. By day 6, Matthew is south the Yucatan and heads east, reaching a point south of Cuba. On day 7, the last day before entering pure fantasy territory, Mathew stalls and drifts back west, strengthening. In fantasy territory, Matthew heads north, crosses west Cuba, crosses Florida, heads up the east coast of the US.
The models that show a smaller byte into Honduras are also quite split. They range from the UKMET which shows the center reaching the Bay of Campeche before reversing course back towards Central America, to the NGFDL which gets it halfway across the Yucatan when a 180 reversal in course takes it back almost to where it is now. And of course the GFDL and HWRF aren't showing a crossing at all. The GFDL starts to go around the peninsula the long way round, by water, while HWRF is almost a perfect match for my usually unreliable general weather model which has Mathew reaching the peninsula and then heading east to dither around south of Cuba for awhile.
The official weather service prediction splits the difference on how big a byte to take out of Honduras and Nicaragua and then shows a rather slow crossing of the Yucatan. But in their discussion, the weather service admits the crossing of the Yucatan is mostly to allow for continuity with their previous forecast and that they have a very low confidence in the forecast's accuracy.
Interesting times. Posted September 23 at 10:05 AM |
|
|
| Tropical Update |
[Sep. 23rd, 2010|03:00 am] |
Lisa is now downgraded to a tropical depression and the prediction it will go away is looking even better. For the next 7 days I show it heading northwest and getting progressively weaker.
Last post I kept mentioning Invest 97 which was a typo for Invest 95. Having mistyped it once I kept referring back to it. Sorry for any confusion I caused. Some confusion was possible because there is an Invest 97 out in the Pacific.
We no longer have to worry about the invest number in any case because we have Tropical Depression 15 and according to my latest run, we will very shorty have Tropical Storm Matthew. Matthew should then graze the Honduras and then head for the thick end of the Yucatan. Essentially the same as last run, but gaining strength a bit faster than expected. What is different is that yesterday the official models were scattered and now we are in the middle of the pack. In 5 days the remains of Matthew should be almost across the peninsula when my storm model shows it reversing course. At the extreme end of any model reliability, it's heading back south.
Switching back to the general weather model, a new low develops south of the peninsula from the storm remnants left behind. This low is joined by the low from the returning Matthew and the joined low moves just south of Cuba where it stalls and restrengthens. Well in fantasy territory, the tropical storm crosses Cuba, crosses Florida, and then moves up the east coast of the US.
How seriously should we take the course reversal of Matthew? Well, it's at the extreme edge of my model's accuracy. Looking at the official models, the GFDL actually shows something similar but has the return trip happening faster when Matthew in only half way across the peninsula. Interestingly the NOGAPs model never shows Matthew entering the peninsula at all, and is generally consistent with my general weather model run from yesterday. The GFS and BAM models point Matthew at the Pacific after crossing and the HWRF is the lone reliable model taking it into the Gulf, though five days shows a east turn towards Cuba.
So our predictions are:
1) Lisa has no future and is effectively dead.
2) Tropical Depression 15 is shortly to be Tropical Storm Matthew.
3) Tropical Strom Matthew will briefly come ashore, but will continue on to the Yucatan Peninsula.
4) The area south of the Yucatan and south of Cuba bears watching. |
|
|
| The Joy of Tropical Prediction |
[Sep. 21st, 2010|10:51 pm] |
Invest 97 is currently sitting up above Venezuela a tad east of the island of Aruba, fairly harmless at the moment. As it moves west into the Carabbean sea it starts picking up strength. I'm expecting a tropical depression perhaps today, but certainly in the next 4 days, with a 50% chance of a tropical storm in that interval. My storm model shows it coming across and grazing Honduras before crossing the peninsula and coming out a bit south of Campeche aimed firmly at Mexico as we enter pure fantasy territory.
My general weather model, which is generally less accurate, is painting an altogether different picture. The last run of the general weather model shows invest 97 taking a slower trip across the Carabbean and a little north, and then stalling out and gathering strength just before it crosses the Yucatan peninsula. It then heads off to the northeast, crosses Florida and becomes the hurricane which travels up the east coast of the US.
I have to say the general weather model is certainly being persistent about a hurricane traveling up the east coast one way or another. But this is persistently in fantasy territory way beyond any degree of accuracy for my model. When our current weather was in fantasy territory, what became invest 97 was shown as a tropical storm well on it's way to hurricane status by now.
Just for the heck of it, I imposed the storm data for invest 97 onto the general weather model up to the point where the invest is approaching the peninsula. With this change, the storm still stalls, drifts to the east picking up strength, heads north through Havana, crosses Florida, and on hitting the Atlantic curves back into the US, a state up, and inland rather than following the coast up.
Back in the Atlantic, Tropical Storm Lisa did edge a little east as predicted, but is now on a northwest course. I still show it dying out in about 5 days time. Because of the northward movement, this is looking a little more plausible to me than in previous runs. At this point I'm picking up a little more confidence that Lisa could die away in the short term, but it certainly isn't the consensus of any of the official models. But the model shows what the model shows. So I'm predicting Lisa as history by this time next week. |
|
|
| Tropical Update |
[Sep. 20th, 2010|09:34 pm] |
Igor is now fading fast and looking at Bermuda in the rear view mirror. It won't be a hurricane long and is starting to turn northwest. Igor actually started it's fade as it approached Bermuda, dropping in strength to a category 1 hurricane by the time it was there. This was amuch faster drop in strength than my model predicted. I would have to go back and look at the official models. With such a rapid fade to cat 1, it's a bit surprising that Igor is still hanging on as a hurricane.
Igor passed about 40 miles west of Bermuda, so Bermuda had about a full 24 hours of tropical storm strength winds along with plenty of water.
Julia is history. The low is still there and unlike Igor the remnants aren't clearing out. I show Julia's remnants circling around and on the southern part of the swing picking up a low again, but not really becoming a tropical system.
Invest 94 is now Lisa right on schedule, or close enough. It's path is really unstable right now. The official models are all over the place. I'm actually showing it edging east a day or two before resuming a path west and a bit north. My model is not the only one showing a slight eastward movement but we're in a small minority. I'm also showing it petering out in a 5 day period which doesn't seem very likely to me. We'll see.
The start of the fantasy hurricane in the Caribbean/Gulf should have stopped being fantasy now as the date of tropical storm formation approaches. The news there is that the official weather service has started following it and it is Invest 95. This means I get detailed data and it looks like it should be a tropical depression in 24 hours and a tropic storm in 3 days. With more detailed information, I show a path that takes it into the base of the Y peninsula where it barely makes it into the Gulf whereupon it gains a little strength and eventually recapitulates Karl, but not as strong.
Meanwhile the general weather model has a much stronger version of Invest 95 crossing the penensula and in fantasy territory slamming into Mexico. But this is where it gets interesting. Two lows come northward, one north out of central America and the other northwest out of south America that form two strong tropical systems in the area south of Cuba, One of them heads west and into Mexico and the other heads northeast and becomes the hurricane that follows the coast of the US up.
What can we conclude from this? Well, Invest 95 is likely to be the next tropical storm and may barely make it into the Gulf. Secondly the gulf looks to potentially be unusually active for this time of year and bears watching.
Posted at 9:30 on Monday, September 20.
Hoping that my next model run is on my new netbook, but with unexpected problems, I'm not holding my breath. |
|
|
| Tropical Update |
[Sep. 18th, 2010|10:03 pm] |
I'm showing Igor not quite coming through Hog Bay and Somerset Village, but not missing them enough to the west to matter. My latest run shows Igor's center not actually touching land. But Bermuda is a small place and Igor doesn't have to switch much to take it completely east of Bermuda. To be perfectly clear, Bermuda is still small compared to the standard deviations in my models prediction's one day out. But what I show is it almost hitting the west coast. On the other hand, Igor is huge and all parts of Bermuda are going to know it came through, regardless of the small details of its path.
Karl is gone and so is the associated data feed. Invest 94 is strengthening. It shouldn't be too much longer before it's a named storm. The official models are divided into two camps as to how northward the path should be. My model is falling barely within the northern cluster. It doesn't really mater because Invest 94 is definitely a fish storm.
On different runs, the fantasy hurricane is jittering between entering the Gulf and going up the east coast, but it's going up the east coast more often. |
|
|
| Tropical Update |
[Sep. 18th, 2010|09:25 am] |
The official models are still showing a lot of scatter, but the scatter is no longer large enough that Bermuda can get off scott free from Igor, I'm still showing Igor grazing Bermuda on the west side. The official weather service prediction now calls for a head-on crossing, which is where I was about a week ago. A big question is whether Igor will be back down to a cat 2 by the time it hits. Not that there's a lot of difference between a high end cat 2 and a low end cat 3, which is what we're talking about.
There is a new invest off Africa that my not-very-good general weather model is showing as an tropical storm in three days. Since it's an invest now I should have data now to run my pretty-darned-good storm model, but the data I depend on just isn't out there yet.
The new invest should not be confused with the system I've been watching that becomes the fantasy hurricane in a bit less than two weeks. In the latest run, the fantasy hurricane hesitates and then goes up the east coast of Florida instead of entering the Gulf. As always, the fantasy stuff is just that, fantasy, and not worth worrying about - just something to keep an eye on. |
|
|
| Tropical Update |
[Sep. 17th, 2010|09:40 am] |
We have three hurricanes in the Atlantic. Karl is heading into Mexico on track. My model now has Igor grazing Bermuda on the west side. The official weather service discussion of Igor finally admits that Igor will pass close to Bermuda.
Nothing much new. The fantasy hurricane in two weeks hits in south Texas in the latest run. |
|
|
| navigation |
| [ |
viewing |
| |
most recent entries |
] |
| [ |
go |
| |
earlier |
] |
| |
|
|
![[info]](http://l-stat.livejournal.com/img/userinfo.gif?v=92.1){kind=small}
clif_davis's journal