| Ida surprisingly strong, but not for long. |
[Nov. 9th, 2009|08:09 am] |
Yesterday, after a day in Pasadena, I came back and made our usual grocery shopping run with M. So it wasn't till last night that I could take a look at Ida. She had come into the Gulf, not only stronger than I had anticipated, but also stronger than weather service had anticipated. After a new model run I found the path had shifted East. This morning's run takes it even closer to the Florida panhandle. The official prediction and my own are not that far apart, but the ever eastward movement of predictions leads me to privately suspect Ida will come ashore in the west part of the Florida panhandle, even though I am officially saying it will still miss.
The data still shows Ida a cat 2, but between the wind shear and the cool water I don't see any way Ida can hold out as a cat 2. My model is showing it coming ashore as a cat 1, but my strength predictions are less reliable than my position predictions. It's not impossible it could be a tropical storm again.
Once ashore, the weather service is now showing Ida moving east just north of the Florida panhandle. My latest run shows Ida clipping the panhandle and then running north of the weather services line, very close to the ensemble model, a weighted average of the reliable official models. Or at least that's what I show the low pressure system doing. The bad weather I show continuing along on a NE path along the lines of my previous predictions. Over water this is a sign that my model has hit a point where it's going flaky, but over land this split in the storm is actually plausible.
Finally Houston's 24 hours of rain (based on applying my model to the weather system traveling offshore up the coast) has been more like 24 hours of overcast, at least so far.
I have a ton of things to take care of this morning before heading back to Pasadena, so I won't be able to look at things till late tonight. So there will likely be one more run before Ida hits land early tomorrow morning. With luck there will only be 1 more tropical weather post this year.
|
|
|
| A miss is as good as a mile. |
[Nov. 8th, 2009|09:40 am] |
Arg. The computer wiped out an extended post. when I hit the wrong key.
Short version -- Ida is a minimal hurricane but not for long once into the gulf. And possibly just before landfall.
The weather service and official models win this one as Ida misses the Yucatan Peninsula, but not by much and a lot of Ida is over land. The weather service is now in agreement with my model that Ida will hit between Lousiana and the Florida panhandle on Tuesday morning. After coming ashore the weather service moves Ida east into the Florida panhandle while my model moves it to the northeast. Half of the official models go bonky and show Ida reversing back to the south once onshore and the UKMET model has it making an eastward turn into mid-Florida rather than follow the pack north. The rest now show landfall anywhere from East Louisiana to West Florida panhandle.
From our little storm offshore our 24 hours of rain should start in the next 6 to 12 hours. |
|
|
| Tropical Storm in the Gulf |
[Nov. 7th, 2009|08:14 am] |
I have been giving a lot of attention to my mother's move, and then a good friend of many years died unexpectedly, so I have not been giving tropical weather a lot of attention. This is the first time I've done a storm run since tropical storm Ida appeared. Yesterday Ida became a minimal hurricane and cut through eastern Nicaragua, dropping back to tropical storm status, and according to my model, will pretty much stay there until it hits the US.
Having an Atlantic/Gulf hurricane touch land this late in the season during an El Nino year is pretty unusual -- the last time it happened was back in 1925, which is so far back, as a minimal hurricane, it didn't even get a name.
The official weather models are all over the place. Predictions range from the GFDL which takes it up towards Louisiana but then bends east and stays offshore following the coast back down off Florida until it finally crosses over the Keys, to the NOGAPS model which brings up to the coast of Louisiana and then bounces back to the Southwest. Of curiosity is the NGFDL which departs substantially from the GFDL. The NGFDL takes the storm almost up to the eastern border of Louisiana and then stalls out, drifting first east and then sustained west in our direction. Most of the models take the storm through the Yucatan Channel between the Yucatan and Cuba.. The weather service is predicting that as well, but are leaving a wide enough cone to allow hitting Cuba or the Yucatan. From their the official weather service prediction is surprisingly close to the GFDL with increasing margins of error that would let it do anything from hitting eastern Louisiana on Tuesday to the southern tip of Florida on Friday.
My model shows the storm crossing the Yucatan within the weather service's cone of error in about 30 hours. But then I show it recovering loss strength from the crossing and continuing towards Houston outside of their expected margin of error. Then on Monday pressure pulls it back to the northeast back into the weather services cone of error and I show it hitting between Louisiana and the Florida panhandle on Tuesday. The weather service shows it having a 30% chance of resuming hurricane status in the next two days, not unreasonable if it cuts through the Yucatan Channel, but based on surface water temperatures, the odds of it hitting the coast of the US as a hurricane are very slim. As I get the chance I'll be rerunning my model to make sure it continues to curve away from Houston on Monday.
There is another storm system with no chance of development which will be coming up the Texas coast. Here in Houston we should probably get rain from it from about 16GMT on Sunday for a little over 24 hours.
So, one of those interesting case studies coming up. The weather service and all of the reliable models saying the Yucatan Channel and Clif saying the Yucatan. Where all of the official models agree, their batting average is pretty darn good, but my model's prediction has beaten such agreement before, notably with Ike. We'll see. |
|
|
| It'll grow back |
[Oct. 14th, 2009|09:33 am] |
I'm not real sure that someone didn't hit me with a Chinese curse.
I've been needing a haircut for a while now, and yesterday I finally took the time to get one. Tried out a new barbershop. As I've done every time I've gotten a haircut for the last jillion years, I said, "Regular haircut, take some off the top." The barber, who I don't think has English as a native language heard, "Take everything off the top." So he did.
Yes, I did see quite a bit of hair coming off and I thought, he's doing a good job. Actually I thought that right up to the point where he turned the chair around to the mirror for my approval. It was not forthcoming, though I did start laughing. He was very apologetic and the other barbers razzed him unmercifully.
I don't think that I had quite that short a cut even during my entry into the military. Let's just say that combing my fuzz won't be a problem for awhile.
M. says it makes me look like a grizzled general. So anyway, if you see me walking around with next to no hair or with a hat on for all two weeks of cold weather we get here, you know why.
I wonder what happens next.
:-)
|
|
|
| This country is nuts |
[Oct. 11th, 2009|12:43 am] |
The national weather service is part of the National Oceanic and Atmospheric Administration (NOAA), which has a budget of about four billion, with about a billion tied up in the national weather service. A lot of the other three billion is tied up in fixed costs. The most flexible item in their budget is satellites. Considering how complicated the weather is, they do an amazing job and provide a public service that translates back into real dollars and cents.
The QuikSCAT satellite, which was launched in '99 and designed to last 2 to 3 years is now 10 years old. It measures the surface wind speed and direction and over most of the worlds oceans, the two QuikSCAT passes each day are the only wind measurements available. A big part of the improvement in tropical weather prediction has come from QuikSCAT. I'm told that the windspeed measurements save the container and bulk shipping industry alone about $135 million each year by reducing their exposure to hurricane-force wind conditions in non-tropical storms. For tropical storms the QuikSCAT measurements are the only effective way of getting reliable strength and position measurements when the storms lay outside the range of the Hurricane Hunter operations.
My models don't depend directly on QuikSCAT data, but they do depend on data derived from the QuikSCAT data, so loss of the satellite would effect even my little attempt to beat out the weather service. And the loss of the satellite is probably imminent, possibly in the next couple of months. It's bearings are in the process of going bad. Torque loss due to friction in the bearing system is causing the satellite to slow its rotation, causing the calibration to deteriorate past the point of reasonable correction. The odds of QuikSCAT surviving another two years are slim and the odds of surviving until a replacement is up are effectively none as it would be 2015 at the earliest if Congress allocated the money today. Congress is not allocating the money today. Congress is voting on whether to cut the NOAA budget today (Tuesday).
Budget cuts to NOAA hit it in the satellite program because that's where the flexibility is. But NOAA has a host of aging satellites at this point. The next most critical is the TRMM satellite program (Tropical Rainfall Measurement Mission (I think)) which is probably going to fail in 2 or 3 months. And there is a whole row of dominoes after that. So what is the response to this? Today congress votes on our (Texas') Senator Hutchison's amendment to the Commerce State Justice Appropriations Act which lops off another 172 million off the NOAA budget and gives it to the the State Criminal Alien Assistance Program. I'm not inherently opposed to assisting illegal aliens, nor am I opposed to catching them (knowing Sen. Hutchison the money goes to the later) but doing so at the cost of one of the few government programs that provably produces monetary benefits in excess of what is spent, to be spent instead on playing whack-a-mole with illegals? Are they completely nuts?
Since the Democrats control congress, you might think that Republican Hutchison's amendment is dead in the water, but sadly that doesn't seem to be the case.,
Another day in a representative democracy.
|
|
|
| Not as usual |
[Oct. 4th, 2009|12:12 am] |
My car was rear-ended and pretty near totaled today in Austin. How did your day go?
Mark Hall and Kim Kofmel got me back to Houston making a huge detour out of their way. AT. Campbell gave me the shirt off his back. (Well ok, not actually off his back but a dry shirt (it was raining after the accident.)) It's wonderful to have friends that are good to you when you really need it.
I understand they caught the hit-and-run driver. Probably uninsured. |
|
|
| Still alive |
[Sep. 18th, 2009|08:54 am] |
My weight is finally back to where it's below what it was before ArmadilloCon -- 20 lb. down for the year. Took long enough to recover from one weekend.
My last weather post back on the 9th. I was predicting Fred would head back to the NE and then stall while weakening, even though most of the official models were showing a path west of mine. As it turned out the stall was accurate, but I missed the eventual east turn. After Fred was no longer a tropical depression it resumed a westerly path and in fact you can still pick out the remnants of Fred, though not for much longer and at least one of the models is still showing it developing back into a tropical depression. The odds are slim and none.
There is a new low, an invest, well off the coast of Africa. I show no development over the next week. I don't think it will make it to the G-storm, grace I think. In fact I show next to nothing over the next week as far as tropical weather goes. Lots of rain here and there and the oceans are significantly warmer than normal for this time of year, but that shows up mostly in the Pacific.
I have been keeping a very sketchy look at the weather the last two weeks. For the first time ever, my revised process to pull in the weather data recognized when the main data feed went south and switched to the secondary feed. Not that I used the data after I had it. This morning was the first model run I've made in a week, and I didn't really have the time to do it.
Time has been even shorter than normal. My mother's move is a major time sink. Spent most of an entire day earlier this week dealing with her insurance on her old place after you add all the pieces up, and that's after my sister did all the investigative work. She has been paying increasing amounts every year for ever increasing deductibles. Turns out it's way out of line with what she should have been paying. Had the joy of canceling her old insurance yesterday. It will have to be done all over again after she moves out, but still... The builder has said her house will be through in mid-November. A full month earlier than what they originally said. There is no possible way we can have her ready to move by then. In the last month we've filled about a third of her new storage room on this side of town and we haven't made a dent in her stuff. I have too many obligations to go into panic mode on it and (one of the great tautologies) my body isn't as young as it used to be.
Today I'm heading back over to her place in Pasadena to drive her to a Dr. appointment and then sort and load some more. I may get some time to work on ALAMO minutes during the wait. The minutes have been the second largest worry. At the point I thought I was almost done based on recorded segments, I found out the last segment was an hour long. And instead of working on the minutes in odd times between other things I've been reducing my total guilt load by taking care of some other things that could be done relatively quickly to move things along. Well that and looking at Facebook. I haven't had time to let Facebook become the time sink I see it could be. As is, I'm calling a 1-week moratorium on commenting and most of reading Facebook. I need the short blocks of time.
Have had a paper the last two weeks I really need to read. Data is missing all the time in the weather feed and when it does I fall back on the use of proxies, usually averages from when things were similar. It doesn't slow down the model run, but the yearly establishment of optimal proxies is fairly compute intensive. There is a new technique used with Bayesian learning called Collapse and Bound (to the best of my memory) that takes extreme assumptions, collapses them to a single point, can take information about the distribution of missing data into account and is much faster than what I'm doing. My machine learning technique is considerably more convoluted than Bayesian learning, but odds are the underlying idea is probably adoptable. Scorn not to stand on the shoulders of giants and all that. Trouble is, I located the paper on the web weeks ago and still haven't had time to sit down and read it. :-(
Well, there's another chunk of time gone. |
|
|
| Hurricane Fred |
[Sep. 9th, 2009|01:21 am] |
As of the latest data feed, Fred has an eye and is a hurricane much faster than I expected. It has continued to pull west as predicted by my model and more recently the GFDL and HWRF models. Over the next week though, they remain to the west of my model predictions. In the latest runs, the NGFDL and NOGAPS models have joined mine in showing an eventual pull to the NE. They do not, however, show the subsequent stall that I do. The UKMET is producing the most westerly predictions. but is the only other of the reliable models showing a stall -- in fact much earlier than my stall,
Well, we'll see what morning and new data brings, I show Fred continuing to build in intensity over the next 24 hours. Tomorow and Thursday I'll be spending most of the day in Pasadena so probably will not be posting. |
|
|
| Tropical Update - TS Fred |
[Sep. 8th, 2009|03:01 pm] |
The system now well away from Africa is now Tropical Storm Fred right on schedule and is heading west as predicted. I show it hitting hurricane status Thursday morning. It should go west for another 24 hours yet before starting it's swing to the NW. This should continue as a swing to the north and then a brief swing the NE before sitting in place for a couple of days as it weakens. Most of the official models predictions are to the west of mine and are showing that the western course resumes but my model isn't showing that, at least not over the next 7 days that my model has some reliable relation to reality.
I show a minor chance, less than 2%, of a tropical system forming off the coast of North Carolina and about a 1/2% chance of a tropical storm forming in the western Gulf this weekend.
|
|
|
| F storm? |
[Sep. 6th, 2009|08:38 am] |
The odds of Invest 95 becoming a tropical storm in the mid-Atlantic are much worse this morning. The likely candidate then is the system I show coming off of Africa now, the center completely in the Atlantic in 24 hours and then my model show it becoming a major storm in a 4 day period. This last was mentioned in my last blog post. It's interesting to compare with today's blog from Dr. Jeff Master's on Weather Underground.
He says:
A strong tropical wave with plenty of rotation is emerging off the coast of Africa this morning, and this wave is a good candidate to show some development this week as it heads west-northwest over the Atlantic. The wave is under about 15 knots of wind shear, and is being developed by several models, including the GFS and ECMWF. However, the models show that this new wave will be pulled northwestward by a strong trough of low pressure this week, and it appears unlikely that the wave will make the long crossing of the Atlantic necessary to threaten any land areas. Another wave with plenty of spin will emerge from the coast of Africa two days from now, and will also have a chance to develop.
By comparison I'm showing generally westward movement for a few days before an increasingly northward movement takes place. In any case, it's a fish storm. I seem to be the only one saying hurricane status at this point though. I'm also not seeing anything else of note coming off of Africa before Friday.
|
|
|
| F and G |
[Sep. 5th, 2009|11:54 am] |
Wednesday and Thursday were supposed to be packing and loading in Pasadena for storage on this side of town, and Wednesday night working with Bill on projects.. But Thursday morning would be available for a brief set of weather runs and another brief stint at ALAMO minutes. That was the theory.
Wednesday, dealing with mother's bank took longer than expected, so the packing Wednesday was kind of light. Got in at a reasonable hour from Bill's, so Thursday morning was looking good. Thursday morning, I got an emergency call from my mother. Her garage was flooding. So much for anything but heading to Pasadena. It turned out to be a valve on her water-heater. It was before the cut-off so all I could do was minimize the flooding until a plumber could get there which turned out to be Friday. Did get in a long day of packing in Pasadena and after unloading in my mother's storage room on this side of town, barely beat Margaret home when she got off work at 9PM.
Friday morning it was back to Pasadena to let the plumber fix the valve on the water-heater. Got in a bit of packing, but mostly it wasn't an intense day. Got back in plenty of time to fix a Chinese supper for Margaret and I to enjoy. Friday night is date night and our science fiction shows. Warehouse 13 (light and fluffy science fantasy action adventure), Eureka (very light and fluffy sf comedy adventure, but with an interesting story.) , and then the first episode from DVD of Torchwood: Children of Earth (NOT light and fluffy - rather unsettling actually) and then a superior Voyager episode from tape (It's really about an alien race and the conflict between science and doctrine, truth and power, and not really Voyager - ignore the standard Trek background and it's a very compelling SF story).
So, I hadn't looked at the weather much past my last post. It looks like the weather service was right and I was wrong and the Invest made it up to Tropical Storm Erika. At which point it faded away and is gone now. I assume that my prediction of the hurricane hitting Baja California and then drifting away was right, because there is no sign of it in my data. Though it could be just off the edge of my model and I could go back through the data to see what happened. I'm not though, for a Pacific storm.
Now, there is an Invest 95 heading NW to nowhere that I show making tropical depression any time and becoming the F storm in about three days. More impressive there is a storm coming off Africa that I show hitting the water tomorrow and strengthening up to hurricane status over the next 5 days. It should definitely be the G storm (or the F storm if Invest 95 doesn't quite make it). After 5 days it bends NW and probably aims towards being a fish storm as well.
I'm loving this hurricane season in the Gulf.
-------
Ha. Flipped back through Jeff Master's blog and he was also predicting Jimena would hit Baja California and then dissipate after 5 days, but he was mentioning it a day after my last post.
Still - Pacific - doesn't count positive or negative.
|
|
|
| Weather |
[Sep. 1st, 2009|12:47 am] |
Invest 94L is still hanging on. The HWRF model is showing it as a tropical storm in 5 days and the National Hurricane Center is giving it a greater than 50% of being a tropical depression by Wednesday. My model shows it barely hanging on for the next 5 days, but then starts to perk up a little. I'm wondering if my model went from being insensitive to shear two years ago to being overly sensitive to shear proxies. Well I'm not going to try to outguess the Machine learning algorithm.
The major storm I mention moving north in the Pacific along the coast is the strongest huricane so far this year. I don;t score Pacific hurricanes or pay them much attention. But it is in my model area. I show it ramming into Baja California and tearing everything up good for the next 5 days until it reaches the end, runs out of steam, and just drifts away.
Sleepy. Gunnu post this tonight and spell-check tomorow if ever. |
|
|
| No News is good news. |
[Aug. 29th, 2009|08:21 am] |
Danny is no more and our Invest is probably dying. Absolutely nothing of interest in the Atlantic or Gulf for the next week. That's not true of the part of the Pacific that lies in my model area, where I show a strong storm system offshore moving up the coast from S. America to N. America. But for my primary area of interest - no news worthy of note.
All together we didn't do badly on Danny, even if 1 or 2 of the official models did better overall, if less consistently. We certainly did better than the National Hurricane Center which was constantly predicting Danny would strengthen more than it did. My model undershot, but the weather service way overshot. My last post had 3 of the official models that I have results for striking land in New England and three of them taking it further offshore. If I had had time yesterday morning to post, I would have noted that the models eventually converged on the path my model had been consistently showing, though they diverged again by the time Danny would have reached Nova Scotia. But the weather service maintained a cone of error that still showed a possibility of clipping the US, right up to the end. The last set of runs before Danny was declared ex-tropical, all of the official models had shifted to show Danny would hit Nova Scotia further north than my still pretty much consistent-from-run-to-run prediction showed. It won't be possible to make an actual comparison, because it's no longer possible to pick out a center of circulation and the low has been subsumed into the general weather pattern. It's impossible to track Danny's remains in the general mish-mash.
The weather service is picking up a lot of criticism in the background chatter for their handling of Danny -- for moving it up to tropical storm status way too early, for reporting maximum winds that were measured far from the center of circulation without noting that fact in their discussions, for over-relying on the SHIPs model which has a known tendency to over-estimate future storm intensity. Something I noticed early, but didn't mention here, was that my model showed most of the weather associated with Danny to the east of the low pressure system and center of circulation. But the actual fact seemed to catch people who rely on the discussions for information by surprise -- though I would argue that under a close reading the information was there.
In any case, Danny was a hard system to predict being influenced by an unusually large number of factors. This could have easily thrown my machine learning based system into territory where it made bizarre predictions. And it didn't. Which is good. I think.
Something I've now noticed is that when I use my primary and secondary data sources, that the one-week projections are a bit better with my primary data source (when I can get it - it disappears at the most inconvenient times) but when I extend to two week model-runs the predictions from the secondary data source is much better. This is not so much in terms of accuracy as plausibleness. The two week predictions from the secondary data source usually look like something that could actually happen. Those from the primary data source will show hurricanes forming over land and other things that make no sense. I think the explanation for this is that the secondary data is less detailed but covers a larger area, particularly to the north, but also a bit further east and west. In theory I could combine the data to get the best of both worlds, and I may yet, but it opens a whole can of worms that it would take 5 or 6 of these posts to make reasonably intelligible.
---------
An edit to note that we are halfway through hurricane season.
---------
Enough weather. I'm calling my sister to see if she needs me and if not it's time to eyeball my short story and then get back to working on meeting minutes. |
|
|
| Danny and Beyond |
[Aug. 27th, 2009|09:52 am] |
The tropical depression became tropical storm Danny even faster than the weather service expected. I'm showing no further intensity build and a decrease in intensity after 48 hours, while the weather service is expecting it to build a bit more, up to 65 knots and marginal hurricane status. Since I've been showing no intensity build up all along, the weather service is more credible. The official models have shifted their path predictions to center on mine which is holding stable. The NAM, NOGAPS, and Canadian models are showing it to the west of my path and are actually bring it ashore centered around Cape Hatteras and from there over New England. The HWRF, GFS and GFDL are showing it to my east, agreeing that it won't hit land until it comes to Nova Scotia. The weather service in their predictions are now allowing just enough margin of error to allow the westward models, but is now centered on a path just barely east of my prediction.
The situation off Africa seems to have stabilized and I am seeing stable results now out to 6 or 7 days. It is now the lead system that will develop (I'm expecting tropical depression in the next two days and possibly sooner) and not the next storm coming off. It should move fairly much west for the next 5 days and start swinging north on the 6th. The weather service has called it an Invest, number 94, which gives me better data and more models to compare to. These vary from the BAMS model which takes it steaming due west to the BAMD which start it northward almost immediately. When I reran with the better data, my predicted path hardly varied. Extending beyond 7 days into fantasy territory, I show it eventually swinging sharply north and coming ashore on Long Island.
There is a minor possibility something could develop off the east coast of the US and follow the path of Bill up. There is an even more minor possibility of something developing off the east coast of Mexico and following the coast up to Texas before coming inland. I'm not really expecting to see either of these.
Went back to the run I made for Bill on August 11, back when it wasn't that far off Africa, and the fantasy projection I made for it then really didn't turn out too wide of the mark. Of course it was followed by fantasy predictions that had it going as far afield as Mexico. Altogether we didn't do too badly on Bill compared to most of the rest of the models. The Canadian model did best on Bill overall though. Supposedly the European model did best on Bill and was the most accurate over the 2008 season, but they don't make their results of their model public on a timely basis, which I would think defeats the purpose.
|
|
|
| Confusing update |
[Aug. 26th, 2009|08:39 am] |
When the weather service started mentioning 92L I couldn't even see it. When I could see it, I showed it wouldn't develop any. It did. I'm still showing it developing no further. The weather service has it becoming tropical storm Danny on Thursday and many of the models are showing it hitting hurricane strength. I'm showing it losing organization as it approaches land in 3 days. Needless to say my confidence in my models intensity prediction is low.
On the other hand, as usual, I think we're in better shape on path prediction. On nearing the Carolinas, the storm will turn and more or less follow the path of Bill up the coast without landfall until it troubles Canada's eastward extensions of Nova Scotia and Newfoundland. As was the case with Bill I'm showing it a bit closer to the coast than the pack of reliable models, except for the BAM model which brings it ashore in the eastern extension of New York, about the end of I495, and then on to Connecticut.
It''s normal for storms to come west out of Africa this time of year and every so often another one comes off. But currently it always seems to be the next storm that looks interesting and then doesn't, and then does, and then doesn't. Normally my model's predictions after 7 days are very unstable shifting from run to run, but right now in the middle Atlantic they're shifting from run to run after three days. One run a storm will turn north, another it heads towards S. America and the Gulf, then it intensifies, then it doesn't, then it shifts north into a fish storm, each model run very different. I have no idea what's causing this confusing instability.
Yesterday I made quite a bit of progress on switching weather stuff on to my netbook, so that I can make runs even when away from home. Data downloading won't be as automatic, but should be livable. With one more good day's work I think I'll be able to do it, although I still won't have the flexibility to recover when my feeds go bad like I do on my main computer. And of course I won't get one more good days work for awhile.
My back has recovered just in time to start the packing/moving to storage Wed/Thurs schedule. Once Friday is here, I'll be trying to clean up my short story for writers circle and hit some of the other things I have backed up that I should have done yesterday. Heavens knows when I'll touch the music again.
Breakfast calls - gotta go... |
|
|
| Tropical update |
[Aug. 22nd, 2009|11:44 pm] |
Bill's eastward movement is now underway. He's not cutting as close to the east coast as I expected (came through quite close to Bermuda actually), but on the other hand a glancing blow is all the Canadian "islands" are going to get. So I'm counting it 1 out of 2 if all goes as expected. Its holding together better than I expected, but it"s still winding down.
Tomorrow a new system moves west across Africa. In the next week it should become an invest, In two weeks it could be reaching Louisiana, but I never show it developing into a tropical storm.
Now that hurricane Bill is unwinding, my normal data-feed resumes. Figures. As usual there's no explanation.
---
I'm currently working on a short story, a door-swings-both-ways plot. Just as a jet doesn't have to carry it's own reaction mass, just fuel to accelerate the surrounding air, darkmatter, if it could be made to serve as reaction mass, could combine with relativistic time dilation effect to make travel throughout the galaxy feasible. Not so much between galaxies though. As always, how do I tell a tech dependent story and minimize info-dump?
We have about 1/8th of my mother's storage unit full. Good we've made so much progress, but not good it's filling so fast. They should be pouring concrete soon. |
|
|
| Near Miss |
[Aug. 21st, 2009|09:20 am] |
I'm still showing Bill as a fish storm that never comes ashore, but each run I do takes it just a little closer to New England and then Canada.. If the sharp eastward movement I expect to kick in 88 hours from now is a bit late, then things are going to be a lot messy.
I bring Bill closer to New York and Boston than the official models do, and then I show a sharper eastward movement so that it just clips Nova Scotia and Newfoundland. Many of the official models are taking it right through the middle of Nova Scotia and Newfoundland. Normally I've developed some faith in the path predictions of my storm model, but it's dependent on input from my general weather model and this is far enough north it's near the edge of my digital world where things go flaky first.
My intensity predictions are nowhere near as confidence inspiring for what it's worth, but from the latest set of data, I would say that Bill is down to a cat 2 storm, though the weather service may still have it as a cat 3. But, I show Bill going back up to a cat 4 in 24 hours before starting to diminish for real. We'll see.
|
|
|
| ArmadilloCon Blues |
[Aug. 17th, 2009|08:11 am] |
I'm going to have to rethink my plan of not worrying about weight or what or how much I eat over major holidays and SF cons. Over ApolloCon it wasn't a problem, but at ArmadilloCon I gained 5 1/2 freaking pounds! At my gradually-nudging-my-weight-down-without-really-dieting pace, (the only plan I'm likely to follow for anything less than avoiding immediate death) it takes a good month and a half to drop 5 pounds. It wipes out about a quarter of the progress I've made since Christmas.
Had a blast at ArmadilloCon though. Howard's reading at the end makes me aware I need to find a way to read Vance's Dying Earth stuff. Wound up missing about half the stuff I wanted to do there, including all of the readings (except Howard) I would have like to have made - Utley, Moon, Wells - No Neal Barrot reading this time, though I know he was there. I don't consider that a good sign as it indicates he isn't writing.
Came back with about 5 story idea-lets suggested by panels or conversations at ArmadilloCon which will go on index cards this morning to be stored until needed. Two of them came from the City panel that was surprisingly and amazingly good that I almost skipped at the last minute - and was really glad I didn't. Most disappointing panel was the Research panel, which wasn't actually bad, but no-one said anything that wasn't obvious. The best panel I had originally meant to skip was the Vampire panel and the panel which was the most fun was the Dr. Who - Speciest panel.
Talked to Martin Wagner a little bit, which was fun. Met A.T.'s good friend and she seems extreemly nice. She got me started on using machine learning techniques to predict weather and didn't run away screaming in horror. Didn't have a few conversations I wanted to have but had the important ones. Came back with two new t-shirts, one for the Texas in 2013 world-con bid. I asked Bill Parker if the Horns on the UFOs were meant to alienate the Aggies and he gave me the look I deserved. :-D
The new weather data is in, but nothing significant seems changed from last night. Projecting 2 weeks ahead on the general model during current runs and Florida gets an awful lot of rain the next two weeks. But fortunately not the pan-handle which is what is getting pounded now.
|
|
|
| Post ApolloCon Tropical Update |
[Aug. 16th, 2009|11:21 pm] |
Wow. My last entry here was Thursday. At that time TD2 looked like it was going to slowly fade away without ever reaching tropical storm status. It looked like the next tropical storm was still a wave, but a wave I was finding very worrying.
Friday morning we headed to Austin for ArmadilloCon, Austin's Science Fiction Convention. I managed a short bit of Internet access on Saturday. I couldn't run a my models on my netbook, but I could look at what was happening. TD2 had strengthened after all becoming Ana. The wave had, as expected graduated all the way up to a tropical storm. Tropical storm Bill looked as if it was moving consistent with the path I was expecting, but I of course I had no way to extend the prediction. Meanwhile there was a new tropical depression off the coast of Florida and a new wave off Africa.
Tonight we are finally home. Taking a look at the situation, the weather service has Ana on its last legs. The tropical depression off the coast of Florida is now a Tropical storm Claudette. And the weather service models are showing it curving up and missing the east coast.
My run is still using the alternate data since the primary data has been out for a week and a half. And my model run produces no surprises. I agree that Bill is looking like a fish storm and that Ana will go away soon. I don't show the new wave from Africa. And Claudette is clearly going inland to its death. At this point it looks like the excitement will soon be over for another little bit.
|
|
|
| Tropical Update |
[Aug. 13th, 2009|02:09 am] |
TD2 is still headed west, however I'm seriously starting to doubt it will make it to the A-named tropical storm. I'm not showing any intensification in it's future and the weather service is starting to comment about wind-sheer and dry air. A week from now I show it hitting the coast of Florida.
What is destined to maybe be the A-storm instead of the B-storm (Bill this year, I think) is now considered a wave. I'm still expecting it to make invest status in the next couple of days and make hurricane status before it's over. At the one week mark, it should be a major storm NE of South America, aimed at the Gulf. At this point, now that it's a wave, most of the reliable official models are also showing it's development to one degree or another, with maybe the exception of the NOGAPs model. But I still seem to be a good 2 days ahead of the pack in picking things up.
When we go more than one week out, into fantasy territory, the fate of the storm varies wildly from one model run to the next. One run will take it across the gulf into Mexico, one will bring it in around Corpus. Coming in and following the west coast of Florida and then up NE across the US back into the Atlantic is very popular (happened twice). At the moment, the fantasy predictions have it back to swinging east of Florida and staying offshore up the east coast of the US, as was the case the last time I posted. Mostly this is entertainment value at this point.
If the storm was not so far out (it's still pretty close to Africa) I'd think about staying up for the next batch of data, but I'm tired. Stayed up late last night working on minutes and traded days to finish it today which means I'm off to Pasadena tomorrow to load boxes.
You know the whole idea of being retired is to trade financial income against freedom to explore things that are interesting without being distracted with constant deadlines. I wonder why it doesn't quite seem to work that way?
|
|
|
| navigation |
| [ |
viewing |
| |
most recent entries |
] |
| [ |
go |
| |
earlier |
] |
| |
|
|
![[info]](http://l-stat.livejournal.com/img/userinfo.gif){kind=small}
clif_davis's journal