Well, Hurricane Ida did develop last week, but it turned out to be not very much of a threat to south Florida.  It did become only the fifth hurricane since 1851 to enter the Gulf of Mexico during the month of November.  Of course, it’s quite likely that a number of such hurricanes over the southern Gulf of Mexico were not properly identified prior to modern satellite and reconnaissance.   Still, Ida was a rarity in that it was a late-season Category 2 hurricane in the Gulf, and it looks like it regained hurricane strength very close to southeast Louisiana on Monday, November 9th.

About the time that Ida was making landfall as a tropical storm rapidly transitioning to an extratropical storm, we received word that at least one platform in Mississippi Canyon (an offshore lease area southeast of the mouth of the Mississippi River) had reported hurricane-force winds for a 2-3 hour period before the two wind sensors were blown away.  At that time, around 3PM CST Monday, November 9th, Ida was classified as a tropical storm with 70 mph winds by the National Hurricane Center.  This new data suggested that Ida may have briefly regained hurricane intensity that afternoon, so we asked the person in charge of the platform to take a look at the wind report from Monday afternoon.

The next day, the data arrived in the form of a very long text file.  Observations were taken at one second intervals, so there was plenty of data to sort through.  With the help of my coworker, Joseph Spain, we were able to import the data into MS Excel and create graphics of the wind trace.   The winds were certainly well above hurricane strength, but the data were recorded at a height of 97 meters.   A hurricane’s winds are measured at 10 meters above the surface, and we use 1-minute average wind, not 1-second winds.  We contacted the National Hurricane Center to let them know of our findings, and to get some assistance in properly converting the 97 meter winds down to the 10 meter level.  James Franklin, a forecaster with the NHC was quite helpful.  He was also quite interested in receiving the data, as he and Lixion Avila were just starting to prepare the post-storm report on Ida.  This new data could mean that Ida would be reclassified as a hurricane just southeast of the Louisiana Delta in the final report.

Below is a plot of Ida’s track across the northern Gulf on the 9th of November.  Note the location of the Mississippi Canyon report and notice that at the very same time as the high winds there, a recon plane just happened to drop an instrument package to measure Ida’s vertical wind profile.  The instrument measured a 73 kt (83 mph) surface wind in Ida’s southwest quadrant at the same time that the platform was reporting hurricane force winds.  But this surface wind measurement was discounted by the NHC because flight level winds were only 53 kts (62 mph) at the time.  Generally, winds at the surface are only about 80-90% of flight level winds, so it didn’t make sense to the NHC that surface winds would be higher than flight level winds.  The 73 kt measurement was discounted as a brief wind gust near a squall and Ida was kept a tropical storm on the 3PM CST advisory that afternoon.

We received permission to share the wind data with the NHC a few days ago  While we cannot divulge the owner/operator of the platform, we can post the wind data here.   Like I said, there were 2 sensors atop the platform, both at a height of about 300 feet above the surface of the water.  The first plot is the actual, unreduced 1-second wind trace.   Below that is the same wind trace but reduced to the 10-meter level (88% of the speed at 97 meters).  Finally, we had to take the 1-second reports and construct a trace of 1-minute average winds.    It is this trace that clearly identifies a period of sustained hurricane force (64kt / 74 mph) winds to the west of Ida’s center that Monday afternoon.

It’s unfortunate that the sensor (both of them) failed.  But the trace does indicate the winds to be above hurricane force from about 2059Z (2:59PM CST) to 2126Z (3:26PM CST), before the instrument was blown away.  Looking back at a satellite loop of Ida over the platform, it appears that hurricane-force winds might have lasted another 30-45 minutes at the platform.  But very shortly thereafter, Ida encountered quite strong wind shear.  The wind shear ripped all the newly-formed squalls from its center, beginning a rapid transformation to a non-tropical or extratropical storm before it reached the Alabama coast later that evening.  Wind reports from land-based stations in Ida’s path were only in the 25-35 mph range.  It doesn’t appear that Ida produced even tropical storm sustained winds across Mississippi, Alabama or the Florida Panhandle.

With just under 10 days left in the 2009 hurricane season, I think it’s safe to say that the season is over.  Strong wind shear dominates the tropics, and that wind shear will only be increasing through the end of the month.   The 2009 hurricane season will go down as a very quiet one, with the possible exception of a small footnote about Hurricane Ida in the Gulf of Mexico in November.

A few weeks ago, a strong disturbance developed in the southwest Caribbean Sea.  Models had been predicting development weeks in advance, indicating a possible Florida threat.  But the disturbance dissipated just about as quickly as it developed.  Now, a new disturbance has developed in the same area of the southwest Caribbean north of Costa Rica.  The National Hurricane Center has identified the disturbance as Invest 97L.  At ImpactWeather, we’ve identified the disturbance as the 70th disturbance of the 2009 hurricane season.

The satellite image above shows the relative location of the disturbance.  One can clearly see a fairly strong rotation evident in satellite loops.  Surface observations do indicate the presence of a low-level circulation.  That, and the presence of thunderstorms near the center, is nearly enough to classify this low pressure area as a tropical depression.  But the NHC is waiting to see if the thunderstorms are able to persist for another 12-24 hours before upgrading the disturbance.  A recon plane is scheduled to investigate the disturbance on Wednesday afternoon.

Computer models typically have a hard time forecasting the movement of such weak systems.  And the model have a very hard time predicting (correctly) development in the southwest Caribbean Sea this time of the year.  It seems that the models get confused by the ever-present thunderstorms and general low pressure in the region.  So we see a lot of false alarms both early in the season (May/June) and late in the season (October/November).   But out of curiosity, let’s look at where some of the non-tropical “global” models are taking this feature.

Both the American (GFS) model and the European (ECMWF) model take the system northwestward toward the Yucatan Peninsula in about 6-7 days.  The Canadian model, which isn’t shown above, also takes the system northwestward generally toward the Yucatan Peninsula by early next week.  The only pure tropical model that’s being run, the GFDL, takes the low slowly northward for 5 days to just south of central Cuba - as a hurricane.     We’ve seen considerable differences in the models, and even in the same model from run to run over the past 3-4 days, so confidence in the track of this disturbance is on the low side.

Sometimes, climatology can give us a clue as to where a system will  track.   Looking back at all of the November tropical cyclones in the Atlantic Basin that developed during the month of November from1851-2006, we can see that there are two most likely tracks.  Most storms developing in the southwest Caribbean either moved west or northwestward into Central America and/or Mexico and dissipated or they tracked north and northeastward, threatening Cuba, south Florida, or the Bahamas. As an interesting side note, my coworker, Dante Diaz informs me that only 4 hurricanes have struck the U.S. during the month of November.   Where did they strike, you ask?   All four struck Florida.

It appears that the various model runs are pretty much in line with climatology. I’m not really sure where this disturbance will track.  I do think that the chances of it becoming at least Tropical Storm Ida are greater than 50%.  And I’m fairly confident that it’s not going to go anywhere very quickly.  I think that a Northwest Track toward the Yucatan, similar to the American GFS model may be most likely.  Climatology suggests such a track.   On that track, interactions with land combined with increasing wind shear north of 18N would very likely lead to dissipation.  But I cannot completely discount the GFDL model which takes the low due north for 5 days to just south of Cuba as a hurricane next Sunday.  If such a track does materialize, I’d think that interactions with Cuba along with increasing wind shear north of Cuba would lead to considerable weakening prior to any impact on Florida or the Bahamas.

Check back here later this week for an update on this disturbance.  With any luck, this will be the last tropical threat of the 2009 hurricane season.

On Monday, models were in good agreement that the disturbance in the southwest Caribbean would track northward, develop into a tropical cyclone, and possibly impact south Florida and/or the Bahamas.   The National Hurricane Center identified this disturbance as Invest 94L early yesterday morning and gave it a less than 30% chance of developing over the next 48 hours.    I agreed with that assessment, thinking that the greater chance of development would come beyond that 48 hour time frame, probably this Friday or Saturday.   All that was left was to wait and watch for signs of organization.

By late last night, it was looking less likely that the system would develop in the western Caribbean.   Thunderstorms were no more organized than on Monday, and the mid level rotation center had moved to the coast of Nicaragua.    It was the European model (ECMWF) which had been most bullish on development over the past 3-4 days.   In my last post here, I’d shown some model projections valid next Monday evening.  Let’s take a look at what the models are indicating now.  First, the ECMWF:

As you can see - nothing!  Just a cold front across the eastern Gulf.  No storm threatening Florida.  The European model has given up on the idea of development in the western Caribbean.   Not so for the American GFS, model, though:

The American GFS model still develops a storm in the NW Caribbean, but it’s much slower with development compared to previous runs.  And beyond next Monday evening, the GFS has a cold front moving down into the southern Gulf, killing off the storm before it enters the Gulf of Mexico.  The GFS often does such things (developing spurious lows), particularly beyond about 5-7 days.  It’s now the only model indicating any development of the system, and then it dissipates it almost as quickly as it develops the storm.   I don’t think that the GFS model has a good handle on what’s going to happen.

What is going to happen down there in the southwest Caribbean?  I created to demonstrate what I think is happening.  Looking at the loop, below, rotation is clearly evident over southern Nicaragua and northern Costa Rica, but it appears that the main rotation center is shifting westward out into the East Pacific, west of Costa Rica.  I identified the area where I think a tropical depression may form with the yellow circle.

Such development is common in the region.  Tropical waves often flare up in the West Caribbean only to finally develop in the East Pacific.   Occasionally, a storm will develop on either side of Central America.  But I don’t think that will be the case this time.  I think that the disturbance will re-focus in the East Pacific and develop there.   If you live in south Florida or the Bahamas, then you can breathe a little easier now.  But don’t completely turn your back on this system until it’s fully inland into Central America.

I thought that title might get your attention.  What am I talking about, there aren’t even any “Code Yellow” areas out there in the tropics today?   That’s true, but the NHC only identifies potential threats over the next 24-48 hours, not in the long range.  What I see in the western Caribbean today reminds me of the conditions that I witnessed prior to the formation of past late-season hurricanes.   The first really strong cold front of the season has made it all the way to the western Caribbean. A tropical wave interacting with the cold front is helping to enhance thunderstorm activity in the region.  Wind shear in the southwest Caribbean is low, and oceanic heat content is quite high.  Basically, all the ingredients for hurricane development are coming into place now.  But will something develop?

Let’s look at what some of our favorite computer models are predicting.   The American GFS model was probably first to sniff “something” out.  But it’s hard to give credit to a model that sees “something” developing in the long range on almost every run, four times a day.  Recently, the European model has picked up on possible development in the southwest Caribbean this coming weekend.  That’s what has me more concerned.  Oh, the Canadian model also predicts development, but it predicts just about everything in the tropics to develop year-round, so I don’t pay much attention to it.   First, let’s look at the current American GFS model.

The map above is a GFS forecast for next Monday morning, October 26th.  It shows a tropical cyclone (hurricane?) in the southwest Caribbean Sea just off the coast of Nicaragua.  Note the big high pressure area to its north.  With high pressure to the north and clockwise flow around the high, the storm would have to track to the west and move inland into Nicaragua.  Bad for Central America, not so bad for Florida.   Problem is, the European model has another idea as to where this potential tropical cyclone would track, as demonstrated by the map below.

Quite a difference!   That “blocking” high pressure is over Texas, not the southeast U.S.  And with high pressure also east of Bermuda, there’s a clear path for anything developing in the southwest Caribbean to track straight north toward Florida.  We’ve seen such paths before on many occasions in mid to late October.  Remember Michelle in 2001?  Or maybe Wilma in 2005?  Both started out in the southwest Caribbean, were slow to move at first, then accelerated north then northeast across or very near south Florida.  What about beyond next Monday?  The European model goes out to 10 days.   I went to the ECMWF home page (http://www.ecmwf.int/) and made the loop below.   Note that if you look at the last map, you can see an approaching cold front over the central U.S.   Question is, will it get to the east coast before this possible hurricane passes Florida?

That’s a scary track if you happen to live in western Cuba, Florida, or along the East U.S. Coast.  What’s even scarier is that this potential hurricane will be tracking over the one region in the Atlantic Basin that has the greatest heat potential.  Hurricanes need lots of deep warm water to intensify (among other things).  This system could track right across some very warm water, as indicated in the latest oceanic heat potential graphic, below.

Of course, the big question is whether or not this area of thunderstorms will actually be able to get its act together and develop into a tropical cyclone.    The NHC is not mentioning the disturbance in their outlooks, which means they see little chance of development within the next 48 hours.  But the models don’t really develop this disturbance for another 72-96 hours at the earliest.   I think that I might give this system a 30-40% chance of developing into a tropical storm or hurricane in the long range (by this weekend).    As for where it might track, I’m certainly not confident in either the GFS or the European’s track.  However, I have noticed that the European model has tended to do better with the overall pattern in the 5-10 day time frame than the GFS.  If that’s the case, then any developing storm in the southwest Caribbean could well take a northward track that might threaten Florida.  I’d pick such a track as being more likely than a westerly track into Nicaragua at this point.

It’ll be interesting to see how this turns out later this week.  I’ll make another post or two as things develop (or don’t develop).   If you live in Florida, don’t stow away your hurricane survival kits just yet.

As I look out across the tropics today, I see what appears to be a transition to what I’ll call a winter pattern developing.  By that I mean dry air and increasing wind shear are evident across the Atlantic Basin.   Tropical cyclones will have a very difficult time developing in such a pattern.   But since the NHC is now naming what some might argue as questionable low centers over cold water in the far north Atlantic, I won’t absolutely rule out the “development” of perhaps one or two more named storms this season.

Take a look at the animation below from the University of Wisconsin.  It’ s a special microwave image that measure the amount of water in the air.  Red areas represent high water content, blue areas are relatively dry.  The dry air is taking over the tropics now:

The Intertropical Convergence Zone, the band of storms north of the Equator from which tropical cyclones often develop, is starting to shut down between Africa and the eastern Caribbean.  There are still some waves moving off the west coast of Africa, but they’re weak compared to the waves that moved off the west coast of Africa a few weeks ago.  And they’re now surrounded by lots and lots of dry air.  Add to that increasing wind shear across the Tropical Atlantic and I think it’ll be hard for a system to develop east of the Caribbean for the rest of the season.  And that dry air is moving across the Caribbean Sea, too.    The only area with a considerable amount of moisture is the western Gulf of Mexico, but wind shear is on the increase in the Gulf, so development is unlikely.  That strong cold front which will push out across the Gulf this coming weekend may be the signal that development in the Gulf may be shut down for 2009.

Besides the lack of moisture, the jet stream is becoming more and more active, digging farther south with each weak.  Cold fronts are becoming stronger and pushing farther out across the Gulf and off the East U.S. Coast.   This means that wind shear will be on the increase this week.  Wind shear has been an issue across just about all of the Atlantic Basin in 2009.  Even though the analysis charts say that the wind shear in 2009 wasn’t far from average, I think that we really don’t have a good handle on how strong the wind shear really is out over the oceans most of the time.  It’s clear that strong wind shear caused the demise of most of the storms (and disturbances) in 2009.    Here’s a montage of 2009 storms/disturbances from Chris Velden at the University of Wisconsin.  Note how many of this year’s storms were fighting strong wind shear - their centers exposed with thunderstorms way off to one side or the other.

Click to Enlarge

We still have about 6 weeks remaining in the 2009 hurricane season.   I think there’s a very good chance that we won’t see any more named storms this year.  Except, perhaps, another stray storm out in the middle of the Atlantic - far from land.  The Caribbean Sea is climatologically the place to look for late season development (Omar and Paloma in October and November of 2008), but the Caribbean has been a hostile environment in 2009, and I see no signs of that changing.   Just to be safe, give it another 2-3 weeks then we can all breathe a huge sigh of relief that the 2009 season is over, and we escaped the death and destruction that has been so common in recent years.

As the Atlantic Basin remains very quiet, today’s blog update looks at tropical cyclones which impacted the southwestern U.S. It was prepared by Dante Diaz, a member of my hurricane team.

When it comes to tropical storms and hurricanes, the headlines are often focused on some community between Brownsville, TX and Eastport, ME or a Caribbean vacation destination. You almost never hear about hurricanes and tropical storms impacting California. However, there are a few rare ones that manage to find their way to California and the Southwest, at least at tropical storm strength. There are two good reasons why such systems are so rare.

First, the sea surface temperatures off the coast of California are much too cold to support a tropical cyclone. Second, the vast majority of systems forming in the East Pacific are driven westward by a large area of high pressure to their north. To make it to California or the Southwest as a tropical storm, several factors must come together. There needs to be a break in the ridge of high pressure to the north. In addition, the tropical system needs to have both the speed and strength to overcome the cold waters off California or the high and dry terrain of the Desert Southwest.

Fig. 1. Typically, tropical cyclones need sea surface temperatures above 26C to maintain their intensity. Cold water in the vicinity of California causes many tropical cyclones to dissipate before they can reach the coast.

The list of those that have made it to the southwestern United States is short to say the least. Four systems are known to have produced sustained winds to tropical storm force winds here (Chenoweth and Landsea 2004).  They are a September 1939 tropical storm that hit California, Joanne, 1972 that impacted Arizona, Kathleen that struck California and Arizona in 1976, and Nora that reached Arizona in 1997. In addition, Chenoweth and Landsea (2004) have done a study of a system in 1858. Their findings indicate that this system brought winds to hurricane force over coastal San Diego on October 2, 1858. Although there are a few other systems indicated to be at tropical storm strength over the Southwest in the National Hurricane Center best track database (such as Lester, 1992), only the aforementioned five systems were observed to have produced winds to tropical storm force in the United States. Numerous other systems reached the southwestern United States as weakening tropical depressions. Only the September 1939 tropical storm made landfall.  Joanne, Kathleen and Nora made landfall in Mexico before entering the United States. The San Diego Hurricane of 1858 remained just offshore.

Fig. 2. Tracks of Hurricanes Joanne (1972), Kathleen (1976), and Nora (1997) that brought tropical storm force winds to parts Arizona and California.

Fig. 3. A weather map at 1400 LT 2 October 1858 depicting when the hurricane was making its closest approach to the San Diego area. From Chenoweth and Landsea 2004.

The main threat from any tropical cyclone in the Southwest is the potential for destructive flash floods. If the 1939 tropical storm was to strike again today, Chenoweth and Landsea (2004) estimate it would cause as much as $200 million (in 2004 dollars) in damages. Although unique in the historical record, the San Diego Hurricane shows that even California is not immune to the effects of hurricanes. The high tides, heavy rains, and winds to hurricane force would bring even greater damage totals, on the order of several hundred million dollars today in coastal southern California.

Weakening tropical cyclones may bring heavy rain and flash floods to the Southwest, but there is still the potential for a greater hazard on the California coast. Once in a while, the weather patterns manage to bring a hurricane out of the Tropical Pacific to the shores of southern California, inflicting the kind of damage usually seen over the Gulf and Atlantic Coasts of the United States.

Autumn arrives across the Northern Hemisphere at 4:18PM CDT today, signaling a transition from the hot, steamy days of summer to the arrival of cold fronts across the Gulf of Mexico.   In fact, the first such cold front will reach the northern Gulf today.   It’s weak, but it’s a sign of things to come.   In another week or two, cold fronts will begin moving out across the Gulf of Mexico and off the Southeast U.S. Coast weekly.   With each passing cold front, wind shear and dry air will spill out across the Gulf of Mexico and Southwest Atlantic, diminishing the chances of tropical development in the region and gradually bringing the 2009 hurricane season to an end.

To say that the 2009 hurricane season has been quiet would be an understatement.  Except for a brief burst of activity in late August and early September, the Atlantic Basin has been quite a hostile environment for tropical cyclones.  Only  Bill and  Fred would be able to find a small area of favorable conditions to allow each to become a hurricane.  The other four named storms, Ana, Claudette, Danny and Erika, would struggle as weak, sheared tropical storms.  Both hurricanes reached major hurricane intensity, but only Bill made landfall in a much-weakened state over eastern Newfoundland.

Is the 2009 season over?  I think it’s a little too early to make such a statement.  Last year, major hurricanes Omar and Paloma developed in the Caribbean Sea in October and November.  But the Caribbean has been dominated by high wind shear and sinking air since June, and there’s no sign that the environment is changing as of late September.  There may be periods over the next month when wind shear will relax enough across the Gulf of Mexico or the Southwest Atlantic east of the Bahamas to allow for something to develop.    The steering currents of late September through October typically would take late-season developments either toward the northeast Gulf (Florida), the East U.S. Coast, or the northern Caribbean.

I think that another 1 or 2 named storms will probably develop over the next 3-4 weeks and that will be it for 2009.   The 2009 hurricane season will be one of those seasons that no one remembers, and that’s not a bad thing.

September 10th marks the climatological peak of the hurricane season, as indicated by the graphic below.  In a typical hurricane season, about half the tropical storms and hurricanes form after September 10th.  So far this season, we have seen 6 named storms form, two of them hurricanes, and both hurricanes reached major hurricane strength.   A closer look at the graphic below reveals that there are typically more tropical storms which form after September 10th as opposed to hurricanes.   That is, the hurricane to tropical storm ration is typically higher prior to September 10th.    So much for climatology, what’s happening out across the tropics now?

We’ve had a burst of tropical development over the past 3-4 weeks, starting with Ana in mid August and ending with Fred, currently way out in the East Atlantic.   A look at the tracks of the 2009 storms as of today (below) reveals that our seasonal prediction has proven to be quite good.    Back in May, we thought that a combination of analog years and the currently developing El Niño would limit activity in the Caribbean Sea this season.  That has certainly proven to be the case.   Only two storms entered the Caribbean this year, Ana and Erika, and  both storms were torn apart by a combination of wind shear and dry air soon after entering the Caribbean Sea.  We thought that about half of the  storms would form and track farther to the north between Africa and the East Caribbean, turning northward east of the Caribbean.   Three of the six storms took such a track (Bill, Danny and Fred).  We did warn about the possible formation of a storm in the Gulf vs. a long-term track toward the Gulf.  Claudette developed right in the eastern Gulf of Mexico with very little warning.     So far, most of our predictions have come true.  But what about the rest of the season?

As I look out across the tropics today, I see a pattern that is more typical of late in September or October in terms of the amount of wind shear present.  Atmospheric conditions across most of the tropics are simply hostile toward development.  Long-range projections are for a considerable amount of sinking air (no thunderstorms) across the Caribbean Sea and the Main Development Region east of the Caribbean through the rest of the month.   All the wind shear and dry air won’t necessarily prevent development completely.   Additional strong disturbances will occasionally move off the west coast of Africa for another 2-4 weeks.   Wind shear across the western Caribbean Sea and Gulf of Mexico could relax at times over the next month, allowing for a few more storms to develop.

I think that the pattern we’ve seen all season will continue through October.  That pattern would be storms forming a bit farther north, skirting near or north of the eastern Caribbean Sea, then turning northward - somewhere.  So far that “somewhere” has been east of the U.S.  But with the pattern across the U.S. forecast to change over the next few weeks, it’s a little more likely that a developing storm could impact Florida northward through the Carolinas during the last half of September and the first half of October.  So don’t let your guard down yet.  And I’m still not ready to rule out another developing storm in the Gulf of Mexico this season.  Sure, wind shear is currently too high there.   But there are now thunderstorms in the Gulf, a vital ingredient for hurricane development.   If the wind shear drops off for a few days, then we have to watch the Gulf for close-in development.

As for totals by the end of season, our May/June prediction of 10 named storms, 4 hurricanes (2 of them major hurricanes) looks pretty good.   But if I were to make any changes, I’d say that we may well see less than 10 named storms by the end of November.  We’ll need to keep a close eye on the tropics for about another month.  After that, I think it’ll be all over - at least for the Caribbean Sea, the Gulf of Mexico, and the East U.S. Coast.

In my last post, I compared the 10-day forecasts from the American GFS, Canadian model and the European model for then Tropical Disturbance 33 which became major Hurricane Bill.  Of all the models, there was only one model that kept Bill safely northeast of the Caribbean and well off the East U.S. Coast.  That model was the European.  In 2008, the European model topped all models for each time period from the 12-hr forecast to 120 hours.  So far, the European model is off to a very good start with Hurricane Bill.

   
European Model Forecast                            Bill’s Actual Track

As you can see from the 10-day European forecast on the left and Bill’s actual track on the right, Bill passed safely northeast of the Caribbean and well off the coast of New England, just as the European model forecast.  The Canadian and American GFS models were just wrong, as were most of the truly tropical models (GFDL and HWRF).   So let’s keep an eye on the European model this year.  Speaking of which, the European model does see another possible development in the tropics later this week.

Tropical Disturbance 35, an easterly wave that’s interacting with an upper level low a few hundred miles northeast of the Caribbean may well develop into the next named storm.   The GFS, Canadian and European models all indicate some development of this disturbance east of the Bahamas in 2-3 days (Wednesday or Thursday).  All three models indicate a similar track for now.  The European and Canadian models indicate a threat to North Carolina with the storm moving northward up the East U.S. Coast on Friday and Saturday.  The GFS indicates a weaker system tracking to the north, well off the East U.S. Coast.   None of the models is indicating a very strong storm, though.

Let’s see how this one plays out.  Can the Canadian and GFS models beat the European model this time?  We’ll see!

After having virtually nothing to talk about for the past two months, it appears that the Atlantic Basin is finally starting to show signs of life. Tropical Depression Two formed (then weakened) earlier this week near the coast of Africa.  The remnant low of this depression is threatening a comeback this afternoon.  It might just make it over the next few days.   But east of this remnant low is the big disturbance that the models have been predicting would become a large hurricane for more than a week.  I’m not sure that what’s left of Tropical Depression Two will survive for the next five days, though it may well become a weak tropical storm at some point.  So I won’t focus on this system.  Instead, I’d like to take a good look at the disturbance to its east, Invest 90L, or “Disturbance 33″ as we’ve identified it at ImpactWeather.

In 2005, we were curious as to how many tropical disturbances there were in a typical season so we began counting them in our daily tropical outlook.  We just started giving each disturbance a number beginning in May and  ending on November 30th.  What we’ve found over the past four seasons is that there’s a fairly consistent number of disturbances each season, regardless of how many named storms develop.  That number is somewhere between 65 and 75 disturbances. In fact, in 2005 with 28 named storms, there were only 65 disturbances identified.

Back to Disturbance 33, the one in the far eastern Atlantic.  Like I said, the models have been extremely bullish on development for the past week or more.  Now that it’s out over the water, we’ll see if the models were correct.  As of this afternoon, the disturbance appears to have a well-defined circulation, but the thunderstorms aren’t strong enough near the center to classify it as a tropical depression.  I think that will change by tomorrow, and we should have Tropical Depression Three over the weekend.  The next question is where will it go?

All of the models, tropical and non-tropical are in quite good agreement as far as the first 5-6 days, with the exception of the European model. If you remember from last season, the European did an excellent job with Gustav and Ike, but failed miserably with Hanna.  Currently, all the tropical models and the American GFS and Canadian models take Disturbance 33 across the northeast Caribbean next Thursday, August 20th.  The European model’s forecast is to the right/north of every other model, taking the  center about 300 miles to the north of the northeast Caribbean next Thursday then northward and out to sea. I’ve created graphics to illustrate the forecast positions for 7AM CDT next Thursday below:

GFS, Canadian and European Models Valid 7AM CDT Thursday, August 20th

As you can see, the GFS and Canadian models are in great agreement in taking what they’re forecasting to be a hurricane into the northeast Caribbean just to the east-southeast of Puerto Rico next Thursday morning.   The GFS and Canadian models are quite in line with the suite of tropical models, below:

Tropical Models

So, we have reasonably good agreement out for the next week.  The purely tropical models don’t forecast beyond 5 days, but the dynamic models like the GFS, Canadian and European models go out as far as 15 days.  So let’s see what they’re forecasting after this predicted hurricane passes the eastern Caribbean.  The GFS and Canadian are in remarkable agreement today, taking the center to near southern Florida by 7AM CDT on Sunday, August 23rd, while the European is quite different, taking the center well out to sea east of the U.S.

GFS, Canadian and European Models Valid 7AM CDT Sunday, August 23

Beyond Sunday, the GFS and Canadian models turn the hurricane northward up the west coast of Florida and drive it inland across northern Florida and Georgia.  That’s today’s runs, anyway.  The problem is that though these two models are amazingly close in their predictions today, that hasn’t always been the case.  Just yesterday, the GFS model took the center of the hurricane inland into southeast Louisiana.  The day before that, the GFS took it northward along the East U.S. Coast.  The Canadian model has been more consistent, taking the storm into south Florida for the past 2-3 days.  And the European model has been quite consistent in taking the storm northward and safely out to sea after passing well north of the Caribbean.

Where will it go (assuming it actually does develop)?  Climatology suggests that the European model may be on the right track.  However, about 20-30% of historical storms that have developed in the same region have threatened or impacted the East U.S. Coast.  Given the projected pattern in advance of the system, I’m inclined to think that this potential storm may eventually impact the U.S. between Florida and North Carolina.  It’ll be interesting to see how this turns out over the next 10-14 days.