Harvey drenching Belize; 97L a threat to the Caribbean and U.S.

Tropical Storm Harvey is closing in towards a landfall this afternoon in Belize, and is dumping very heavy rains on northern Honduras, northern Guatemala, and Belize as it steadily moves west near 12 mph. A personal weather station on Roatan Island on the north coast of Honduras has received 6.68" of rain as of 10am EDT this morning from Harvey, and had a peak wind gust of 42 mph. The Roatan airport has received 3.55", and had a peak wind gust of 40 mph. The first significant spiral band from Harvey moved over Belize City at 7am local time, dropping nearly an inch of rain on the city. Belize National Meteorological Service radar shows that Harvey has appeared to close off an eyewall as of 11:30am EDT, which may allow the storm to intensify another 10 - 15 mph before landfall. The 11am NHC wind probability forecast gave Harvey a 3% chance of making it to hurricane strength, but the discussion noted that it wouldn't be that hard for Harvey to gain another 10 - 15 mph before landfall. I estimate there is a 30% chance that the winds along a 10-mile stretch of Belize coast where the eyewall makes landfall will reach hurricane force.


Figure 1. Radar image of Harvey taken at 11:30am EDT on Saturday, August 20, 2011, a few hours before landfall in Belize. A small closed eye is visible just south of the offshore islands of Belize. Image credit: Belize National Meteorological Service.


Figure 2. True-color MODIS image of Tropical Storm Harvey taken at 12:25pm EDT on Friday, August 19, 2011. An hour after this picture was taken, Harvey became a tropical storm with 40 mph winds. Image credit: NASA.

An exceptionally active early part of hurricane season
It's been a strangely hyperactive season for weak storms in the Atlantic so far this year. Tropical Storm Harvey is the 8th named storm this year, and its formation date of August 19 marks the 4th earliest date on record for the Atlantic's 8th storm. Only 2005, 1933, and 1936 had the 8th storm of the season form earlier. All eight storms this year have stayed below hurricane strength, making 2011 the first hurricane season since record keeping began in 1851 to have more than six consecutive tropical storms that did not reach hurricane strength. As I discussed in yesterday's post, a major reason for this is the lack of vertical instability over the tropical Atlantic so far this year. We've had a large amount of dry, sinking air over the tropical Atlantic, and the usual amount of dry, dusty air from the Sahara, both helping to keep the atmosphere stable and stop this year's storms from intensifying into hurricanes. Hurricane activity typically ramps up big-time by August 20, with more than 80% of all the hurricanes and 65% of all the tropical storms occurring after that date. With 97L looking like it will become a named storm in the next few days, at our current pace, 2011 will become the second busiest Atlantic hurricane season on record, with 24 - 27 named storms. There are only 21 names in the list of names for a hurricane season, so we may have to break out the Greek alphabet again in late October this year, as occurred in 2005. Ironically, we are using the 2005 list of names this year, so 16 of this year's 21 names are repeats of 2005. I'm not too happy about seeing a hurricane season challenge the Hurricane Season of 2005 in any way, and let's hope we don't retire another five names this year, like occurred in 2005! With vertical instability much lower this year than in 2005, and that year having already seen one storm (Dennis) retired by this point in the season, I doubt that will happen, though.


Figure 3. The annual cycle of average hurricane frequency in the Atlantic. Historically, about 35% of all the tropical storms and 15% of all the hurricanes will have occurred by August 20.

Invest 97L likely to become a tropical storm by Tuesday, could threaten the U.S.
A tropical wave near 14°N 56°W, about 450 miles east of the Lesser Antilles Islands, is moving west to west-northwest at 15 - 20 mph. This wave, designated Invest 97L, has built a respectable amount of heavy thunderstorm activity over the past day, but remains disorganized. Dry air to the north and west is slowing development, as well as moderate wind shear of 10 - 20 knots, as analyzed by the University of Wisconsin CIMSS group. An impressive amount of large-scale spin is obvious in visible satellite loops, but there is no sign of a well-defined surface circulation. An ASCAT pass at 9:04am EDT this morning showed a strong wind shift, but no closed circulation. Ocean temperatures are about 28.5°C, about 2°C above the threshold needed to support a tropical storm. A hurricane hunter aircraft will investigate 97L this afternoon.


Figure 4. Morning satellite image of the tropical disturbance Invest 97L.

The computer models have shifted southwards since yesterday, and now take 97L south of Puerto Rico on Monday, and along the south shore of the Dominican Republic on Tuesday. On Wednesday, 97L should pass near or over southern Haiti, Eastern Cuba, and Jamaica. On Wednesday and Thursday, the models agree that a trough of low pressure will dip down over the Eastern U.S., which is likely to turn 97L to the north. The exact timing and strength of this trough varies considerably from model to model, and will be critical in determining where and when 97L will turn to the north. We can expect that 97L will impact Central Cuba, the Bahamas, and the Florida Keys on Thursday, but it is uncertain whether 97L's turn to the north will take the storm into the Gulf of Mexico or not.

The computer models continue to enthusiastically develop 97L, and all the ingredients seem to be in place for a tropical storm to form by Monday or Tuesday as 97L crosses the Northeast Caribbean. The atmosphere is expected to be moister over the Caribbean, wind shear will remain a low 5 - 10 knots, and sea surface temperatures will increase to near 29°C. The main impediment for development will likely be two-fold: too much dry, stable air, and proximity to land.

There has been an unusual amount of dry, stable air in the Atlantic this year, due to a combination of dry air from Africa, and upper-atmosphere dynamics creating large areas of sinking air that dry as they warm and approach the surface. This stable air has been largely responsible for the fact that none of our seven tropical storms so far this year has made it to hurricane strength, despite the presence of sea surface temperatures that are the 3rd warmest on record across the tropical Atlantic. Tropical Storm Emily in early August encountered problems with dry air when it crossed the Northeast Caribbean, and 97L may have similar difficulties. There will be some moderate wind shear of 10 - 20 knots to the north of 97L over the next week, and this shear may work in concert with the dry air to slow development.

Given 97L's current disorganization and problems with dry air, I believe it is unlikely the storm will be stronger than a 55 mph tropical storm on Tuesday morning, when it will be close enough to the mountainous island of Hispaniola that a good portion of its circulation will be over the island, disrupting the storm. 97L may also make a direct hit on the Dominican Republic or Haiti sometime Tuesday or Wednesday morning, which could even destroy the storm, like happened to Tropical Storm Emily in early August. However, there is at least a 30% chance that 97L will miss Hispaniola, and slide through the waters between Jamaica and Eastern Cuba, allowing the storm to intensify into a hurricane south of Cuba. At this point, it appears there are too many hurdles for 97L to negotiate for it to arrive in the Florida Straits as a hurricane, since the storm has to cross Cuba and/or Hispaniola, plus contend with dry air and wind shear. However, 97L hasn't even developed a well-defined circulation yet, making it difficult for the models to zero in on a solution for where the storm might go. The average error for an official 5-day forecast from NHC for a developed storm is 200 - 250 miles; the error will be much higher for a 6 to 7-day forecast of an Invest that hasn't developed yet. Given the uncertainties, this weekend would be a good time to go over your hurricane preparedness if you live anywhere in the Caribbean, Bahamas, or Florida, since 97L could well be paying you a visit as a tropical storm or hurricane sometime in the next week.


Figure 5. Skill of computer model forecasts of Atlantic named storms during 2010. OFCL=Official NHC forecast; GFS=Global Forecast System model; GFDL=Geophysical Fluid Dynamic Laboratory model; HWRF=Hurricane Weather Research Forecasting model; NOGAPS=Navy Operational Global Prediction System model; UKMET=United Kingdom Met Office model; ECMWF=European Center for Medium Range Weather Forecasting model; TVCN=one of the consensus models that lends together all (or most) of the above models. Image credit: National Hurricane Center 2010 verification report.

Which model should you trust?
Wunderground provides a web page with computer model forecasts for many of the best-performing models used to predict hurricane tracks. So which is the best? Well, the best forecasts are made by combining the forecasts from three or more models into a "consensus" forecast. Over the past decade, NHC has greatly improved their forecasts by relying on consensus forecast models made using various combinations of the GFS, GFDL, NOGAPS, UKMET, HWRF, and ECMWF models. If you average together the track forecasts from these models, the NHC official forecast will rarely depart much from it, and the NHC forecast has been hard to beat over the past few years. The single best-performing model over the past two years has been the ECMWF (European Center model). This model out-performed the official NHC forecast in 2010 for 1-day, 2-day, 3-day and 4-day forecasts, and in 2009 for 4-day and 5-day forecasts. You can view ECMWF forecasts on our wundermap with the model layer turned on. The European Center does not permit public display of tropical storm positions from their hurricane tracking module of their model, so we are unable to put ECMWF forecasts on our computer model forecast page that plots positions from the other major models. As seen in Figure 5, over the past two years, the GFS and GFDL model have been the next best models, with the UKMET model not far behind. Last year, the NOGAPS model did very poorly, forcing NHC to come up with some new consensus models this year, the TCOA and TVCA, that do not include the NOGAPS model. For those interested in learning more about the models, NOAA has a great training video (updated for 2011.)

Invest 98L near the Cape Verde Islands
A tropical wave near Cape Verde Islands off the coast of Africa, Invest 98L, is spreading heavy rains and strong gusty winds to those islands today. So far this morning, top sustained winds measured in the Cape Verde Islands were 23 mph at Mindelo. Water temperatures are warm, near 27 - 28°C, and wind shear is low, 5 - 10 knots, so 98L should continue to organize today before running into more hostile conditions on Sunday. NHC gave the storm a 40% chance of developing into a tropical depression by Monday morning in their 8am advisory. Once 98L passes to the east of the Lesser Antilles, it has a long stretch of ocean to cross before it could affect any other land areas. Approximately 70 - 80% of all tropical cyclones that pass this close to the Cape Verde Islands end up curving out to sea and not affecting any other land areas, according to Dr. Bob Hart's excellent historical probability of landfall charts. The latest set of long-range model runs go along with this idea, and I'd be surprised if 98L threatens any land areas.

Dismal Swamp fire creating dangerous air pollution in Virginia
Lightning from a thunderstorm on August 4 sparked a fire in Southeast Virginia's Dismal Swamp, which continues to burn out of control. Yesterday, air quality alerts for Code Purple pollution--the worst category of air pollution--were posted for Suffolk, Virginia and continue today. The region, including the cities of Norfolk and Hampton Roads, have seen an increase in hospital admissions for people with breathing problems, plus an increase in traffic accidents due to low visibility conditions on area roads. The fire has burned 6100 acres and is 15% contained. Given that it is burning more than 1 foot underground, it will be difficult to put out unless heavy rains raise the water table. The region is under "Abnormally dry" drought conditions, the lowest category of drought on the five-category drought scale.

Jeff Masters