Matthew Vaults to Category 4 Strength; Uncertainty Clouds the Forecast

The spectacular three-day strengthening of Hurricane Matthew continued on Friday afternoon. Not yet even classified as a depression on Wednesday morning, Matthew was packing Category 4 winds of 140 mph as of the 5 pm EDT Friday advisory from the National Hurricane Center. Satellite imagery on Friday put a spotlight on Matthew’s rapid intensification, as a 15- to 20-mile-wide eye developed during the day (see Figure 1 below). Air Force Hurricane Hunter flights on Friday afternoon found peak flight-level winds of 117 knots (135 mph) around 2:20 pm EDT. Top near-surface winds, estimated by radiometer outside of rain-contaminated areas, were 111 knots (128 mph) around 3:45 pm EDT.

Hurricane Matthew is already being blamed for two deaths. One occurred on St. Vincent on Thursday, when a teenager was trapped against a house by a boulder that was dislodged by the storm. The other came on Friday in Colombia, where a drowning death in a rain-swollen river occurred in Uribia, La Guajira.


Figure 1. Enhanced infrared satellite image for Tropical Storm Matthew as of 2015Z (4:15 pm EDT) Friday, September 30, 2016. Image credit: NOAA/NESDIS.

How did Matthew get so strong so quickly?
Vertical wind shear of up to 20 knots has plagued Matthew for most of the last two days, yet the storm has not only maintained its structure but grown at a ferocious rate. Dissertations may be written on how this happened! Working in Matthew’s favor has been a steadily moistening atmosphere along its westward path, which means that the shearing winds didn’t push too much dry air into Matthew. Once it developed a central core, Matthew was able to fend off the wind shear much more effectively. In addition, water temperatures are unusually warm throughout the Caribbean (and the entire western North Atlantic), with an area of high oceanic heat content directly beneath Matthew’s path. Such deep oceanic heat allows a storm to strengthen without churning up cooler waters from below that could blunt the intensification.

Matthew’s ascent highlights the nagging challenge of predicting hurricane intensity. NHC statistics for the past few years show a steady improvement in track forecasts and much more erratic progress in intensity forecasts (see Figure 2 below). The typical 48-hour track error has been cut in half since the late 1990s, dropping from around 150 nautical miles (170 miles) to around 75 nautical miles today. Meanwhile, the 48-hour intensity forecast error has averaged about 12 knots (15 mph) in the last several years, which is not much better than the 15-knot errors that were typical in the mid-to-late 1990s. Much of that error is the result of just a few rapidly intensifying storms, such as Matthew.


Figure 2. Trends in track and intensity forecasts from the National Hurricane Center for Atlantic hurricanes through 2015. Units are nautical miles (left) and knots (right); add 15% to obtain miles and miles per hour. Image credit: NHC.

Although Matthew strengthened far more quickly than projected in the official outlook--and expected by most observers--there were signs that rapid intensification was possible, as we discussed on Wednesday afternoon. The 18Z (2:00 pm) Wednesday run of the SHIPS statistical model included a 44% chance that Matthew’s strength would increase by 55 knots in 48 hours. In fact, this is exactly how quickly Matthew intensified: from 50 knots (60 mph) at 18Z Wednesday to 105 knots (120 mph) at 2:00 pm Friday. SHIPS is only one tool used by forecasters to assess potential intensity change. Dynamical forecast models were generally less gung-ho on rapid intensification on Wednesday, and even subsequent SHIPS runs pulled back a bit.


Figure 3. Radar image of Hurricane Matthew as seen from NOAA’s P-3 Orion hurricane hunter aircraft at 1:40 pm EDT September 30, 2016. At the time, Matthew was a rapidly intensifying Category 3 hurricane with 120 mph winds. Image credit: NOAA/AOML/HRD, via tropicalatlantic.com

Figure 4. Visible satellite image of Matthew (1-kilometer resolution) at 1945Z (3:45 pm EDT) Friday, September 30, 2016. The Colombian and Venezuelan coastlines are outlined in yellow. Image credit: CSU/RAMMB/CIRA.


Matthew poses a major threat to Jamaica
Matthew is moving just south of due west at 9 mph. Its location about 75 miles north of Punta Gallinas, Colombia, puts it about as close to South America as any major hurricane is known to have gotten (even about 50 miles closer than 2004’s Hurricane Ivan).

Although Matthew’s westward track will keep it offshore of Colombia and Venezuela, a Tropical Storm Warning is in effect for the Colombian coast from the Venezuelan border west to Riohacha. The Columbian coast will remain on the less-intense left-hand side of Matthew, reducing the odds of hurricane-force winds and limiting the heaviest rains. Riohacha’s Almirante Padilla reported sustained winds of less than 20 mph on Friday afternoon.

Models agree strongly that Matthew will begin taking a fairly sharp right turn on Saturday, heading north-northwest through the central Caribbean. Conditions will be even more favorable for strengthening by this point. Wind shear is projected to drop dramatically (perhaps below 10 knots by Sunday), the deep atmosphere will moisten further (close to 80% relative humidity), and Matthew will be passing over waters with extremely high oceanic heat content. Now that Matthew is a major hurricane, we can expect ups and downs in its strength from day to day as internal processes such as eyewall replacement cycles kick in, but we can’t rule out the possibility Matthew will become a Category 5 while in the Caribbean.

The threat to the Greater Antilles from Matthew is becoming increasingly worrisome. The most immediate concern is for Jamaica, where a Hurricane Watch has been posted. The latest NHC outlook brings Matthew over the eastern tip of Jamaica on Monday afternoon. A westward shift of just 50 miles--well within the range of uncertainty at this point--would put the city of Kingston in Matthew’s dangerous right-hand side. A major hurricane striking Jamaica from the south would be a virtually unprecented event. Figure 5 shows the tracks of all major hurricanes passing over or very near Jamaica since 1851. All of the prior events involved storms tracking on a classic west-northwest path except for an unnamed 1912 hurricane that crossed the northwest tip of the island on a northeast path, then made a 180-degree turn. Among all hurricanes since 1851 (not shown), the only one to have crossed Jamaica on a primarily northward track during the last 80 years is Sandy (2012), which struck eastern Jamaica at Category 2 strength. Sandy caused an estimated $100 million in damage in Jamaica and knocked out power to most of the island. Matthew could be much stronger than Sandy, and a northward-oriented path through central Jamaica could bring a severe storm surge into the highly vulnerable Kingston area.

Next in Matthew’s sights on the NHC-predicted track would be Cuba, whose excellent history of hurricane awareness and preparation would likely reduce potential impacts. Matthew may also be weakened by any direct passage over mountainous Jamaica, although it could easily strike Cuba as a major hurricane.

If Matthew were to trend eastward rather than westward, the risk to western Haiti will rise dramatically. Model guidance has trended gradually west over time, which gives some hope that Haiti will escape the worst of Matthew. Still recovering from the devastation of the 2010 earthquake, and plagued with deforestation and poverty, Haiti would be highly vulnerable to the impacts of a major hurricane.


Figure 5. Tracks of all major hurricanes passing within the shaded circle encompassing Jamaica during the period from 1851 to 2015. Each of these was moving from right to left (east to west), except for the 1912 hurricane, which moved east-northeast and then backtracked toward the west-southwest as it weakened. Image credit: NOAA.


Figure 6. WU depiction of National Hurricane Center track and intensity forecast for Matthew as of 5:00 pm EDT Friday, September 30, 2016.

Long-term outlook for Matthew
If anything, the prospects for Matthew later next week have become more uncertain over time. Models continue to take Matthew north through The Bahamas, but then we have major divergence among our top models. As just one example, the 12Z Friday operational run of the GFS model pulls Matthew almost due north, slamming it into eastern Maine as a significant hurricane or very intense post-tropical storm by next weekend. In stark contrast, the 12Z run of the ECMWF model strands Matthew in the Bahamas, where it lingers through next weekend and into the following week as a major hurricane. The 12Z run of the UKMET, our other top track model, also stalls Matthew in the Bahamas, then angles it northwest toward the Southeast U.S. coast.

Why such profound disagreement? The simplest explanation is that track errors increase markedly over time, and there is little skill beyond about 5-7 days. In this case, there is a great deal of uncertainty over how the mid-latitude steering features over the United States and the western Atlantic will evolve over the next week. NOAA’s Gulfstream-IV jet has been flying regular missions to sample the environment around Matthew, which has likely led to improvements in the short-term track forecasts. The problem is that the upper-level trough that will be a key influence on Matthew’s track next week is still thousands of miles away--moving through the northeast Pacific, where observations are scarce. It is far too soon to know with confidence how the upper-level features will evolve next week, so we need to keep our expectations very modest for confidence in any East Coast forecast.

The bottom line:

--Matthew poses a very serious risk to the western Greater Antilles early next week.

--A trek over the mountainous terrain of Jamaica, Cuba, and/or Haiti would dramatically weaken Matthew. At least some restrengthening would be possible over the Bahamas.

--Matthew could affect any part of the U.S. East Coast from Florida to Maine at some point from the middle of next week into the weekend.

--Long-range forecasts will vary, perhaps several times each day. Because the key features that will steer Matthew are very uncertain at this point, any given model shift may not mean much until the evolution of these features becomes better defined, which could take several days.

Jeff Masters and I will be posting regular updates through the weekend, typically between 10 am and noon EDT and between 6 and 8 pm EDT. For those new to our blog, the comments section is packed with valuable insights from our many members, including meteorologists as well as dedicated laypeople. Wunderblogger Steve Gregory posted a Friday afternoon update on Matthew, MAJOR HURRICANE MATTHEW – FLORIDA TO MAINE ON ALERT (INTERIM UPDATE).

Bob Henson


Figure 7. The 70 forecasts from the 12Z Friday European (ECMWF) model ensemble (left) and 18Z GFS model ensemble (right) continued to show a wide variety of solutions for the track of Matthew that pose various threats along the U.S. East Coast. (The tracks from the ECMWF that previously targeted the Gulf Coast have almost completely disappeared.) Image credit: Climate Forecast Applications Network (CFAN).