Above: If the 06Z Sunday forecast from the GFS model were to verify exactly—with Florence centered in far eastern NC and far southeast VA from Friday through Monday, September 14-17—these are the rainfall amounts that would result, as predicted by the model. Widespread 10”+ amounts would extend from Cape Hatteras to the Washington, D.C., area; the Norfolk/Hampton Roads area of southeast VA would receive more than 20”. This is just one model solution: actual rainfall amounts at any point could be far less (or even greater) depending on Florence's actual track and speed. Image credit: tropicaltidbits.com. |
Florence has re-intensified into a hurricane, and is expected to rapidly intensify into a very dangerous Category 4 hurricane by Tuesday as it heads towards the Southeast U.S. Coast. Florence is likely to make landfall on Thursday evening or Friday morning on the Southeast U.S. East Coast, or remain just offshore. The odds have increased that Florence will stall on Friday and meander near or over the coast for several days, making the hurricane a huge rainfall and coastal flooding threat.
Florence was about 750 miles southeast of Bermuda late Sunday morning, moving west at 6 mph. Sea surface temperatures (SSTs) were a warm 28.5°C (83°F), but Florence was embedded in an atmosphere with somewhat dry air (a mid-level relative humidity of 50%). The high wind shear of 20 – 25 knots that had been driving this dry air into the core of the storm had abated significantly on Sunday morning, and was a light 5 - 10 knots. Satellite images on Sunday afternoon showed that the storm was significantly more organized, with a prominent eye, a more symmetrical shape, and impressive spiral banding. Florence was a medium-sized hurricane, with tropical storm-force winds that extended out up to 115 miles from the center. A NOAA hurricane hunter aircraft flying out of Bermuda on Sunday morning found 75 mph surface winds. No other hurricane hunter missions are scheduled for today; regular hurricane hunter missions by the Air Force will begin on Monday. The NOAA jet will fly another dropsonde mission on Sunday night.
Figure 1. Visible-wavelength satellite image of Hurricane Florence at 1450Z (10:50 am EDT) Sunday, September 9, 2018. Image credit: tropicaltidbits.com. |
Intensity forecast for Florence: expect a Cat 4 by Tuesday
Florence’s environment has grown much more conducive for intensification, so much so that over the past day, NHC has made their most aggressive rapid intensification forecasts since at least 1998 for an Atlantic storm, according to an analysis by Sam Lillo. The SHIPS model predicts shear will remain low through Wednesday night. SSTs will gradually increase to 29°C (84°F) during this period, and ocean heat content will more than double, from 20 to nearly 50 kilojoules per square centimeter. Our top intensity models unanimously predict strengthening of Florence into a Category 3 or 4 hurricane by Tuesday, and the storm is also expected to increase in size. Florence will still be embedded in a relatively dry atmosphere, so it is possible the storm could suffer from dry air intrusions that would interfere with the intensification process, but this will likely not occur while the wind shear is light.
Figure 2. HWRF model winds and pressure forecast for 11 pm EDT Thursday, September 13, from the 6Z Sunday run of the model. The HWRF model was our top intensity model from 2017, and predicted that Florence would make landfall near Wilmington, North Carolina as a Category 3 hurricane. Image credit: Levi Cowan, tropicaltidbits.com. |
Big waves from Florence
Florence’s power and expanding size are generating swells that were affecting Bermuda on Sunday, with waves of 3 - 5 feet in the outside waters. These waves are predicted to ipeak at 5 – 15 feet for Tuesday and Wednesday for Bermuda. There are no watches for Bermuda at this time, and the island is well north of Florence’s cone of uncertainty.
Large swells from Florence will begin reaching portions of the U.S. East Coast by Sunday night, and also propagate to north and northeastward-facing coasts of the Lesser Antilles, Puerto Rico, Hispañola, the Turks and Caicos and the Bahamas into this weekend. These swells will produce life-threatening surf and rip current conditions at beaches.
Invest 94L not likely to impact Florence
An elongated area of low pressure that developed in the waters a few hundred miles west of Bermuda on Friday, which was designated 94L by NHC, has lost most of its spin, and Sunday morning satellite imagery and Bermuda radar showed 94L was just a smallish clump of heavy thunderstorms located a few hundred miles southeast of Bermuda. In their 8 am EDT Sunday Tropical Weather Outlook, NHC no longer mentioned 94L.
As Florence passes to the south of the remnants of 94L on Monday, the hurricane will likely absorb what is left of 94L, resulting in a small boost to Florence’s size and intensity.
Figure 3. The 0Z Sunday, September 9, 2018 track forecasts (left) and 0Z Saturday, September 8, 2018 track forecast (right) by the operational European model for Florence (red line, adjusted by CFAN using a proprietary technique that accounts for storm movement since the time of the model run), along with the track of the average of the 50 members of the European model ensemble (heavy black line), and the track forecasts from the “high probability cluster” (grey lines)—the four European model ensemble members that have performed best with Florence thus far. The forecasts show that the uncertainty in the European model forecast has shrunk considerably, which is likely due to the inclusion of data from last night’s dropsonde mission by the NOAA jet. All of these forecasts predict a U.S. East Coast landfall. Image credit: CFAN. |
The model uncertainty for Florence’s future track has shrunk considerably since yesterday, thanks, in part, to the dropsonde data from last night’s mission by the NOAA jet. Of the five major models we use for tracking hurricanes—the European, GFS, UKMET, HWRF, and HMON—the best-case scenario from their 0Z Sunday runs was given by the UKMET model, which was the only model to predict that Florence would miss making landfall on the U.S. East Coast this week. Even that forecast, if it came true, might be dire for North Carolina, since the model predicted that storm would stall out for several days just offshore, dumping massive amounts of rain over coastal North Carolina. The 12Z Sunday GFS model run also showed this scenario, but the 12Z Sunday UKMET model now has landfall occurring in North Carolina.
Track forecast for Florence: very dangerous for the U.S. East Coast
Florence is being steered by the clockwise flow around the Bermuda High, the semi-permanent high-pressure system centered over the North Atlantic Ocean. The edge of the Bermuda High will lie along the U.S. East Coast during the middle of next week, and if and when Florence rounds the southwestern flank of the Bermuda High, the hurricane would turn to the northwest and then north. Thus, the exact position and orientation of the southwestern flank of the Bermuda High, in combination with the strength of the trough of low pressure that will be just to its west, will be the keys to the timing and location of Florence’s turn. Since this turn is still four days away, there is substantial model disagreement on this, as the average error in a 4-day official NHC forecast in recent years has been about 150 miles. A direct hit from Florence could happen anywhere along the swath of coast from Georgia to Massachusetts, with North Carolina being at greatest risk.
The main scenarios as I see them, with some rough odds as to their likelihood:
1) Florence will continue heading west-northwest to northwest until landfall occurs between Charleston, South Carolina and Cape Hatteras, North Carolina (40% chance). Recent operational runs of the European and GFS models, along with many of their ensemble members, have championed this scenario. Of our top five models for predicting hurricane tracks, the GFS model has been the best-performing model for Florence’s long range 4 – 5 day track forecasts thus far, with errors less than half of that of the worst-performing model, the European (see stats on model performance at Brian Tang’s excellent website here). This is switch from last year’s forecasts for Hurricane Irma, for which the European model performed much better than the GFS model.
2) The Bermuda High intensifies as Florence approaches the coast, resulting in a more westerly motion, with a landfall in Georgia or southern South Carolina (10% chance). A number of members of the European ensemble forecast have suggested this outcome.
3) Florence turns to the northwest or north just off the coast of North Carolina, making landfall in the mid-Atlantic or New England between Virginia and Massachusetts on Sunday (20% chance). Many members of the GFS ensemble model forecasts have made this forecast.
4) Florence will turn to the north or stall just off the coast of North Carolina, then head northeast out to sea, without making any landfall (30% chance). The UKMET model and members of the GFS and European model ensemble forecasts have made this forecast.
Figure 4. Analyzed departure of precipitation from the 1981-2010 average for the 60-day period ending at 8 am EDT Sunday, September 9, 2018. The totals are calculated from a multi-sensor blend of rain-gauge, radar, and satellite reports. Parts of coastal NC/VA and a large swath from the Washington-Baltimore area through central Pennsylvania have received 8" to 12" more rainfall than normal over the last two months, increasing the flood danger from rains that may fall from Florence. Image credit: NOAA/NWS Advanced Hydrologic Prediction Service. |
Florence: the “Harvey of the East Coast”?
The models are increasingly on board with a scenario where the trough of low pressure that was expected to turn the hurricane to the north late this week will be too weak to do so, as a strong ridge of high pressure builds over the Mid-Atlantic. This “blocking ridge” would block Florence’s forward progress. Florence might then stall and wander near or over the coast for as many as six days, becoming the “Harvey of the East Coast”, dumping prodigious amounts of rain. The analogue would be especially strong if Florence slams inland as a major hurricane with a first round of impacts—as did Harvey near Rockport, Texas—and then stalls to produce a multiday flood disaster, as Harvey ended up doing in southeast TX.
Figure 5. Total soil moisture anomaly (departure from the 1971-2010 average, in millimeters) for the period from August 28 to September 4, 2018. Moist soils over the Mid-Atlantic will increase the liklihood of flooding from Florence. Image credit: NWS/NOAA/CPC. |
A soggy late summer has set the stage for major flood concerns. Much of the period since mid-July has been exceptionally wet over parts of the mid-Atlantic. The cities of Hatteras, NC; Washington, DC (Reagan and Dulles airports); and Wilkes-Barre and Williamsport, PA, had their wettest summers on record. For the week ending Tuesday, 4 September, soil moisture was running up to 100 millimeters (4 inches) above normal over large parts of the mid-Atlantic (see Figure 5). Preexisting saturated soils greatly exacerbate the risk of flooding, although they are not an absolute requirement. A weather.com feature notes that, for the period 1962–2012, rainfall was the culprit behind 27% of all tropical cyclones fatalities in the U.S. from 1963 to 2012.
If you live in a hurricane-prone location, now is a good time to make sure you have a preparedness plan in place!
Our next update today will be on the four other serious tropical cyclone threats in the Atlantic and Pacific.
Bob Henson contributed to this post.