Atlantic Hurricane Season Forecast Update for August 2018


Outlook Overview

  • WeatherTiger’s August update to our 2018 hurricane season presents only minor changes from our initial forecast issued in late May for overall activity of between two-thirds and three-quarters of an average year. Cooler than average waters in key regions of the Atlantic, a developing El Niño in the Pacific, and much stronger than average trade winds in the Caribbean this summer confirm these signals.
  • U.S. landfall risks are also diminished relative to normal, though one or more U.S. hurricane landfalls are still possible. Analog years suggest a much quieter than normal Cape Verde season, particularly reducing U.S. major hurricane landfall odds.

Updated Seasonal Outlook for Tropical Cyclone Activity
As the Atlantic remains only a whisper louder than silent today, early August is a good time to update WeatherTiger’s outlook for the remainder of the hurricane season. In our initial forecast issued in late May, we called for a generally below average amount of Accumulated Cyclone Energy (ACE) in 2018, driven by cooler than average waters in the tropical Atlantic and an El Niño developing in the fall. Since then, there have been no alarms and no surprises in how the best predictors have evolved. With few major changes to our reasoning since May, we’ll try to keep this update short, as the hurricane season itself will hopefully be as well.

It is true that the first two months of the season have been above normal, or more active than about four out of every five years. However, it is important to remember that as of August 8th, only 10% of the season behind us in ACE terms, and there is no relationship between how active May, June, and July are, and how the rest of the year will proceed. Above, the 14 most similar seasons to 2018 through 8/7 are plotted in red over the rest of the season. Six of these years finished above normal, four near normal, and four below normal—a nearly random distribution. August TC has a relationship with the remainder of the season, but so far, August has been mostly quiet.

Much stronger indications of future activity are found in the winds and waters of the Atlantic and Pacific. These relationships are more significant when looking at aggregate summer conditions than spring, so forecast confidence usually is much increased in early August relative to late May. One particularly useful relationship is with the strength of the low-level summer trade winds that blow from east to west across the tropical Atlantic. The faster these winds, the more difficult it is for tropical storms to organize in the August-September peak of the Cape Verde season as they zip along with the steering currents. In the plot at right, the region of significance between July trades and future ACE is highlighted in black in the top panel. In July 2018, the deep blue shading in that contour in the lower panel means these easterly trades were much stronger than average, or highly unfavorable to an active season.

Sea surface temperatures over the summer also showed a pattern likely to keep these unfavorable winds mostly in place through August and September, as well as to diminish the available heat potential to TCs. The tripolar “sandwich” of cold/warm/cold SST anomalies observed in the Atlantic in July 2018 is a classic manifestation of a negative Atlantic Meridional Mode, again linked to relatively inactive Atlantic hurricane seasons. Our analog-matching algorithm has consistently picked very quiet seasons as the best matches for 2018 conditions this summer, with top Atlantic SST analogs ranging from 30-65 additional ACE units onward from mid-August, as well as being mostly bereft of high-impact U.S. landfalls.

Focusing more directly on landfall risks, the strongest predictor of how much ACE will occur annually near or over the continental U.S. (since 1900) is wind shear over the eastern Pacific Ocean and to a lesser extent, the Caribbean Sea. At left, more shear in the eastern Pacific in July is significantly linked to fewer and weaker U.S. TC landfalls in August-October. This is basically a proxy for effective El Niño or La Niña amplitude over the Tropics. This year, as the Pacific heads towards weak or moderate El Niño conditions by the end of fall, Pacific wind shear is much diminished (blue) relative to normal, indicative of lower than normal chances of U.S. hurricane landfalls in the 2018 season.

Overall, I’m happy to report that these key factors and others have evolved just how we thought they would over the last two months, so our August projections are little changed from those issued in May. These predictors were assimilated into WeatherTiger’s proprietary seasonal forecasting algorithm, resulting in a calibrated projection of seasonal totals of 72.5 ACE units, 12 tropical storms, 6 hurricanes, and 2 major hurricanes, including the 16 units of ACE, 4 tropical storms, and 2 hurricanes that have already occurred. This is generally between two-thirds and three-quarters of the average amount of tropical cyclone activity since 1950. As shown above, replicated real-time skill increases markedly when using July predictors, and our forecasts for August onward hurricane count have roughly a 60% decrease in forecast error relative to climatology over 1984-2017.

Still, caution is warranted here, as within the last ten days the Main Development Region of the Atlantic has warmed and is now less sharply cooler than normal. Coupled with a relatively slow evolution of a Central Pacific-based El Niño, that trend certainly needs to be watched. Of course, it always is worth remembering that even inactive seasons can yield individual storms that are serious threats to life and property, and this year may present particular risks of storms waiting to develop until they are closer to the U.S. coastline. One year that jumps out of the analog set as evincing this tendency is 2002 (at left, below), in which six tropical storm and one U.S. hurricane landfall (Lili) occurred after August 15th, while the eastern Atlantic remained basically dead. While much of this close-in TC activity was of the rainy nuisance variety, Lili very nearly notched a major hurricane landfall in Louisiana before falling apart at the very last second.

Overall, our objective algorithmic guidance for U.S. landfall risks is of a piece with our overall activity forecast. WeatherTiger projects 1-2 additional tropical storm landfalls, a 45% chance of one or more hurricane landfalls, and a 15-20% chance of a major hurricane landfall for the remainder of the 2018 hurricane season. Compare those odds to the climatological 75% and 35% odds of one or more hurricane/major hurricane U.S. landfalls, respectively, after August 15th: this year’s risks are not negligible, but certainly lower than average across a broad slate of skillful predictive metrics. Nevertheless, as we head into the most active ten weeks of hurricane season, we’ll certainly keep watching the skies, and recommend that you do so as well.

Forecaster: Dr. Ryan Truchelut (

Like what you read? Get every data- and science-driven WeatherTiger weekly tropical outlook and individual storm forecast direct to your inbox as soon as it is released. Sign up now for our tropical subscription package to get all of our expert, hype/drama-free forecasts sent straight to you, plus a free one-on-one consultation with Dr. Truchelut. You can also follow us on Twitter and Facebook for the latest as the 2018 season rolls on.

A downloadable PDF of this forecast is available here. Redistribution of this forecast is permitted with proper attribution.

Like WeatherTiger's tropical coverage? What if I told you... you could get it in your inbox? Get every data-driven WeatherTiger weekly tropical outlook and individual storm forecast for the 2020 hurricane season as soon as it is released. Sign up now for our hurricane forecasting subscription package to get all our forecasts sent straight to you, plus tons of subscriber exclusives.