Hurricane Arthur formed off the southeast U.S. coast in July 2014 and made landfall in coastal North Carolina. (NASA)
Hurricane Arthur formed off the southeast U.S. coast in July 2014 and made landfall in coastal North Carolina. (NASA)

The start of the 2016 Atlantic hurricane season is just three weeks away, and initial seasonal hurricane predictions have ranged from near-average (by Colorado State University, Tropical Storm Risk and several private groups) to above-average (by AccuWeather and the Weather Company) activity.

Some meteorologists are pointing to warmer-than-normal sea surface temperatures (SSTs) near the East Coast to advertise an increased hurricane threat to the U.S. coastline. Our analysis does not support this argument. We actually find the risk of landfalling hurricanes is highest when temperatures near the East Coast are cooler.


April SST anomalies across the North Atlantic Ocean. The black box in the figure highlights our “close-in” region. (National Oceanic and Atmospheric Administration)

To investigate this issue in more detail, we examined the number of named storms, number of hurricanes and Accumulated Cyclone Energy that were generated in a broad “close-in” region off the East Coast (20 to 40 degrees north, 80 to 60 degrees west, as highlighted by the black box in the above figure) and explored their relationship with SSTs in the same region.

(One important caveat before moving ahead is that just because SSTs are warm off the East Coast now does not guarantee that they will remain warm in the box for the peak of the Atlantic hurricane season. The correlation between April SSTs and August to October SSTs in our close-in region is 0.52, which means that more than 70 percent of the variability in August to October SSTs is not predicted by April SSTs in the same region. In other words, what you see now is historically only a modest predictor of what we’ll see during the peak of the hurricane season.)

Warm waters do seem to favor more close-in storm development or “home-grown storms.” In the 10 warmest seasons since 1950, there were 39 home-grown storms, compared with only 20 in the 10 coldest years.

Tropical cyclone activity close-in to the United States East Coast for the 10 warmest vs. 10 coldest Augusts-Octobers since 1950.
Tropical cyclone activity close-in to the United States East Coast for the 10 warmest vs. 10 coldest Augusts-Octobers since 1950.

The relationship held with hurricanes but was a bit weaker: 20 home-grown hurricanes developed in the 10 warmest seasons compared with 14 in the 10 coldest seasons.

The relationship between home-grown storms and ocean temperature tends to break down, however, when we look at all seasons from 1950-2015, not just those with much warmer- and colder-than-normal waters off the East Coast. This is demonstrated in the scatter plot below comparing close-in average SSTs with home-grown tropical cyclones.


Scatterplot of TC formations compared with close-in SSTs. Very little relationship is seen.

Several of the most active home-grown storm seasons had below-average SSTs, while several of the warmest seasons had normal to below-normal tropical cyclone activity.

(The correlation between the two time series is only 0.17, which means that over 95 percent of the variance in tropical cyclone formation off the East Coast is explained by other factors. If there were a significant relationship, the dots would roughly line up from the lower-left to the upper-right.)

While warmer-than-normal temperatures off the East Coast show some weak correlation with increased home-grown storms, when we consider both home-grown storms and those arriving from the deep tropics, cooler close-in waters actually tend to result in more tropical cyclone activity overall. We see this when we compare the Accumulated Cyclone Energy (ACE) in the 10 coldest seasons near the East Coast with the 10 warmest seasons (see the ACE comparison on the bar chart above).

Why would warmer SSTs lead to less ACE being generated off the East Coast?

ACE tends to pile up in major (Category 3 to 5 on the Saffir-Simpson wind scale) hurricane activity. Major hurricane activity in the Atlantic basin typically originates in the deep tropics, and deep tropical major hurricane activity is more common when the Atlantic Multi-Decadal Oscillation (AMO) is positive. But when the AMO is positive, temperatures off the East Coast tend to be cool and vice versa.

For example, over the past few years, the AMO has been tending negative, reflected by colder than normal water in the far North Atlantic. Associated with this colder water is a stronger-than-normal subtropical high, which drives above-average SSTs off of the East Coast. This same stronger-than-normal subtropical high forces subsidence (sinking and drying motion) over the tropical Atlantic, suppressing deep tropical tropical cyclone formation.

Consequently, fewer hurricanes with deep tropical influence tend to affect the East Coast when the AMO is negative and water just off the coast is warmer.


SST differences from August through October (2013 through 2015) minus August through October (1995 through 2012). (NOAA)

We see this concept that more significant hurricane activity (from both home-grown storms and those from the deep tropics) tends to affect the East Coast when close-in waters tend cool, when we analyze the record of landfalling Atlantic hurricanes.

The 10 warmest close-in SST seasons had seven hurricanes and one major hurricane impact the East Coast, while the 10 coldest seasons had 11 hurricanes and three major hurricanes. In other words, historically, the East Coast has been affected by more hurricanes and major hurricanes when “close-in” SSTs during the peak of the season are cooler than average.


Tracks of the seven hurricanes impacting the US East Coast in the 10 warmest close-in SST seasons compared with the eleven hurricanes impacting the US East Coast in the 10 coldest close-in SST seasons.

Overall, our conclusion is that while above-average close-in SSTs may favor weak tropical cyclone formation off the East Coast, it may actually decrease the odds of the region being impacted by hurricane activity.

As we always point out, though it is interesting and helpful to look at climatology, statistical relationships, and probabilities in nature, a single major hurricane landfall on the East Coast this season would be a significant event. A lower-than-average probability does not mean zero probability!