First Look at 2016 Hurricane Season: Unusually Big Question Marks

Even more than one might expect, the 2016 Atlantic hurricane season is shaping up with both high- and low-end possibilities, based on preliminary thoughts released Thursday by a team at Colorado State University. The CSU project, founded by Dr. William Gray with Dr. Phil Klotzbach now serving as lead author, has issued seasonal outlooks since 1984 for the anticipated amount of hurricane activity in the Atlantic basin. In 1993 the group began issuing outlooks each December for the following year’s activity, but in 2011 those outlooks were dropped due to lack of any demonstrated skill. In its place, the CSU team now releases what they call a “qualitative discussion,” which highlights the factors at play and includes a more generalized sense of what we might expect.

Although it’s not presented as a quantitive outlook, the Thursday release (PDF) does include a set of four potential scenarios for 2016, each rated in terms of probabilities that we will see a given amount of seasonally-averaged accumulated cyclone energy (ACE). These scenarios hinge on two factors: how quickly the current El Niño will diminish, and how the Atlantic Multidecadal Oscillation/thermohaline circulation (AMO/THC) will evolve.

1. AMO/THC becomes above average in 2016 and no El Niño impacts remain (resulting in an ACE of ~ 170) – 25% chance
2. AMO/THC is above average in 2016 but some El Niño impacts remain (ACE ~ 120) – 35% chance
3. AMO/THC is below average and no El Niño impacts remain (ACE ~ 80) – 20% chance
4. AMO/THC is below average and some El Niño impacts remain (ACE ~ 50) – 20% chance

CSU relates the ACE values shown above to these general ranges of activity:

170 ACE – 14-17 named storms, 9-11 hurricanes, 4-5 major hurricanes
120 ACE – 12-15 named storms, 6-8 hurricanes, 2-3 major hurricanes
80 ACE – 8-11 named storms, 3-5 hurricanes, 1-2 major hurricanes
50 ACE – 5-7 named storms, 2-3 hurricanes, 0-1 major hurricane

The upshot is that we have a 45% chance of falling into one of the two more extreme scenarios: either #1 (a very busy season) or #4 (a very quiet season). Looking back at each of the last four years, CSU placed the combined odds of the highest- and lowest-end scenarios at only 20% to 25%. So if CSU is right, the Atlantic season of 2016 has an considerably better chance than the last four seasons of being either unusually active or uncommonly tranquil. Want more evidence for this split verdict? CSU points out the wildly contrasting outcomes for the Atlantic hurricane seasons that followed our two other “super El Niño” events since 1950 (1982-83 and 1997-98):

1983: Atlantic ACE = 17% of average
1998: Atlantic ACE = 182% of average

This stark difference is largely because the 1982-83 El Niño decayed gradually, holding on till midsummer, whereas the 1997-88 El Niño quickly shifted into a moderate La Niña by late summer. The most recent monthly outlook for El Niño, issued Thursday by NOAA, calls for a transition to ENSO-neutral conditions by late spring or early summer. Computer models are in agreement on this shift, though the timing varies somewhat. More often than not, a strong or very strong El Niño event is followed by La Niña in the next summer or autumn. It’s certainly possible that La Niña could be in place by late summer (see Figure 1 below), which would favor an active Atlantic hurricane season.


Figure 1. The December outlook issued by a collaborative group of forecasters at NOAA’s Climate Prediction Center and the International Research Institute for Climate and Society (IRI). The odds of neutral or La Niña conditions rise considerably by next summer. Image credit: NOAA/IRI.

How do El Niño and the AMO/TCH influence hurricanes?
In a relationship that’s well understood, El Niño tends to suppress Atlantic hurricane activity, mainly by enhancing upper-level wind shear, while La Niña favors more Atlantic activity by reducing wind shear. These factors come and go from year to year. Meanwhile, the ups and downs of the AMO/THC modulate hurricane activity over much longer periods by affecting sea-surface temperatures and other conditions over the tropical and subtropical Atlantic. “These changes are natural and have been occurring for at least the last 1,000 years,” notes a NOAA FAQ on the phenomenon.

The AMO/THC has shown large multidecadal trends over the last century, favoring reduced hurricane activity from the 1970s to the mid-1990s and increased activity from that point until the early 2010s. Over the last couple of years, the AMO/THC has slowed down, in tandem with a reduction in Atlantic hurricane activity. However, the AMO/THC can vary on the shorter term as well, so it’s not yet certain that we have entered a new multi-decade era of reduced activity--although Phil Klotzbach made the case for this in a paper published in Science this past September and summarized in a writeup by Klotzbach at Capital Weather Gang. (The Science paper can be accessed through a link at the bottom of the CWG article.) Even if we are indeed in a new era, shorter-term factors could strengthen the AMO/THC during a given year. This includes the train of weather features and oceanic effects that El Niño is expected to produce over the next few months from the Pacific to the Atlantic.

“The uncertainties related to the AMO for 2016 are enormous,” said Eric Blake (National Hurricane Center) in an email, “because we lack a reliable way to track the feature and because of possible effects of El Niño countering the longer-term cycle.”

CSU will issue its full seasonal outlooks for Atlantic hurricane activity in 2016 on April 14, June 1, and August 3. As one would expect, the skill of these outlooks steadily improves as the hurricane season nears. Even if it’s too soon right now to expect an accurate forecast for 2016, the latest thoughts from CSU make me even more eager to see how this very uncertain hurricane season will unfold.

Bob Henson


Figure 2. The largest patch of below-average ocean temperatures for 2015 (January - October) has been in the North Atlantic, where record-cold values have been observed southeast of Greenland. These values are consistent with a weaker Atlantic multidecadal oscillation/thermohaline circulation. Image credit: NOAA/NCEI.