As we wrap up September, there have been just two short-lived Category 1 hurricanes in the Atlantic. Yet seasonal forecasts predicted an extremely active season. What’s going on?
Before diving into the seasonal forecasts, let’s take inventory on where the season stands.
In an average season, 8 tropical storms, 4 hurricanes, and 1 major (category 3 or higher) hurricane form by this date. This year, we’ve experienced 10 tropical storms, 2 hurricanes, and no major hurricanes.
Though we’ve had close to the average number of total storms, most have been short-lived and/or weak. If you went out for a cup of coffee at any time this hurricane season, you would’ve missed many of them.
Aside from tropical storm Andrea’s modest impact in the Southeast, none of these storms has made landfall in the U.S.
Total storm energy *much* below normal
I have referenced ACE (Accumulated Cyclone Energy) as a common measure of seasonal tropical activity before, but as a refresher, it’s the sum of the squares of all of the storms’ peak wind speeds at 6-hourly intervals. It is proportional to the kinetic energy of a storm, based on its peak wind value (not the size of the storm or distribution of its winds).
(A storm can achieve high ACE values if it lives a long life and/or is intense. For example, last year, long-lived but relatively weak Nadine meandered in the tropics for 21 days and racked up an ACE of 26.3, while Katia in 2011 was only around for 11.5 days but was very intense – reaching category 4 intensity – and ended up with an ACE of 27.)
This year’s ACE is 23.1, 28 percent of average for this date or around the average on August 27, prior to the peak of hurricane season, which has now passed. Of all storms this year, Humberto leads the pack contributing a lowly ACE of 8.3 to the total.
The low activity so far this year is not unprecedented, but unusual.
According to meteorologist Ryan Maue’s Web site, only four other years have had lower ACE totals as of this date (since 1950): 1962, 1977, 1983, and 1994. The highest end-of-season ACE among those years is just 35.6. So it would be a tremendous accomplishment if the 2013 hurricane season finished up with even HALF (52) the ACE of an average season (104).
Mysterious lack of activity
But just before the season began, every group making seasonal forecasts was calling for an above average or “very active” season… so what happened?
Prior to the hurricane season through today, the factors favoring lots of storm activity have included low surface pressures, warm sea surface temperatures, a strong African easterly jet (which enhances disturbances that enter the Atlantic and can potentially grow into storms), and the lack of an El Niño (which can promote hostile westerly winds).
Given that long list of storm-enhancing factors, what does the list of suppressing factors look like?
One signal that jumps out across the heart of the tropical Atlantic is very dry air. While some previous active seasons have had drier-than-normal air present (2004 is a great example), it has not reached the extreme of this year.
The plot below shows the departure of mid-level relative humidity from average, and peak values are around 24 percent. This anomalously dry air has also existed at lower and higher altitudes. Furthermore, the same areas that show up as dry here are also characterized by large-scale subsidence (sinking air) which acts to squash organized thunderstorm activity necessary for hurricane formation. That signal was not present in previous active years.
The jury is still out as to WHY so much subsidence and dry air has dominated the basin the past couple months, but clearly, it wasn’t very predictable.
Colder than normal ocean surface temperature in an area west of Portugal is another suppressing factor that wasn’t generally accounted for, but that hindsight has shown according to Phil Klotzbach, hurricane researcher at Colorado State University.
“[There is] very strong correlation between June sea surface temperatures in that area and Atlantic ACE values,” Klotzbach said. “I suspect that those cold anomalies propagate and directly impact stability and circulation patterns in the tropical Atlantic during the peak of the season.”
This season will probably teach forecasters to look at a previously-overlooked predictor(s).
But the suppressed activity is not just specific to the Atlantic. Tropical cyclone activity in the entire northern hemisphere is down this year – about 46 percent of average, with all basins (including the east and west Pacific) falling well short of average. There seems to be a combination of global-scale (or at least hemispheric-scale) atmosphere and/or ocean patterns that is causing this drop in activity.
Sure, the season isn’t over yet, but with August and September behind us, it’s REALLY hard to just catch up to average, let alone reach above average. Climatologically, roughly three-quarters of the season is behind us, though October is nothing to be dismissed (see figure below).
Keep in mind that ACE and seasonal activity should not be confused with U.S. landfalls or destructive storms. There are plenty of examples of very active seasons having minimal impact on the U.S. and quiet seasons having a large impact on the U.S. Of course, if you live near the coast, you should be prepared for every hurricane season, regardless of long-range predictions, it only takes one!
Why bother forecasting seasonal activity at all?
After what will likely be a major forecast “bust” for the groups making seasonal hurricane predictions, critics will jump on the opportunity to attack such forecasts. Here are several reasons why seasonal forecasts are worth making and developing.
1) All forecasters have faced a “bust” at least a few times, even if it’s a forecast for a high temperature made a day in advance! Weather forecasting by its nature will not always be accurate, and if you haven’t tried it, give it a go. It is humbling. But just because you’re sometimes wrong doesn’t mean you should stop trying.
2) Seasonal forecasts of tropical Atlantic activity have been around for 30 years, going back to the pioneering work of Bill Gray and colleagues at Colorado State University in the early 1980s. Several other agencies and groups have joined in on the challenge since then, including NOAA.
The seasonal forecasts are often better than climatology alone, meaning that on average, there is skill in the forecasts. As veteran hurricane forecaster Bill Gray once told me, “in the majority of years we are better than climatology and in some years we are worse. These forecasts were originally developed to give the public something better than seasonal climatology and to focus public attention to hurricanes.”
(Note: contrary to some rumors, seasonal hurricane forecasts are not utilized at all by insurance companies to justify rate changes; their data come from complex catastrophe modeling.)
3) It challenges us to better understand how the atmosphere works.
Phil Klotzbach says “if it were easy, then everyone would do it!” There are many complex oscillations in the atmosphere and ocean that exist on multiple scales. Trying to unravel the patterns and relationships to yield a skillful prediction of hurricane activity months in advance is not a trivial task. It also shows us that there are still some missing pieces of the puzzle or that that we don’t know what all of the pieces are – keeping scientific curiosity alive.
4) If there are key players that drive overall activity which are currently unpredictable (like this year’s large-scale subsidence, etc), how can there ever be any skill? Again, the skill comes out in the average, not necessarily in every season. Current methodologies are published in peer-reviewed literature, so this is not some roll of the dice. Seasonal forecasters do the best they can with the data and knowledge available. They’d be the first ones and the most qualified ones to point out the shortcomings. And why settle for the unpredictable always remaining unpredictable?
If your only interest is where and when a hurricane will make landfall, seasonal forecasting isn’t for you. That falls into the daily weather spectrum of forecasting… out to a few days or a week.
Seasonal forecasting is largely an academic exercise, but with an end result that has been and continues to be of interest to the public. If you disagree with it and its premise, you are entitled to do so, but for everyone else’s benefit, it will remain an active area of research and learning.
* Brian McNoldy is a senior researcher at the University of Miami’s Rosenstiel School of Marine and Atmospheric Science.
Extreme weather gallery 2013
You may also like: