The 2013 Atlantic hurricane season was the least active in thirty years.  The season officially ends on Saturday, but in this lackluster year, it looks like we can wrap it up a little early.

On average (1981-2010), 12 tropical storms form during an entire season, 6-7 of those go on to become hurricanes, and 2 of those reach major hurricane intensity (category 3+). In 2013, there were 13 tropical storms, 2 hurricanes, and no major hurricanes. Even the two hurricanes that formed were just briefly minimal category 1 storms.

The last time a season ended up with only two hurricanes was 1982, and the last season to have zero major hurricanes was 1994. The most intense storms this year had maximum sustained winds only reach 75 knots (Humberto and Ingrid). Dr. Phil Klotzbach, a researcher at Colorado State University, noted that is the weakest maximum intensity for a hurricane during a season since 1968. And you may recall that the first hurricane (Humberto) formed just hours before the record latest first hurricane formation date on September 11.

In terms of Accumulated Cyclone Energy, or ACE, the seasonal total stands at 31.1 (the units are 10,000 knots squared if anyone is curious), the lowest seasonal total since 1983… and that’s just 30% of average.  Looking back to 1950, only four other years come in with lower ACE totals: 1983, 1982, 1977, and 1972.

To put that in perspective, Hurricane Igor (2010) alone racked up about 140% of this season’s entire total.  Dr. Ryan Maue at WeatherBell keeps thorough statistics on ACE around the world, and this chart displays values of accumulated daily ACE for each season going back to 1950.

High expectations

Heading into the season, all of the typical signals and precursors pointed forecasters to a fairly uniform conclusion: the season will be very active.  Forecast teams from CSU, NOAA, FSU, UK Met Office, Tropical Storm Risk, etc released their seasonal outlooks in late May to early June and every one of them was predicting about 7-9 hurricanes, and about 130-165% of an average season’s ACE.   Only ECMWF {European Center for Medium-range Weather Forecasting) predicted a slightly below-active season at about 80% of average ACE, although they also predicted 6-7 hurricanes.  Recall that the season has ended with just 2 hurricanes and 30% of the average ACE.

Even among the groups that issue periodic mid-season updates and forecasts, the success rate was not any higher.  The core months of August, September, and October slipped by with only minimal activity, despite key factors continuing to appear favorable.

While such an inactive season is not unprecedented, usually the forecasts would do a decent job of anticipating the upcoming slow season. Of course, the advantage of such a quiet season is there were not any devastating landfalls anywhere in the basin.

Major forecast bust

Why was this season so inactive? What wasn’t accounted for in the forecasts? While there are some hypotheses, the jury is still out. To analyze the large-scale patterns over the last few months, we look at “global climate reanalyses”… and there are a few sources of those data (ERA-Interim, NASA-MERRA, NCEP/NCAR). The problem is that they don’t all agree even on the fundamental sign of some of the possible explanatory factors.

Prof. Mark Saunders and Dr. Adam Lea at University College London have begun to explore this issue, and mention that their preliminary findings “question a number of the suggestions for why the 2013 hurricane season was unexpectedly quiet”. We may have to wait another couple months or so before all of the reanalysis data are processed by the various agencies and a more robust survey can be completed.  But in the meantime, I’ll offer some possible explanations

First, major signals such as the El Nino Southern Oscillation (ENSO), surface pressure, and sea surface temperature were all indicative of an average to above-average season.  But let’s take a look at some of the possible suppressing factors (I say “possible” not only because of the aforementioned uncertainties in reanalysis data, but because the system is too complex to merely point fingers at something without understanding how everything is interconnected).

1) Dry air

In the plot below, the anomalous mid-level relative humidity is shown, averaged over Aug-Sep-Oct.  Even over this long three-month window, the vast majority of the basin is dominated by drier-than-normal air, especially in the deep tropics off the coast of Africa.  Dry air can quickly weaken or dissipate a tropical cyclone (or inhibit its formation in the first place).

2) Stable air

Besides being dry at low-mid levels, the vertical instability in the tropical Atlantic during the core months was also reduced.  That means that the average temperature profile in that region was less conducive to thunderstorm growth and development.  These data are derived from GFS analyses in real-time, not from one of the reanalyses I mentioned previously.

3) Weak African Jet

Tropical waves, the embryos of many tropical cyclones, have their origins over continental Africa.  There is a persistent feature called the African Easterly Jet which is a fast-moving river of air in the low-mid levels of the atmosphere extending from Ethiopia westward into the tropical Atlantic Ocean.  This jet breaks down into discrete waves, and every few days, another wave exits the coast.  Some are barely noticeable, while others become tropical storms shortly after leaving the coast.  During the core of hurricane season, most tropical cyclones have an easterly wave pedigree, so they’re a big deal when it comes to overall seasonal activity.

Winds in the jet normally cruise along at 20-25 mph at an altitude of 700 mb (10,000 feet) when averaged over this period (Aug-Oct), but this year, they were about 40% weaker… or roughly 12-17 mph.  One would expect that to have a big impact on the amplitude of easterly waves, and thus, hurricane season.

Links to global warming?

One question that inevitably gets asked is how this (in)activity relates to climate change. It is not accurate to associate any particular season (and definitely NEVER a specific storm!) with climate change. Whether the season happened to be very active, average, or very inactive, that does not allow us to draw any conclusions about what role climate change played.  The reason is that intra- and inter-seasonal variability is so large that any subtle signals that climate change might be contributing are completely overwhelmed.

According to IPCC5 that compiles recent research and presents the state of the science, “studies that directly attribute tropical cyclone activity changes to anthropogenic greenhouse gas emission are lacking”.  Gabriel Vecchi at GFDL summed up the connection very well in a recent interview with NPR: “given the data that we have available and the signal that we expect global warming to give us, which is still of uncertain magnitude and sign, we weren’t able to give it anything beyond low confidence. We didn’t have confidence to say whether we have seen trends yet.”

The 2014 Atlantic hurricane season begins in 188 days (June 1), and the first few names on the list are Arthur, Bertha, and Cristobal.