With the storm-nurturing effects of a warm Atlantic likely to be counterbalanced by a robust El Niño, a slightly below-average Atlantic hurricane season is likely in 2023, the Colorado State University (CSU) hurricane forecasting team says in its latest seasonal forecast, issued April 13.

Led by Dr. Phil Klotzbach, with co-authors Dr. Michael Bell and Alexander DesRosiers, the CSU team is calling for 13 named storms, 6 hurricanes, 2 major hurricanes, and an Accumulated Cyclone Energy (ACE) of 100. In comparison, the long-term averages for the period 1991-2020 were 14.4 named storms, 7.2 hurricanes, 3.2 major hurricanes, and an ACE of 123.

The CSU outlook predicts the odds of a major hurricane hitting the U.S. this year to be 44% (long-term average: 43%). It gives a 22% chance for a major hurricane to hit the East Coast or Florida Peninsula (long-term average: 21%), and a 28% chance for the Gulf Coast (long-term average: 27%). The Caribbean is forecast to have a 49% chance of having at least one major hurricane pass through (long-term average: 47%).

The CSU forecast uses a statistical model honed from 40+ years of past Atlantic hurricane statistics, plus dynamical model output from four groups: ECMWF (European model), UKMET, Japan Meteorological Agency, and Centro Euro-Mediterraneo sui Cambiamenti Climatici. These models produced a wide range of possibilities for the 2023 Atlantic hurricane season, ranging from a slightly below-average season to a well above-average season.

A map of sea surface temperatures for mid-April 2023. The Gulf of Mexico is much warmer than usual.
Figure 1. Departure of sea surface temperature (SST) from average for April 13, 2023. In the hurricane Main Development Region (MDR) between Africa and Central America, SSTs were near average in the Caribbean, and much above average in the eastern tropical and subtropical Atlantic. Virtually all African tropical waves move through the MDR, and these tropical waves account for 85% of all Atlantic major hurricanes and 60% of all named storms. Above-average SSTs in the MDR during hurricane season generally lead to an active season, in the absence of an El Niño event. Conversely, when MDR SSTs are cooler than average, a below-average Atlantic hurricane season is more likely. (Image credit: tropicaltidbits.com).

Analogue years

Eight years with similar preseason January, February, and March atmospheric and oceanic conditions were selected as “analogue” years that the 2023 hurricane season may resemble. These years had neutral or La Niña conditions the previous winter, warm-neutral to significant El Niño conditions during the peak of the Atlantic hurricane season (August–October), and near- to above-average sea surface temperatures in the tropical Atlantic. The analogue years were:

1969 (18 named storms, 12 hurricanes, and 5 major hurricanes);
2002 (12 named storms, 4 hurricanes, and 2 major hurricanes);
2004 (15 named storms, 9 hurricanes, and 6 major hurricanes);
2006 (10 named storms, 5 hurricanes, and 2 major hurricanes);
2009 (9 named storms, 3 hurricanes, and 2 major hurricanes);
2012 (19 named storms, 10 hurricanes, and 2 major hurricanes);
2014 (8 named storms, 6 hurricanes, and 2 major hurricanes); and
2015 (11 named storms, 4 hurricanes, and 2 major hurricanes).

The average activity for these years was 12.8 named storms, 6.6 hurricanes, 2.6 major hurricanes, and an ACE of 105 – slightly below the long-term average.

The CSU team cited two main reasons, addressed below, for their forecast:

1) A significant El Niño seems likely 

NOAA issued an El Niño Watch in its April 13 monthly discussion of the state of the El Niño/Southern Oscillation, or ENSO. Neutral conditions are expected to persist through the Northern Hemisphere spring, with a 62% chance of a transition to El Niño during the May-June-July period. El Niño is then expected to continue through the end of 2023 into 2024.

NOAA’s and Columbia University’s International Research Institute for Climate and Society ENSO forecast for the peak portion of the Atlantic hurricane season (August-September-October) is for a 1% chance of La Niña, a 17% chance of ENSO-neutral, and an 82% chance of El Niño. Most of the El Niño models are predicting a weak to moderate strength El Niño event forming by late summer, with the dynamical model consensus favoring a borderline weak/moderate event for the peak of hurricane season.

El Niño conditions favor a slower-than-usual Atlantic hurricane season as a result of an increase in the upper-level winds over the tropical Atlantic that can tear storms apart (higher vertical wind shear). In contrast, when neutral or La Niña conditions are present, an active hurricane season becomes more likely.

2) The current SST pattern correlates with active Atlantic hurricane seasons

The eastern tropical Atlantic, subtropical Atlantic, and Gulf of Mexico currently have sea surface temperatures (SSTs) that are well above average, while the Caribbean is near average. “Overall, the current SST anomaly pattern correlates relatively well with what is typically seen in active Atlantic hurricane seasons,” the forecasters said. “This anomalous warmth is one of the reasons why our forecast is for just a slightly below-normal season, despite the potential for a robust El Niño.”

As is its practice, the CSU team included this standard disclaimer:

“Coastal residents are reminded that it only takes one hurricane making landfall to make it an active season for them, and they need to prepare the same for every season, regardless of how much activity is predicted.”

Last year (2022) serves as a good reminder of this caveat. Overall, it was a near- to slightly-below-average season – not unlike what is predicted by CSU for this year. Yet the season spawned Hurricane Ian, one of the nation’s most expensive hurricanes on record, which had catastrophic impacts on Florida.

Two graphs showing that early spring hurricane forecasts are not great predictors of final season impact.
Figure 2. Comparison of the percent improvement in mean square error over climatology for seasonal hurricane forecasts for the Atlantic from NOAA, CSU and Tropical Storm Risk (TSR) from 2003-2022, using the Mean Square Skill Score (MSSS). Values less than zero indicate that pure climatology does a better job than the forecast. The figure shows the results using two different climatologies: a fixed 50-year (1951-2000) climatology, and a 10-year 2013-2022 climatology. Skill for forecasts issued in April is close to or below zero, is modest for June forecasts, and is moderate-to-good for August forecasts. Using this methodology, TSR has had the best seasonal forecasts in general, although NOAA’s June forecasts were slightly more skillful than TSR’s. (Image credit: Tropical Storm Risk, Inc.)

A caveat: April hurricane season forecasts have little or no ‘skill’

On average, April forecasts of hurricane season activity have had no “skill,” or even negative skill, when computed using the Mean Square Skill Score (Figure 2). This does not mean a particular April forecast will be incorrect – just that, on average, a forecast simply using climatology would do as well or better. April forecasts must deal with the so-called spring predictability barrier. In April, the El Niño/La Niña phenomenon commonly undergoes a rapid change from one state to another, making it difficult to predict whether El Niño, La Niña, or neutral conditions will be in place for the coming hurricane season.

CSU’s April 2022 forecast a year ago had predicted an above-average Atlantic hurricane season for the year, with 19 named storms, 9 hurricanes, 4 major hurricanes, and an Accumulated Cyclone Energy (ACE) of 160. That forecast fell short since the 2022 season ended up with 14 named storms, 8 hurricanes, 2 major hurricanes, and an ACE of 95. The April CSU forecasts for the 2020 and 2021 seasons were quite accurate, though.

The next CSU forecast, due June 1, is worth close attention, as late May/early June forecasts have shown considerable skill over the years. Among other outlooks already issued or in the works:

  • NOAA is set to issue its first seasonal hurricane forecast for 2023 in late May.
  • The British private forecasting firm Tropical Storm Risk, Inc. (TSR) issued its first 2023 Atlantic hurricane season forecast on April 6, calling for a below-average season with 12 named storms, 6 hurricanes, 2 intense hurricanes, and an ACE index of 84.
  • The U.S. forecasting firm Atmospheric G2 is predicting 15 named storms, 7 hurricanes, and 3 major hurricanes, noting that “The combination of an incipient El Nino and very warm [North Atlantic sea surface temperatures] is unprecedented in the recent record.”

Also see: How to make an evacuation plan

Bob Henson contributed to this post.

Website visitors can comment on “Eye on the Storm” posts (see below). Please read our Comments Policy prior to posting. (See all EOTS posts here. Sign up to receive notices of new postings here.)

Jeff Masters

Jeff Masters

Jeff Masters, Ph.D., worked as a hurricane scientist with the NOAA Hurricane Hunters from 1986-1990. After a near-fatal flight into category 5 Hurricane Hugo, he left the Hurricane Hunters to pursue a...