North Atlantic tropical storm seasonal forecast 2023

This forecast was issued on 01 August 2023 and supersedes the pre-season forecast issued in May 2023. Please see the link below for the pre-season forecast.

Forecast for the whole 2023 season

Issued 01 August 2023 

All values given below represent observed activity for the 6-month period from February to July 2023 plus the 6-month GloSea6 forecast activity from August 2023 to January 2024. Note, the official tropical storm season, which usually includes most activity, is from June-November.

The most likely number of named tropical storms (winds of at least 39 mph) predicted to occur in the North Atlantic is 19, with a 70% chance that the number will be in the range 14 to 24. The 1991-2020 long-term average is 14. 

The most likely number of hurricanes (winds of at least 74 mph) predicted to occur in the North Atlantic is 9, with a 70% chance that the number will be in the range 6 to 12. The 1991-2020 long-term average is 7. 

The most likely number of major hurricanes (winds of at least 111 mph) predicted to occur in the North Atlantic is 6, with a 70% chance that the number will be in the range 3 to 9. The 1991-2020 long-term average is 3.  

An ACE index of 215 is predicted as the most likely value, with a 70% chance that the index will be in the range 120 to 310. The 1991-2020 average is 123. 

Note: The unnamed subtropical storm which occurred in January 2023 fell outside the period covered by this prediction (February 2023-January 2024).

Background

'Tropical cyclone' is the generic term for a low-pressure system over tropical or subtropical waters, with intense convective activity (e.g. thunderstorms) and winds circulating in an anticlockwise direction in the northern hemisphere (clockwise in the southern hemisphere). A tropical storm is a tropical cyclone with mean wind speeds of at least 39 mph. The terms hurricane and typhoon are region-specific names for strong tropical cyclones with wind speeds of at least 74 mph. For further details about tropical cyclones see Met Office tropical cyclone facts

The North Atlantic tropical storm season usually runs from June to November. The degree of activity over the whole season varies from year to year and is measured in several ways. 

  • Total number of tropical storms. The number of named tropical cyclones with winds of at least 39 mph observed over the season. This is the best known measure of the level of storm activity. In this forecast the number of tropical storms also includes those that reach hurricane and major hurricane strength (see below). 

  • Total number of hurricanes. The number of tropical cyclones that reach sustained winds of at least 74 mph. 

  • Total number of major hurricanes. The number of tropical cyclones that reach sustained winds of at least 111 mph.  

  • The Accumulated Cyclone Energy (ACE) index. A measure of the collective intensity and duration of all named tropical storms, hurricanes and major hurricanes during the season. 

The table below shows the number of tropical storms, hurricanes, major hurricanes and ACE index observed in recent years. 

North Atlantic tropical storm activity (February to January the following year)

Year Tropical storms Hurricanes Major hurricanes ACE index
2013 14 2 0 36
2014 8 6 2 67
2015 11 4 2 63
2016 15 7 4 141
2017 17 10 6 225
2018 15 8 2 129
2019 18 6 3 130
2020 30 13 7 180
2021 21 7 4 146
2022 15 8 2 97

Historical Forecasts

At the start of each North Atlantic season the Met Office forecasts the number of tropical storms, hurricanes, major hurricanes and ACE index.  Seasonal forecasts of tropical storm activity have been issued annually since 2007. See our previous seasonal tropical storm forecasts issued in either May or June for 2007, 2008, 2009, 2010 , 2011, 2012, 2013, 2014, 2015, 2016201720182019202020212022 and 2023.

Method

The forecast is made using information from the Met Office seasonal prediction system, GloSea6, which is an ensemble prediction system based on a fully coupled ocean-atmosphere general circulation model (GCM). This system simulates the ocean-atmosphere processes and interactions that determine tropical storm development. Multiple forecasts are made (using ensemble forecasting methods) to allow estimation of the range of likely outcomes.  The ensemble is then used to produce a best estimate (or most likely) predicted value (the ensemble mean) and forecast range (using the spread in the ensemble outcomes). Further details on the forecast method can be found in Camp et al. (2015). 

Skill

Recent studies have shown that dynamical models have considerable skill predicting the number of tropical storms - for example successfully predicting the change from the exceptionally active season of 2005 to the below-normal activity of the 2006 season. In 2017, the Met Office seasonal forecast system successfully predicted the above-average activity that was observed in the North Atlantic, including the high frequency of tropical cyclone tracks around the northeast Caribbean (Camp et al., 2018). Full details of the skill to predict tropical storm activity around the world has previously been documented by Camp et al. (2015) and for real-time forecasts in Klotzbach et al. (2019).  

Factors Influencing the Current Seasonal Forecast

There is a strong signal in our forecasts for El Niño to develop and persist through the peak of the Atlantic hurricane season (August to October 2023). This would normally be expected to increase vertical wind shear across the tropical North Atlantic, suppressing tropical storm activity. However, as shown above, the 2023 forecast predicts most likely high tropical storm activity in the Atlantic. There are several factors which are likely counteracting the influence of El Niño resulting in this forecast of high activity, including:

  • Vertical wind shear through the troposphere over the tropical Atlantic is predicted to be lower than previously seen during El Niño events. 
  • Sea surface temperatures are predicted to remain well above average across the tropical North Atlantic during the whole hurricane season.

Current El Niño forecasts suggest that the transition from neutral to El Niño conditions is likely to take place over the next few months. However, the speed of transition, its intensity and the time it takes to influence Atlantic hurricane activity will all have a direct bearing on the season ahead with the upper end of the Met Office forecast providing a reasonable worst-case scenario.

Stakeholders are advised to follow the advice and analysis of the official World Meteorological Organisation Regional Specialised Meteorological Centre for the Atlantic region, the US National Hurricane Center (NHC), during the hurricane season. NHC provides forecasts for all active tropical cyclones and the likelihood of storm formation up to seven days ahead in the Atlantic based on a consensus from all major forecast centres.

Verification

The table below shows the seasonal forecast values issued on 01 August 2023 for the period February 2023 to January 2024 together with the end of season observed values. This season the observed number of tropical storms, hurricanes, major hurricanes and ACE Index were all within the 70% forecast range.

 

Most likely number 70% forecast range Observed number
Tropical storms 19 14-24 19 *
Hurricanes 9 6-12 7
Major hurricanes 6 3-9 3
ACE Index 215 120-310 144 *

* The unnamed storm which formed in January 2023 was not included since it occurred outside the February 2023 to January 2024 forecast period.

A public forecast for the June to November 2024 period in the Atlantic will be issued in May 2024.

References

Camp, J., Roberts, M., MacLachlan, C., Wallace, E., Hermanson, L., Brookshaw, A., Arribas, A., Scaife, A. A., (2015). Seasonal Forecasting of Tropical Storms Using the Met Office GloSea5 Seasonal Forecast System. Q.J.R. Meteorol. Soc, 141: 2206-2219. 

Camp, J., Scaife, A. A., Heming, J. (2018). Predictability of the 2017 North Atlantic hurricane season. Atmos Sci Lett. 2018; 19:e813.  

Klotzbach, P. et al. (2019). Seasonal Tropical Cyclone Forecasting. Tropical Cyclone Research and Review, 8 (3), 134-149.  

To view PDF documents you will need to download Adobe Reader