To the Editor —
The Atlantic hurricane seasons in 2013 and 2014 were quieter than average, and there are indications that hurricane activity in 2015 will also be below normal. Here we investigate whether the active Atlantic hurricane era that began in 1995 may have ended. To this end, we assess hurricane variability in the Atlantic since 1878, along with a proxy for the Atlantic multidecadal oscillation (AMO), whose positive phases have been noted to be closely linked to active periods for Atlantic hurricanes. We find that the AMO proxy values are currently at their lowest values since the early 1990s, when Atlantic hurricane activity was well below average.
The Atlantic entered an active hurricane period in 19951,2. From 1995 to 2012, accumulated cyclone energy (ACE)3 in the Atlantic basin averaged 140 × 104 knot2 compared with 68 × 104 knot2 from 1970 to 1994, a statistically significant increase at the 5% level. High ACE is also observed from 1878–1899 (108 × 104 knot2) and 1926–1969 (101 × 104 knot2) compared with low values from 1900–1925 (65 × 104 knot2) and 1970–1994 (68 × 104 knot2) (Fig. 1a). A limited observational network may have led to an underestimate in ACE during the earlier part of the record.
The AMO, an indicator of sea surface temperature (SST) variations in the North Atlantic, has been argued to arise from natural climate variations in the thermohaline circulation1,2,4. Alternatively, it could be primarily driven by alterations in levels of sulfate aerosols5. We argue that the weight of the evidence points towards natural oceanic variability being the principal driver of the AMO1,4,6. The AMO phase was classified as being positive from 1878–1899, 1926–1969 and 1995–2012, and negative from 1900–1925 and 1970–19942. Positive AMO phases are characterized by above-average far North and tropical Atlantic SSTs, below-average tropical Atlantic sea level pressures (SLPs), and reduced levels of tropical Atlantic vertical wind shear. All three of these conditions are known to create a more favourable environment for Atlantic hurricane formation and intensification1.
Following an active 2012 Atlantic hurricane season, most predictions indicated another active season in 20137. A strong temporary weakening of the AMO was observed during 20138, which may have been one of the reasons for the season being quieter than expected. Activity during the 2014 hurricane season was also below normal. The combined ACE of 103 × 104 knot2 registered in 2013–2014 was the lowest value that had been recorded since the ACE of 71 × 104 knot2 in 1993–1994. And following this marked reduction in activity over the past two years, a below-average season is now expected for 20159.
Random resampling of three-year periods of positive AMO phase gives a 1% chance of an ACE of 155 × 104 knot2 or lower during the positive AMO phase. It is therefore unlikely, if 2015 turns out to be as quiet as predicted, that we are in a positive AMO phase. Borderline weak El Niño conditions in 2014 and strong El Niño conditions predicted for this year's hurricane season may, however, also have contributed to observed below-average hurricane activity in 2014 and predicted below-average hurricane activity in 2015.
A proxy using a combination of North Atlantic SSTs from 50–60° N, 50–10° W and SLPs from 0–50° N, 70–10° W has been utilized to monitor the strength of the AMO in real time10. When the AMO is positive, SSTs in the far North Atlantic tend to be warmer, while SLPs throughout the tropics and subtropics tend to be lower2. This index has decreased since 2012 (Fig. 1a): SST anomalies in the tropical and far North Atlantic have become cooler and SLP anomalies throughout most of the Atlantic have increased. The decrease in far North Atlantic SSTs in the past three years has been associated with a weaker thermohaline circulation6. Annual mean SSTs in the North Atlantic have cooled in 2013 and 2014 compared with values averaged from 1995 to 2012 (Fig. 1b).
During the most recent 25 positive AMO years (1963–1969 and 1995–2012), 204.5 major hurricane days (MHD) were observed in the Atlantic, compared with 63.25 during the most recent 25 negative AMO years (1970–1994), a ratio of 3.2 to 1. Marked differences are also observed in tropical cyclone tracks in the deep tropics between the positive versus negative AMO phase (Fig. 1c). For example, in the Caribbean (south of 20° N, west of 60° W), 16.25 MHD occurred during the most recent 25 years in the negative AMO, compared with 46.75 MHD during the most recent 25 years in the positive AMO (a ratio of 2.9:1). Similar large differences are seen when examining MHD in the tropical Atlantic (south of 20° N, east of 60° W). A total of 30.75 MHD occurred during the positive AMO, compared with only 9.5 MHD during the negative AMO (a ratio of 3.2:1).
A transition of the AMO to its negative phase implies that the frequency of storms that make landfall along the east coast of the US should decrease significantly1. We subdivide the US coastline into the Gulf Coast/Florida Panhandle on the one hand and the Florida Peninsula/East Coast on the other hand (breaking at Tarpon Springs, Florida). The frequency of landfalling major hurricanes, based on historical data from 1878–2012 along the Florida Peninsula and East Coast decreases from 0.35 to 0.16 events per year between a positive and a negative AMO phase, while the frequency of landfalling major hurricanes remains virtually constant (0.32 in positive AMO vs. 0.35 in negative AMO) for the Gulf Coast and Florida Panhandle. The lack of change along the Gulf Coast may be due to the minimal impact of the AMO on tropical cyclone formation in the Gulf of Mexico (not shown).
A large variety of other climatic factors have also been shown to be linked to phase changes of the AMO including frequency and intensity of El Niño10 and likelihood of Sahelian drought. Consequently, the impacts of a potential phase change of the AMO extend well beyond its impacts on Atlantic basin hurricane activity.
Goldenberg, S. B., Landsea, C. W., Mestas-Nuñez, A. M. & Gray, W. M. Science 293, 474–479 (2001).
Klotzbach, P. J. & Gray, W. M. J. Clim. 21, 3929–3935 (2008).
Bell, G. D. et al. J. Clim. 81, S1–S50 (2000).
McCarthy G. D., Haigh, I. D., Hirschi, J.-M., Grist, J. P. & Smeed, D. A. Nature 521, 508–510 (2015).
Booth, B. B. B., Dunstone, N. J., Halloran. P. R., Andrews, T. & Bellouin, N. Nature 484, 228–232 (2012).
Rahmstorf, S. et al. Nature Clim. Change 5, 475–480 (2015).
Vecchi, G. A. & Villarini, G. Science 343, 618–619 (2014).
Fogarty, C. T. & Klotzbach, P. J. Bull. Am. Meteorol. Soc. 95, S106–S107 (2014).
Klotzbach. P. J. & Gray, W. M. Extended Range Forecast of Atlantic Seasonal Hurricane Activity and Landfall Strike Probability for 2015 (Colorado State Univ., 2015); http://tropical.atmos.colostate.edu/Forecasts/2015/june2015/jun2015.pdf
Gray, W. M., Sheaffer, J. D. & Landsea, C. W. in Hurricanes, Climate Change and Socioeconomic Impacts: A Current Perspective (eds Diaz, H. F. & Pulwarty, R. S.) 1–49 (Westview Press, 1996).
P.K. and W.M.G. acknowledge support from the G. Unger Vetlesen Foundation.
About this article
Cite this article
Klotzbach, P., Gray, W. & Fogarty, C. Active Atlantic hurricane era at its end?. Nature Geosci 8, 737–738 (2015). https://doi.org/10.1038/ngeo2529
Tropical cyclone precipitation regimes since 1750 and the Great Suppression of 1843–1876 along coastal North Carolina, USA
International Journal of Climatology (2021)
Atlantic‐Pacific SST Gradient Change Responsible for the Weakening of North Tropical Atlantic‐ENSO Relationship due to Global Warming
Geophysical Research Letters (2019)
A Review of the Role of the Atlantic Meridional Overturning Circulation in Atlantic Multidecadal Variability and Associated Climate Impacts
Reviews of Geophysics (2019)
Adaptation and sustainability of water management for rice agriculture in temperate regions: The Italian case‐study
Land Degradation & Development (2019)
Climate Dynamics (2019)