The hurricane season 2017: a cluster of extreme storms

Within a span of four weeks, the hurricane trio of Harvey, Irma and Maria made the 2017 hurricane season the costliest ever. Overall losses reached around US$ 215bn, according to preliminary estimates, and insured losses are expected to be around US$ 100bn.


In terms of original values, the 2017 losses were higher than in the previous record year of 2005 that included hurricanes Katrina, Rita and Wilma (overall losses US$ 170bn, insured losses US$ 85bn). The question implicitly raised by the events of 2017: What are the causal factors involved in making a record-breaking loss year?

Only a portion of the 2017 hurricane season was extremely active – the second half of August through early October. During this period, several meteorological preconditions for above-normal activity were met, such as sea surface temperatures in the tropical main development region substantially above average, very low levels of difference in winds at low levels and aloft, an active West African monsoon producing tropical cyclone seedlings, and – last but not least – sufficient moisture levels in the mid to upper troposphere. Given the confluence of such conducive conditions, the season “turned on” and produced a cluster of exceedingly strong storms. The favourable conditions led to six major hurricanes (cat 3–5) out of an overall ten hurricanes (cat 1–5), and 17 named storms in total. In comparison, note that the period of elevated activity since 1995 averaged 3.4 major hurricanes per year. Accumulated kinetic energy released by the storms approximately tied the totals of 2004 and 1995, with only 2005 being stronger since 1950. All of these high activity years fall into the period of elevated levels of tropical North Atlantic sea surface temperatures since 1995.

Meteorological preconditions for above-normal activity

It is noteworthy that major hurricanes Harvey, Irma, and Maria of 2017 all underwent rapid intensification, reached record or near-record intensities, produced record rainfalls and – in the case of Irma – kept its extreme intensity for a record-breaking amount of time.

Within a span of four weeks, the hurricane trio of Harvey, Irma and Maria made the 2017 hurricane season the costliest ever. Overall losses reached around US$ 215bn, according to preliminary estimates, and insured losses are expected to be around US$ 100bn.

Similar tracks of Irma and Maria through the Caribbean

A broad spectrum of mechanisms accounted for the widespread destruction that now stretches from the Texas Gulf coast to Florida to the northeast Caribbean: In the case of Harvey, high pressure blocked the storm’s path and reduced its forward speed, resulting in days of torrential rainfall over roughly the same region in eastern Texas and Louisiana. With the other storms, high winds and storm surge caused most of the damage in the Caribbean and Florida. Due to the similar tracks of Irma and Maria through the Caribbean, some islands were hit twice, with Maria destroying what Irma had left standing. Taken together, many places and coastal stretches in the region were severely affected, including in some places significant long-term outages of essential infrastructure, such as electricity and communications. These conditions challenged governmental budgets and required specialised labour to repair them, leading to shortfalls in repair capacity. The time taken to repair these systems further exacerbated damage levels, drove prices up, and triggered an economic downturn and emigration from some locales.
Climate conditions definitely had an influence on this season’s activity. Although the northern North Atlantic saw a period of remarkable surface cooling between 2014–2016, this cooling signal never reached the tropical North Atlantic. Hence, the level of enhanced sea surface temperatures since the onset of the current warm phase in the mid-1990s is still unabated in the tropics, although the last few hurricane seasons before 2017 were not very active. In fact, the August–October 2017 sea surface temperature anomaly in the main development region of the tropical North Atlantic in 2017 was the third highest since 1995. This, together with the other conducive meteorological settings, rendered 2017 another season of the current active era in the tropical Atlantic. This means that high activity periods reveal a temporal structure of high season-to-season variability with occasional extremely active seasons which are characterised by a cluster of very intense storms. Hence, one of the important outcomes of the 2017 season should be to make sure that the risk models employed by the industry improve their capability to simulate seasonal clusters of strong storms.
Besides the influences from natural multidecadal climate variability such as warm and cold phases, climate change may also already have played a role, although this cannot be attributed with any statistical significance. Current projections of future conditions expect almost unchanged or stagnating overall tropical cyclone numbers in most ocean regions for the mid-21st and end-21st century. By contrast, the frequency of the extreme storms (Cat 4–5) is projected to increase in most areas with continued climate change. Also, rainfall rates within 100 km of the storm’s centre will increase due to higher evaporation rates, and mean maximum intensities will rise a little. Against the background of these projections, the 2017 season looks like a foretaste of the future. Indeed, we suspect that the future projections of increased numbers of extreme storms may materialise in terms of a higher frequency of exceptional seasons such as 2004, 2005 and 2017.

Munich Re Experts
Eberhard Faust
Eberhard Faust
Head of Research: Climate Risks and Natural Hazards
Munich Re (until 01.11.2020)
Mark Bove
Mark Bove
Nat Cat Solutions Manager
Munich Re US