In the 20 or so years that span my reinsurance career the subject of the influence of El Niño–Southern Oscillation (ENSO) on Atlantic hurricanes has often come up. And the topic invariable leads to the following comment “Yes, but didn’t hurricane Andrew hit in an El Niño year?”.
But the truth as usual is more complicated.
Why do I bring this up now, when Andrew, the most destructive hurricane in U.S. history at the time of occurrence, hit land all the way back in 1992? Well, the evolution of the ENSO 3.4 index this year is predicted to be similar to 1992.
The actual development of the ENSO in 1992 is shown in the plot below, (large dots). It’s clear that the year started out while an El Niño event was underway, however by the start of the hurricane season, that El Niño event was clearly over, and conditions had entered a neutral phase.
This highlights how careful we need to be when evaluating the potential impact of ENSO on losses, since the assignment of full years to one of three ENSO states is likely to be ambiguous at best.
There is a wide spread of forecasts for ENSO, just as there have been a range of forecasts for hurricane activity. A common criticism of all these forecasts is that they don’t say anything about losses. To resolve this Willis Re has developed a model to estimate the maximum likely loss variation between El Niño and La Niña states.
Recent work with Willis Research Network partners at Exeter University identified links between the level of observed variability in the counts of hurricanes in the Atlantic, and climate indices such as ENSO. Indeed, we recently published a few blogs exploring how a potential La Niña will affect 2016’s hurricane season and specifically what a La Nina could mean for insured property exposure along the U.S. coast.
By making a set of reasonable assumptions, it is possible to estimate the likely upper bounds of ENSO influence on aggregate and event losses consistent with this RMS U.S. hurricane model. Which combined with the ENSO forecasts is a good step towards loss forecasts.