Impact of the 2017 North Atlantic Hurricane Season on the RMS Medium-Term Rate

In my years of contributing to this blog, I have written extensively about the long-standing debate about the current state of hurricane activity in the North Atlantic Basin. This debate has become no clearer following the 2017 hurricane season; one of the busiest and costliest seasons on record. 2017 followed a stretch of below- to near-average seasons that began in 2012 and it is unclear whether future seasons will remain active or return the recent level of relative quiet.

Figure 1: Comparison of North Atlantic Basin major hurricane count and its five-year average with climatology, 1970-present

Last year, I wrote about the Version 17 Medium-Term Rate (MTR) update, in which the MTR registered below the long-term rate for the first time in the history of the forecast. You will recall that the MTR analyzes climate models reflecting three main theories of Atlantic hurricane variability and produces a five-year outlook of hurricane landfall frequency. In some cases, the output of these models shows conflicting signals, much like the ongoing scientific debate.

The most statistically skillful of these models in version 17 of the RMS® North Atlantic Hurricane Models identified a shift to a future below-average period of hurricane activity, in part based on the 2012-16 decrease in Atlantic Basin major hurricanes. However, uncertainty exists in identifying phase shifts near the end of a data record and it may be shortsighted to update the MTR forecast on the back of potentially anomalous season.

Indeed, scenario analyses performed by RMS modelers show that an inactive 2018 would produce an MTR forecast that remains below the long-term average. As a result, RMS will not be issuing an MTR forecast update in 2018.

What Factors Drove the Version 17 Update?

As a reminder, it was the “shift” climate models, which view changes in basin phase as natural oscillations, that produced below-average landfall forecasts in version 17. Conversely, the sea surface temperature and “active baseline” climate models did not identify a transition to a less active phase, in part based on Atlantic sea surface temperatures (SSTs) that have remained warmer than average since the mid-1990s.

We take a weighted average across all 13 models, based on tests made of each model’s skill in predicting hurricane activity in sample periods from the past. In Version 17, the shift models demonstrated greater skill than their counterparts and the higher weight allocated to these models decreased the MTR below the long-term rate.

Regionalization Accounts for Above-average Tail Risk

However, understanding the story on this headline alone ignores an extra layer of intelligence built into the RMS forecast. The MTR relies on regionalization — that is, the impact of projected SSTs in determining where along the coastline, and at what strength, hurricanes are likely to make landfall.

Atlantic SSTs measured during the peak of hurricane season have remained warmer than average for many years. Thus, while the Version 17 MTR calls for fewer than average hurricanes, the energy provided by warm sea surface temperatures may lead to hurricanes that are stronger than average.

Comparisons of industry exceedance probability (EP) curves in the U.S. and the Caribbean demonstrate this unique severity distribution. In the U.S., the likelihood of exceeding industry losses beyond the 150-year return period is higher in the MTR than in the long-term rate (LTR). In the Caribbean, this threshold is even lower: the five-year return period. This point corresponds to a US$4 billion industry loss, an amount well exceeded in 2017 by Hurricanes Irma and Maria.

Figure 2: Ratio of RMS MTR and LTR 2017 Industry Occurrence Exceedance Probability Curves for the U.S. (left) and the Caribbean (right)

This analysis demonstrates that the MTR still considers the possibility of highly damaging events and seasons, such as 2017, to be greater than the long-term average. As a result, RMS believes the Version 17 MTR forecast to remain a valid view of hurricane risk.

The Future of the Medium-term Rate

RMS will update the MTR forecast next year in Version 19 of the RMS North Atlantic Hurricane Models. This forecast will consider the same three groups of climate models mentioned previously.

We expect the sea surface temperature and active baseline models to produce forecasts above the long-term average, based in part on warmer than average sea surface temperatures. Thus, the Version 19 MTR, and its position relative to the long-term rate, will heavily depend upon the output of the shift models. Will they still project the recent transition to a less active phase?

Anticipating activity in future seasons allows us to add data points to a statistically limited dataset. To do this, RMS modelers ran our systematic and objective forecast methodology with several possible outcomes for the upcoming 2018 season. Three of these scenarios include:

  • A 2018 season with fewer than three major hurricanes in the basin, with none making landfall in the U.S., maintains an MTR forecast below the long-term average
  • A 2018 U.S. major hurricane landfall removes the phase change in the shift models, producing a slightly above-average forecast
  • A repeat of the hyperactive 2017 season in 2018 produces an above-average forecast, but below the version 15 forecast

Figure 3: The outcome of several Version 19 MTR scenario analyses performed by RMS modelers

These scenario analyses prove beneficial in a number of ways. First, they allow RMS modelers to test the sensitivity of the MTR climate models during an uncertain time of vigorous scientific debate.  Second, and more importantly, RMS clients can follow the progress of the next hurricane season in real time with an advance understanding of how each major hurricane formed will impact the next update to the RMS model.

Senior Product Manager, Model Product Management
Tom is a Senior Product Manager in the Model Product Management team, focusing on the North Atlantic Hurricane Model suite of products. He joined RMS in 2009 and spent several years in the Client Support Services organization, primarily providing specialist peril model support. Tom joined RMS upon completion of his B.S. and M.S. degrees in meteorology from The Pennsylvania State University, where he studied the statistical influence of climate state variables on tropical cyclone frequency. He is a member of the American Meteorological Society (AMS).

Leave a Reply

Your email address will not be published. Required fields are marked *