Insurers, reinsurers, financial institutions, and policymakers are being asked increasingly pointed questions about their preparedness for the possible effects of a changing climate. AIR takes an active role in modeling climate change and is committed to helping the global community understand, measure, and manage the associated uncertainties.

AIR recently released a report on the potential impacts of climate change on hurricane-induced losses in the United States. The study that led to the report assumed a high greenhouse gas emissions scenario of the Intergovernmental Panel on Climate Change (IPCC) Representative Concentration Pathway (RCP) 8.5. The choice of RCP 8.5 caught the interest of some on Twitter, who argued that RCP 8.5 is no longer considered plausible. That conversation provided the motivation to further explain our choice in this blog.

A Baseline View

As we noted in the report, the RCP 8.5 scenario is labeled a business-as-usual scenario, but it is not meant to be interpreted as an outcome-is-likely one. In fact, it was chosen for inclusion in the IPCC’s 2014 climate change report because it was at the 90th percentile of the business-as-usual scenarios available at the time. “Business as usual” means that this scenario assumes there are no climate change policies in place and is thus a baseline, or reference, view. But as there are three other RCP scenarios—2.6, 4.5, and 6.0—all representing lower emissions pathways, why not focus on one of them instead?

It’s important to point out that none of the four RCP scenarios are intended as forecasts. For this reason—and because they depend on underlying assumptions regarding the rate of technological progress, policy implementation, and population and economic growth—it’s impossible to make a statement along the lines of “RCP 8.5 has an x-percent likelihood of occurring,” or that RCP 4.5 has a y-percent chance of occurring.

Worst-Case Scenario

There are good reasons for addressing the RCP 8.5 scenario. In the world of insurance, assessing extreme event risk is critical to ensure solvency. While a focus on a worst-case scenario may not be optimal for determining public policy, it is appropriate in catastrophe modeling. Plan for the worst and hope for the best is sound advice.

Interestingly, the RCP 8.5 scenario may not be such an outlier. In scientific peer-reviewed literature, RCP 8.5 is noted as a still-plausible scenario—even for late into this century, although with  decreasing likelihood, we hope. A recent article by Schwalm et al. (2020) suggests that it could be more than just still-plausible by 2050; it points out that from 2005 to 2020 RCP 8.5 is most closely tracking observed emissions. Looking at emissions projections from the International Energy Agency out to 2030 and even 2050, RCP 8.5 unfortunately remains a good match, according to the authors.

There are other reasons for not discounting RCP 8.5. While some assumptions underlying this  scenario include significant growth in the use of coal (a circumstance that looks increasingly unlikely), other assumptions can be combined to produce a similar—or even worse—outcome, such as the thawing of the permafrost, which is expected to release vast amounts of methane.

A Focus on 2050

The focus of the report AIR recently published was on 2050, rather than the end of the century. This was considered a time horizon not so far out that it would be of no practical application to our clients making capital expenditure decisions, but at the same time far out enough that climate change impacts could be more easily distinguished from natural variability. The task then was to find published research on the climate change impacts of a 2°C increase in atmospheric temperatures over pre-industrial levels—the increase reflected in RCP 8.5 by 2050—on U.S. hurricane landfall activity. This information could then be used to create a catalog of events representing that climate.

Such a highly specific requirement—published research providing information about hurricane risk under an RCP 8.5 emissions scenario in 2050 that could be modeled—presented challenges. Many studies of hurricane risk focus on late-century impacts. Even then, one may find that only studies based on RCP 4.5 and RCP 8.5 are available. The solution came in the form of a GFDL study of projected Atlantic hurricane activity in the late century. GFDL was looking at RCP 4.5, which also produces a roughly 2°C warming, but by 2090. What was even more useful about the study is that it gave some visual indication of how the number of Category 4 and Category 5 hurricanes was expected to change under that amount of warming.

Helping Stakeholders Better Understand

The fact that the 2°C threshold of RCP 4.5 in 2090 is crossed by the RCP 8.5 scenario in 2050 made the GFDL study ideal for our purposes. While it may be true that other 2050 RCP 8.5 factors that are important for hurricanes may not equate to those projected to exist under 2090 RCP 4.5, the underlying atmospheric temperature is likely still a very important one.

Atmospheric temperatures are highly correlated with sea surface temperatures, which are, in turn, closely associated with hurricane activity, as evidenced by the Atlantic Multidecadal Oscillation (AMO). (It should be noted that the report makes no attempt to account for potential changes in the phase of the AMO that may—or may not—come to pass in the decades ahead. Nor does it account for any impacts that climate change might have on the AMO itself.)

This report provides a unique view of how climate change could affect hurricane activity in the U.S. by mid-century. It is designed to help stakeholders better understand and prepare for a situation that climate change may bring.

​ Explore the sensitivity of catastrophe risk to a changing climate