September’s Tropical Storm Imelda was the fifth-wettest tropical cyclone on record in the U.S. and the third slow, wet storm to impact the U.S. in as many years. Last year, Hurricane Florence set the record for the wettest tropical cyclone to impact the Carolinas—its crawling motion contributing to its precipitation totals. In August 2017 Hurricane Harvey was the wettest storm to flood Texas and now, only two years later, Imelda is the second-wettest storm to flood the state. Imelda’s impact might have been even worse had it stalled, as Harvey did; instead, Imelda continued moving northward across western Louisiana, albeit slowly.
A Slowing Trend
A growing list of peer-reviewed scientific articles has noted that, in recent years, tropical cyclones are moving more slowly over water and even more so over land. Some have suggested that this slowing down might be the result of climate change. A physically plausible explanation consistent with expected climate change–induced impacts on large-scale steering mechanisms would go a long way toward bolstering confidence in, and an understanding of, their findings.
One very general explanation is that climate change is reducing the pole-to-equator temperature gradient. That weaker gradient may be slowing the large planetary–scale atmospheric waves that move around the globe, which means that weather systems, including tropical ones, will move more slowly or even stall. But the specifics of the finding in terms of whether different intensities or different locations make a difference is worth exploring.
Inspired by Imelda and curious about the academic research, I computed the 24-hour average speed post-landfall of U.S. hurricane–strength storms since 1979 (all landfalls with at least one per storm at hurricane strength), which is more or less the start of the satellite era. I looked at four different groups by combining two regions and two intensity bands. Spoiler alert: my results largely confirm the reported slowing down of storms, but it is not a slam dunk.
For storms making landfall along the Gulf Coast (including the west coast of Florida) with central pressure (Cp) of 967 mb or less (e.g., strong storms), forward motion has slowed by a little more than 4 mph on average, or almost 30%, since 1979 (1979-2016, to be precise—a period of 38 years). How robust was the data? Twelve years out of the 38-year period produced qualifying landfalls (again, Gulf Coast and Cp <= 967 mb) equally split between the first 19 years and the second 19. Five of the six contributing years in the first half had annual average forward speeds of more than 10 mph at landfall. In contrast, only three of the six years in the second half had annual average forward speeds above 10 mph.
For weak tropical cyclone landfalls (with Cp > 967 mb) along the Gulf Coast, the trend actually goes the other way at twice the magnitude. Even though there were more landfalling years for this group, a big gap between Humberto in 2007, which made landfall at 11 mph, and Hermine in 2016, which made landfall at 21 mph (highest forward speed in the 38-year period for storms with landfall Cp > 967 mb), greatly influenced the overall trend.
My results for weak storms making landfall along the East Coast were more mixed. Although the trend shows an overall drop of about 3 mph over the period (a change of about 28%), the last weak storm to make landfall along the East Coast before 2017 was Ophelia in 2005 that grazed North Carolina at less than 4 mph.
Variability and Gaps
It turns out that, while some of the magnitudes of the trends from the three groups I just mentioned look impressive, there is so much interannual variability—not to mention the multi-year gaps between landfalls and the brevity of the record—that it may take another 40 years in some instances for the signal of the trend to exceed the noise of the interannual variability of forward speed (assuming trends and variability don’t change).
But of course, I have saved the most significant group for last (as any good storyteller would, I suppose). For strong storms making landfall along the East Coast, the data show a slowdown of a whopping 11+ mph, from 26 to 15 mph. That’s an over 40% decrease, and that’s significant—statistically significant, in fact.
And even though the 30% slowdown I found for strong storms along the Gulf Coast is not as robust, the notion of strong storms making landfall at increasingly slower speeds along highly populated sections of the U.S. coastline is a cause for concern, in terms of increasing losses from wind, precipitation-induced flooding, and storm surge. The fact that Imelda was a tropical storm in the Gulf and seemingly contradicted the trend for that group (which is one that I could not evaluate within the AIR U.S. Hurricane Model), makes it even more concerning that very slow and very wet storms of all intensities have occurred for the last three years in a row.