Hurricane Harvey studies: Yesterday’s 100-year storm is today’s 30-year storm


The story explaining the incredible flooding in Houston during Hurricane Harvey has many chapters, ranging from meteorology to the history of groundwater use and development zoning. The chapter on climate change has already had a few pages filled in, thanks to a study quickly published by MIT hurricane scientist Kerry Emanuel. This week, two complementary studies flesh the chapter out a little more.

The first paper comes from a group of scientists who have worked to rapidly analyze a number of extreme weather events over the past few years, including flooding in Europe and Louisiana last year. The general strategy for this type of undertaking is not entirely dissimilar from tracking the home run hitting of steroid-using baseball players. You can’t really know if an individual home run would have occurred sans steroids, but that’s not the point. Instead, you work out whether home runs like the one you just witnessed are generally being juiced.

In this case, the researchers were able to build on their analysis of the nearby Louisiana deluge from 2016. As in that study, they analyzed the history of rainfall measurements in the region to work out just how unusual the incredible rainfall totals from Harvey were—and whether the chances of an event like that have changed over time.


Even in today’s climate, Harvey’s rainfall was at leasta 1,000-year event, meaning there’s just a one-in-a-thousand chance it happens in any given year. But the researchers see evidence that the odds would have been even smaller a century ago. They estimate that there has been a (roughly) 18-percent increase in the intensity of extreme rain events along this stretch of the Gulf Coast since the late 1800s.

One of the simplest relationships in climate change is how the water-vapor capacity of air increases about 6-8 percent for every degree Celsius of warming. The Houston area has warmed just about 1°C, so that’s how much you might expect rainfall extremes to change if nothing else was going on. One thing that can amplify that simple relationship, though, is the fact that water vapor added to the atmosphere also releases heat when it condenses into droplets. This causes the air to keep rising until even more water is wrung out, which can mean stronger storm events.

The rainfall records seem to support something like that scenario. Storms around Houston are producing more rain than they used to.

To test the cause of this change in rainfall intensity, the researchers also simulated the region with several climate models. Model simulations like these would include the occurrence of “blocking high” pressure setups like the one that kept Hurricane Harvey pinned over Houston for so long. That is, after all, one way to get high rainfall totals.

They threw out one model because it couldn’t do a good-enough job matching past extreme weather patterns in this specific area. Interestingly, the two models that passed that test produced slightly different answers. One model showed a 17-percent increase in extreme rainfall between 1861 and 2017—quite similar to the trend found in the actual rainfall observations. However, the other model simulated a smaller increase of 8 percent, which is close to the unamplified effect of water vapor capacity in warmer air.

In the end, the researchers go with an average of the different estimates, concluding that our best understanding is that human-caused climate change has made modern storms—including Harvey—dump about 15 percent more rain. Another way to describe that is to say that a storm of Harvey’s magnitude is about three times more likely now than it was in the late 1800s.

Double check

A second study, published by Mark Risser and Michael Wehner at Lawrence Berkeley National Laboratory, performed a similar analysis of past weather station records. Their methods were a little different: they focused on 7-day rainfall totals rather than 3-day totals, for example, and checked for the influence of El Niño/La Niña conditions. But their results were basically the same.

They estimate that Harvey’s rainfall was probably almost 19 percent higher due to global warming, which also means the probability of a storm of Harvey’s size is now 3.5 times more likely.

The researchers use the previous record-high 7-day rainfall as an example. In 1950, that would have been about a 200-year storm. But in 2017, it is closer to a 30-year storm.

Although a 1,000-year storm is bound to happen somewhere in the world fairly regularly, it’s still ultimately “bad luck” when it happens to you. A number of chance events had to line up to make Hurricane Harvey as extreme as it was. But on the list of things humans did that made Harvey and its effects worse than it otherwise would have been, climate change cannot be ignored.

Environmental Research Letters, 2017. DOI: 10.1088/1748-9326/aa9ef2
Geophysical Research Letters, 2017. DOI: 10.1002/2017GL075888   (About DOIs).