Conor here: Seems timely with Tropical Storm Chantal hitting North Carolina this weekend.
By Brian J. Yanites, Associate Professor of Earth and Atmospheric Science and Professor of Surficial and Sedimentary Geology at Indiana University. Originally published at The Conversation.
Hurricane Helene lasted only a few days in September 2024, but it altered the landscape of the Southeastern U.S. in profound ways that will affect the hazards local residents face far into the future.
Mudslides buried roads and reshaped river channels. Uprooted trees left soil on hillslopes exposed to the elements. Sediment that washed into rivers changed how water flows through the landscape, leaving some areas more prone to flooding and erosion.
Helene was a powerful reminder that natural hazards don’t disappear when the skies clear – they evolve.
These transformations are part of what scientists call cascading hazards. They occur when one natural event alters the landscape in ways that lead to future hazards. A landslide triggered by a storm might clog a river, leading to downstream flooding months or years later. A wildfire can alter the soil and vegetation, setting the stage for debris flows with the next rainstorm.

Satellite images before (top) and after Hurricane Helene (bottom) show how the storm altered landscape near Pensacola, N.C., in the Blue Ridge Mountains. Google Earth, CC BY
I study these disasters as a geomorphologist. In a new paper in the journal Science, I and a team of scientists from 18 universities and the U.S. Geological Survey explain why hazard models – used to help communities prepare for disasters – can’t just rely on the past. Instead, they need to be nimble enough to forecast how hazards evolve in real time.
The Science Behind Cascading Hazards
Cascading hazards aren’t random. They emerge from physical processes that operate continuously across the landscape – sediment movement, weathering, erosion. Together, the atmosphere, biosphere and the earth are constantly reshaping the conditions that cause natural disasters.
For instance, earthquakes fracture rock and shake loose soil. Even if landslides don’t occur during the quake itself, the ground may be weakened, leaving it primed for failure during later rainstorms.
That’s exactly what happened after the 2008 earthquake in Sichuan Province, China, which led to a surge in debris flows long after the initial seismic event.
Earth’s surface retains a “memory” of these events. Sediment disturbed in an earthquake, wildfire or severe storm will move downslope over years or even decades, reshaping the landscape as it goes.
The 1950 Assam earthquake in India is a striking example: It triggered thousands of landslides. The sediment from these landslides gradually moved through the river system, eventually causing flooding and changing river channels in Bangladesh some 20 years later.
An Intensifying Threat in a Changing World
These risks present challenges for everything from emergency planning to home insurance. After repeated wildfire-mudslide combinations in California, some insurers pulled out of the state entirely, citing mounting risks and rising costs among the reasons.
Cascading hazards are not new, but their impact is intensifying.
Climate change is increasing the frequency and severity of wildfires, storms and extreme rainfall. At the same time, urban development continues to expand into steep, hazard-prone terrain, exposing more people and infrastructure to evolving risks.
The rising risk of interconnected climate disasters like these is overwhelming systems built for isolated events.
Yet climate change is only part of the equation. Earth processes – such as earthquakes and volcanic eruptions – also trigger cascading hazards, often with long-lasting effects.
Mount St. Helens is a powerful example: More than four decades after its eruption in 1980, the U.S. Army Corps of Engineers continues to manage ash and sediment from the eruption to keep it from filling river channels in ways that could increase the flood risk in downstream communities.
Rethinking Risk and Building Resilience
Traditionally, insurance companies and disaster managers have estimated hazard risk by looking at past events.
But when the landscape has changed, the past may no longer be a reliable guide to the future. To address this, computer models based on the physics of how these events work are needed to help forecast hazard evolution in real time, much like weather models update with new atmospheric data.

A March 2024 landslide in the Oregon Coast Range wiped out trees in its path. Brian Yanites, June 2025

A drone image of the same March 2024 landslide in the Oregon Coast Range shows where it temporarily dammed the river below. Brian Yanites, June 2025
Thanks to advances in Earth observation technology, such as satellite imagery, drone and lidar, which is similar to radar but uses light, scientists can now track how hillslopes, rivers and vegetation change after disasters. These observations can feed into geomorphic models that simulate how loosened sediment moves and where hazards are likely to emerge next.
Researchers are already coupling weather forecasts with post-wildfire debris flow models. Other models simulate how sediment pulses travel through river networks.
Cascading hazards reveal that Earth’s surface is not a passive backdrop, but an active, evolving system. Each event reshapes the stage for the next.
Understanding these connections is critical for building resilience so communities can withstand future storms, earthquakes and the problems created by debris flows. Better forecasts can inform building codes, guide infrastructure design and improve how risk is priced and managed. They can help communities anticipate long-term threats and adapt before the next disaster strikes.
Most importantly, they challenge everyone to think beyond the immediate aftermath of a disaster – and to recognize the slow, quiet transformations that build toward the next.
I posted this in Links, but it’s relevant here as well. These links are from a study done by the University of Exeter for the Institute and Faculty of Actuaries in Britain:
Risks and Recommendations
Economic Impacts
Supplementary graphs (some real shockers)
The economic models had to be modified to take into account the cascading hazards problem.
Thanks HMP. Very comprehensive. Maybe too comprehensive to gain the attention of ‘leaders’.
The article above misses a key element of stream/river flooding: impervious surfaces. Even in rural areas (Texas’ Hill Country) as little as 10% increase in impervious surface (buildings, roadways, parking lots) will change stream/river hydrology adversely (rapid rise). Of course, any concentrated downpour is going to add to adverse river flow, as well.
The Texas Hill Country is an impervious area as it mainly on a limestone base with thin layers of soil. Flash floods are a known risk in the area. Buildings may exacerbate the situation.
Comprehensive in terms of bracketing outcomes, but not necessarily useful. Yanites does an excellent job here laying out the nonlinearities in terms of path analysis. The particulars matter in a way that cannot be predicted. We can predict at the level of saying there is likely to be increased erosion, and the sediments will need to be handled downstream.
But the patchiness means that we can’t predict where the catastrophic local effects will occur. A couple of days ago we were forecast to get a quarter inch of rain. We got two & a half in a couple of hours, but ten minutes across town was dry. So the average smeared out, but the kurtosis spike looks like a singularity if it hits you.
This post dovetails with Financial Times on Coming Climate-Change-Driven Meltdown in Real Estate: Gradually, Then Suddenly? It’s too expensive to mitigate everywhere. That’s fine for an insurance company assessing individual risks per policy. But trying to run an economy based on mortgages requiring insurance is a recipe for ruin.
Stream flow models can incorporate radar data when calibrating to better account for convective rainfall irregularities that rain guages miss. However, this feature isn’t especially well known (I don’t think it’s advertised) even though the feature is built in to sone models and takes time to master and run.
I just learned this a few months ago and its a game changer.
After months they reopened the important I-40 crossing from NC to TN. And then a few weeks ago a heavy rain sent rocks downslope to block the temporary two lane.
However that blockage was cleared in less than two weeks. The bad news is that restoring the full four lane interstate as it once was may take years.
My town is getting over Helene even as a few trees continue to fall. Steep sloped North Carolina may take a lot longer. What we all hope is that nothing like Helene will happen again. A faint hope?
It’s significantly more likely another Helene will occur in due course, although I’m sure you already knew that…
Out of curiosity, where are you located? I am considering buying land south of Black Mountain near some friends. Flood-resilience has been a big topic of conversation. I got real lucky my land in Yancey county wasn’t obliterated like some friends’ places were.
And I find it helpful not to hope. Prepare, monitor and then don’t worry. Of course, the old habit of worrying is hard to release!
I live in upstate SC. As you know Black Mtn got slammed by Helene. But this was the first time we’ve ever had a direct hit by a hurricane–at least in my lifetime–and one has to believe the odds of another are not that great. There have been lots of hurricanes down on the coast which is over 200 miles away.
Where I live there are still millions of trees and the ones that went down were often shallow rooted hardwoods or oaks without a windbreak. Often the bendy pines took it in stride. Sloped ground undoubtedly makes the danger worse.
I’ve lived in the Chapel Hill & Durham portion of NC and left there in 2006; and since 2016 I have lived here in upstate region of South Carolina. I don’t need to repeat what a fellow resident has already shared late Monday. I find much to like about living here, but zoning and central planning / organized codes might seem like a mish mash, just depends on which county you live or drive through.
There are likely several counties in NC where you can find available land to own or pursue, either further east of or possibly south of Black Mountain. I know Greensboro pretty well ( where family has lived for nearly 40 years ) and also some of the counties west and southwest of Guilford county.
Have lived off and on in NC since 1954 (age 6), just in time to see Hurricane Hazel devastate most of eastern NC, our new home. Returned in 1968 and again in 1989, when Hurricane Hugo clobbered Charlotte a week or two after we moved there. Last time was 1996-2010, but we kept a vacation home within the City of Hendersonville until this year, just long enough to see Hurricane Helene make Henderson County #2 statewide in Helene fatalities. Looks to me like damaging hurricanes reaching beyond NC’s “fall line” (its western half) are increasingly frequent, even if not yet “normal.”
In my opinion, anyone interested in western NC should avoid residing anywhere that would become a flood plain of ANY degree (i.e., not even a “1000-year” flood) and make sure that in the case of such a flood you will still have access to food, fuel, shelter, and escape routes. I was not in Hendersonville when Helene hit, but my neighbors were stranded without groceries, gas, electricity, and comms for about 10 days even though our places were not harmed. You might also want to keep a close watch on any tropical storm whose name begins with “H.”
With the amount of money in their floats, insurance companies could afford to be proactive in reducing risk to their bottom line by being proactive in mitigation of potential disasters to save themselves from downstream claims. Of course it won’t happen, cheaper to pull out of the region and socialise the cost.
You do realize that your demand of the insurers subsidizing the bad living habits of some people is socializing the costs since everyone that they insure will have to shoulder cost of the few.
.
There are a bunch of people demanding to live in disaster areas. They also demand that they don’t have to pay more then the more prudent people in the state (Or there is a law that prevents that kind of selective premium based on the risk zone one lives in). Why should an insurer subsidize them?
And that is what happened, insurers left instead of subsidizing those people.
So now those people instead of taking the hint that it might be smarter to not live in an disaster area are going to demand that the government subsidizes them. And the government will likely cave instead of taking the hint and saying that they aren’t going to subsidize living in a disaster area.