Climate Risk's Cracked Rearview Mirror
Risk Market News has relaunched the Risky Science Podcast, with in-depth discussions with scientists, insurers, investors, portfolio managers, and others about the evolving science of predicting and modeling risk across both natural and man-made perils.
In the first episode of we sit down with Dr. Daniel Swain—climate scientist, meteorologist, and founder of the popular Weather West blog—for a deep conversation on the breakdown of historical assumptions in weather and climate modeling.
As Dr. Swain explains, we’re now operating in a non-stationary world—one where the past is no longer a reliable guide to the future. From increasingly destructive wildfires and hurricanes to uncharted extremes in heat and precipitation, traditional catastrophe models and historical loss patterns are proving insufficient in assessing today’s evolving climate risks.
The discussion explores how high-resolution, physics-based models are beginning to close that gap, and why relying solely on legacy statistical tools or coarse climate models can underestimate real-world exposure. Dr. Swain also shares his views on the promise and limitations of artificial intelligence, the role of hybrid modeling approaches, and why public-private partnerships are essential to maintaining trusted and transparent risk tools in an era of scientific disruption.
Finally, we dive into the growing role of the private sector in climate risk modeling—and the potential dangers of fragmentation and opaque methodologies as public funding in science faces increasing pressure.
Some key discussion points...
🔁 The Limits of Historical Catastrophe Models
"The historical way of doing things, looking at historical losses to predict future losses is not going to cut it—not only because of climate change, although definitely because of climate change—we live in a non-stationary world."
📉 Underestimating Risk Due to Sparse Event Histories
"There are some stretches of coast where it just appears that by some miracle, a big hurricane just happened not to make landfall in a highly hurricane-prone area for 100 years. And in some of these catastrophe models, historically, that section of coast would be listed as being at lower risk... even though from a broader geophysical perspective, it was just as much risk."
🔥 Wildfires as a Modern Example of Model Failure
"In California and any number of other places in the last 10 or 15 years, there have been so many wildfire loss events that were so far above anything that had been seen anywhere on earth previously... We didn’t really have contemporary historical examples of [that severity] at all, as it turns out."
📊 The Shortcomings of Past Risk Assumptions
"We didn’t do a good enough job... being sufficiently wide-eyed—not being too myopic about what the natural world could have thrown at us even absent climate change. And now we’re making that problem much worse by broadening the envelope of what is possible."
🌍 The Broader Scientific Implication
"We live in a non-stationary world. Stationarity is even more dead than it was a decade or so ago, when that turn of phrase was first coined."
You can listen to the full discussion by clicking on the podcast logo below.
RMN paid and free members can access the podcast the full podcast transcription.