In February, after torrential rains led to partial failure of a spillway on Northern California’s Oroville Dam, forcing evacuation of nearly 200,000 people downstream, Stanford climate scientist Noah Diffenbaugh wrote an op-ed in The New York Times entitled “What California’s Dam Crisis Says About the Changing Climate.”
My colleague, Upmanu Lall, director of the Columbia University Water Center, commented on the article on Climate Feedback (a website that organizes climate scientists to comment on the accuracy of mainstream media articles on climate): “Misleading title—the dam crisis is primarily about the aging infrastructure and the neglect of investment in engineering and R&D as to how to plan, design, and build a new generation of water and energy infrastructure, while maintaining what we have. It tells us nothing about climate change.”
So is Oroville a climate change story?
“Natural” disasters that threaten infrastructure (as well as lives) are occurring more frequently. There’s no question that much of the problem is just the aging of the infrastructure itself, but the changing climate does mean that the conditions any future dams, roads, and bridges must face will be to some extent different from those in the past. In principle, any upgrades should account for those changes. The question is how big the changes are, compared to the various sources of uncertainty and subjectivity that would be involved in building public works to withstand weather even in the absence of global warming.
The climate attribution studies that can assess quantitatively the role of human influence on the rains that caused the Oroville failure haven’t been done yet. But it’s almost certain that at most, they’ll find a modest, partial influence, a “loading of the dice” to make an event of this magnitude more probable. In other words, a rainfall event of the same magnitude almost certainly could have happened anyway, even without any human influence on climate, just with lower probability. And those probabilities themselves are not known that accurately. Random chance and natural variability play large roles. Truly extreme weather events are rare by definition, and historical data records are often too short to characterize their statistics well.
And even if we had perfect knowledge of the statistics of weather, there would be a subjective element in engineering design. With limited resources, nothing can be made completely indestructible; a dam can only be built to handle some specific level of flooding, within reason. What “within reason” means is not something that science alone can say. Accounting for global warming, in practice, may be equivalent to just making a more conservative judgment call.
So one doesn’t have to reject the science on climate change to argue that, at least in some cases, it isn’t the main motivation for fixing our aging infrastructure. Much of it would need attention even if global warming weren’t happening. And because chance and natural variability will dominate the statistics of many kinds of extreme events for some time yet, preparing for climate change may not be that different from preparing for extreme events that would have been possible even in the pre-industrial climate.
Any decision process should use the best scientific information, including information about the human influence on climate. But because of the uncertainties in the statistics of extreme weather, the human element involved in deciding how conservatively to build anything, and the fact that many facilities like Oroville would be in bad shape today even if the climate were static, it should be possible to agree well enough on how to fix them without needing to see eye to eye on how big a part of the problem global warming is. Maybe we can save the big arguments for energy policy.
— Adam Sobel is a professor at Columbia University’s Fu Foundation School of Engineering and Applied Sciences and Lamont-Doherty Earth Observatory, and director of Columbia’s Initiative on Extreme Weather and Climate