[Note: this post is second in a series that describes the Columbia Global Flood Initiative, a new Earth Institute initiative that attempts to better understand, predict and mitigate the impact of extreme floods around the world. See part one here.].
Traditionally, floods were thought of as essentially random, unpredictable events. When designing and planning for extreme floods, managers and engineers would look at the historical data to find and estimate how large an “extreme” flood in the area would be. Extreme floods, also called 100-year floods, are floods of a magnitude that have a 1 percent chance of being reached or exceeded in a given year for a given location.
This is an important number for engineers, water managers and policymakers as it frequently determines how floodplains are regulated and infrastructure will be designed—how large levees, floodwalls or reservoirs should be, for example.
Lately, however, a lot of people are wondering just how helpful the 100-year flood benchmark really is, as places seem to be getting hit by 100-year floods all the time. “I thought this was only supposed to happen once every 100 years” is an understandable reaction, if slightly misleading—the 100-year flood actually has a 1 percent chance of happening in any given year, which means that a 100-year flood happening two years in a row is certainly possible, if statistically unlikely.
After Iowa was hit with devastating floods in 2008, the cities of Cedar Falls and Iowa City revised their flood guidelines to say that new building had to happen above the 500-year, or .2 percent, benchmark.
So why do we keep getting so many 100-year floods? And given what we know about global warming, when we see the devastating floods of Pakistan, Australia, Brazil and elsewhere, a natural question arises: are floods getting worse? And is human-caused climate change to blame?
At first glance is seems that there are reasons to think so. Consider this chart from Improving American River Flood Frequency Analyses, National Research Council publication from 1999. The chart looked at flood magnitude for the American River at Fair Oaks, near Sacramento, California.
According to the chart, extreme flood magnitude for the area has sharply risen over the course of the 20th century—which seems to suggest that climate change is increasing the risk of flooding.
The problem with this reasoning is that it still relies on the assumption that flood risk is “stationary” (that the past is normally representative of the future). If we look at flow records for the American River going back to the 8th century, A.D. (reconstructed from tree ring data), it suggests that the severity of 100-year floods from this river follows a cyclical or “regime-like” pattern.
What this chart and many others increasingly show is that inter-annual, inter-decadal and longer-term variations in the planet’s climate impact the timing and magnitude of extreme floods. Many people are aware the El Nino/La Nina, in which a tropical Pacific climate pattern that cycles about every five years on average affects weather around the world. At a longer time scale, the Pacific Decadal Oscillation is a pattern that shifts on average about every 20 to 30 years.
Such changes can have complex interactions with other factors—if the primary direction of atmospheric moisture flow shifts even a small amount, for example, it can hit a mountain range from a different angle, resulting in much more or much less flooding of a given area over a certain period of time.
In other words, if designers of water infrastructure fail to anticipate longer-term climate cycles, they are misinformed as to the risk of flood. As a result, engineers may find themselves perpetually over- or under-designing infrastructure—either running a greater risk of catastrophe from unexpectedly strong storms that overwhelm infrastructure, or spending capital and resources and potentially causing unnecessary ecological damage by building for an event that is unlikely to occur over the lifetime of the project.
A better understanding of flood regimes, on the other hand, could allow scientists not only to design better infrastructure up front, but to better weigh the odds of shifting into a different climate regime at any given time as events dynamically unfold.
Calculating the odds of moving into a different climate regime, coupled with more accurate seasonal climate forecasts, could also help reservoir managers make more informed decisions about how much water to store or release ahead of a predicted flood season–thus decreasing flood risk while also allowing for ecologically and economically healthier stream flows.
But for a truly integrated flood management plan, designers need to think beyond infrastructure design and emergency management to help those who suffer from extreme floods rebuild their lives and livelihoods.
Understanding the climate/risk/flood connection can help governments and private groups design better financial mechanisms such as CAT bonds or index insurance to not only help people and institutions rebuild after an extreme flood hits, but to help reservoir managers and others better manage overall flood risk in innovative ways. I’ll discuss some of the ways that might work in a later post.