The Science of Predicting Climate Change and Future Flooding

A study of stream gauge records for the past 100 years shows that precipitation has been increasing in the Northeastern U.S., and that heavier precipitation events have been getting heavier.

by / May 27, 2014
Mathias Collins is a hydrologist with the National Oceanic and Atmospheric Administration. Brian Smith

Mathias Collins is a hydrologist with the National Oceanic and Atmospheric Administration (NOAA). He has participated in a number of studies on flooding and climate change, including one that examined stream gauge records for the past 100 years and found a trend toward more frequent floods with a heavier magnitude. We talked to him about this study and predicting flooding and climate change in the coming decades.  

Emergency Management: You studied stream gauge records of the past 100 years in the Northeast and found an upward trend in frequency and magnitude of floods. Tell me what that study told you and if it was alarming.

Mathias Collins: I would start by saying I don’t consider it alarming, but what we found was interesting in light of two things.

On one hand, the result wasn’t unexpected because of the precipitation trends in the region. Precipitation has been increasing in the Northeastern U.S., particularly so in the upper ends of distribution. The heavier precipitation events have been getting heavier.

So in one sense, the findings weren’t surprising, but in another they were. Flooding doesn’t always directly reflect changes in precipitation. That’s because the land surface modifies what happens to the precipitation that’s delivered to it. Our studies were trying to get a read on if there are trends and if there is a climate signal to those trends.

We have to carefully select stream gauges to try to control for that change in land use. Urbanization also increases flooding so we want to make sure we’re not measuring that. All that is to say we go through an extensive process to identify gauges that measure stream flow that’s as natural as possible. That’s hard to do.

There are large areas of the eastern U.S. that have been fairly dramatically reforesting over the last 100 years. So we have urbanization concentrating and at the same time less forestry and farming. When you look at the New England landscape compared to 100 years ago, it’s much more forested. That would tend to, over a 100-year record, damp flooding. Forested cover tends to do the opposite of urbanization, which is damp flooding.

That’s where our findings were surprising. We didn’t expect to see a direct translation of the precipitation increases into flooding increases, but we still have a fairly strong signal of flooding increases. Not at every gauge but a majority of gauges, and we have a number that are statistically significant.

EM: It sounded like an ominous trend. But you wouldn’t call it alarming?

MC: I wouldn’t, but I would say it’s important to look at the work in the context of other things as well. We tried to control for urbanization, but of course much of the area where we have these finding we do see urbanization.

Where the changes get more dramatic is when you consider urbanization on top of climate impacts. If anything is alarming, that’s probably where you would see it. People in those urbanizing watersheds may be seeing their flooding issues as mostly urbanization, but they’re partly climate. And the other thing is the frequency aspect, because that’s where we find our strongest findings. ... The strongest findings are in the numbers of floods per year; they’re increasing and that’s purely an atmospheric phenomenon.

And so areas where you have urbanization, you not only have the magnitude enhanced by the changes in the climate but you also have the frequency of flooding changing. Whether that’s alarming or not I think is on a case-by-case basis, but certainly our work suggests people should be considering this when they’re thinking about design or emergency management. Certainly frequency is increasing.

EM: Explain what the North Atlantic Oscillation (NAO) is and how it relates to climate change.

MC: In our work when we look for causal agents for these changes, one thing to look at are agents of cyclic climate variability. Basically in the atmospheric system, there are semi-permanent features of the atmosphere and they tend to oscillate between semi-stable states.

All of that is to say it influences storm tracks. The NAO has been known to affect whether we have wet or dry conditions over long periods, years to decades. So we explored whether there was some association between the NAO phase and flooding, and we did find some relationship although it’s complicated and weak. It’s not a direct relationship, but it’s persistent and we believe it’s impacting flooding, both in magnitude and frequency.

We also believe it’s not the only thing impacting the flooding. Floods are complex phenomenon that tend to rely on a certain set of conditions [converging] at a certain time. We have to start thinking of it as a suite of things that all point toward greater magnitude and greater frequency.

So we still believe the NAO is important and is the function of cyclic climate variability as opposed to climate change. It’s a natural variability. It doesn’t explain everything. But when you look at some other phenomenon, those too point to a greater likelihood of flooding.

For example, a colleague did a study on how “closed low” storms have been increasing in the Northeast. A closed low is a category of storm type, a measure of the strength of the low-pressure system.

The frequency of closed lows in the Northeast and the quantity of precipitation associated with them increased significantly since the late 1940s and this is consistent with what we’d expect from human-induced increases in global mean temperature.

The conclusion is there isn’t one explanation. We should be looking for a number of things that produce flooding. They’re episodic phenomenon, and they require a certain number of things all happen to cause them. You need the rain; you need wet soil and antecedent. And then add the blocking phenomenon on top of that. It’s a grouping of phenomenon that happen at once.

EM: How should we mitigate these trends? Are we on the right track?

MC: The question as I see it is that when people are designing projects on a project-by-project basis, are they taking this kind of information into account? Oftentimes these processes default to existing information, existing studies, existing data and they don’t automatically include updating recent studies.

It’s almost a case-by-case thing but our design teams take it upon themselves to re-examine existing flood frequency estimates and information like that to reflect how our rivers have been flooding for the last 30 years. We document these changes as taking place over the last 30 years, like a step change starting in 1970, which is a date a lot of natural phenomenon have been identified as changing.

But even if you update to reflect the most recent decades, there still is a question of does the recent data give us a guide to what we can expect in the coming decades? We don’t really know. If we believe these changes are a function of anthropogenic [human-caused] warming, which we’re frequently calling climate change, is the last 30 years a good guide to the next 30 years? We don’t know, but that’s how design is done.

We estimate what’s likely to occur in the future based on the past. And that’s the fundamental nut of climate change as applied to planning. Is the past going to be a reasonable guide to the future?

EM: That leads into my next question. Is it possible to look into the future with any kind of accuracy?

MC: I think it’s really hard. We’ve been wrestling with that as we take some of these findings and try to think of them in a context for design and how can we give recommendations on what to do for design.

Extrapolating into the future and assuming climate change is generally a modeling exercise, and you take model estimates from circulation models. There are a number of ways you can use those model estimates, but you can then project the future based on model estimates for precipitation and so forth, but those are fraught with problems.

It’s a real difficulty because in the design world, we often feel like our best guide is experience on the waterway. What’s it done in the past? And we’ve been saying with our research that if you’re not looking at the recent past you’re definitely out of date, but is even that good enough? We’re not sure.

This story was originally published by Emergency Management

Jim McKay Contributing Editor

Jim McKay is the editor of Emergency Management magazine. He lives in Orangevale, Calif., with his daughter, Ellie, and son, Ronan. He relaxes by fly fishing on the Truckee River for big, wild trout.

Platforms & Programs