Search Magazine  

Multi-faceted forecasting

Extending the reach of climate change research


Some of the effects of Hurricane Sandy on Mantoloking, New Jersey. Photo: National Oceanic and Atmospheric Administration,
Remote Sensing Division
Some of the effects of Hurricane Sandy on Mantoloking, New Jersey. Photo: National Oceanic and Atmospheric Administration, Remote Sensing Division

Climate simulations predict that storms like Hurricane Sandy, which ravaged swaths of the New York and New Jersey coastline last year, will become more frequent and intense in coming decades. Increases in sea level prompted by a generally warming climate would only add to their impact. Such unsettling projections have focused attention on the vulnerability of densely populated coastal communities and the need to plan for the effects of climate change.

This sort of foresight is the province of ORNL's Climate Change Science Institute and CCSI climate researcher Ben Preston. Preston explains that bringing together the range of scientific talent needed to address the challenges of a changing climate is what the CCSI is all about.

Putting the pieces together

Several years ago ORNL had separate groups studying climate modeling, ecosystem science, climate data management, and the consequences of climate change.

"We had internationally known research capabilities and expertise in each of these areas," he says, "but we hadn't put the pieces together. The CCSI is designed to bring these different capabilities together and build synergies among them."

Preston, who also serves as the institute's deputy director, said collaboration has already produced results through the CCSI's Next-Generation Ecosystem Experiments. "This study of the effects of warming on Alaska's terrestrial ecosystems relies on both ecosystem scientists and climate modelers to understand how the ecosystems of the high Arctic are affected by climate change," he says.

The effort, involving dozens of researchers from several national labs and universities, is generating a lot of data—in the form of observations of surface and subsurface ecosystem processes as well as modeling data. As a result the CCSI is also calling on data management specialists to organize and manage the data to make it available to the rest of the research community.

"The NGEE project illustrates the kind of interdisciplinary research the institute is doing," Preston says. "Anyone interested in the consequences of climate change— whether he or she is working on a global scale or regional scale—would benefit from having access to climate modelers who can tell them something about how the climate is going to change, what's going to happen with extreme weather, and on what timescales. This broad-based expertise enables our studies to say something about what the consequences of a changing climate could be for areas as diverse as energy production, agricultural yields or the risks to coastal communities."

Energy impacts and more

Since the CCSI's primary sponsor is the US Department of Energy, its research on the societal consequences of climate change naturally focuses on the US energy sector. To understand how climate affects the nation's energy infrastructure, scientists need to know how it affects basic environmental factors such as water availability—because producing energy often involves using a lot of water—or how changes in the climate might affect the productivity of different types of bioenergy crops.

"Our major focus is on energy," Preston explains, "but in order to do that we really have to consider a range of energy systems, which means we have to look at many different types of impacts."

The institute also studies how changes in the global climate could affect homeland security (as a result of disruptions such as crop failures or food shortages) or critical infrastructure (through the impact of drought on hydroelectric power generation, for example) and the possible impact of storms and rising sea levels on coastal communities. "Hurricane Sandy brought the vulnerability of certain areas of the coast into sharp focus," Preston says. "These storms may not happen frequently, but when they do, densely developed areas have a huge amount of exposure."

He explains that places such as New York and New Jersey are growing quite rapidly, particularly along the coast. So if a storm like Sandy hits the same area 50 years from now, the damage is going to be that much greater.

"A lot of my work focuses less on what's happening with the climate or with physical systems like buildings and infrastructure, and more on understanding what's happening with society—how it changes over time in terms of demographics, economics, and risk management," he says. "We want to know how those factors interact with the effects of a changing climate."

The CCSI uses a number of computational tools to better understand this relationship. These include many computer-based climate models mixed with scenarios that CCSI scientists develop by analyzing trends in population growth and economic development. "To consider the possible future social and economic impacts of changes in the climate, we need to know where people and resources will be located," Preston says.

In harm's way

A recent study by Preston published in Global Environmental Change makes the point that climate change and changes in society can't be viewed as unrelated. "Everyone wants to know how the climate is going to change," Preston says. "Can we predict future climate change? Can we predict changes in the frequency or intensity of extreme weather events like heat waves or hurricanes or flooding? If we only think about the climate, we see that a lot of our understanding of future consequences is dominated by what we know or don't know about these climate variables. However if you look back at the last century, you'll see that economic losses related to extreme weather events have been rising—not necessarily because of climate change, but because we have been developing hazardous landscapes and putting more and more people in harm's way."

To illustrate this point, he notes that before the proliferation of air-conditioning, Florida was not a place where too many people wanted to live. However, since the 1960s the population of Florida, as well as the rest of the US Southeast, has been booming—despite the fact that it's particularly prone to hurricanes and flooding.

The study considered how changes in population and wealth across the US will influence exposure to leading causes of climate-related damage such as hurricanes, wildfires, tornadoes and flooding.

"We know losses related to these events have a particular geographic pattern," Preston says. "Losses related to hurricanes, for example, are related to coastal areas, particularly in the Southeast. Wildfire damage is specific to the Southwest. Tornadoes and floods mainly strike the Midwest."

Turning this general information into detailed maps required plowing through a great deal of historical loss information, mostly federal government records. Once Preston had national maps of the distribution of losses, he projected future patterns of "societal exposure"—that is, how many people will be living in these areas in coming decades and how well-off they will be.

"To do that we developed a demographic model based on historical information," he explains. "We generally know what the population size is for every US county at present, and we know the birth rates and death rates for various age groups and ethnicities. Using this information we built a model of the future population for every county. We also made the assumption that the economies of the counties will continue to grow over the next few decades as they have in recent decades. This assumption is open to question, but at least it allows us to project the implications if we continue on our current path."

Multiplying the average wealth of an individual in each of these counties by the number of people gave researchers a measure of the economic exposure for each county. Then they predicted the change in population of each county over time. The model predicts, for example, that the population of the Great Plains is expected to decrease over coming decades, while the urban and coastal areas of the South are likely to continue to experience increases in population that are higher than the national average.

By integrating climate and demographic information in this way, researchers came to a number of conclusions about what areas of the country were most at risk for climaterelated damage. For example, Florida's losses resulting from extreme weather events are expected to grow up to a factor of 5 by 2050 due to increasing population and wealth.

"This kind of information allows us to place the discussion about climate change in some kind of context," Preston says. "Yes, we're concerned about climate change. Yes, we're concerned about how it might influence hurricanes and other extreme events, but a lot of the big changes in such events are projected to arise in the latter half of the century. If you're talking about the next few decades, losses in these rapidly developing areas might increase by a factor of 2 or 5 or even more, simply because we're putting more people in harm's way."

Pressure for predictions

Preston acknowledges increasing pressure to move from understanding how climate systems work to making predictions—whether it's projecting how the climate will change over the next several decades or predicting agricultural yields next year.

"That expectation is putting new demands on our science and modeling capabilities, and it might send us back to the drawing board to develop the next generation of computational tools," Preston says. "Of course, that's understandable; it's the iterative nature of science. We want to continually bring our best tools to bear on any question."

Newly acquired data may be able to provide some of the information needed to predict changes in the climate.

"Our Next-Generation Ecosystem Experiments are a good example of that," Preston says. "We actually don't know a whole lot about the dynamics of Arctic ecosystems, how carbon dioxide is released from permafrost, or how these things might influence the global climate system. Large-scale experiments like NGEE allow us to gain a better understanding of what's going on at the process level. We can use that knowledge to improve our models."

He also notes that there's a growing recognition within the climate science community that some simulation tools may not be up to the challenge of providing longterm predictions of climate change impacts. For example, agricultural crop models that are used to simulate the effects of climate change on agricultural productivity often do not capture the range of crop responses to climate, particularly climate extremes, observed in the real world.

"There is a move in the crop modeling community to start from scratch and build new models that are designed from the ground up to simulate, capture and represent the long-term climate change processes we are interested in," Preston says.

"Scientists have been studying the climate for a long time—centuries. We've made a lot of progress, and we've got a lot of knowledge. Now that our research is having a broader impact and is of interest to a broader audience, the pressure to extend the reach and utility of climate predictions is particularly intense. I think you see this across the climate change arena. The CCSI will play a key role in meeting that challenge." —Jim Pearce