New Findings Point Towards Higher Climate Sensitivity

Eric Ellman

For the second time in a month, scientists at Yale University have found evidence that current models of global climate may be underestimating how much warming will occur as greenhouse gases in the atmosphere continue to rise. Trude Storelvmo, associate professor in the Yale Department of Geology & Geophysics, is principal investigator on two studies that independently point to higher values of climate sensitivity than the estimates provided by the Intergovernmental Panel on Climate Change (IPCC) in its last assessment report published in 2013. 

The concept of climate sensitivity has been used by scientists for many years to describe how Earth’s climate system will respond, on average, to increasing atmospheric concentrations of greenhouse gases such as carbon dioxide (CO2). At the start of the Industrial Revolution, around 1750, the atmospheric concentration of CO2, the most important greenhouse gas, was about 300 ppm. Today it is above 400 ppm and rising by about 2 ppm per year, mainly because of COthat is released by the burning of fossil fuels in the world’s energy systems. 

In 1979, MIT scientist Jule Charney led a study for the National Research Council that made the first modern estimates, using global climate models, of what is called the “equilibrium climate sensitivity”—a quantity that is intended to represent Earth’s average temperature response, in the long run, to a doubling of the concentration of atmospheric CO2. Here, the “long run” means after the climate system has stabilized, or reached equilibrium, at the higher concentration; since CO2 remains in the atmosphere for thousands of years, this can require extremely long time scales by human standards.

The Charney report estimated that a doubling of CO2 would result, at equilibrium, in an average global temperature rise between 1.5 and 4.5 degrees Centigrade (between 2.7 and 8.1 degrees Fahrenheit). Nearly forty years after the Charney report, the range of estimates for equilibrium climate sensitivity is about the same. For example, the 2013 assessment report of the IPCC (the Fifth Assessment Report, usually called AR5) uses the following wording: “Equilibrium climate sensitivity is likely in the range 1.5°C to 4.5°C… The lower temperature range is thus less than the 2.0°C in AR4, but the upper limit is the same.

The work by Storelvmo’s group suggests that scientists may actually be close to designating the lower values in the currently accepted range as being “unlikely.” According to Michael Mann, Distinguished Professor of Atmospheric Science at Penn State University, the latest studies “provide sobering evidence that Earth’s climate sensitivity may lie in the upper end of the current uncertainty range.”

The research published today in Science by Storelvmo, her graduate student Ivy Tan, and Mark Zelinka of Lawrence Livermore National Laboratory (Tan et al, 2016) analyzes satellite measurements to conclude that the middle and upper troposphere contain a higher ratio of water to ice than was previously believed. The water is present in the form of super-cooled droplets; the ice, in tiny crystals. The ratio of water to ice is critical because tiny water droplets are highly reflective of incoming sunlight and ice crystals are not. A higher liquid fraction in the upper atmosphere means greater reflection of incoming solar radiation: less sunlight and heat reach the Earth’s surface and there is less warming. This is a classic example of negative feedback in the climate system: as the atmosphere starts to warm because of increased greenhouse gases, a mechanism sets in to dampen the effect—in this case, suspended ice crystals in the upper troposphere melt into tiny water droplets that reflect solar energy back into space.

Current climate models correctly simulate this effect by replacing ice with water as the world warms, but these models appear to have overestimated the actual availability of ice. Less ice to melt means that less water will form to brighten the atmosphere. The latest findings by Storelvmo and her colleagues indicate that much of the incoming solar radiation that today’s models predict will be reflected back to space (as tropospheric ice converts to water) will actually pass through the atmosphere and heat the Earth.

To calculate how much additional heating will occur, the authors configured the National Center for Atmospheric Research’s Community Atmosphere Model (CAM 5.1) to calculate equilibrium climate sensitivity using a range of atmospheric water-to-ice ratios based on the latest satellite estimations. The newly constrained model predicted an equilibrium climate sensitivity (ECS) of between 5.0 and 5.3 degrees, compared to an estimate of 4.0 degrees using the conventional water-to-ice ratio. Storelvmo expects that when other researchers adjust their respective models to reflect appropriate proportions of ice and water in mixed-phase clouds, all of their ECS values will shift upward.

The implications of higher ECS are profound, and might already be making themselves felt were it not for a different phenomenon described by Storelvmo earlier this month in a paper in Nature Geoscience (Storelvmo et al 2016). That study used ground-based measurements of incoming solar radiation to determine that highly reflective sulfate aerosols in the atmosphere—which have both natural (volcanic activity) and man-made (burning of high-sulfur fuels) origins—can mask the full warming effect of carbon dioxide. Using mathematical techniques from econometrics, Storelvmo and her colleagues, Thomas Leirvik  of the University of Norland (Norway), Ulrike Lohmann and Martin Wild of ETH Zürich, and Peter Phillips of Yale, attempted to disentangle the competing influences of carbon dioxide and sulfates in the record of recent temperature changes. Their analysis determined that existing levels of sulfate aerosols have counteracted about 0.5 degrees Centigrade worth of warming that would otherwise have occurred in the last few decades. Since sulfate aerosols can have a severe impact on health and the environment, causing smog and acid rain, along with asthma and other respiratory diseases, efforts are underway in many countries to reduce their presence in the atmosphere, for example, by switching to low sulfur fuels. These efforts may have the unintended consequence of accelerating global warming.

Higher values of climate sensitivity would help resolve some apparent disagreements between climate models and evidence from the geologic record says Yale professor Mark Pagani: “Evaluations of ancient Earth’s sensitivity to CO2 are difficult to make, but the equilibrium climate sensitivities obtained are often higher than model estimates for the modern climate system.”

Are Economic Models Misinterpreting Climate Sensitivity Too?

While scientists use global climate models to identify how greenhouse gases will affect average global temperature, economists use integrated assessment models (called IAMs) to estimate how those rising temperatures will impact society. And just as Storelvmo and her colleagues have argued that current climate models may be underestimating warming, Bob Litterman, former director of risk management for Goldman Sachs, who now chairs the World Wildlife Fund’s Investment Committee, suspects that today’s assessment models may be underestimating the economic damages.  

IAMs are where the rubber meets the road in efforts to assess the impacts of climate change. Their output attempts to estimate the cumulative damages to society in a number called the “social cost of carbon”—a quantity intended to represent the average dollar cost to society of each additional ton of CO2 that is released to the atmosphere. IAMs estimate the cost using models that correlate economic damages with the increased frequency of storms, heat waves, droughts, lost crops, diseases, etc., that are expected in a warmer world. The U.S. government is bound by law to use the social cost of carbon when promulgating rule-makings. Thus, getting the best value for Earth’s actual climate sensitivity—which predicts future average global temperature, which in turn helps to predict future damages—is much more than an academic exercise. For example, getting the right values for climate sensitivity and the social cost of carbon is necessary to set the right value for a fee or tax on carbon that many experts see as the ultimate solution for curbing carbon emissions.

As someone who sees climate change as a risk management problem, Litterman feels that the IAMs don’t incorporate risk aversion adequately. Nor does he feel that they take the climate science component seriously. In an effort to reconcile both failings, he recently helped to organize a conference at Arizona State University that convened ten climate scientists and ten economists. Storelvmo participated in a similar forum with climate scientists and economists organized by the Yale Climate & Energy Institute two years ago at Yale, and had similarly concluded that IAMs were deficient, but for different reasons.

“Integrated assessment models,” she says, “calculate damages based solely on anticipated increases in global annual mean surface temperature” rather than using a larger suite of parameters, including sea-level rise and precipitation changes, all of which can profoundly impact economic output. Equally problematic is that economists’ models typically utilize the equilibrium climate sensitivity, which estimates long-term warming, rather than a related quantity called the “transient” climate sensitivity, which represents the Earth’s more or less immediate response to increased greenhouse gases and is therefore more appropriate for calculating short-term economic effects.

What the climate science world needs now, both experts agree, is a new approach: collaboration between scientists and economists to take advantage of available computational resources to run integrated assessment models alongside full-scale global climate models and thereby overcome the shortcomings of today’s piecemeal approach.


Subscribe to