Will Runaway Water Warm the World?

It was hot. Hotter than any record in the books. Instead of photographing picturesque fountains in the towns of southern France, tourists were soaking in them. In London, trains sat quietly in the stations; officials were too afraid that the metal tracks would buckle to allow a speeding engine to race over them. Sparked by hot, dry conditions, wildfires raged across France, Spain, Portugal, and Italy. Swiss mountain glaciers thinned more than any other year in the past decade. The doomsday-like heat wave that engulfed Europe in July and August 2003 also carried a darker toll. In France alone, 14,802 more people died that August than in the same month the previous year; for all of Europe, the unofficial death toll reported in the media soared to 19,000. Were these unusually high temperatures—up to ten degrees Celsius hotter than 2001—a result of global warming? It’s not clear, but some fear that the summer’s heat may be an ominous harbinger of some future climate.

  Page 2
  Rhone Glacier

Across the globe, temperatures are slowly creeping up. According to the U.S. National Climatic Data Center, the global average surface temperature has gone up 0.4 degrees Celsius (plus or minus 0.1 degree) in the past 25 years alone. While the extra heat may not have you sweating yet, larger increases are predicted, and that has some people tugging at their collars. The Intergovernmental Panel on Climate Change, a policy advisory group made up of members of the World Meteorological Organization and the United Nations Environment Programme, estimates that the average global surface temperature could climb anywhere from 1.4 to 5.8 degrees Celsius by the year 2100.


Once a thick tongue of ice that poured into the Gletsch valley (inset), the Rhone Glacier has shrunk dramatically since 1850. In 2003, the Rhone Glacier and other Swiss glaciers retreated more than any other year since scientists began taking measurements in the 1800s. While the summer’s extreme temperatures caused the glaciers to thin more than usual, scientists say that the glaciers retreated in response to long-term warming. (Photograph copyright bigfoto.com, inset courtesy Library of Congress)

Graph of Predicted Temperature Rise

Part of the reason the predicted temperature range is so great is that scientists don’t entirely understand whether the atmosphere will become more humid as it warms, and humidity is one of the primary factors that will influence how much the climate will warm over the next century. If the humidity of the atmosphere does indeed increase, it can as much as double the warming from carbon dioxide alone. Thus, an understanding of how the humidity of the atmosphere will change is of fundamental importance in predicting future climate. The problem is one that Ken Minschwaner and Andrew Dessler, researchers at NASA Goddard Space Flight Center, have worked to remedy using data from the Upper Atmosphere Research Satellite (UARS).


Predictions of warming in the next 100 years vary by about 5 degrees Celsius, from a low of 1.4 degrees to a high of 5.6 degrees. The wide variation is due in part to uncertainty in the magnitude of the feedback between warming and increased rates of evaporation. In this graph, dark green areas represent predictions based on the averaged results of multiple climate models, while light green areas represent the predictions of single climate models. (Graph adapted from Climate Change 2001: The Scientific Basis)

  Artist's Rendering of UARS

Minschwaner, also a Professor of Physics at the New Mexico Institute of Mining and Technology, and Dessler, also a researcher with the University of Maryland’s Earth System Science Interdisciplinary Center, formulated a simple, one-dimensional model to describe how the humidity of the atmosphere will change as the Earth heats up in response to carbon dioxide emissions from burning of fossil fuels. Surprisingly, their model predicted smaller increases in humidity in the upper atmosphere than large global climate models do, and data collected by the Microwave Limb Sounder and the Halogen Occultation Experiment on NASA’s UARS satellite support their model. Their findings imply that the Earth will warm significantly, but probably not as much as most global climate models predict. Their results appeared in the Journal of Climate on March 15, 2004.


Using instruments aboard NASA’s Upper Atmosphere Research Satellite (UARS), scientists measured humidity high in the atmosphere. The researchers then compared those humidity measurements with sea surface temperature records. Using these observations, the researchers quantified the feedback between rising temperatures and increasing concentrations of water vapor in the atmosphere. This crucial variable in climate change estimates had previously been based on speculation and modelling, but not direct observations. (Rendering by Jesse Allen, NASA GSFC)


Earth’s Steamy Blanket


But what does water have to do with global warming? When asked that question, Dessler turned to a white board in his office and began drawing pictures of clouds in a swath of space he had wiped clean of equations. Soon the figure was accompanied by arrows and a graph plotting water and temperature. Waving his dry-erase marker in the air, he explained why a little bit of extra water in the atmosphere is such a big deal.

At the root of global warming are greenhouse gases. The atmosphere acts as a global thermostat, letting sunlight in, but trapping outgoing heat. In this way, it keeps the Earth’s surface temperature in a range suitable for life. In theory, we could turn up the global thermostat by increasing the proportion of greenhouse gases in the atmosphere. “As humans add carbon dioxide, and carbon dioxide absorbs outgoing radiation,” Dessler says, “the Earth warms up.” As it does, more water evaporates from the oceans and into the atmosphere. “Since water vapor is also a greenhouse gas, the additional water in the atmosphere further heats the surface, leading to even more water evaporating,” he explains. And even though carbon dioxide is the greenhouse gas that gets all of the attention, it can’t compete with water vapor in heat-trapping power.

  Page 1Page 3
Graph of Atmospheric Absorption

“Water vapor is the most important greenhouse gas in the atmosphere,” says Steven Sherwood, a professor in the Department of Geology and Geophysics at Yale University. As anyone who lives in a humid climate can attest, water traps heat being radiated from the Earth. In effect, water vapor envelops the Earth in a thick, steamy blanket. Warming due to carbon dioxide emissions from fossil fuel combustion evaporates even more water, increasing the thickness of the blanket, which leads to more heating, which leads to more water vapor… The loop is called the water vapor feedback, and it has the potential to be a serious problem. Sherwood explains. “If you have enough of this positive feedback, then of course the whole climate system would be unstable.” Today’s climate, he quickly adds, is not unstable. “But as you pile on more and more of this sort of thing, you get closer and closer to an unstable situation. So if the climate is unstable, small differences in how strong these feedbacks are can become relatively important, more important than you might think.”


Atmospheric gases, especially carbon dioxide and water vapor, prevent some heat (thermal infrared radiation) from escaping from the Earth into space. In this graph, the dotted line represents the energy that would be emitted from the Earth without an atmosphere, and the solid line shows the effects of water vapor, ozone, and carbon dioxide. (Graph adapted from Walter Roedel, Physik unserer Umwelt Die Atmosphäre, 2000)

  Photograph of Ocean and Haze

Clearly, detailing how the water vapor feedback works is essential in predicting and mediating future climate change. “I was drawn to this problem by its importance,” says Dessler. “As a scientist, you want to work on the most important problem that you can find, and I think everyone would agree that water vapor feedback falls into this category.” To understand how much the water vapor feedback could boost the Earth’s temperature in the future, Dessler and Minschwaner decided to focus on finding out how much water will enter the atmosphere as the temperature climbs.


Although carbon dioxide in the atmosphere traps some heat near the Earth’s surface, its effect is much less than that of water vapor. The small amount of warming caused by carbon dioxide may be greatly magnified by increased evaporation from the ocean surface as global temperatures rise. (Photograph copyright Corel)


It’s not the Heat, it’s the Humidity


The abundance of water vapor in the atmosphere is usually expressed as “relative humidity”: the percent of water in the air relative to the amount of water the air can hold. Just as an 8-ounce cup holding 4 ounces of water is 50 percent full, air that contains half the water it can hold is said to be at 50 percent relative humidity. But if you pour the 4 ounces of water into a 16-ounce cup, the cup is only 25 percent full, even though you still have the same amount of water. The same principle applies to the percentage of water in the atmosphere. As temperatures increase, the air becomes capable of holding more water, and the percent of water in the air drops unless more water is added.

  page 2Page 4
Graph of Specific Humidity vs. Temperature

In climate modeling, scientists have assumed that the relative humidity of the atmosphere will stay the same regardless of how the climate changes. In other words, they assume that even though air will be able to hold more moisture as the temperature goes up, proportionally more water vapor will be evaporated from the ocean surface and carried through the atmosphere so that the percentage of water in the air remains constant. Climate models that assume that future relative humidity will remain constant predict greater increases in the Earth’s temperature in response to increased carbon dioxide than models that allow relative humidity to change. The constant-relative-humidity assumption places extra water in the equation, which increases the heating.

Many have questioned whether this prediction of a wetter future atmosphere is right, including Dessler and Minschwaner. “There’s no theoretical, simple line of reasoning that should say that it [relative humidity] should be constant,” says Ian Folkins, an associate professor of atmospheric sciences at Dalhousie University in Halifax, Nova Scotia, Canada. Critics of the constant-relative-humidity assumption have said that compensating effects will prevent large quantities of extra water from entering the atmosphere, explains Dessler. “The atmosphere is very efficient at generating dry air. Increases in these processes could balance increased evaporation in a warmer climate, leading to little change in the humidity in the atmosphere.” Like air running over the cooling coils in an air conditioner, he adds, air that rises to high altitudes cools off and water condenses out, leaving the air drier.

Water Woes: Predicting the Humidity of the Future

To start to pin down the relationship between humidity and temperature, Minschwaner and Dessler modeled how water in the atmosphere around 11 to 14 kilometers from the surface of the Earth reacts to changes in temperature. They chose to focus their study on the upper troposphere over the tropics because it is a physically simple system compared to other sections of the atmosphere. For example, “things like evaporation of rain don’t have much of an effect,” Dessler says. While there is very little water in this section of the upper atmosphere, the climate is quite sensitive to the amount of water that is there because, closer to the cold of space, water cools off and becomes far more reluctant to let go of any heat it absorbs. The higher the altitude, the more efficiently water vapor traps heat.

Minschwaner and Dessler’s model describes how the humidity of the upper troposphere changes as the surface warms. As the Earth warms, more water is expected to evaporate from the surface. At the same time, thunder storms are expected to become more severe and extend to higher altitudes in the atmosphere. Since temperature decreases with altitude, warm, humid air rising to higher altitudes in such storms will encounter colder temperatures, and therefore more water is ’freeze dried’ out.” These two factors oppose each other, and the overall change in water vapor in the upper troposphere is a combination of these opposing forces. In order to predict changes in humidity, you have to predict both increased evaporation from warmer temperatures and increased freeze-drying from convection to higher altitudes. Minschwaner and Dessler’s model shows that these two factors are closely coupled, and in fact, the two can not vary independently.

As air warms it becomes capable of holding more and more water vapor. This graph shows the maximum amount of water in air at temperatures ranging from -40 to 40 degrees Celsius. (Graph courtesy Quantitative Environmental Learning Project)

Graph of Model Results

Within these constraints, the model does predict that there will be a net increase in the water content of the upper troposphere as the Earth’s surface temperature rises, but not so much that the relative humidity remains constant. That means that water vapor will cause the Earth to warm, because the feedback is positive, but it won’t warm as much as it would if constant relative humidity were maintained—a result that contradicts the assumptions put into big global climate models. “I don’t think too many people would have expected a simple model like this to give a result other than the one that people have been assuming will happen,” Sherwood notes.


Minschwaner & Dessler’s model results showed an increase in water vapor in the upper troposphere (grey) as temperatures rose (left), but not rapidly enough to maintain constant relative humidity (right). Therefore, their model predicts increasing temperatures to increase humidity, but not to the degree assumed by many climate models. The blue, green, and yellow lines represent progressively increasing temperatures. (Graph adapted from Minschwaner & Dessler)


Support from the Skies


Unexpected though the results may be, they are supported by satellite data. By choosing to model the upper troposphere, Dessler and Minshwaner were able to test their model against the data collected by two instruments on the Upper Atmosphere Research Satellite. The Microwave Limb Sounder (MLS) and the Halogen Occultation Experiment (HALOE) have been taking regular measurements of water vapor in the upper troposphere since late 1991, giving the scientific community its first look at what was actually going on over time in the upper troposphere. HALOE observes the way that sunlight passes through the atmosphere at sunrise and sunset. As the Sun rises, its light slices horizontally across the atmosphere. Stationed opposite the Sun, HALOE measures how the light changes as it passes through the atmosphere. This gives a vertical profile of the make-up of the atmosphere, including water vapor concentrations. The Microwave Limb Sounder measures naturally occurring microwave thermal emissions from the limb of Earth’s atmosphere to make a similar vertical profile of the atmosphere. Minschwaner and Dessler correlated these profiles of the upper troposphere with surface temperatures to determine the effect of temperature on relative humidity. “We are the first to use direct observations of water vapor in this region,” says Dessler.

  page 3
  Astronaut Photograph of Sunrise through the Earth's Limb

They found that the water vapor content of the upper troposphere measured during the 1990s climbed as the Earth’s surface temperature rose, but not enough to maintain constant relative humidity, just as their model predicted. These observations give Dessler and Minshwaner’s results far more weight than they might otherwise have had. Richard Lindzen, the Alfred P. Sloan Professor of Meteorology at the Massachusetts Institute of Technology (Cambridge, Massachusetts) and a pioneer in the study of the water vapor feedback, acknowledges, “Regardless of the model [which he calls ‘not very conclusive’], the observations do make a case that the water vapor feedback above 200 millibars [12 kilometers] is likely to be somewhat positive.”


Instruments aboard UARS measure the concentration of gases in the atmosphere by looking through the Earth’s “limb”. This is the portion of the atmosphere that rises above the curvature of the Earth’s surface as seen from space. This vantage point allows instruments to measure the atmosphere at different altitudes. (Astronaut photograph ISS007-E-17719 provided by the NASA-JSC Gateway to Astronaut Photography of Earth)

Graph of Measured Variation of Humidity with Change in Sea Surface Temperature

Water in a Changing Climate

But what do these results mean for the larger picture of climate change? “The climate implications are very limited,” Lindzen says. One of his main criticisms is that the upper troposphere doesn’t have much influence over the water vapor feedback of the entire atmosphere. The region between three and ten kilometers—where weather occurs—has a far greater impact on the Earth’s climate. “At the levels Dessler [and Minschwaner] are concerned with, there simply is not much water vapor,” says Lindzen. In response, Dessler argues that their conclusions almost certainly apply at 10 kilometers, where water does have a significant climate impact (although they only can only confirm the model’s behavior at 12 kilometers altitude).

Still, Sherwood agrees with Lindzen. “The levels where they found something unexpected are making a relatively small contribution, so we might be talking about something like a ten percent weaker feedback effect than we thought.” But, when other influences on the climate are factored in, “that could still make a difference,” he adds.

Folkins believes that Minschwaner and Dessler’s results might be used to refine the scientific understanding of water vapor feedback and the models that predict climate change. “It is important to understand water vapor both using simple process approaches and data, plus global climate models.” Unlike big climate models, Minschwaner and Dessler’s model can be tested and confirmed with satellite data, Folkins points out, and that makes it valuable to the scientific community. “I think it’s a pretty provocative and good paper because it should get people thinking more seriously about their assumptions on how water vapor will change in the upper troposphere.”

“There is a certain sense of complacency that water vapor feedback is understood. And that comes from the fact that a lot of these global climate models agree with each other,” Folkins observes. “But just because they agree, doesn’t mean they are all right.”

  • References:
  • Intergovernmental Panel on Climate Change, Climate Change 2001: A Synthesis Report
  • Minschwaner, K. and Dessler, A. “Water Vapor Feedback in the Tropical Upper Troposphere: Model Results and Observations” Journal of Climate, March 15, 2004
  • National Oceanic and Atmospheric Administration, “Global Warming,”

Measurements of water vapor taken by HALOE (green crosses) and compared with sea surface temperatures show a similar relationship to that predicted by Minschwaner & Dessler’s computer model (dotted line). Specific humidity rises with rising sea surface temperatures, but not fast enough to maintain constant relative humidity (angled dashed line). The pale green area is a best fit of the data, including uncertainties. (Graph adapted from Minschwaner & Dessler)