Primer on Climate Change
Robert E. Dickinson
University of Arizona, Institute of Atmospheric Physics
The climate system is expected to change substantially over the next century as the warming effect of increasing greenhouse gas concentrations is realized. Yet, the general circulation models (GCMs) used to project the future climate generate a description of the future that is rich in physical details but applicable only over areas with spatial dimensions of several hundred kilometers or more. Furthermore, these regional-scale descriptions are poorly understood, vary from model to model, and depend on aspects of the models that have received little or at least inadequate attention. Past studies have shown global warming could have major effects on western water resources. More certain answers require further improvements in the models.
THE BASIS FOR CONCERN ABOUT CLIMATE CHANGE
This paper discusses what the scientific community knows and does not know about how the climate and western water resources will change in the future because of increasing greenhouse gas production from human activities. This is a problem with a decades-to-centuries lead time, similar to the long lead times required for planning for water in the West. Although changes in human behavior could make a very substantial difference in the potential for climate change, they cannot eliminate the problem. Future climate change is inevitable. Because the magnitude and details of this change are still so uncertain, even if we respond to reduce it, we must also still begin to prepare ourselves for the
possibility of a different future climate, especially with regard to water resources. At least, we must get ready for a future more uncertain than that suggested by past records.
Most of our present scientific perspective on climate change has been around for more than a decade, and the first scientific discussions of the possible effect on climate of increasing carbon dioxide concentrations date back more than a century. Half a dozen or more assessments of the current scientific consensus have been conducted since 1979, authored by prestigious National Research Council committees and international workshops. It is really striking how little divergence there has been in the conclusions of these studies.
The latest assessment by the scientific community of climate change, just now being released, was carried out by hundreds of scientists and was orchestrated by an organization established just for this task: the Working Group Number One of the Intergovernmental Panel for Climate Change (IPCC) (IPCC, 1990). The IPCC report addresses four topics related to climate change:
factors that may affect climate change during the next century, especially those related to human activity;
responses of the atmosphere-ocean-land-ice system to climate change;
current capabilities for modeling global and regional climate changes; and
the past climate record and presently observed climate anomalies.
The IPCC report looks especially at the following questions:
What factors determine global climate?
What are the greenhouse gases, and how and why are their concentrations increasing?
Which greenhouse gases are the most important in climate change?
How much do we expect the climate to change?
How much confidence do we have in our predictions?
Will the climate of the future be very different from today's climate?
Have human activities already begun to change the global climate?
How much will the sea level rise if the climate changes?
How will climate change affect ecosystems?
What should be done to reduce uncertainties about climate change, and how long will reducing uncertainties take?
This paper examines some of the above questions, focusing on those most relevant to water supplies in the West.
THE CLIMATE SYSTEM AND THE GREENHOUSE EFFECT
The climate system consists of the atmosphere, oceans, land, and ice surfaces and biological processes that interact with these physical elements (Figure 4.1). Many factors determine climate, but foremost among these is solar radiation (Figure 4.2). The amount of solar energy reaching the earth at the top of the atmosphere is almost constant, so variations at the surface are largely controlled by season, latitude, and cloudiness.
Day-to-day changes of weather, as well as climate variations and change, especially the temperatures we feel locally, are all strongly dependent on the disposition of this solar heating. How much is reflected back to space? How is it balanced by terrestrial thermal emissions (that is, by the heat energy that leaks to space to cool our planet to balance the absorbed solar radiation)? How is the internal energy of the atmosphere and oceans sloshed back and forth as these fluids act as heat engines to drive winds and currents?
A less obvious factor, but one very important everywhere, especially for temperatures, is the ''greenhouse effect.'' Thermal radiation cools all terrestrial surfaces: oceans, soils, leaves, rooftops, and so on. Without an atmosphere, this radiation would all escape directly to space. However, over most of the wavelengths in which this energy travels, atmospheric gases and clouds absorb and reradiate it. What is crucial to the greenhouse effect is the fact that atmospheric temperatures are lower at greater heights; that is the reason that mountains are generally colder than lowlands. The colder gases radiate less energy upward than they absorb, and they also radiate downward. The downward radiation directly reduces nighttime radiative cooling at the ground, more so when the sky is moist and especially when it is cloudy.
It is the weakening of the upward radiation to space by high, cold layers that warms the overall climate system. By reducing net atmospheric cooling, the "greenhouse" gases in these cold layers warm the atmosphere, which, in turn, warms the earth's surface. Various greenhouse gases radiate to space mostly over about the
lowest 15 kilometers of atmosphere, with an average radiating height of about 6 kilometers. They protect us from having a climate that would otherwise be about 33°C colder—a climate much more like that of the Himalayas.
The two most important natural greenhouse gases are water vapor and carbon dioxide, although several other natural constituents and an increasing number of entirely man-made greenhouse gases make significant and generally growing contributions to the greenhouse effect (Figure 4.3). Climatologists have several methods for evaluating the effects of greenhouse gases. Our most accurate information about the greenhouse effect is provided by atmospheric radiation models, derived from fundamental physical theory and measured strengths of the various radiating gases. We also can look at radiation leaving the earth and compare it with the theory. We see the exiting thermal radiation reduced from its surface flux according to characteristic wavelength signatures of the absorbing greenhouse gases and calculable according to greenhouse theory. Another way to appreciate the possible impacts of extreme greenhouse effects is to look at the very hot surface of Venus and the very cold temperature of Mars, also calculable from greenhouse theory.
The two best established factors of the theory of global climate warming are: (1) how much heating results from given concentrations of greenhouse gases (Figure 4.4), and (2) how much concentrations of these gases have increased over the last 100 years, and especially the last 20 to 30 years. We worry especially about methane, nitrous oxide, ozone in the upper troposphere, and the chlorofluorocarbons. The relative number of carbon dioxide molecules has increased from 280 to 355 parts per million (27 percent), methane from 0.7 to 1.7 parts per million, and chlorofluorocarbons from nothing to total concentrations of all species of about a part per billion. The small concentrations of the latter have a surprisingly large effect on greenhouse warming (as well as on stratospheric ozone), contributing about 11 percent of the total increase over the last 100 years (IPCC, 1990). Still, carbon dioxide has contributed somewhat more than half of the total and will continue to dominate increases in greenhouse warming.
Our good records of the international consumption of fossil fuels (coal, oil, and natural gas) indicate that about twice as much carbon each year is put into the atmosphere as that which remains in the atmosphere. What happens to the carbon that does not remain in the atmosphere is still relatively poorly understood. It presumably goes somewhere into the large reservoirs for carbon
provided by the oceans, land vegetation, and soils (these compartments each hold more carbon than the total in the atmosphere). Difficulty accounting in detail for this lost carbon results from the potentially large additions from burning of tropical forests, the relatively small amounts that can be established as going into the oceans (Tans et al., 1990), and the lack of evidence for carbon going into land reservoirs. Will the fraction of fossil fuel carbon going into the atmosphere increase in the future and so accelerate the atmospheric buildup? We can't say without a better understanding of the carbon cycle.
HOW MUCH WILL OUR HOUSE WARM UP?
Given the well-established increasing concentrations of greenhouse gases and reasonably accurate estimates of the resulting increased heating, what does this tell us about future climate change? Suppose you were to go home and turn on your heating system (but remove the thermostat switch). It is fairly certain that your house would warm up to some new temperature. But what temperature? It would be useful in estimating this new temperature to know how much energy your heating system puts out. This answer would also depend on the temperature outside, how well insulated your walls and roof may be, what ventilation there may be from cracks and open windows (depending on the wind outside), how much additional heat the house may be receiving from the sun, whether or not it is raining, whether the temperature will become warm enough to set off an automatic sprinkler system, and so on. How long it would take to warm up would depend on the thermal mass of your walls.
To estimate what might be the new, warmer temperature of our heated house, we would develop equations for all its energy losses, which depend on the difference between the temperature outdoors and that in the house (Figure 4.5). This model of the house energy exchanges could then be used to infer a new temperature, one for which the energy losses from the house would just balance the heat put out by our furnace. This is what we do or would like to do with climate models. By analogy, we know that we are putting excess heat into our global home (or, equivalently, we are adding to the insulation), and we even know reasonably well how much. It is a fairly obvious conclusion that the planet will warm up, but how much? This depends on how easy it is for this extra heat to leak back out again and on what other climate factors the warmer temperatures will change, either to remove some of the excess heat being supplied or to add yet more. How much will the thermal inertia of the system, especially the oceans, delay the warming?
As with our heated house analogy, the easiest task is to get an overall idea of how much temperature will increase. Determining how this temperature increase is distributed from region to region is more difficult. Again turning to the house calculation, for us to know how temperature might vary from room to room, we would have to understand much about the convection currents carrying the heat between the rooms. Likewise, using a climate model to calculate the change of temperature and water resources in the
western United States is much more difficult than just determining how much the world as a whole might warm up. This is the challenge the climate modeling community will be attacking over the next decade. The best we can do now is describe a plausible range of possible futures and say what tools must be developed, what research done, to narrow this range of possibilities.
CONSTRAINING POSSIBLE FUTURES THROUGH MODELS
A standard measure of the excess global heat from human activities is the heating that would result from doubling the atmospheric carbon dioxide concentration. This change, expected some time near the end of the next century, adds about the same or a little less heat than would come from a 2 percent increase in solar energy output. This amount is several hundred times the energy added directly from the combustion globally of fossil fuels. If we were to add 2 percent to the sunshine received on a single day, we would see only a slightly warmer temperature, perhaps a few tenths of a degree at most. However, applied over many years, temperatures would rise between 1.5 and 5°C over past normal conditions. We develop our sense of the possible and probable by running the most detailed three-dimensional models of our climate system now available and by exploring how the models' outputs depend on poorly described parts of the system. We are currently especially stuck on what clouds might do to change the reflection of solar energy or to change greenhouse warming as climate changes.
Climate models treat a large number of physical processes and the linkages between these processes. It is only through these models that we are able to combine the knowledge of experts about individual processes into an overall synthesis of how the system behaves and how it might change in the future. The more advanced general circulation climate models give a self-consistent simulation of the atmospheric and surface hydrological cycles (Figure 4.6). These models generate day-to-day weather in half hour time steps. Atmospheric moisture is carried around by three-dimensional winds generated by numerical solution of the equations for atmospheric hydrodynamics on a sphere. Where atmospheric moisture is determined to be in excess of saturation, this excess is removed from the model atmosphere as rain or snow, depending on the temperature of the lowest model layers.
GCMs use a water budget approach to balance soil moisture change with the difference between precipitation and runoff.
Thus, the GCMs compute both a runoff and an evapotranspiration. The most developed components of the GCMs are their treatments of large-scale atmospheric hydrodynamic and temperature patterns (treated the same as in weather prediction models) and their modeling of atmospheric radiation. Less well developed are the treatments of atmospheric humidity, precipitation, and clouds (radiative properties and convection), the dynamics of the oceans and sea ice (especially how the atmosphere couples to these systems), variations in atmospheric chemical composition, and the climate aspects of land vegetation and soils, including surface hydrology.
From the viewpoint of western hydrology, we wish to know quantitatively how patterns and amounts of precipitation and evapotranspiration will change. Precipitation will depend on transport of moist air from the Pacific or the Gulf of Mexico, or dry air from elsewhere, and the presence of large-scale upper-air movement and convective destabilization. Evapotranspiration will depend on the amount of solar radiation received by the vegetation, the temperature and relative humidity of overlying air, surface roughness, the net effect of vegetation stomatal resistances, and details of surface runoff processes. All these considerations are already included in current climate models. Large-scale geographical patterns of precipitation as seen on a smooth global map compare quite well with reality. The concern is that some of the treatments are not yet realistic enough and that so far we have not adequately checked their operation against observations.
PRESENT GUIDANCE FROM CLIMATE MODELS AND OTHER ARGUMENTS CONCERNING WESTERN WATER
GCMs generate an extraordinarily detailed description of climate processes. In principle, they convert scenarios for increasing trace-gas concentrations into changes of all the more important surface hydrological parameters: precipitation, surface radiation, surface temperature, evapotranspiration, soil moisture, and runoff. Yet, because the real system is even more detailed and complex than the models, there are still severe limitations on the confidence we can place in our abilities to model future change. Since there is no other plausible method to ascertain future precipitation and surface air temperatures, we must continue to improve and better understand the GCMs.
Two approaches have been used to determine possible future changes of water resources in the West: (1) the use of climate models to project precipitation and temperature, followed by the inclusion of these projections as input into detailed regional hydrological models; and (2) the direct use of climate model calculations of surface hydrological parameters.
Qualitatively, the projections of global average increasing temperature and precipitation are well founded in basic physical principles and are given by all GCMs. However, a wide range of possibilities for regional anomalies has been suggested by simulations for the changes over regions as small as the western United States (Kellogg and Zhao, 1988). Furthermore, surface hydrologi-
cal processes are much more complex than what is presumed in the GCMs, which, for example, largely or entirely ignore the effects of topography on runoff and infer dynamic effects on precipitation using only a very smooth topography.
The question of how runoff will change given changes in precipitation and temperature provided by GCMs has now been explored by a number of authors. The simplest such approach is to use the mean measured runoff in different basins (Stockton and Boggess, 1979) or the year-to-year variations in a single basin (Revelle and Waggoner, 1983) to infer statistically how runoff might change with changes in precipitation and temperature. This statistical correlation approach does not allow for an examination of changes in the seasonality of runoff or for a distinction between water supplies from snowpack or rainfall. More recent studies of the sensitivity of runoff to precipitation and temperature (Gleick, 1987; Schaake, 1990) have used simple water-balance models. These models resemble the treatments of surface hydrology in GCMs in following a soil moisture budget but differ in their use of monthly (rather than hourly or smaller) time steps and in their use of an empirical relationship between evaporation, soil moisture, and temperature based on the concept of potential evapotranspiration. They employ a physically based formulation for runoff that is comparable in complexity and limitations to that used in the GCMs. However, they have been adjusted to observed runoff records. Gleick (1987) modeled the Sacramento basin, apparently adjusting only a few physical parameters and validating the model against independent data. Schaake (1990) fit data from 52 basins in the southeastern United States by adjusting five parameters. Both these models include snow accumulation and melt.
The water balance studies of Gleick (1987) and Schaake (1990) show that warming of the projected magnitudes would have a large impact on the seasonality of runoff, with heavier winter flows and reduced summer flows (Figure 4.7). The smaller annual changes they find are comparable to those found in the earlier correlative studies.
The examination of surface hydrology modeled explicitly by GCMs to explore the impacts of climate change on water resources has not yet been carried very far. Emphasis has been primarily on the changes in soil moisture (Manabe and Wetherald, 1987; Meehl and Washington, 1988) and on possible implications for mid-continental drought. Kellogg and Zhao (1988) have discussed the wide divergence between past model simulations (Figure 4.8).
Climate models are our only tools to project into the future. However, current climate modeling results are significant only in a qualitative sense for surface hydrology as illustrations of physically consistent possible change. They show a wide divergence in their future projections of the atmospheric quantities most important for determining surface hydrology—in particular precipitation and net surface radiation. Furthermore, little attention has yet been given to developing aspects of the models important for surface hydrology. Improvements in these features and means to quantify the relative confidence we can place in these different projections must be a high priority for future work.
Gleick, P. H. 1987. Regional hydrologic consequences of increases in atmospheric CO2 and other trace gases. Climatic Change 10:137-161.
Henderson-Sellers, A., and K. McGuffie. 1987. A Climate Modelling Primer. Chichester, England: John Wiley & Sons.
Intergovernmental Panel for Climate Change (IPCC). 1990. Climate Change: The IPCC Assessment. Cambridge, England: Cambridge University Press.
Kellogg, W., and Zong-Ci Zhao. 1988. Sensitivity of soil moisture to doubling of carbon dioxide in climate model experiments, part I: North America. J. Climate 1:348-366.
Manabe, S., and R. T. Wetherald. 1987. Large scale changes in soil wetness induced by an increase in carbon dioxide. J. Atmos. Sci. 44:1211-1235.
Meehl, G. A., and W. M. Washington. 1988. A comparison of soil-moisture sensitivity in two global climate models. J. Atmos. Sci. 45:1476-1492.
Revelle, R. R., and P. E. Waggoner. 1983. Effects of a carbon dioxide-induced climatic change on water supplies in the Western United States. In Changing Climate. Washington, D.C.: National Academy Press.
Schaake, J. C. 1990. From climate to flow. In Climate Change and U.S. Water Resources. New York: John Wiley & Sons.
Schneider, S. H. 1987. Climate modeling. Scientific American 256:72-80.
Stockton, C. W., and W. R. Boggess. 1979. Geohydrological Implications of Climate Change on Water Resources Development. Ft. Belvoir, Virginia: U.S. Army Coastal Engineering Research Center.
Tans, P. P., I. Y. Fung, and T. Takahashi. 1990. Observational constraints on the global atmospheric carbon dioxide budget. Science 247:1431-1438.