Click for next page ( 16


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 15
1 INTRODUCTION The consensus view of scientists is that human activities are changing Earth's climate and that this could have very important consequences for human life and natural ecosystems (IPCC, 2001b; NRC, 2002~. Projections of how climate might change as a result of human activities remain uncertain, however. The International Panel on Climate Change (IPPC) Third Assessment Report (TAR) projects that annually and globally averaged surface temperature will increase by 1.4C to 5.8C during the interval between 1990 and 2100 (Cubasch et al., 2001~. The large range of possible warming results in approximately equal measure from two sources. First, the rate at which humans will release greenhouse gases and make other changes in the natural environment of Earth in the fixture is difficult to predict. The future rates of human modification of the environment depend on social, economic, and political processes as well as technological innovation and diffusion, and are unknown. Policy makers may make different choices if scientists provide credible information about the magnitude and structure of the climate response to greenhouse gas releases. The second source of uncertainty is how the climate system of Earth will respond to human forcing. Interactions among physical, chemical, and biological processes that determine the response of the climate system to human activities are not fully understood. If the carbon dioxide concentration in the atmosphere were doubled and the climate were allowed sufficient time to come into a new equilibrium, the projected uncertainty in the warming of the global mean surface temperature would still be large ~ In this document we have generally tried to follow the IPCC practice of using the word "projection" when referring to estimates of future climates that are hypothetical in the sense that they depend on an assumption of a particular scenario for emissions (and hence radiative forcing). We use the word "prediction" when the answer is not contingent on a climate-forcing scenario or the climate-forcing scenario is considered fixed, such as in the problem of calculating the equilibrium response to doubled CO2. 15

OCR for page 15
16 UNDERSTANDING CLIMATE CHANGE FEEDBACKS (1.5C to 4.5C according to the Intergovernmental Panel on Climate Change) (IPCC, 2001a). Our inability to reliably determine the influence of various feedback processes is one of the most important reasons why projections of possible future climate change show such wide variations. Scientific research can provide knowledge that will help refine and focus these projections so that they become more accurate over time. Climate scientists often separate influences on climate change into forcings and feedbacks. Climate forcings are changes that initiate outside of the naturally evolving climate system, and can be either natural or human- caused (See Table 1.1~. Processes in the climate system that can either amplify or damp the system's response to changed forcings are known as feedbacks. Feedbacks are interactions in the climate system between the variables defining the state of the atmosphere, ocean, and land surface. The range of possible outcomes in climate change projections that results from the internal dynamics of the climate system is the result of feedback processes and our inability to capture these adequately in models. A feedback process is a process whereby a change in one variable, such as carbon dioxide concentration, causes a change in temperature, which causes a change in a third variable, such as water vapor, which in turn causes a further change in temperature. Climate models suggest that the temperature change enhancement associated with feedback processes is greater than the temperature change resulting from the direct effect of the carbon dioxide doubling without feedbacks (IPCC, 2001a). Stott and Kettleborough (2002) find that the magnitude of global warming over the next 40 years is insensitive to the rate of greenhouse gas releases; in their study the range of possible warmings is determined by the range of estimates of the strength of climate feedbacks and not by the range of estimates of climate forcing. Therefore, study of climate feedbacks and climate sensitivity is very important for projecting climate changes over the next 40 years. Even in a simple linear analysis the temperature response is not linear in the strengths of the feedbacks, because all the other feedback processes modify the temperature change associated with one feedback process (Hansen et al., 1984~. In a system with a strong positive feedback, such as water vapor feedback in the climate system, the strong positive feedback process amplifies the changes associated with weaker feedback processes (See Box 1.1~. The integrated effect of climate feedback processes on climate sensitivity can be estimated by using the observed record of global mean temperature over the past 120 years (IPCC, 2001a). This method requires estimates of the climate forcing, climate sensitivity, and the uptake of heat

OCR for page 15
INTRODUCTION TABLE 1.1 Climate Forcing Variables Compared to Climate State Variablesa Climate Forcing Variables Climate System State Variables Solar irradiance Volcanic eruptions Greenhouse gas production by humans Aerosol production by humans Reactive gas production by humans 17 Land surface modification by humans Temperature of air, land, and water Precipitation and snow cover Humidity, clouds, and winds Ocean currents, salinity, and ice cover Soil moisture and vegetation properties Aerosol distribution Atmospheric trace gas concentration aThe variables in the led column are natural arid human-caused climate forcings that are defined to be outside the climate system for the purposes of this report. The processes that couple the climate system variables in the right column can result in climate feedback that will determine the response of climate to forcing. by the climate system, and each of these factors is uncertain. Consequently, the range of probable future climates is only loosely constrained by models fitted to the instrumental record of global mean temperature. Andronova and Schlesinger (2001) used a Monte Carlo simulation with a simple climate system model to estimate a probability distribution function for climate sensitivity. Climate sensitivity is here defined to be the equilibrium response of global mean surface temperature to doubling carbon dioxide. They concluded that there is a 54 percent likelihood that the actual climate sensitivity lies outside the range of 1.5-4.5C and that the 90 percent confidence interval for climate sensitivity is 1.0-9.3C. Knutti et al. (2002) found a 40 percent probability that the warming will exceed the IPCC estimates, but only a 5 percent probability that the warming will be less than the IPCC lower limit. Forest et al. (2002) found similarly that the 5 percent and 95 percent confidence limits on the climate sensitivity are 1.4K to 7.7K, compared to the 1.5-4.5K range stated by IPCC. Use of the instrumental record of global mean temperature cannot constrain climate sensitivity to a narrow range because the climate-forcing magnitude, amount of heat storage, and even the temperature record itself are not known with sufficient precision. An enhanced effort to understand and model the most important climate feedback processes is needed to improve our fundamental knowledge and will lead to better characterizations of the climate system, potentially reducing the wide ranges now seen in climate change projections. Improved understanding, combined with more rigorous comparison of observed and modeled feedback processes, should lead to more confidence in climate

OCR for page 15
18 UNDERSTANDING CLIME TE CHANGE FEEDBA CKS model projections. This approach should be pursued in parallel with the system level approach based on the global mean temperature record. Thus a key finding of this report is that an enhanced research effort is needed to better observe, understand, and model key climate feedback processes. Research on climate feedback processes should be designed to integrate observational and modeling efforts toward understanding and modeling of climate feedback processes; integrate the subdisciplines of climate science for a comprehensive study of the key climate feedback processes; and integrate different time scales of weather and climate variability into studies of climate feedback processes. Although observations are used to test the climatological statistics derived from climate simulations, more attention needs to be given to using data to test the simulation of feedback processes in these models and their role in determining climate sensitivity. To do this will require greater synergy between the efforts of observational scientists and modelers. In addition, because climate change feedbacks often incorporate processes from different disciplines, such as sea-ice processes and ocean circulation, or land surface processes and cloud processes, climate feedbacks research will also require greater synergy between traditional subdisciplines in climate science. Many climate feedback processes operate on time scales short enough to be tested effectively by comparing numerical weather forecasts with instantaneous data. For example, the ability of models to simulate the occurrence of frontal clouds in middle latitudes can be better understood by comparing instantaneous fields observed from satellites with instantaneous fields simulated in weather prediction models. Similar use can be made of seasonal forecasts, which bring slower feedback processes into play. Systematic biases in seasonal forecasts of climate often reflect problems with the treatment of climate feedback processes in the forecast models. For example, Li and Philander (1996) found that the improved simulation of marine boundary layer clouds was important in simulating the annual cycle in the tropical Pacific and its relation to the El Nino phenomenon. The interannual variations associated with ENSO events can also be used to better understand climate and carbon cycle coupling in the ocean and on land, since the growth rate of carbon dioxide in the atmosphere is highly correlated with interannual variations in tropical Pacific sea surface temperature (e.g., Jones et al., 2001~.

OCR for page 15
INTRODUCTION 19 BOX 1.1 Classical Treatment of Climate Sensitivity and Feedback Processes One can write a simple linear expression that relates the change in equilibrium temperature ATeqtO the magnitude of the applied forcing, AT Wm . ~Teq= ~ ~0 (1) The climate sensitivity parameter ~ measures the ratio of the temperature change to the applied climate forcing. Feedback processes alter the relationship between the magnitude of forcing and the magnitude of the climate response. The most fundamental feedback in the climate system is the temperature dependence of radiative emission. As objects get warmer they emit more radiant energy, as expressed by the Stefan-Boltzmann law of blackbody emission, Irradiance = AT . If a linear model is assumed, and only the temperature dependence of blackbody emission is considered, then the sensitivity parameter is TO = ~ 4~Te3 ~ . Assuming an emissivity of one and an emission temperature of 255K, this gives a basic sensitivity parameter of JO = 0.26K ~ Wm 2 ~ . From (1) then we could write Otto = 20 I (2) If a forcing of 4 Wm is applied to this system, then the expected equilibrium surface temperature change is about 1 K. The gain factor, g, is the fraction of the equilibrium climate change associated with feedback processes in addition to basic blackbody feedback. ATeqATo ~Tfeedbacks ATeq ATeq (3)

OCR for page 15
20 UNDERSTANDING CLIMA TE CHANGE FEEDBACKS It can be shown that Teq 1 _ ~ g i (4) where a number of different feedback processes with feedback factors, gi ~ are assumed to be linearly additive. If the gain is zero, the response is just ATo' and as the gain approaches one the response becomes very large. If the feedbacks are considered to act independently, then the gain factors for individual feedback processes are additive and their importance can be measured by their relative contributions to the total gain. g = "water vapor + becloud + "surface ice + "lapse rate + "other (5) The gain factor for water vapor feedback is about 0.5, which according to (4), will double the temperature response to climate forcing, changing the equilibrium response to doubled carbon dioxide from 1C to 2C. If an additional feedback only half as strong as water vapor feedback is added to the system, with a gain factor of +0.25, then the temperature response will be 4.0C if the weaker feedback is positive, and 1.3C if the weaker feedback is negative. Thus, once a strong positive feedback is present in the system, the effects of the other feedback processes are amplified. These equations assume small perturbations of the equilibrium climate and (5) assumes that the feedback processes are independent and additive. Climate feedback processes do interact with each other in important ways. Moreover, the climate will not be in equilibrium for the next several centuries, but rather will be responding in a transient way to changing conditions. For these reasons the formalism of linear feedback analysis described here can be used only as a rough guide to the relative importance of feedback processes.