The Physics of Energy, page 153
Ocean and land biomass CO2 uptake At the present time, Earth’s oceans are absorbing roughly one quarterof excess human CO2 emissions, slowing the rate of growth of atmospheric CO2. Paleoclimate data suggest, however, that in earlier interglacial warming periods, an increase in radiative forcing from changes in orbital parameters and associated warming of the planet led to a net release of CO2 from the oceans. The mechanisms that maintain the ocean/atmosphere CO2 balance are not well understood, but they are key to understanding the longevity of CO2 in Earth’s atmosphere and the extent and duration of climate change caused by anthropogenic CO2. Terrestrial biosystems currently absorb approximately another quarter of anthropogenic CO2 emissions.Some studies suggest that both ocean and land uptake will slow or even reverse direction as CO2 levels increase further; in particular, the fractional rate of ocean uptake of excess CO2 above pre-industrial levels will likely decrease as CO2 saturation and temperatures increase [268, 269], and may already be slowing [270], but there is as yet no scientific consensus on the precise future trajectory of either the ocean or land carbon sinks over the coming decades or centuries.
Any estimation of future human activity related to climate effects involves assumptions about social behavior, economics, and politics, which are outside the scope of this book. While the per capita rate of energy use from fossil fuel sources is clearly an important component in determining total human greenhouse gas emissions, as discussed in §1 (Introduction), the rate of population increase is also an important factor. All these ingredients are subject to substantial uncertainty and are difficult to estimate. Nevertheless, it is worth attempting to estimate future climate change based on a range of possible scenarios in order to estimate how human behavior over the next few decades will affect Earth’s climate for several centuries into the future. Since methane and aerosols are removed from the atmosphere on a much shorter time scale than CO2, anthropogenic CO2 is a primary driver of century- to millennium-scale climate effects, and provides a single simple parameter on which one can base a discussion of some future effects on the time scale of the next century or two.
In the remainder of this section we focus on the time scales over which CO2 emissions, atmospheric CO2, and global temperature change may reach their respective peaks, and put this in the context of scenarios for future climate considered by the IPCC [245]. We must emphasize the heuristic nature of this discussion given the many uncertainties involved; the goal of this analysis is to give a sense of the relevant time scales involved, not to make specific predictions.
CO emissions and atmospheric concentration From Figure 34.12, it is clear that anthropogenic CO2 emissions have increased rapidly in the first decade of the twenty-first century. Whether government regulation, economic imperatives, and other human choices or actions are leading to a leveling out or will result in a decrease in emissions in the coming decades is impossible to know now. In the most optimistic scenario, emissions might peak in the coming decade and drop precipitously well before the end of the century. In a pessimistic scenario, emissions could rise for on the order of one hundred years, and tail off slowly as all available fossil fuel resources are used. Obviously, the difference in resulting radiative forcing between these scenarios is enormous. The IPCC in 2013 considered four overall scenarios for radiative forcing [245]. These were based not on specific socioeconomic scenarios but rather on internally consistent scenarios of radiative forcing and atmospheric concentrations of greenhouse gases and aerosols. These Representative Concentration Pathway scenarios, RCP 2.6, 4.5, 6.0, and 8.5, are labeled by the peak or stabilization level of radiative forcing reached in the twenty-first century. The RF levels in these models are graphed in Figure 35.14, and the CO2 emissions in each model are shown in Figure 35.15.
Figure 35.14 Total radiative forcing for Representative Concentration Pathway (RCP) scenarios used in IPCC5 report [245].
Figure 35.15 Scenarios of (a) CO2 emissions, (b) atmospheric CO2, (c) projected surface temperature change, and (d) ocean thermal expansion for Representative Concentration Pathway (RCP) scenarios used in IPCC5 report. Adapted from [245].
In addition to the uncertainty in human activity and the rate of greenhouse gas emissions, the poorly understood nature of the land and ocean carbon sinks further complicates predictions. As mentioned in §34.4, part of the carbon uptake from terrestrial biomass is associated with reforestation of previously cleared regions in the Northern Hemisphere, and part is from a general increase in plant growth due to increased accessibility of CO2. It is unclear for how long either of these processes will continue to remove CO2 from the atmosphere at current rates. Similarly, our understanding of oceanic absorption of CO2 is insufficient to make reliable predictions for the future. We emphasize that as the climate has been warming over the last 50 years oceanic uptake of CO2 has increased, whereas the geological data on the last several hundred thousand years shows that warming in the past appears to have led to a net increase of atmospheric CO2 associated with decreased oceanic levels of carbon. Since neither process is yet well understood, predictions of future oceanic CO2 uptake are subject to substantial uncertainty. As a result, even were the rate of CO2 and other greenhouse gas emissions in the future known with complete certainty, there would still be considerable uncertainty as to the effects on radiative forcing over the next several centuries. The IPCC5 RCP scenarios focus on the future trajectory of radiative forcing, and do not make specific assumptions regarding anthropogenic emissions and the future behavior of ocean and land sinks.
Time scale of atmospheric CO2 Even after net CO2 emissions peak, atmospheric levels will continue to rise for quite some time. Atmospheric CO2 concentrations will only begin to drop when the rate at which CO2 is removed by natural systems such as oceanic absorption exceeds the anthropogenic excess in CO2 emissions. For example, since currently (circa 2015) roughly one half of the excess emissions are removed by natural systems, emissions would need to drop by a factor of two for atmospheric levels to begin to decrease. Given the uncertainties mentioned above, the most optimistic assumption is that uptake by both terrestrial biosystems and the oceans will continue to remove a similar fraction of the CO2 excess above the pre-industrial equilibrium concentration. A projection of future atmospheric CO2 based on this assumption is probably an underestimate, since recent studies suggest that both land and ocean sinks will slow or even reverse with increased CO2 levels [268, 269]. Slowing of terrestrial biomass or ocean uptake would lead to a longer persistence of atmospheric CO2 levels and higher levels in the future.
The time scale for excess CO2 to be removed from the atmosphere is thus not well understood. Estimates range from several decades to on the order of centuries or possibly thousands of years. The terrestrial biomass carbon sink has a finite capacity that may be saturated well before the pulse of excess CO2 from human fossil fuel usage has been removed from the atmosphere. And even as CO2 levels begin to drop again biological systems such as terrestrial plants will presumably begin to release net carbon to the atmosphere again to restore equilibrium. While chemical weathering would presumably eventually restore atmospheric CO2 levels to their current natural equilibrium even in the absence of other mechanisms, this process occurs over very long time scales.
In the short term the primary mechanism for removing most anthropogenic CO2 from the atmosphere is assumed to be ocean absorption, leading to deep ocean mixing. Assuming that the mixed layer is eventually saturated at a higher CO2 level in equilibrium with the atmosphere, the rate for further CO2 removal from the atmosphere depends in part upon the rate of deep water formation in areas such as the North Atlantic. It is possible that this process may slow as, for example, melt water from the Greenland ice pack is released into the North Atlantic, reducing the salinity of that water. Notwithstanding these uncertainties, a rough estimate based on the (optimistic) assumption that current rates of ocean and land biomass absorption of excess CO2 will continue, suggests that once emissions have ceased, the atmospheric lifetime of excess atmospheric CO2 will be on the order of 50 years or more. Thus, even in an optimistic scenario such as RCP 2.6, the effects of human fossil fuel emissions on atmospheric CO2 levels will almost certainly last at least into the next century. In a less optimistic scenario, if the ocean and land sink absorption does not continue indefinitely, the lifetime of CO2 in the atmosphere may be closer to the deep ocean mixing time of on the order of one thousand years.
Time scale for temperature change Different parts of the climate system have very different time scales for reacting to externally forced climate change. As discussed above in the context of orbital-scale climate change, large ice sheets react to changes in temperature or radiative forcing on a time scale of hundreds to thousands of years. Thus, for example, it is expected that the large Antarctic ice sheets will remain largely intact over the several hundred or thousand years during which the anthropogenic pulse of added atmospheric CO2 impacts the climate. The climate system, therefore, will not have time to come into a new global equilibrium. Instead, relatively faster-reacting parts of the system will shift dramatically in response to the rapid changes in radiative forcing, while other parts of the climate system will be less affected. Since there is no clear geological record of such rapid changes in external climate forcing at any point in the past, paleoclimate records provide only a partial analogy. Thus, climate modeling, while imperfect, may be our best guide for what to expect in the next several hundred years for a given anthropogenic climate forcing. Since climate models are generally used to simulate geologically short periods of time evolution, slower-reacting components such as the Antarctic ice sheets are often incorporated as fixed boundary conditions. As discussed in §34.5, climate models suggest that the change in global surface temperature will be related to the change in radiative forcing through the climate sensitivity parameter , so that a doubling of CO2 would lead in steady-state to an increase in temperature of something like . This change will lag by a time on the order of many decades following a given increase in radiative forcing. For example, a simple computation (see Problem 35.6) shows that it would take several decades for a radiative forcing of 3.7 W/m to raise the temperature of the oceanic mixed layer by 3.2°C, even if all extra energy input from radiative forcing contributed to this warming.
Time Scales of Anthropogenic Climate Effects
Simple considerations suggest that atmospheric CO2 levels will continue to rise for many decades after anthropogenic CO2 emissions peak, and that global mean temperatures will continue to rise over even longer time scales. This, together with the logarithmic response of radiative forcing to CO2 levels, suggest that past and near-term future emissions will lock in climate changes that will persist over at least the next century.
Putting all these pieces together, the picture we arrive at is that for a given profile of anthropogenic CO2 emissions, atmospheric CO2 levels will continue to rise for some time after the emissions peak, and temperatures will in turn continue to rise for some time after the atmospheric CO2 level itself peaks. Thus, there will be a time lag of perhaps 50–100 years or more between the time when human emissions begin to decrease and the time of maximum deviation in global surface temperature. A very simple model of these effects and time scales is developed in Example 35.1. The same effects can also be seen clearly in the more detailed projections used in the IPCC5 report, shown in Figure 35.15.
Example 35.1 A Simple Model of Time Scales for CO2 Concentrations and Climate Response
A highly simplified model of the relationship between CO2 emissions, atmospheric CO2 levels, and global temperature change illustrates time lags between these processes and provides a qualitative understanding of some of the processes underlying the simulation results illustrated in Figure 35.15. CO2 emissions are modeled as a logistic distribution (see Box 33.6) with peak fixed in 2050 and width and height chosen to fit emissions data from the last 50 years. CO2 emissions are measured in units of 2010 emissions; CO2 levels are graphed as a multiple of pre-industrial levels (280 ppm); and temperature is plotted in °C above the 1961–1990 mean. The graphs in the figure are based on the assumption that land and ocean absorption remove 2% of excess atmospheric CO2 yearly (roughly the current rate), and that radiative forcing is controlled by the temperature of the oceanic mixed layer, which in turn absorbs 50% of excess incoming radiation (as in Problem 35.6), assuming that the radiative forcing associated with doubling of CO2 is , as in eq. (34.36).
This model is intended to give a heuristic sense of the relative time scales involved for peaking of CO2 emissions, atmospheric concentration, and climate effects, and should not be taken as quantitatively accurate. The object is to formulate a simple calculable model that can illustrate the scale of the time lags involved and that anyone can easily reproduce. The reader is invited to improve the assumptions and to compute and analyze the results him/herself (Problem 35.7). Note that the assumption made for the land and ocean sinks is almost certainly overly optimistic, so the true lag of atmospheric concentration and climate response is likely longer than this model suggests.
Note that the logarithmic dependence of radiative forcing and temperature change on CO2 levels (as in eq. (34.35)) means that the effect of increasing CO levels on global temperature is greatest for the initial deviation from pre-industrial levels. While doubling atmospheric CO2 content by adding 280 ppm above pre-industrial levels may lead to 3.2°C of temperature change in a new steady-state configuration, to generate a mean 6.4°C temperature increase would require going to four times pre-industrial levels, or 1120 ppm. This sublinear growth can be seen clearly in the different RCP scenarios shown in Figure 35.15. This logarithmic dependence of climate on greenhouse gas content can be regarded either as a reassuring sign that the anthropogenic increase in CO2 emissions is unlikely to lead to a catastrophic runaway climate situation, or as a call to immediate action because the emissions with the greatest relative climatic impact are the ones occurring now.
It is well established that global mean surface temperatures will rise as a consequence of past and future anthropogenic greenhouse gas emissions. Nevertheless, many uncertainties remain in current analyses of the expected future consequences of these emissions. Climate models, which provide the best available estimates of climate sensitivity, are imperfect in a number of ways. The models are unable to fully incorporate important feedback such as from clouds with certainty, may be missing some important features such as increased ice melt from particulate emissions, and disagree in some ways (such as latitudinal temperature distribution) with paleoclimate data about some particular past climatic conditions. As a result, as emphasized in the reports of the IPCC, the assessment of the uncertainty in projections of future climate is at least as important as the specific predictions made. An illustration of the relative uncertainties in the RCP models and simulations is shown in Figure 35.16. The figure shows several sources of uncertainty in projected temperature for RCP models [245]: internal variability refers to uncertainty within individual models, arising from natural variability in the climate system, and represents a fundamental limit on the precision of predictions; model spread indicates differences among different simulations, and gives some indication of uncertainty due to imperfect modeling of natural systems; and RCP scenario spread refers to the variation in input parameters, including uncertainty about future human actions, and is responsible for the greatest uncertainty in temperature predictions.
Figure 35.16 IPCC estimate of uncertainties in temperature projections through the end of this century [245].
It is also possible that abrupt changes in radiative forcing may generate climate change that cannot be anticipated by near-equilibrium modeling. At times in the past, such as the glacial–interglacial transition points, fairly small changes in external forcing seem to have switched the global climate from one class of local equilibria to another qualitatively different class of local equilibria. Such changes in Earth’s climate cannot be ruled out over the next several hundred years in which anthropogenic forcing may play a dominant role in determining short-term climatic change.
35.3Effects of Climate Change
We summarize some of the main consequences that may be associated with anthropogenic climate change over the next 100–200 years, keeping in mind the caveats in the final paragraphs of the preceding section.
35.3.1 Global Distribution of Temperature Change
While we have focused so far primarily on global average temperature differences, climatic change will not be uniformly distributed around the planet. Greater warming will occur over large continental land masses than over oceans. This effect arises only in small part from the different heat capacities of land and ocean surface waters, and has more to do with atmospheric and surface effects, such as clouds and soil moisture changes [245]. Paleoclimate data from warmer periods such as the Cretaceous and Eocene, as well as global climate models, indicate that as global temperature rises the temperature gradient between equatorial and polar regions decreases. Thus, high latitudes are expected to be more strongly affected by global warming than lower latitudes. Ice albedo feedback makes this effect particularly pronounced in the Northern Hemisphere. While, as mentioned above, climate models and paleoclimate data are not in complete agreement regarding the extent of heat transport to high latitudes in warmer climates, combining these perspectives makes some predictions possible.
