4
Physical Climate Change in the 21st Century

4.1
REGIONAL PATTERNS OF WARMING AND RELATED FACTORS

An approximation on which we will rely in this report is “pattern scaling”, in which a robust pattern of climate change is assumed to exist that describes the geographical and seasonal structure of the temperature response that does not depend on climate sensitivity or on details of the forcing or emission scenario. All of the scenario and sensitivity dependence is captured within the time evolving global mean surface temperature, TG(t). Letting the symbol ξ stand for the two horizontal spatial coordinates and the time of year, the assumption is that

The pattern τ(ξ) has a spatial-annual mean of unity by definition. We focus on temperature here and discuss pattern scaling for precipitation in Section 4.2.

The validity of this approximation is discussed by Santer et al. (1990), Mitchell et al. (1999), and Mitchell (2003). It has been used extensively for regional temperature (and precipitation) change projections (Dessai et al., 2005; Murphy et al., 2007; Watterson, 2008) and impacts studies, as a substitute to running fully coupled simulations under different scenarios or with different models with a range of climate sensitivities.

The pattern is derived from experiments with fully coupled Global Climate Models (GCMs), with validation from efforts to isolate the well-mixed greenhouse gas signal from the historical temperature record. The value of the method relies on the pattern remaining fairly constant during a simulation, across different concentration pathway scenarios and across different model settings. Regionally and temporally differentiated results under different scenarios or climate sensitivities can be derived by first characterizing the stable geographical pattern of warming (and its spatial variability



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 105
4 Physical Climate Change in the 21st Century 4.1 REGIONAL PATTERNS OF WARMING AND RELATED FACTORS An approximation on which we will rely in this report is “pattern scaling”, in which a robust pattern of climate change is assumed to exist that describes the geographical and seasonal structure of the temperature response that does not depend on climate sensitivity or on details of the forcing or emission scenario. All of the scenario and sensitivity dependence is captured within the time evolving global mean surface temperature, TG(t). Letting the symbol x stand for the two horizontal spatial coordinates and the time of year, the assumption is that T(t,x) = TG(t) t(x) The pattern t(x) has a spatial-annual mean of unity by definition. We focus on temperature here and discuss pattern scaling for precipitation in Section 4.2. The validity of this approximation is discussed by Santer et al. (1990), Mitchell et al. (1999), and Mitchell (2003). It has been used extensively for regional temperature (and precipitation) change projections (Dessai et al., 2005; Murphy et al., 2007; Watterson, 2008) and impacts studies, as a substitute to running fully coupled simulations under different scenarios or with different models with a range of climate sensitivities. The pattern is derived from experiments with fully coupled Global Cli- mate Models (GCMs), with validation from efforts to isolate the well-mixed greenhouse gas signal from the historical temperature record. The value of the method relies on the pattern remaining fairly constant during a simula- tion, across different concentration pathway scenarios and across different model settings. Regionally and temporally differentiated results under dif- ferent scenarios or climate sensitivities can be derived by first character- izing the stable geographical pattern of warming (and its spatial variability 105

OCR for page 105
106 CLIMATE STABILIZATION TARGETS if some measure of uncertainty is sought) on the basis of available coupled model simulations, then adjusting the profile of transient warming by using a fast and simple energy balance model to predict the global mean surface temperature evolution. There are significant differences between the pat- terns generated by different models—for example, in the amount of polar amplification in the Arctic per degree of global warming. These differences are averaged over in the ensemble-based estimates of mean patterns, but uncertainty can be characterized by the inter-model spread in the pattern t(x). The choice of the pattern in the studies available in the literature are often as simple as the ensemble average (across models and/or across scenarios, for the coupled experiments available) of the spatial change in temperature, normalized by the corresponding change in global average temperature, choosing the end of the simulations (usually last two decades of the 21st century) and a baseline of reference (pre-industrial or current climate). Similar properties and results have been obtained using more sophisticated multivariate procedures that optimize the variance explained by the pattern. There are limitations to this approach. It can break down if aerosol forc- ing is significant, not only because aerosols and greenhouse gases can have different spatial footprints, but also because the effects of aerosols them- selves are more difficult to characterize in this simple way. For example, Asian and North American aerosol production are likely to have different time histories in the future. Our focus in this report is on the greenhouse gas component of climate change, making the pattern scaling assumption more justifiable. Simple pattern scaling is regarded as especially useful for summarizing model projections of transient climate change due to well-mixed greenhouse gas increases on a time scale of a few centuries. But it is less accurate for stabilization scenarios, as the temperature changes approach an equilib- rium response. From the early work of Manabe and Wetherald (1980) and Mitchell et al., (1999) it has been clear that the pattern of temperature response evolves as the slow component of the warming, associated with equilibration of the deep oceans on multi-century time scales, equilibrates. In particular, on these long time scales the warming of high latitudes in the Southern Hemisphere is much larger relative to the global mean warming than in the earlier periods. Held et al. (2010) emphasize that this slow warm- ing pattern is present, but of small amplitude, during the initial transient adjustment phases of the response as well. There are some regions of sharp temperature gradients, near the ice edge

OCR for page 105
PHYSICAL CLIMATE CHANGE IN THE 21ST CENTURY 107 for example, where pattern scaling will break down. As the climate warms, temperature changes will be large as the ice edge moves across a particular location, but then return to small values with additional warming, because the ice edge is now further poleward. Given other uncertainties, we find that pattern scaling is justified for current attempts to link stabilization targets and impacts, keeping in mind limitations due to the evolution of the pattern of warming on long, stabili- zation time scales and the limitations in regions of sharp gradients. On the basis of CMIP3 simulations, Chapters 10 and 11 of IPCC AR4, WG1 analyzed geographical patterns of warming and measures of their variability across models and across scenarios. The executive summary of Chapter 10 reports that “[g]eographical patterns of projected SAT warming show greatest temperature increases over land (roughly twice the global av- erage temperature increase) and at high northern latitudes, and less warming over the southern oceans and North Atlantic, consistent with observations during the latter part of the 20th century …”. Figure 10.8 of the report depicts the patterns of annual average warming across three scenarios (A2, A1B and B1) and three time periods (2011-2030, 2046-2065, and 2080-2099) over which change is computed. Figure 10.9 shows seasonal patterns for DJF and JJA under A1B. Chapter 10 also reports that the spatial correlation of fields of temperature change is as high as 0.994 in the model ensemble mean when considering late 21st century changes between A2 and A1B. A table in the same section (Table 10.5) quantifies the strict agreement between the A1B field, as a standard, and the other scenario patterns using a measure proposed by Watterson (1996) with unity meaning identical fields and zero meaning no similarity. Values of this measure are consistently above 0.8 and increase as the projection time increases (later in the 21st century fields agree better than earlier in the century), with values of 0.9 or larger for the late 21st century. The same table also shows that the agreement deteriorates if considering commitment scenarios. The results are documented as apply- ing to seasonal warming patterns besides annual averages. On the basis of the previous discussion and results, we compute patterns of standardized warming from the available CMIP3 SRES scenario simula- tion and produce maps of the ensemble average warming (these—in their non-normalized version—are available from AR4 WG1 Figures 10.8 and 10.9, and individual models’ maps are available in supplementary material in Chapter 10 (http://ipcc-wg1.ucar.edu/wg1/Report/suppl/Ch10/Ch10_ indiv-maps.html, and Chapter 11 (http://www.ipcc.ch/pdf/assessment- report/ar4/wg1/ar4-wg1-chapter11-supp-material.pdf) along with measures of the regionally differentiated variability of this pattern across models and scenarios (IPCC, 2007a).

OCR for page 105
108 CLIMATE STABILIZATION TARGETS Figure 4.1 shows patterns of warming normalized “per degree C of global annual average warming”. Ranges across models or across scenarios are shown in Figures 4.4 and 4.5. We show maps of global geographical patterns and North American patterns for annual average warming and December-January-February and June-July-August average warming. We highlight here the main features characterizing warming patterns: The annual and DJF mean patterns present a typical gradient of warming that decreases from north to south, with the higher latitude of the Northern Hemisphere seeing the largest increases, and the land masses warming more than the oceans. The JJA patterns’ main characteristics are an enhanced warming of the interior of the continents and the Mediterranean Basin, with a gradient that is generally equator to poles rather than north-south. The largest source of variation resides in the inter-model spread rather than the inter-scenario spread, and it is mainly localized over and at the edge of the ice sheets of the Arctic and Antarctica. We also present in Figure 4.2 a number of scatter plots depicting the relation between the magnitude of global average warming (by 2080-2099 compared to 1980-1999) and the magnitude of regional warming across models (each model one dot) and scenarios (color-coded) for several of the “Giorgi regions” (Giorgi and Francisco, 2000). We chose four regions that subdivide the North American continent (western, central, and eastern North America—WNA, CNA, and ENA respectively—and Alaska, ALA) plus two regions in other climates for comparison, the Mediterranean Basin (MED) and Southern Australia (SAU). The linearity of the relationship and the fairly tight spread around it is clear. As already noted, the source of variation between different models is larger than from the scenarios: averaging each model across the three scenarios would not reduce the scatter as much as averaging all the models within one scenario, as shown by the larger star- shaped marks. Finally, in Figure 4.3 we show December-January-February and June- July-August warming patterns as calculated in models including both an- thropogenic and natural forcing for the 20th century, together with 20th century observations. Figure 4.3 shows that the warming patterns in the models and observations display many common large-scale features, and a comparison with Figure 4.1 demonstrates how 20th century patterns con- tain already many of the large-scale features that characterize 21st century patterns. Among these, the relatively larger magnitude of the warming in December-January-February than in June-July-August in both observed and modeled patterns; the amplification of the warming in the high latitudes of the Northern Hemisphere characteristic of December-January-February

OCR for page 105
FIGURE 4.1 Geographical patterns of warming per 1°C of global annual average warming, over the whole world (top) and over North America 4.1.eps (bottom). From left to right: patterns for average temperature over the whole calendar year or for average temperature in boreal winter (De- cember, January, and February) or summer (June, July, and August). Patterns are obtained by scaling end-of-21st-century changes (compared 6 bitmaps, landscape to end-of-20th-century climatology) by global average annual warming over the same period, using temperature at the surface from the 18 CMIP3 models whose output is available for SRES scenarios A2, A1B, and B1. 109

OCR for page 105
110 CLIMATE STABILIZATION TARGETS FIGURE 4.2 Scatterplots comparing regional average warming versus global average warming (both aver- aged over the calendar year) for western North America (WNA), central North America (CNA), eastern North America (ENA), Alaska (ALA), Southern Australia (SAU) and the Mediterranean (MED). Each point indicates results from an individual model under oneFigure 4-2.epsscenarios (A1B, blue; A2, red; and B1, green). of the three SRES bitmap Stars indicate the multi-model ensemble averages for each of the three scenarios.

OCR for page 105
PHYSICAL CLIMATE CHANGE IN THE 21ST CENTURY 111 Patterns of Warming (1955-2005) Dec-Jan-Feb Jun-Jul-Aug GISS Observataions CRU Models 4-3.eps FIGURE 4.3 Relative patterns of warming (normalized to one for the globe),for December-January-February 6 tiffs + vector labels (left) and June-July-August (right) for 1955-2005 (obtained as differences between 20-year average temper- atures for 1986-2005 and 1955-1974). The top two panels show results from two instrumental temperature records, NASA GISS and CRU. White indicates regions where data are not available. The bottom panels show results for the Climate Modelling Intercomparison Project (CMIP3) multi-model ensemble. Patterns ex- pected from projections for the 21st century are largely similar to those shown here, as seen in Figure 4.1.

OCR for page 105
112 Annual Pattern, Range by Mod. DJF Pattern, Range by Mod. JJA Pattern, Range by Mod. 9.0 9.0 9.0 8.0 8.0 8.0 50 7.0 50 7.0 7.0 50 6.0 6.0 6.0 5.0 5.0 5.0 0 0 0 4.0 4.0 4.0 3.0 3.0 3.0 2.5 2.5 2.5 2.0 −50 2.0 −50 2.0 −50 1.5 1.5 1.5 1.0 1.0 1.0 0.5 0.5 0.5 0.0 0.0 0.0 −150 −100 −50 0 50 100 150 −−150 −100 −50 0 50 100 150 −150 −100 −50 0 50 100 150 JJA Pattern, Range by Mod. Annual Pattern, Range by Mod. DJF Pattern, Range by Mod. 9.0 9.0 9.0 70 70 70 8.0 8.0 8.0 7.0 7.0 7.0 60 60 60 6.0 6.0 6.0 5.0 50 5.0 5.0 50 50 4.0 4.0 4.0 40 40 40 3.0 3.0 3.0 2.5 2.5 2.5 2.0 2.0 2.0 1.5 1.5 1.5 30 30 30 1.0 1.0 1.0 0.5 0.5 0.5 0.0 0.0 0.0 −160 −140 −120 −100 −80 −60 −160 −140 −120 −100 −80 −60 −160 −140 −120 −100 −80 −60 FIGURE 4.4 For the same patterns as in Figure 4.1, the range across the 18 models of the average of each model’s patterns obtained under the three different scenarios is shown as a measure of the variabilitygof rthe4patterns produced by different GCMs. Fi u e -4

OCR for page 105
DJF Pattern, Range by Scen. JJA Pattern, Range by Scen. Annual Pattern, Range by Scen. 1.00 1.00 1.00 50 50 50 0.75 0.75 0.75 0 0.50 0 0.50 0 0.50 0.40 0.40 0.40 0.30 0.30 0.30 −50 −50 −50 0.20 0.20 0.20 0.15 0.15 0.15 0.10 0.10 0.10 0.05 0.05 0.05 0.00 0.00 0.00 −150 −100 −50 0 50 100 150 −150 −100 −50 0 50 100 150 −150 −100 −50 0 50 100 150 Annual Pattern, Range by Scen. DJF Pattern, Range by Scen. JJA Pattern, Range by Scen. 0.40 0.40 0.40 70 70 70 60 60 60 0.30 0.30 0.30 50 50 50 0.20 0.20 0.20 0.15 0.15 0.15 40 40 40 0.10 0.10 0.10 30 30 0.05 30 0.05 0.05 0.00 0.00 0.00 −160 −140 −120 −100 −80 −60 −160 −140 −120 −100 −80 −60 −160 −140 −120 −100 −80 −60 FIGURE 4.5 For the same patterns as in Figure 4.1, the range across SRES A1B, A2, and B1 of the patterns derived as a multi-model ensemble is shown as a measure of the variability of the patterns under different emission scenarios. Figure 4-5 113

OCR for page 105
114 CLIMATE STABILIZATION TARGETS warming; and the strong signal of warming in the semi-arid to arid regions especially in the Mediterranean basin and nearby portions of the Asian continent in June-July-August. Areas of disagreement between models and observations remain, particularly in the form of more homogeneous patterns of warming at the subcontinental scales in the modeled patterns than in observations, with one example being the area of cooling in the southeast of the United States appearing in the HadCRU data and not represented in the models. 4.2 PRECIPITATION RESPONSE As described in Section 4.1, it is reasonable to assume that the local temperature response to an increase in a well-mixed greenhouse gas is proportional to the global mean temperature response, with a well-defined spatial and seasonal pattern. There are also good reasons to assume that the local precipitation response scales with the global mean surface tem- perature response, although the uncertainties are greater, both with regard to the spatial and seasonal structure of the pattern and with regard to the limitations of this pattern scaling assumption. Using the CMIP3 archive and computing the precipitation response, measured as a percentage change, divided by the global mean warming and then averaging over models and scenarios, just as for the temperatures in Figure 4.1, one obtains the pattern shown in Figure 4.6, both globally for the annual average and the summer and winter seasons, and focusing on North America (which is not found in the IPCC AR4 report [IPCC, 2007a]). The patterns are very similar to those shown for a particular scenario and time frame in the Summary for Policy Makers of the AR4/WG1 report (IPCC, 2007e). There is a general increase in precipitation in subpolar and polar lati- tudes, and a decrease in the subtropics, and an increase once again in many equatorial regions. The boundary between the subtropical decrease and subpolar increase cuts through the continental United States, but with the boundary moving north in the summer and south in the winter. As a result, this ensemble mean projection is for an increase in precipitation in much of the continental United States in the winter and a reduction in the summer. In contrast, Canada is more robustly wetter and Mexico drier, being located closer to the centers of the subpolar region of increasing precipitation and the subtropical region of decreasing precipitation, respectively. The Mediter- ranean/Middle East and southern Australia are other robust regions of drying in these projections.

OCR for page 105
FIGURE 4.6 Geographical patterns of percentage precipitation change per 1°C of global annual average warming, for the whole world (top) and over North America (bottom). From left to right: patterns of changes in precipitation averaged over the whole calendar year or over boreal winter (December, January, and February) or summer (June, July, and August). Patterns are obtained by scaling end-of-21st-century percentage changes in precipitation (compared to end-of-20th-century climatology) by global average annual *warming* over the same period for the 18 CMIP3 models whose output is available for all three SRES scenarios A2, A1B, and B1. White regions are where less than Figure 4-6.eps 115 two-thirds of the models agree over the sign of the change. bitmap, landscape

OCR for page 105
148 CLIMATE STABILIZATION TARGETS thermal expansion in the upper 700m thermal expansion in the deep ocean ice sheets of Antarctica and Greenland glaciers and ice caps terrestrial storage } estimated sea levels (from satellite altimeter observations) sum of the contributions one standard deviation error for sea level (est.) NRC rigorous estimates of one standard deviation error for upper-ocean thermal expansion FIGURE 4.22 (a) Total observed sea level rise and its components. The components are thermal expansion in the upper 700 m (red), thermal expansion in the deep ocean (orange), the ice sheets of Antarctica and Greenland (cyan), glaciers and ice caps (dark blue), and terrestrial storage (green). (b), The estimated sea levels are indicated by the black line, the yellow dotted line,4-22.eps dotted line (from satellite altimeter and the red bitmap with vector type observations). The sum of the contributions is shown by the blue line. Estimates of one standard deviation error for the sea level are indicated by the grey shading. For the sum of components, we include our rig- orous estimates of one standard deviation error for upper-ocean thermal expansion; these are shown by the thin blue lines. All time series were smoothed with a 3-year running average and are relative to 1961. Source: Domingues et al. (2008).

OCR for page 105
PHYSICAL CLIMATE CHANGE IN THE 21ST CENTURY 149 FIGURE 4.23 Estimates of the contribution of glaciers and ice caps to global change in sea level equivalent 4.23.eps (SLE), in millimeters SLE per year. Source: Allison et al. (2009b). bitmap land and Antarctica, can potentially contribute a total of approximately 0.7 m to global sea level, and they provide a source of freshwater in many mountain regions worldwide (Bahr et al., 2009). For 1961-2003, glaciers and ice caps contributed 0.5±0.2 mm y–1 to global sea-level rise (Figure 4.22), increasing to 0.8±0.17 mm y–1 for 1993-2003 (Allison et al., 2009b). This new assessment (Figure 4.23) shows ice loss from glaciers and ice caps slightly higher than those reported in IPCC AR4, contributing now about 1.2±0.2 mm y–1 to global sea level rise. Glaciers and ice caps are not in balance with the present climate; glaciers need to decrease in volume by 27% on average, and ice caps need to decrease by 26% to attain equilibrium

OCR for page 105
150 CLIMATE STABILIZATION TARGETS (Bahr et al., 2009). A future equilibrium with the current climate implies a change in sea level of 0.089±0.015 m due to glaciers and 0.095±0.029 m due to ice caps, with a total change in sea level of 0.184±0.033 m. The ratio of accumulation area (AAR) at the end of the melt season and total glacier area dropped from roughly 0.54 in 1961 to 0.44 in 2007, and it is expected to drop further to 0.31 by 2050. This is a conservative estimate— observations indicate a faster than linear decrease in global ice mass balance over the past 40 years (Kaser et al., 2006). Although the actual decrease in AAR may be faster than linear, this conservative estimate represents a 30% decrease from the current value. As a rough approximation, we can assume that the AAR of every glacier decreases by the same percentage, giving an estimate of the fractional volume change for each glacier. In that case, the minimal sea level rise from glaciers and ice caps will be 0.373±0.021 m over the next 100 years (Bahr et al., 2009). Greenland and Antarctic Ice Sheets On the polar ice sheets, there is observational evidence of accelerating flow from outlet glaciers both in Greenland and in west Antarctica. Both inland snow accumulation and marginal ice melting have increased over the Greenland ice sheet, but there is little evidence for any significant ac- cumulation trend over the Antarctic ice sheet. Antarctica and Greenland maintain the largest ice reservoirs on land. For 1993-2003, the estimated contributions for the Greenland and Antarctic ice sheets are 0.21±0.07 mm y–1 and 0.21±0.35 mm y–1, respectively (Bindoff et al., 2007). There is little information to constrain ice sheet contributions for previous decades, but it is thought that the Greenland contribution has increased significantly in recent years (Lemke et al., 2007). Since IPCC AR4, there have been a number of new studies on ice sheet mass budget that have considerably enhanced our understanding of ice sheet vulnerabilities (Figure 4.24) (Allison et al., 2009a). Recent observations have shown that changes in the rate of ice discharge into the sea can occur far more rapidly than previously suspected (e.g., Rignot, 2006). The pattern of ice sheet change in Greenland is one of near-coastal thinning, primarily along fast-moving outlet glaciers. Accelerated flow and discharge from some major outlet glaciers (also called dynamic thinning) is responsible for much of the loss (Rignot and Kanagaratnam, 2006; Howat et al., 2007). Pritchard et al. (2009) used high-resolution satellite laser altimetry to show that dynamic thinning of fast-flowing coastal glaciers is now widespread at all latitudes in Greenland. Figure 4.24 shows estimates

OCR for page 105
PHYSICAL CLIMATE CHANGE IN THE 21ST CENTURY 151 FIGURE 4.24 Estimates of the net mass budget of the Greenland Ice Sheet since 1992 (Allison et al., 2009b). The horizontal dimension of the boxes shows the time period for which the estimate was made, and the vertical dimension shows the upper and lower limits of the estimate. The colors represent the different methods that were used: black is satellite radar altimetry, orange is aircraft laser altimetry, purple is aircraft/ satellite laser altimetry, red is the flux component method, and blue is satellite gravity. of the mass balance of the Greenland ice sheet that have been made since the early 1990s. The horizontal dimension of the boxes shows the time period for which the estimate was made, and the vertical dimension shows the upper and lower limits of the estimate. The colors represent the dif- ferent methods that were used: black is satellite radar altimetry, orange is aircraft laser altimetry, purple is aircraft/satellite laser altimetry, red is the flux component method, and blue is satellite gravity. The dashed green box represents the estimated Greenland balance of the IPCC AR4 assessment. These data indicate that mass loss from the Greenland ice sheet may be in- creasing, although it is also clear that the various estimates are frequently not in agreement. Greenland lost roughly 150 Gt y–1 since 2000, increasing to 180±50 Gt y–1 (0.5±0.14 mm y–1 SLR) for the time period 2003-2007. More than 50% of the current ice loss is caused by increase in ice discharge and ocean interaction of tidewater glaciers, the remaining part can be explained by the increase in surface melt due to warmer summer temperatures (Hanna et al., 2008). The interior of the ice sheet is expected to be less vulnerable to future changes than the edge regions. Current discharge rates may repre- sent a transient instability, and whether they will increase or decrease in the

OCR for page 105
152 CLIMATE STABILIZATION TARGETS FIGURE 4.25 Estimates of the net mass balance of the Antarctic Ice Sheet since 1992 (Allison et al., 2009b). The horizontal dimension of the boxes shows the time period for which the estimate was made, and the vertical dimension shows the upper and lower limits of the estimate. The colors represent the different methods that were used: black is satellite radar altimetry, orange is aircraft laser altimetry, red is the flux component method, and blue is satellite gravity. future is unknown. Currently there is no dynamic ice sheet model that can predict the response of the Greenland ice sheet for a warmer climate. We can constrain a possible upper bound of SLR contribution from Greenland assuming a doubling in ice discharge and a continued increase in surface melt using the AR4 A1B scenario (Pfeffer et al., 2008). The total contribution to SLR would be about 0.16 m by 2100, with 0.09 m contribution from ice dynamics, and 0.07 m from surface melt (Pfeffer et al., 2008). The Antarctic ice sheet shows a pattern of near balance for East Antarc- tica and mass loss from West Antarctica and the Antarctic Peninsula since 2003 (Cazenave et al., 2009). However, the uncertainties of these measure- ments are large (Figure 4.25) and there is no strong evidence for increasing Antarctic loss over the period shown. However, there is a region in West Antarctica that shows increasing ice loss in recent years (Figure 4.25, most of this signal comes from West Antarctica). The Amundsen Coast Basin, including Pine Island ice and Thwaites Glacier, is not confined by large ice shelves, and these marine- based ice masses, which are potentially unstable, contain about 1.5 m sea level equivalent. The average ice velocity in this region is 2 km y–1, which is considerably higher than the average velocity of all Antarctic ice streams (0.65 km y–1) (Pfeffer et al., 2008). If this ice discharge continues to increase, SLR contribution from West Antarctica cannot be ignored. Pfeffer

OCR for page 105
PHYSICAL CLIMATE CHANGE IN THE 21ST CENTURY 153 et al. (2008) assume in their sea level rise sensitivity study for Antarctica a doubling of outlet glacier velocities in Pine Island and Thwaites Glacier within the first decade, and ice loss acceleration from the Antarctic Penin- sula at the same rate than melt increase at present-day rates of surface mass balance change, resulting in a sea level rise of 0.12 m by 2100, just from the increase in ice discharge. Summary of Sea Level Change Greenland and Antarctic ice sheets together would contribute up to 0.285 m sea level under the AR4 A1B warming scenario, assuming a dou- bling in ice discharge for the Greenland outlet glaciers, and the Amundsen Coast Basin in Antarctica. Such an increase in ice discharge has already been observed for several regions in Greenland. Glaciers and ice caps are expected to contribute 0.37±0.02 m sea level rise under the same warm- ing scenario, and thermal expansion is expected to contribute 0.23±0.09 m by 2100. Thus, the sea level rise by 2100 is expected to be at least 0.60±0.11 m from thermal expansion and ice loss from glaciers and small ice caps only. Assuming additional ice loss from Greenland at the rate, the total global sea level rise would be about 0.65±0.12 m by 2100. Doubling in ice discharge for both Greenland and Antarctica, the sea level increase could be as high as about 0.88 ±0.12m by 2100. The estimated range in sea level rise in 2100 is therefore from about 0.5 to 1 m. The dynamic response of ice sheets to global warming is the largest unknown in the projections of sea level rise over the next century. Vermeer and Rahmstorf (2009) made a semi-empirical projection linking sea level to temperatures from past observations; their statistical projection for a tem- perature scenario A1B (IPCC AR4: 2.3-4.3°C increase for 2100) predicts a SLR of 0.97-1.56 m above 1990 by 2100. This is consistent with what hap- pened during warming in the last interglacial time period (LIG) and cannot be ruled out. The LIG warming was caused by perturbations of Earth’s orbit (Overpeck et al., 2006) and arrived much more gradually than is projected for human-induced warming, so a faster sea level rise in the future than in the LIG would not be surprising. On the other hand, changes in the LIG were linked to ice at low elevation, which could behave differently from that at high elevation in the interior of Greenland.

OCR for page 105
154 CLIMATE STABILIZATION TARGETS 4.9 OCEAN ACIDIFICATION The oceanic uptake of excess atmospheric carbon dioxide alters the chemistry of seawater, which may impact a wide range of marine organisms from plankton to coral reefs (Doney et al., 2009a,b; NRC, 2010) (see also Section 6.3). Ocean acidification is in fact a series of interlinked and well- known changes in acid-base chemistry and carbonate chemistry due to the net flux of CO2 into surface waters (Figure 4.26). The chemical shifts include increases in the partial pressure of carbon dioxide (pCO2), the concentration of aqueous CO2, and the hydrogen ion (H+) concentration and decreases in pH (pH = –log10[H+]). The increase in hydrogen ion concentration acts to lower the concentration of carbonate ions (CO32–) through the reaction H+ + CO32– => HCO3–, even though the total amount of dissolved inor- ganic carbon (DIC) goes up (DIC = [CO2] + [HCO3–] + [CO32–]). Declining CO32– in turn lowers calcium carbonate (CaCO3) mineral saturation state, W = [Ca2+][CO32–]/Ksp, where Ksp is the thermodynamic solubility product that varies with temperature, pressure, and mineral form. Ocean surface waters FIGURE 4.26 Schematic indicating the effects on seawater carbonate chemistry due to the uptake of 4.26.eps excess carbon dioxide (CO2) from the atmosphere. Ocean acidification causes increases in some chemical species (red) and decreases in other species (blue). Ocean acidification also causes a reduction in pH (pH bitmap = –log10[H+]) and the saturation states, W, of calcium carbonate minerals in shells and skeletons of plank- tonic and benthic organisms and in carbonate sediments. On millennial and longer time scales, ocean pH perturbations are buffered by external inputs of alkalinity, denoted by calcium ions (Ca2+) and changes in the net burial rate of carbonate sediments.

OCR for page 105
PHYSICAL CLIMATE CHANGE IN THE 21ST CENTURY 155 are currently supersaturated (W > 1) for the two major forms used by marine organisms, aragonite (corals, many mollusks) and calcite (coccolithophores, foramaniferia, and some mollusks,). Because of pressure effects and higher metabolic CO2 from organic matter respiration, W decreases with depth often becoming undersaturated (W < 1), at which point unprotected shells and skeletons begin to dissolve. The controls on seawater pH and saturation state vary with time-scale. In the surface ocean, the pH of seawater varies substantially over annual to interannual time scales due to the net biological formation of organic matter (lowers CO2 and raises pH and W) and CaCO3 shells and skeletons (the reverse). Upwelling of CO2-rich water from below and variations in temperature, salinity, and alkalinity (a measure of the acid buffering ca- pacity of seawater) also influence surface water carbonate chemistry. The saturation state of polar waters is lower in large parts because of colder temperatures. In coastal waters, pH and W exhibit large natural spatial and temporal variations due to the interplay of river runoff, strong biological pro- ductivity, and in, some locations, coastal upwelling (Salisbury et al., 2008). Over decadal to century time scales, ocean carbon chemistry is modulated by net CO2 uptake from the atmosphere and trends in ocean circulation and biological productivity, which tend to redistribute dissolved inorganic carbon and alkalinity within the ocean water-column. On even longer time scales of many centuries to millennia, the weathering of calcium carbonate rocks on land adds alkalinity, a measure of the acid buffering capacity, in the form of calcium ions (Ca2+) and carbonate ions (CO32–), and alkalinity is removed by the burial, on continental shelves and margins, of biologi- cally formed carbonate sediments made of the shells and skeletons of some plankton, corals, and other calcifying organisms (Figure 4.26). Carbonate sediment burial rates are sensitive to seawater chemistry, and on millennial time scales longer, efficient damping feedbacks act to stabilize mean ocean alkalinity and pH. At the small number of available open-ocean time-series sites, signifi- cant secular trends in surface ocean carbonate chemistry are well docu- mented for the past two decades (Figure 4.27). The time-series records document clearly an increase in surface water pCO2 and DIC and a decline in pH that is consistent with the rate of change in atmospheric CO2 (Dore et al., 2009). The WOCE/JGOFS Global CO2 Survey completed in the 1990s provided a global estimate of ocean anthropogenic CO2 distributions and a baseline for assessing changes in ocean chemistry with time (Sabine et al., 2004). Decadal resurveys of a subset of the WOCE/JGOFS ocean transects also exhibit decreasing pH through time over the upper thermocline and

OCR for page 105
156 CLIMATE STABILIZATION TARGETS FIGURE 4.27 Time-series surface seawater carbonate system at Station ALOHA in the subtropical North Pacific north of Hawai’i, 1988-2008. The upper panel displays the partial pressure of CO2 (pCO2) in seawater 4.27.eps calculated from dissolved inorganic carbon (DIC) and total alkalinity (TA) (blue symbols) and in water- bitmap saturated air at in-situ seawater temperature (red symbols). Atmospheric CO2 data is from the Mauna Loa Observatory, Hawai’i. The lower panel displays in-situ surface pH based on direct measurements (green symbols) or as calculated from dissolved inorganic carbon and total alkalinity (orange symbols). Linear regressions (colored lines) and regression equations are reported for each variable. Source: Adapted from Dore et al. (2009).

OCR for page 105
PHYSICAL CLIMATE CHANGE IN THE 21ST CENTURY 157 FIGURE 4.28 Projected evolution of the annual-mean, zonally averaged aragonite saturation, W, plotted as a function of the annual-mean atmospheric CO2 mixing ratio at the ocean surface. The corresponding 4.28.eps years for the SRES A2 and B1 scenarios are given at the top. The largest decreases in aragonite saturation values occur in the tropics. Arctic and Southernbitmap Ocean surface waters transition from supersaturation to undersaturation in the annual-mean beginning at approximately 460 ppm and 550 ppm CO2, respectively. Undersaturated conditions occur for individual months at even lower atmospheric CO2 levels, beginning at approximately 410 ppm for the Arctic and 490 ppm for Southern Ocean. Source: Adapted from Steinacher et al. (2009). across ocean basins (Byrne et al., 2010). Based on ice-core CO2 data and the WOCE/JGOFS Survey, surface ocean pH has already dropped on average by about 0.1 pH units from pre-industrial levels (pH is measured on a logarith- mic scale and a 0.1 pH drop is equivalent to a 26% increase in hydrogen ion concentration) (Orr et al., 2005). The patterns of ocean acidification in subsurface waters depend on ocean circulation patterns; thermocline waters in subtropical convergence regions and deep-waters in polar regions where cold surface waters sink into the interior ocean are affected more than other parts of the subsurface. Future acidification of surface waters can be predicted for a given at- mospheric carbon dioxide level (see Figure 4.28). An additional decline of 0.15 pH unit would occur if atmospheric carbon dioxide increases from cur- rent levels to 550 ppm, and larger pH changes would occur, approximately

OCR for page 105
158 CLIMATE STABILIZATION TARGETS proportionally, for higher CO2 concentrations. Polar surface waters may become under-saturated with respect to aragonite, a key calcium carbon mineral that can affect the ability of organisms to build their shells, for at- mospheric CO2 levels of 400-450 ppm for the Arctic and 550-600 ppm for the Antarctic (Orr et al., 2005; Steinacher et al., 2010). For tropical surface waters, large reductions in calcium carbonate saturation state are expected to occur, but waters are expected to remain super-saturated for projected atmospheric CO2 during the 21st century for current scenario projections. Calcium saturation horizons (W = 1) have been observed to move upward, that is, shoaled (Feely et al., 2004; Orr et al., 2005), and there is evidence that water undersaturated for aragonite is already upwelling onto the con- tinental shelf off the U.S. west coast due to a combination of strong wind- induced upwelling and the penetration of anthropogenic CO2 into off-shore source waters (Feely et al., 2008). For most of the surface ocean, climate change feedbacks are weak, and warming and altered ocean circulation have a limited effect on changing pH and W that are determined primarily by atmospheric CO2. An exception is in the Arctic, where sea-ice retreat and changes in surface freshwater balance amplify atmospheric CO2-driven pH and W declines (Steinacher et al., 2010).