On palaeoclimate time scales, enhanced levels of geological and geomorphological activity have been linked to climatic factors, including examples of processes that are expected to be important in current and future anthropogenic climate change. Planetary warming leading to increased rainfall, ice-mass loss and rising sea levels is potentially relevant to geospheric responses in many geologically diverse regions. Anthropogenic climate change, therefore, has the potential to alter the risk of geological and geomorphological hazards through the twenty-first century and beyond. Here, we review climate change projections from both global and regional climate models in the context of geohazards. In assessing the potential for geospheric responses to climate change, it appears prudent to consider regional levels of warming of 2°C above average pre-industrial temperature as being potentially unavoidable as an influence on processes requiring a human adaptation response within this century. At the other end of the scale when considering changes that could be avoided by reduction of emissions, scenarios of unmitigated warming exceeding 4°C in the global average include much greater local warming in some regions. However, considerable further work is required to better understand the uncertainties associated with these projections, uncertainties inherent not only in the climate modelling but also in the linkages between climate change and geospheric responses.
Changes in many aspects of geological and geomorphological activity have been linked to climate variability on palaeoclimate time scales. This includes glacial lake outbursts and rock-dam failures, submarine and sub-aerial landslides, tsunamis and landslide ‘splash’ waves, seismic and volcanic activity and gas hydrate destabilization. With human influence on climate now becoming significant and potentially very large, an emerging question is whether this could influence any or all of these processes and hence alter the risk of geological and geomorphological hazards. This field of study is very much in its infancy, but as part of the background to other research, this paper assesses climate change projections in the context of climate-related geological and geomorphological processes.
While changes in the global pattern of stress and strain in the Earth’s crust cannot be ignored, the more focused response of the geosphere to post-glacial planetary warming and hydrological adjustment is linked primarily to a relatively limited portfolio of ice-mass loss, rapidly rising sea levels and increased availability of liquid water, the latter arising from either ice melt or elevated levels of precipitation. These in turn drive a change in geological and geomorphological activity through changes in a small number of key environmental parameters acting mainly within the crust, notably variations in load pressure and increases in pore-water pressure. While the former are capable of promoting fault rupture and magma migration towards the surface, most notably in areas of ice-mass loss or sea-level rise, the latter are particularly instrumental in promoting the failure of masses of unstable rock or sediment in both marine and sub-aerial settings. The consequence of these climate-change-forced adjustments in environmental parameters, during post-glacial times, are: increased levels of seismicity (e.g. Davenport et al. 1989; Wu et al. 1999; Wu & Johnston 2000; Hetzel & Hampel 2005) and volcanism, both in and adjacent to formerly ice-covered regions (Sigvaldasson et al. 1992; Jull & McKenzie 1996; Licciardi et al. 2007; Huybers & Langmuir 2009) and also in the marine environment (McGuire et al. 1997; Huybers & Langmuir 2009); a higher incidence of large, submarine landslides along continental margins (Lee 2009); and elevated levels of mass movement in areas of mountainous terrain (Lateltin et al. 1997; Friele & Clague 2004; Hermanns et al. 2004).
Here, we examine the potential, in a warming world, for such responses to changes in environmental conditions to be accentuated at local, regional and global scales, leading to possible changes—via triggering, adjustment or modulation—in hazardous geological and geomorphological activity. Principal relevant climate quantities will be rising atmospheric and ocean temperatures, ice-mass loss, ocean volume increase and a greater incidence of extreme precipitation. Environmental settings that are most susceptible appear to be (i) regions of ice loss at high latitudes and high altitudes, (ii) ocean margins, (iii) mountainous terrain and (iv) regions predicted to experience significantly increased levels of precipitation (McGuire 2010).
2. Climate change research: informing mitigation and adaptation
The Earth’s climate has always changed over time, to a varying degree and on a range of time scales. For most of geological history, these changes were driven by natural forcings, most notably the Milankovitch and solar cycles and degassing in volcanic environments. Since the middle of the nineteenth century, increasingly detailed analysis of our weather and climate has revealed unusual changes that cannot be explained by these natural factors alone. The observed rise in global temperatures over recent decades is very likely to be due to the influence of anthropogenic changes in the concentrations of greenhouse gases (GHGs) such as CO2, CH4 and N2O (IPCC 2007a). Anthropogenic CO2 emissions show a general increase over time (CDIAC 2010) and, in the absence of major changes in energy policy, this increase is expected to continue for some decades (Nakicenovic et al. 2000) and is potentially only limited by the still large availability of fossil fuel reserves. Under such scenarios, increases in global mean temperature between 1.6°C and 6.9°C relative to the pre-industrial period are projected (IPCC 2007a).
Long response times in the climate system mean that the full effect of past anthropogenic climate forcing has not yet been realized, and hence the global climate is already committed to further change even if anthropogenic GHG emissions were to cease immediately. Moreover, given that GHG emissions continue to show an overall upward trend, and the international policy has yet to achieve a binding agreement on a workable mechanism for reduction of emissions, further increases in year-on-year emissions also seem inevitable, and limiting long-term cumulative emissions also appears to be increasingly challenging. In practice, then, the level of ‘committed’ warming appears likely to be above the level suggested by the physical inertia of the climate system alone.
There are therefore two reasons to assess the potential impacts of future climate change under a range of different scenarios. The first reason is to inform mitigation action: how much, and how soon, do GHG emissions need to be reduced in order to avoid harmful impacts? Clearly, this issue requires information on harmful impacts, which could potentially include geological and geomorphological hazards. The second reason is to assess the impacts of climate change that are now unavoidable, in order to inform adaptation planning to help minimize these impacts. Again, if the risks of geological and geomorphological hazards are altered by anthropogenic climate change even within the horizon of committed change, then this is important information for managing exposure to those risks.
One approach to the ‘mitigation question’ is to refer to logical ultimate consequences of ongoing climate warming, such as the loss of major ice sheets (Hansen 2003). Should all land ice melt, global sea levels would rise by up to approximately 70 m—this represents an extreme scenario that could be the ultimate, long-term response of the Earth system to a doubling of atmospheric carbon dioxide levels above pre-industrial levels (Hansen et al. 1997). Such loss of all land ice would require disintegration of the East Antarctic ice sheet, a potentially catastrophic event, but one that is not of immediate concern based on current climate modelling. However, there is evidence that lower levels of warming could lead to an eventual loss of the Greenland and West Antarctic ice sheets, resulting in a possible commitment to several more metres of sea-level rise beyond 2100.
Insight into these large impacts can be gained from palaeoclimate evidence—for example, during the Middle Pliocene, around 3 Myr BP, sea level was 25±10 m above today’s, in long-term equilibrium with a global average temperature just 2–3°C higher than that of the present day (Dowsett et al. 1994). However, the palaeoclimate record of natural climate variability provides no direct analogues for the full range of changes that may occur as a consequence of anthropogenic forcing of climate, owing to the different processes and time scales involved. Therefore, it is valid to use numerical models of the Earth’s climate system to assess potential future changes and their impacts.
(a) Modelling the climate
The past, present and future influences of natural and anthropogenic driving forces on the climate system can be assessed on a planetary scale using a range of models, from simple climate models (SCMs) that focus on key planetary-scale processes such as the global energy balance and carbon cycle to general circulation models (GCMs) that simulate the atmospheric circulation and other atmospheric processes as mechanistically as possible (still requiring a large number of approximations). For a more detailed evaluation at higher temporal and spatial resolutions, regional climate models (RCMs) are ‘nested’ within GCM projections to downscale the results of larger, coarse-resolution models in the context of finer-scale influences such as orography and land use.
Uncertainties in climate projections arise from a number of sources, with the influence of each of these varying through time (Meehl et al. 2007; Hawkins & Sutton 2009). The first area of uncertainty is natural variability—in other words, those fluctuations in our climate that occur on seasonal, annual and decadal time scales. In near-term climate projections, the greatest uncertainties arise from these regional climate responses. The second area of uncertainty arises from the fact that models are only approximations of the real world and hence are not perfect predictors of future change. This uncertainty is extremely difficult to quantify as the limitations of the approximations may not be known. However, one approach beginning to account for this uncertainty is to use a large number of models that deal with the need for approximations in different ways. Under the same driving conditions, individual climate models can give different results depending on their structural nature. This area of uncertainty is important to consider in the short term, alongside natural variability, and its influence increases over time.
In the long term, a third area of uncertainty becomes important, namely the effect of uncertainties in the anthropogenic emissions of GHGs. In the near term, the projected climate responses to the various plausible emission scenarios are not significantly different in comparison with the uncertainties arising from natural climate variability and uncertainties in the climate response to any given emission scenario. This is partly due to the long response time scales of the climate system described above—some change is already locked in, so will occur whichever emission scenario is followed, and the full response of any given emission scenario will take several decades to emerge. However, later in the twenty-first century, the projections due to different emission scenarios start to differ, and the uncertainty in future emissions becomes a major factor. Nevertheless, uncertainties in the climate response to any given emission scenario still remain very significant, especially at regional scales and in precipitation as opposed to temperature.
(b) Emission scenarios
The consequences of different potential future changes in anthropogenic emissions are examined through the use of emission scenarios. These are typically based on internally consistent storylines of future population growth, technology change, economic state and the character of the geo-political system. The most widely used emission scenarios so far are from the IPCC special report on emissions scenarios (SRES; Nakicenovic et al. 2000). These consider a wide range of future socio-economic pathways, but do not consider policies explicitly intended to reduce anthropogenic influences on climate. Therefore, these scenarios can be used to assess the potential consequences of not implementing reduction of emissions. Although these scenarios were developed in the 1990s and can therefore now be compared with actual emissions, it is still regarded as too early to reliably assess whether any of the SRES scenarios are emerging as more realistic than any others (Leggett & Logan 2008; Van Vuuren et al. 2008) More recently, new scenarios accounting for such climate policy have begun to be used in climate modelling studies, in order to assess the level of ‘residual’ climate change that would still occur with reduction of these emissions.
(c) Climate change projections
The SRES scenarios have been used in both SCMs and GCMs to assess the potential range of climate change that could arise from socio-economic futures that do not include an action to reduce anthropogenic influence on climate. These suggest a likely range of global mean temperature increase of 1.1–6.4°C by 2090–2099 relative to 1980–1999 (IPCC 2007a), which equates to approximately 1.6–6.9°C relative to pre-industrial levels. Under the highest emission scenario of the main SRES group, known as A1FI (where ‘FI’ denotes ‘fossil intensive’), a warming of 4°C relative to pre-industrial level would most likely be reached in approximately the 2070s, with a possible earlier date of approximately 2060 (Betts et al. submitted). The lowest of the main emission scenarios, known as B1, results in a central estimate of warming of 1.8°C by the 2090s relative to 1980–1999, which is 2.3°C relative to pre-industrial level.
At present, the most comprehensive set of climate projections with GCMs is that produced for the IPCC’s fourth assessment report (AR4), commonly known as the ‘AR4 multi-model ensemble’. Members of this ensemble of GCMs were used to simulate twenty-first century climate with a subset of the SRES scenarios. This subset included the ‘low’ B1 scenario described above, the A1B scenario (many similarities to the A1FI scenario but with a balanced growth between fossil fuel and other energy sources) and the A2 scenario, which assumes high population growth but results in lower emissions than A1FI from the 2020s onwards. The A1FI scenario was not used in the AR4 ensemble.
The full multi-model ensemble included 23 GCMs, but not all models were used for all scenarios. The most comprehensively studied scenario using all 23 models was A1B, with the central estimate of global warming from the GCMs being 2.8°C by the 2090s relative to 1980–1999. The ‘likely range’ of this warming for A1B from SCMs and expert judgement was 1.7–4.4°C.
Scenarios including global policies to reduce emissions have only recently begun to be used in GCM studies, but SCMs have been used to assess the implications of these for minimizing the level of future global warming. As with the projections using the SRES scenarios, these ‘mitigation’ scenarios produce a range of possible trajectories of global warming that reflect uncertainties in the climate response. Scenarios giving a significant probability of remaining below 2°C warming relative to pre-industrial level feature a peak in global emissions approximately around the year 2020 (IPCC 2007b).
(d) Using climate projections to inform mitigation and adaptation
The available climate projections are relevant for informing both mitigation and adaptation, but different aspects of the projections are relevant to these different issues. A key distinction is the time horizons over which the scenarios remain similar, and at which they begin to diverge, as this distinguishes the avoidable climate change relevant to mitigation from the unavoidable climate change to which the world will need to adapt. Using the SRES-based GCM ensemble, the climate change projected through the 2020s and into the 2030s is insensitive to the range of non-mitigation emission scenarios used in IPCC AR4, with the projections only diverging by 0.05°C from each other in the period 2011–2030 (figure 1; Meehl et al. 2007). It is only after the 2030s that the projections begin to diverge (Meehl et al. 2007). A similar time horizon of divergence is seen for mitigation scenarios assuming deep emission cuts.
This implies that projections after the 2030s are relevant to the question of which impacts of climate change could be avoided by implementing mitigation policies. The SRES-based GCM projections, and others projections using SCMs, can be used to assess the potential impacts of unmitigated climate change as part of the evidence to inform international climate policy. Prior to the 2030s, the projected climate change is not affected by reduction of emissions and the uncertainty arises only from the climate system itself. This means that climate change impacts, identified as taking place prior to this time horizon, may only be addressed through adaptation.
A further issue regarding adaptation concerns the residual climate change after the 2030s under mitigation scenarios. This relates to the commitment to some ongoing climate change even if GHG concentrations are stabilized. To illustrate this, some of the AR4 GCMs were used to simulate climate with GHG concentrations stabilized at the year 2000. About half of the projected rise in global mean temperature from the present day to the 2030s was attributed to the ‘committed climate change’. Fixing GHG concentrations at 2100 for the B1 and A1B scenarios still led to a further ongoing rise in projected global mean temperatures of 0.4–0.8°C by 2300. Moreover, mitigation scenarios with SCMs project that global temperatures could remain near 2°C above pre-industrial level for the next century even with rapid and deep emission cuts (Met Office 2007). Furthermore, even with mitigation measures being implemented within the twenty-first century, thermal expansion of the oceans will continue (as in the 2100 stabilization scenario modelling of figure 2).
(e) Regional climate change
This paper uses regional climate change assessments from AR4, which provide plausible ranges for temperature and precipitation averages by the end of the twenty-first century (2080–2099). They are derived from a set of global models using the SRES A1B emission scenario. Each projection presents the maximum and minimum possible change, alongside the median, although it must be emphasized that these projections do not encompass all possible futures.
The focus on the A1B scenario for regional climate uncertainty estimates meant that the AR4 regional climate change assessment did not examine the state of the climate for higher levels of global warming that would be expected if a scenario of high emissions is followed and/or if feedbacks in the climate system are stronger than expected in the standard models. This is an important omission in the context of geological and geomorphological hazards and is particularly important to consider for high-latitude regions. Some limited regional analysis is available for projections of higher levels of global warming—for example, Sanderson et al. (submitted) assess the 17 members of the AR4 multi-model ensemble that were used to project climate change under the A2 scenario, and distinguish a subset of nine that project warming of 4°C or above by the 2090s relative to pre-industrial level. Sanderson et al. (submitted) refer to these as ‘high-end’ projections, and this name is also used here.
Similarly, there is little information available as yet on the regional impacts of mitigation scenarios, for informing long-term assessment of committed change. The regional climate projections presented here provide some context for other work on potential changes in geological and geomorphological hazards, but considerable further work is required in order to assess the potentially larger changes in regional climate change that may further increase the risks of hazardous geospheric responses, and indeed the implications of committed climate change for these potential hazards.
3. Climate forcing of hazards in the geosphere
The geological and geomorphological hazards introduced in §1 are here evaluated in the context of regional climate change. Table 1 summarizes these hazards alongside relevant climate quantities, projected climate changes, and susceptible regions and environmental settings.
(a) Global oceans
Warmer oceans and increased ocean mass due to input of land ice meltwater have the potential to influence the stability of gas hydrate deposits in marine sediments and, as a consequence, alter the stability of submarine slopes. Increased ocean mass may also elicit volcanic and seismic responses in coastal and island settings. These, in turn, may contribute to the formation of sub-aerial, volcanic landslides, submarine landslides and tsunamis.
During the twenty-first century, global climate models project non-uniform increases in temperature across the surface of the Earth (figure 3). The spatial pattern of projected surface temperature change, presented in the IPCC’s AR4, suggests warming occuring to a greater extent over the land than the sea (Meehl et al. 2007). Observations of land–sea temperatures throughout the twentieth and into the twenty-first century reinforce the existence of a land–sea contrast (e.g. Dommenget 2009). This contrast is, in part, due to the thermal inertia of the oceans in comparison with that of the land when responding to the driving force of anthropogenic climate change. However, other studies have demonstrated that this contrast is also evident in equilibrium climate models and hence cannot be a function of the differential heat capacity alone. Joshi et al. (2007), Sutton et al. (2007) and most recently Dommenget (2009) all argue that the land–sea contrast seen in both equilibrium and transient climate models is driven mainly by the asymmetrical response and feedbacks of the land surface temperatures to the warming oceans.
The greatest ocean surface warming is projected for the Arctic, partly as a result of changes in albedo from melting sea ice (see §3b) and partly due to changes in poleward energy transport. The melting and forming of sea ice also contribute to the meridional overturning circulation of the North Atlantic. Trends in sea-ice cover since the late 1970s have revealed both decreases in the extent of sea-ice coverage of about 0.3×106 km2 per decade (Cavalieri et al. 2003; Stroeve et al. 2005) and a thinning of the icepack by 40 per cent over the 30-year period from 1966, especially in areas where sea ice was initially thickest (Rothrock et al. 1999, 2003). Research shows how the natural drivers of variations in ice extent and thickness alone cannot explain the decreases observed in recent decades—it is only when the influence of anthropogenic GHG emissions is included that the observed changes can be fully understood (Min et al. 2008). Equatorial regions also see large increases in ocean surface temperatures in the AR4 multi-model ensemble, particularly in the eastern Pacific (e.g. DiNezio et al. 2009).
Conversely, some regions must, by definition, see less warming than the global average—one region, in particular, projected to warm less rapidly is the North Atlantic between Greenland and the UK (figure 3). It is suggested that this is a result of a slowdown of the meridional overturning circulation (Meehl et al. 2007). Looking at zonal changes within the oceans, modelling suggests that the warming is limited to the mixed layer through the 2020s, with warming penetrating into the deeper ocean occurring later in the twenty-first century.
The increases in global ocean temperatures outlined above result in an increase in seawater volume, and hence thermosteric sea-level rise. By the end of the century, this thermal expansion could be 1.9±1.0, 2.9±1.4 and 3.8±1.3 mm yr−1 under the scenarios SRES B1, A1B and A2, respectively, using the IPCC AR4 global climate model ensemble (Meehl et al. 2007). According to IPCC AR4, through the twenty-first century, thermal expansion contributes between 0.10 and 0.41 m to sea-level rise, a large proportion of the projected total global mean sea-level rise of 0.18–0.59 m under the full range of SRES scenarios including A1FI. The remainder is from the melting of glaciers, ice caps and the Greenland and West Antarctic ice sheets, which increases the total mass of water in the oceans. The IPCC acknowledges that the contribution from melting ice sheets is largely unconstrained and could increase the upper end of the sea-level rise and ocean mass increase estimates. The rate of global mean sea-level rise has increased from a few centimetres per century in recent millennia, to a few tens of centimetres per century in recent decades (Milne et al. 2009). Some studies forecast a global sea-level rise by 2100 that is significantly higher than that projected in the IPCC AR4. Notably, Rahmstorf (2007) proposes a rise of between 0.5 and 1.4 m, whereas Pfeffer et al. (2008) estimates an upper bound of 2 m by the end of the century. These studies remain controversial as they rely on major approximations, but they do highlight the urgent need for better understanding and quantification of the processes of ice dynamics, which are of huge importance to global sea-level rise projections.
(b) High-latitude regions
The effects of anthropogenic climate change may be greater and occur more rapidly at high latitudes. The polar regions exhibit large inter-annual and decadal variability in their climates. This, alongside difficulties in representing the teleconnections and the interactions of the atmosphere–land–cryosphere–ocean–ecosystem feedback models at high latitudes within global and regional climate, results in a degree of uncertainty in the projections, especially for the Antarctic, that is greater than that ascribed to many other regions (Christensen et al. 2007).
Climate change projections in the polar regions, particularly the Arctic, are sensitive to changes in sea-ice extent due to the ice-albedo feedback mechanism (Ridley et al. 2008). Projections for the twenty-first century suggest a pattern of major declines in summer sea-ice extent around the Arctic and, to a lesser degree, declines in winter extent. Projected changes to sea-ice thickness are largely unconstrained. As discussed, the restrictions in modelling capabilities in the Arctic and Antarctic regions result in a wide spread of possible futures.
In 2007, the IPCC AR4 concluded that the Arctic (north of 60° N) is very likely (greater than 90% probability) to warm over the twenty-first century, with the annual mean warming very likely to exceed the global mean warming, with some seasonal variation. Alongside rises in precipitation of up to 28 per cent, temperature increases in this region under the A1B scenario could be between 2.8°C and 7.8°C, with a median of 4.9°C, by 2080–2099, relative to 1980–1999 (Christensen et al. 2007). Similar temperature projections under A1B for Alaska and northern Asia (including Kamchatka) are 3.0–7.4°C, with a median of 4.5°C, and 2.7–6.4°C, with a median of 4.3°C, respectively. Precipitation in these regions is projected to increase by up to one-quarter. In ‘high-end’ projections under the A2 scenario, projections of global warming of 4°C or more suggest that surface temperatures in much of the Arctic could rise by 15°C by the 2090s (Sanderson et al. submitted).
Not only is warming intimately related to the sea-ice cover in the region, but it would also have a direct impact on the extent, depth and timing of snow cover across the Arctic, and glacial mass in the mountainous areas of Alaska and Kamchatka, in addition to influencing the thaw depth of permafrost areas. Stendel & Christensen (2002) project the thickness of the permafrost active layer to increase by 30–40% across the Northern Hemisphere towards the end of the twenty-first century, with the extent of permafrost retreating polewards. However, modelling the large-scale responses of permafrost to climate change remains a considerable challenge.
When considering ice-mass loss in Greenland, recent observations and studies have reinforced the findings of IPCC AR4. It stated that the margins of the ice sheet are thinning and that, despite inland increases in mass, the overall balance of the ice sheet is negative and is therefore contributing to current, and future, sea-level rise (Lemke et al. 2007; Allison et al. 2009). Antarctica is also undergoing some ice-mass loss in the West Antarctic ice sheet, although there is significant variation across the continent and the much larger East Antarctic ice sheet is actually increasing in mass (Lemke et al. 2007). Throughout the twenty-first century, the projected changes in precipitation in Antarctica (from −2% to +35%) and the increases in temperature (1.4–5.0°C) and sea-level rise may further influence the disintegration of ice shelves and the contribution of the West Antarctic ice sheet to rising sea levels.
(c) Mountain regions
Mountain regions susceptible to geospheric responses to climate change include the European and New Zealand Alpine ranges, the Pyrenees, Caucasus, Andes and the Himalayas.
In all of these regions, mean temperatures are projected to rise, with extreme precipitation and temperature events increasing in both magnitude and frequency. In southern Europe (as defined by Christensen et al. 2007) including the mountainous regions of the Alps and Pyrenees, annual mean temperatures are projected to rise by between 2.2°C and 5.1°C (median of 3.5°C) by the end of the twenty-first century under the A1B scenario. The greatest increases in seasonal mean temperature are projected to occur in summer; this is a key point because the impacts of warming may be more important in some seasons than others, particularly in relation to the seasonality of temperature fluctuations across the freezing point and consequent effects on glacier mass balance and freeze–thaw weakening of slopes. The New Zealand Alps are projected to see smaller rises in average temperature, whereas increases in the high-altitude regions of Asia are projected to be comparable to those of southern Europe (Christensen et al. 2007). In the ‘high-end’ subset of simulations with the A2 scenario, the multi-model average warming projected over the Himalayas reached 8°C or more in December–January–February, and 7°C in June–July–August (Sanderson et al. submitted).
A warming climate is projected to influence precipitation patterns across these mountainous regions. Most global climate models and RCMs are not, however, of sufficiently high resolution to represent the complex topography of these regions, leading to inadequate representation of contemporary precipitation; consequently, many regional projections for the twenty-first century are often regarded as unreliable for use in detailed impact investigations. Nevertheless, analysis of large-scale circulation patterns within these climate models can inform general precipitation projections. Within the European Alps, studies (e.g. Christensen & Christensen 2007; Déqué et al. 2007) have shown that precipitation is expected to intensify over the summer months. Annually, however, precipitation in this region, in common with New Zealand, is projected to decrease. Across Europe, the snow season could be shortened, with a 50–100% decrease in snow depth by 2100, depending upon altitude, and a shift in precipitation from snow to rain as temperatures increase. This could be accompanied by a rise of the snow line by an average of 150 m for every degree Celsius increase (Christensen et al. 2007). Glaciers are also expected to retreat in many regions, such as the Himalayas. However, the snow pack and ice mass of higher-altitude regions and some of the very cold, northern areas of Scandinavia and Russia could be less sensitive to rising temperatures than the lower latitudes and altitudes of Europe where temperatures are already generally nearer the freezing point.
(d) Volcanic landscapes
Both glaciated and non-glaciated volcanic landscapes are susceptible to the aforementioned global and regional projections of climate change, most notably as a consequence of temperature rises and changes in extreme precipitation. The Cascade Range volcanoes of western North America, for example, could see increases in annual mean temperature of 2.1–5.7°C (with a median of 3.4°C), together with a −3% to +14% change in precipitation. For the Andes, warming by the end of the century could reach 4.5–5°C (Vuille et al. 2008), under a SRES A2 emission scenario. As noted by McGuire (2010), close to 60 per cent of all active volcanoes are coastally located or form islands, while most of the balance occurs within 250 km of the coast. As discussed earlier, this means that most active volcanic systems have the potential to be influenced by changes in crustal stress and strain associated with ocean loading caused by future sea-level rise. The disposition of tectonic plates ensures a non-random distribution of active volcanoes, with large concentrations at high latitudes (e.g. Alaska, Kamchatka and Iceland) and in the tropics (including Indonesia, Papua New Guinea and the Philippines).
High-latitude volcanoes are often glaciated and consequently susceptible to many of the same climate-change-driven, deglaciation-related, hazards as mountainous terrains of non-volcanic origin.
Non-glaciated high-relief volcanic regions, including in the Caribbean, Europe, Indonesia, the Philippines and Japan, could also be affected by climate change, principally due to modified precipitation patterns and especially a rise in the frequency and magnitude of severe rainfall events. Of particular concern in many of these regions could be changes to the frequency and intensity of tropical cyclones. Following the active hurricane seasons in the North Atlantic in 2004 and 2005, the World Meteorological Organisation (WMO) made a statement that ‘No individual tropical cyclone can be directly attributed to climate change’ (IWTC 2006). As an active area of research, a number of studies have been conducted in recent years examining the influence of the warming climate on tropical cyclones around the world (Shepherd & Knutson 2007). Within the IPCC’s AR4, the conclusion was reached from a study of coarse-resolution climate models (50–100 km grid spacing) alongside finer-scale projections (down to approx. 9 km grid spacing) that climate change could lead to future tropical cyclones having increased peak wind intensities and more intense precipitation, particularly towards the centres of the storms (Meehl et al. 2007). However, many of these studies also show fewer tropical storms occurring globally, and with varying degrees of change to cyclone intensities (Sugi et al. 2002; McDonald et al. 2005; Chauvin et al. 2006; Oouchi et al. 2006; Yoshimura et al. 2006; Bengtsson et al. 2007; Caron & Jones 2008).
On palaeoclimate time scales, enhanced levels of geological and geomorphological activity have been linked to climatic factors, including examples of processes that are expected to be important in current and future anthropogenic climate change. Planetary warming leading to increased rainfall, ice-mass loss and rising sea levels is potentially relevant to geospheric responses in many geologically diverse regions. Anthropogenic climate change therefore has the potential to alter the risk of geological and geomorphological hazards through the twenty-first century and beyond. Such changes in risk have not yet been systematically assessed.
An appropriate application of climate model projections over a range of time scales, combined with an understanding of geohazards, can go some way towards improving understanding of these changing risks. This requires quantification of possible changes in relevant climate quantities in the regions of interest, considering both the potential climate changes under unmitigated emissions that could still be avoided, and the residual climate changes that would not be avoidable by mitigation and to which adaptation will therefore be necessary. Within these two categories, there are large uncertainties in global and regional climate responses.
Scenarios of unmitigated climate change project global warming of between 1.6°C and 6.9°C by 2100 relative to pre-industrial level. If GHG concentrations are stabilized in the lower part of this range by the end of this century, warming would continue by approximately a further 0.4–0.8°C by 2300 for low- and medium-emission scenarios—the ongoing commitment to warming has not been assessed for higher warming scenarios, but the warming commitment appears to increase with the temperature at which stabilization occurs. Political aims of keeping global warming below 2°C above pre-industrial level may be achievable if global emissions peak during the 2010s and decline by several per cent per year thereafter, but longer delays in peaking emissions further decrease the likelihood of achieving this ambition.
Regional climate changes have only been assessed in detail for a limited number of projections that do not cover the full range of projected global changes, but certain key messages can still be drawn from the available studies. For example, much of the ocean surface is expected to warm by less than the global average, although the Arctic Ocean is expected to warm by more than the global average. Over land, most regions are projected to warm by more than the global average.
Under the unmitigated scenarios, global mean sea level is projected by IPCC to rise by 0.18–0.59 m by 2100 relative to present day, including thermal expansion and ice loss from Greenland and Antarctica but without future rapid dynamical changes in ice flow. IPCC recognized high uncertainty in dynamic ice processes and did not provide a best estimate or upper bound on sea-level rise projections—more recent work has attempted to quantify the implications of rapid ice flow changes, but projections of global sea-level rise of up to 2 m by 2100 remain controversial. Future committed sea-level rise is already significant and stabilization of GHG concentrations by 2100 could lead to further committed rise for the next millennium. A significant factor in this committed rise would be whether major ice sheets become committed to irreversible decline.
Future precipitation changes are subject to even higher uncertainty, and, in the case of annual and seasonal mean precipitation, even the sign of the change cannot be confidently given in many cases. The majority of climate models agree on projections of increased precipitation in the high latitudes, but projections in other regions vary in sign. However, precipitation events are expected to become more intense.
When assessing the potential for geospheric responses to climate change, it appears prudent to consider regional levels of warming of 2°C above pre-industrial level as being potentially unavoidable as an influence on processes requiring a human adaptation response within this century. At the other end of the scale when considering changes that could be avoided by reduction of emissions, scenarios of unmitigated warming exceeding 4°C in the global average include much greater local warming in some regions, such as 8°C in the Himalayas and 15°C in the Arctic. However, considerable further work is required to better understand the uncertainties associated with these projections, uncertainties inherent not only in the climate modelling but also in the linkages between climate change and geospheric responses.
The work of RAB and FL was supported by the Joint DECC and Defra Integrated Climate Programme—DECC/Defra (GA01101)
One contribution of 15 to a Theme Issue ‘Climate forcing of geological and geomorphological hazards’.
- © 2010 The Royal Society