20.6.2 Spatially-explicit methods: global impacts of climate change
Warren (2006) and Hitz and Smith (2004) observe that most impact assessments are conducted at the local scale. It is therefore extremely difficult to estimate impacts across the global domain from these localised studies. A small number of studies have used geographically-distributed impacts models to estimate the impacts of climate change across the global domain. The “Fast Track” studies (Arnell, 2004; Nicholls, 2004; Arnell et al., 2002; Levy et al., 2004; Parry et al., 2004; Van Lieshout et al., 2004) used a con
sistent set of scenarios and assumptions to estimate the effects of scenarios based on the HadCM3 climate model on water resource availability, food security, coastal flood risk, ecosystem change and exposure to malaria. Schroeter et al. (2005) used a similar approach in the ATEAM project to tabulate impacts across Europe using scenarios constructed from a larger number of climate models. Both these sets of studies used a wide range of metrics that varied across sectors. Table 20.4 summarises some of the global-scale impacts of defined climate-change scenarios. Although the precise numbers depend on the climate model used and some key assumptions (particularly the effect of increased CO2 concentrations on crop productivity), it is clear that the future impacts of climate change are dependent not only on the rate of climate change, but also on the future social, economic and technological state of the world. Impacts are greatest under an A2 world, for example, not because the climate change is greatest but because there are more people to be impacted. Impacts also vary regionally and Table 20.5 summarises impacts by major world region. The assumed effect of CO2 enrichment on crop productivity has a major effect on estimated changes in population at risk of hunger (Chapter 5, Section 5.4.7).
Table 20.4. Global-scale impacts of climate change by 2080.
| ||Climate and socio-economic scenario |
| ||A1FI ||A2 ||B1 ||B2 |
|Global temperature change (°C difference from the 1961-1990 period) ||3.97 ||3.21 to 3.32 ||2.06 ||2.34 to 2.4 |
|Millions of people at increased risk of hunger (Parry et al., 2004); no CO2 effect ||263 ||551 ||34 ||151 |
|Millions of people at increased risk of hunger (Parry et al., 2004); with maximum direct CO2 effect ||28 ||-28 to -8 ||12 ||-12 to +5 |
|Millions of people exposed to increased water resources stress (Arnell, 2004) ||1256 ||2583 to 3210 ||1135 ||1196 to 1535 |
|Additional numbers of people (millions) flooded in coastal floods each year, with lagged evolving protection (Nicholls, 2004) ||7 ||29 ||2 ||16 |
Table 20.5. Regional-scale impacts of climate change by 2080 (millions of people).
| ||Population living in watersheds with an increase in water- resources stress (Arnell, 2004) ||Increase in average annual number of coastal flood victims (Nicholls, 2004) ||Additional population at risk of hunger (Parry et al., 2004)1 Figures in brackets assume maximum direct CO2-enrichment effect |
| ||Climate and socio-economic scenario: |
| ||A1 ||A2 ||B1 ||B2 ||A1 ||A2 ||B1 ||B2 ||A1 ||A2 ||B1 ||B2 |
|Europe ||270 ||382-493 ||233 ||172-183 ||1.6 ||0.3 ||0.2 ||0.3 ||0 ||0 ||0 ||0 |
|Asia ||289 ||812-1197 ||302 ||327-608 ||1.3 ||14.7 ||0.5 ||1.4 ||78 (6) ||266 (-21) ||7 (2) ||47 (-3) |
|North America ||127 ||110-145 ||107 ||9-63 ||0.1 ||0.1 ||0 ||0 ||0 ||0 ||0 ||0 |
|South America ||163 ||430-469 ||97 ||130-186 ||0.6 ||0.4 ||0 ||0.1 ||27 (1) ||85 (-4) ||5 (2) ||15 (-1) |
|Africa ||408 ||691-909 ||397 ||492-559 ||2.8 ||12.8 ||0.6 ||13.6 ||157 (21) ||200 (-2) ||23 (8) ||89 (-8) |
|Australasia ||0 ||0 ||0 ||0 ||0 ||0 ||0 ||0 ||0 ||0 ||0 ||0 |
Table 20.6 compares the global impacts of a 1% annual increase in CO2 concentrations (i.e., the IS92a scenario, see IPCC, 1992) with the impacts of emissions trajectories stabilising at 750 (S750) and 550 (S550) ppm (Arnell et al., 2002). The results are not directly comparable to those reported in Table 20.4, because different population assumptions, methodologies and indicators were employed in their preparation. Nevertheless, the results suggest that aiming for stabilisation at 750 ppm has a relatively small effect on impacts in most sectors in comparison with 550 ppm stabilisation. The S550 pathway has a greater apparent impact on exposure to hunger because higher CO2 concentrations under S750 result in a greater increase in crop productivity (but again, note that CO2-enrichment effects are highly uncertain).
Table 20.6. Global-scale impacts under unmitigated and stabilisation pathways. Source: Arnell et al., 2002.
| ||2050 Scenario: ||2050 Scenario: |
| ||Unmitigated ||S750 ||S550 ||Unmitigated ||S750 ||S550 |
|Approximate equivalent CO2 concentration (ppm) ||520 ||485 ||458 ||630 ||565 ||493 |
|Approximate global temperature change (°C difference from 1961 to 1990) ||2.0 ||1.3 ||1.1 ||2.9 ||1.7 ||1.2 |
|Area potentially experiencing vegetation dieback (million km2) ||1.5 to 2.7 ||2 ||0.7 ||6.2 to 8 ||3.5 ||1.3 |
|Millions of people exposed to increased water stress ||200 to 3200 ||2100 ||1700 ||2830 to 3440 ||2920 ||760 |
|Additional people flooded in coastal floods (millions/year) ||20 ||13 ||10 ||79 to 81 ||21 ||5 |
|Population at increased risk of hunger (millions) ||-3 to 9 ||7 ||5 ||69 to 91 ||16 ||43 |
Each of these tables present indicators of impact which ignore adaptations that will occur over time. They can therefore be seen as indicative of the challenge to be overcome by adaptations to offset some of the impacts of climate change. Incorporating adaptation into global-scale assessments of the impacts of climate change is currently difficult for a number of reasons (including diversity of circumstances, diversity of potential objectives of adaptation, diversity of ways of meeting adaptation objectives and uncertainty over the effectiveness of adaptation options) and remains an area where more research is needed.
Aggregation of impacts to regional and global scales is another key problem with such geographically-distributed impact assessments. Tables 20.4 to 20.6, for example, keep track of people living in watersheds who will face increased water-related stress. Of course, many people live in watersheds where climate change increases runoff and therefore may apparently see reduced water-related stress (if they see increased risk of flooding). Simply calculating the ‘net’ impact of climate change, however, is complicated, particularly where ‘winners’ and ‘losers’ live in different geographic regions, or where ‘costs’ and ‘benefits’ are not symmetrical. Watersheds with an increase in runoff, for example, are concentrated in east Asia, while watersheds with reduced runoff are much more widely distributed. Similarly, the adverse effects felt by 100 million people exposed to increased water stress could easily outweigh the ‘benefits’ of 100 million people with reduced stress.
The Defra Fast Track and ATEAM studies both describe impacts along defined scenarios, so it is difficult to infer the effects of different rates or degrees of climate change on different socio-economic worlds. A more generalised approach applies a wide range of climate scenarios representing different rates of change to estimate impacts for specific socio-economic contexts. Leemans and Eickhout (2004), for example, show that most species, ecosystems and landscapes would be impacted by increases of global temperature between 1 and 2°C above 2000 levels. Arnell (2006) showed that an increase in temperature of 2°C above the 1961 to 1990 mean by 2050 would result in between 550 and 900 million people suffering an increase in water-related stress in both the SRES (Special Report on Emissions Scenarios, Nakićenović and Swart, 2000) A1 and B1 worlds. In this case, the range between estimates represents the effect of different changes in rainfall patterns for a 2°C warming.