The impacts of climate change across the globe: A multi-sectoral assessment
- 8.1k Downloads
- 35 Citations
Abstract
The overall global-scale consequences of climate change are dependent on the distribution of impacts across regions, and there are multiple dimensions to these impacts. This paper presents a global assessment of the potential impacts of climate change across several sectors, using a harmonised set of impacts models forced by the same climate and socio-economic scenarios. Indicators of impact cover the water resources, river and coastal flooding, agriculture, natural environment and built environment sectors. Impacts are assessed under four SRES socio-economic and emissions scenarios, and the effects of uncertainty in the projected pattern of climate change are incorporated by constructing climate scenarios from 21 global climate models. There is considerable uncertainty in projected regional impacts across the climate model scenarios, and coherent assessments of impacts across sectors and regions therefore must be based on each model pattern separately; using ensemble means, for example, reduces variability between sectors and indicators. An example narrative assessment is presented in the paper. Under this narrative approximately 1 billion people would be exposed to increased water resources stress, around 450 million people exposed to increased river flooding, and 1.3 million extra people would be flooded in coastal floods each year. Crop productivity would fall in most regions, and residential energy demands would be reduced in most regions because reduced heating demands would offset higher cooling demands. Most of the global impacts on water stress and flooding would be in Asia, but the proportional impacts in the Middle East North Africa region would be larger. By 2050 there are emerging differences in impact between different emissions and socio-economic scenarios even though the changes in temperature and sea level are similar, and these differences are greater in 2080. However, for all the indicators, the range in projected impacts between different climate models is considerably greater than the range between emissions and socio-economic scenarios.
Keywords
Flood Risk Coastal Flood Cool Energy Demand Coastal Flood Risk Climate Model Scenario1 Introduction
The assessment reports of the Intergovernmental Panel on Climate Change (IPCC) review hundreds of studies into the potential impacts of climate change (e.g. IPCC 2007, 2014). Two key conclusions can be drawn from these assessments. First, the distribution of impacts across space and between regions is as relevant as the global aggregate impact when assessing the global-scale impacts of climate change; the distribution of impacts is highlighted in IPCC reports as one of the five integrative ‘reasons for concern’ about climate change alongside aggregate impacts. Second, impacts occur across many dimensions of the environment, economy and society and therefore need to be expressed in terms of multiple indicators. However, there have still so far been few consistent studies of the impact of climate change across sectors and the global domain. Most global studies have concentrated on one sector, and different studies have used different climate and socio-economic scenarios. The few multi-sectoral studies (Hayashi et al. 2010; van Vuuren et al. 2011; Piontek et al. 2014) have used few climate models and a small number of indicators. It has therefore been difficult to produce consistent assessments not only of the global-scale impacts of climate change, but also of the potential for multiple impacts across several sectors. Such assessments are of value not only to global-scale reviews of the potential consequences of climate change, but also to organisations concerned with the distribution of impacts across space. These include development, disaster management and security agencies, together with businesses or organisations with international coverage or supply chains.
This paper presents for the first time an assessment of the multi-dimensional impacts of climate change across the global domain for a wide range of sectors and indicators, using consistent climate and socio-economic scenarios and a harmonised methodology. Impacts are estimated under four different future world scenarios using up to 21 different climate model patterns to characterise the spatial pattern of climate change. The assessment constructs a set of coherent narratives of impact across regions and sectors, and also includes a representation of some of the major sources of uncertainty in potential regional impacts. It complements other global-scale assessments that used the same methodology and models to identify the relationship between amount of climate forcing and impact (Arnell et al. 2014) and the impacts avoided by climate mitigation policy (Arnell et al. 2013).
2 Methodology
2.1 Overview of the approach
The assessment involves the application of a suite of spatially-explicit impacts models run with scenarios describing a range of emissions and socio-economic futures. These emissions and socio-economic futures are here represented by the A1b, A2, B1 and B2 SRES storylines (IPCC Intergovernmental Panel on Climate Change 2000). Scenarios characterising the spatial and seasonal distribution of changes in climate and sea level around 2020, 2050 and 2080 are constructed from up to 21 global climate models (Meehl et al. 2007a) in order to assess the climate-driven uncertainty in the projected impacts for a given future. The period 1961-1990 is used as the climate baseline.
Summary of the impact indicators
| Indicator | Description | Drivers of change | Further details |
|---|---|---|---|
| Water | |||
| Population exposed to a change in water resources stress | A change in exposure to stress occurs where runoff in water-stressed watersheds changes significantly, or watersheds move into or out of the stressed class. Water-stressed watersheds have less than 1000 m3/capita/year. Runoff is estimated using the MacPDM.09 hydrological model. | Change in runoff due to climate change Change in population | Gosling and Arnell (2013) |
| River flooding | |||
| Flood-prone population exposed to a substantial change in frequency of flooding | A substantial change occurs when the frequency of the baseline 20-year flood doubles or halves. River flows are estimated using the MacPDM.09 hydrological model. | Change in runoff due to climate change Change in population | Arnell and Gosling (2014) |
| Coastal | |||
| Change in coastal wetland extent | Calculated using DIVA v2.04 | Change in relative sea level rise | Brown et al. (2013) |
| Additional average annual number of people flooded from extreme water levels | Calculated using DIVA v2.04. It is assumed that the level of coastal protection increases with population and wealth | Change in relative sea level rise Change in population Change in income | Brown et al. (2013) |
| Agriculture | |||
| Cropland exposed to change in crop suitability | A substantial change occurs where a crop suitability index changes by more than 5 % | Change in climate | Index defined in Ramankutty et al. (2002) |
| Change in spring wheat, soybean and maize productivity | Productivity is simulated using GLAM. Adaptation is incorporated by selecting the variety (from three) with the greatest productivity and varying planting dates with climate | Change in climate Change in CO2 concentration | Osborne et al. (2012) |
| Environment | |||
| Proportion of (non-cropped) region with a substantial change in Net Primary Productivity (NPP) | Calculated using JULES/IMOGEN. A substantial change is greater than 10 %. | Change in climate Change in CO2 concentration | Model summarised in Huntingford et al. (2010) |
| Change in total regional forest extent | Calculated using JULES/IMOGEN. Change in area under forest plant function types. | Change in climate Change in CO2 concentration | Model summarised in Huntingford et al. (2010) |
| Change in soil organic carbon (SOC) in mineral soils | Calculated using RothC, and aggregated over all land cover types. | Change in climate Change in CO2 concentration | Gottschalk et al. (2012) |
| Infrastructure | |||
| Change in regional residential heating and cooling energy demands | Energy requirements are based on heating and cooling degree days, population size and assumptions about heating and cooling technologies | Change in climate Change in population Change in income Change in energy efficiency | Model based on Isaac and van Vuuren (2009) |
The socio-economic impacts of climate change in a given year are expressed relative to the situation in that year in the absence of climate change (i.e. assuming that the climate remains the same as over the baseline period 1961-1990). For the ‘pure’ biophysical indicators—crop productivity, suitability of land for cropping, terrestrial ecosystems and soil organic carbon—impacts are compared with the 1961-1990 baseline. Impacts are presented at the regional scale (Supplementary Table 1).
Most of the indicators represent change in some measurable impact of climate change, such as the average annual number of people flooded in coastal floods or crop productivity. Three of the indicators (water scarcity, river flooding and crop suitability), however, represent change in exposure to impact. The extent to which exposure translates into impact depends on the water management and agricultural practices in place, but these are so locally diverse and dependent on local context that it is currently not feasible to represent them numerically in global-scale impacts models. The indicators do not incorporate the effects of adaptation to climate change, with the exception of crop productivity where the crop variety planted varies with climate (see Supplementary Information).
Impacts can be expressed in either absolute or relative terms, and there are advantages and disadvantages in both when comparing impacts across regions. Large percentage impacts in a region may represent small absolute numbers and therefore make a small contribution to the global impact, but may indicate substantial impacts in the region itself. In contrast, a small percentage impact in another region may produce large absolute impact—and thus contribute substantially to the global total—but the implications for the region itself may be smaller. Most of the impacts in this paper are expressed in absolute terms, but relative changes can be calculated from the data in the tables.
The distribution of impacts between regions and across sectors varies with different spatial patterns of change in climate, as represented by different climate models. One possible way of summarising the global and regional impacts of climate change would be to show the ensemble mean (or median) impact for a given sector and region across all climate model patterns, perhaps with some representation of uncertainty through identifying consistency between the different models (as is often done for climatic indicators such as temperature and precipitation). However, this is problematic when the concern is with multiple indicators of impact and comparisons between regions for two main reasons. The calculation of an ensemble mean makes assumptions about the relative plausibility of different climate models, but more importantly the ensemble mean impact does not necessarily represent a plausible future world. Calculating the average reduces the variability between regions and the relationships between sectors and indicators.
An alternative approach is therefore to treat each climate model as the basis for a separate narrative or story, describing a plausible future world with its associations between indicators and regions. Uncertainty in potential impacts is then characterised for each region and indicator by comparing the range in impacts across different climate models, but recognising that aggregated uncertainty—across regions or indicators—is not equivalent to the sum of the individual uncertainty ranges.
2.2 Climate and sea level rise scenarios
Climate scenarios were constructed (Osborn et al. 2014) by pattern-scaling output from 21 of the climate models in the CMIP3 set (Meehl et al. 2007a: Supplementary Table 2) to match the changes in global mean temperature projected under the four SRES emissions scenarios A1b, A2, B1 and B2. These global temperature changes were estimated using the MAGICC4.2 simple climate model with parameters appropriate to each climate model (Meehl et al. 2007b: Supplementary Fig. 1a). Pattern-scaling was used rather than simply constructing climate scenarios directly from climate model output partly to better separate out the effects of underlying climate change and internal climatic variability, and partly to allow scenarios to be constructed for all combinations of climate model and emissions scenario. Rescaled changes in mean monthly climate variables (and year to year variability in monthly precipitation) were applied to the CRU TS3.0 0.5×0.5o 1961-1990 climatology (Harris et al. 2014) using the delta method to create perturbed 30-year time series representing conditions around 2020, 2050 and 2080 (Osborn et al. 2014). The terrestrial ecosystem and soil carbon impact models require transient climate scenarios, and these were produced by repeating the CRU 1961-1990 time series and rescaling to construct time series from 1991 to 2100 using gradually increasing global mean temperatures (Osborn et al. 2014). Pattern-scaling makes assumptions about the relationship between rate of forcing and the spatial pattern of change, which have been demonstrated to be broadly appropriate for the averaged climate indicators used here (e.g. Tebaldi and Arblaster 2014), but which do constitute caveats to the quantitative interpretation of results.
Sea level rise scenarios were constructed for 17 climate models. Spatial patterns of change in sea level due to thermal expansion were available for 11 of the models, and for the other six globally-uniform thermal expansion scenarios were calculated using MAGICC4.2. Uniform projections of the contributions of ice melt were added to these patterns, and the patterns were rescaled to correspond to specific global temperature changes using the same methods as applied in Meehl et al. (2007b). Ice melt contributions from Greenland and Antarctica, as well as ice caps and glaciers were calculated following the methodology of Meehl et al. (2007b), with additional data to calculate ice sheet dynamics from Gregory and Huybrechts (2006) (see Brown et al. 2013). Global average sea level rise scenarios are shown in Supplementary Fig. 1b; note that the highest change is produced by one model which is considerably higher—by around 20 cm in 2100—than the others. The effects of changes in the Greenland and Antarctic ice sheet dynamics are not incorporated, but the range in sea level rise across the models is large compared with the potential magnitude of the dynamic effect.
2.3 Socio-economic scenarios
Future population and gross domestic product at a spatial resolution of 0.5×0.5o were taken from the IMAGE v2.3 representation of the SRES storylines (van Vuuren et al. 2007). The population living in inland river floodplains was estimated by combining high resolution gridded population data for 2000 (Center for International Earth Science Information Network CIESIN 2004) with flood-prone areas defined in the UN PREVIEW Global Risk Data Platform to estimate the proportions of grid cell population currently living in flood-prone areas. Cropland extent was taken from Ramankutty et al. (2008). It is assumed that river floodplain extent, cropland extent and the proportion of grid cell population living in floodplains do not change over time.
3 Exposure in the absence of climate change
Regional and global exposure to impact in 2050 in the absence of climate change
| Total population (millon) | Water-stressed population (million) | Flood-prone people (million) | Cropland (thousand km2) | Average spring wheat yield (kg/ha) | Average soybean yield (kg/ha) | Average maize yield (kg/ha) | Total Soil Organic Carbon (SOC) (Pg C) | Regional average NPP (kg C m-2 years-1) | Regional forest area (thousand km2) | Coastal wetland (thousand km2) | Average annual people flooded in coastal floods (thousand/year) | Heat energy demand (PJ) | Cool energy demand (PJ) | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| W. Africa | 454 | 42 | 33 | 832 | 793 | 119 | 584 | 13 | 1.0 | 1,166 | 42 | 5 | 0 | 12 |
| C. Africa | 179 | 5 | 12 | 216 | 525 | 518 | 910 | 20 | 1.2 | 3,445 | 10 | 2 | 7 | 4 |
| E. Africa | 314 | 97 | 13 | 268 | 891 | 702 | 1,191 | 9 | 0.8 | 432 | 4 | 2 | 103 | 1 |
| Sn Africa | 230 | 11 | 14 | 422 | 979 | 1,484 | 1,460 | 23 | 0.8 | 2,651 | 33 | 101 | 128 | 22 |
| S. Asia | 2,085 | 1,466 | 357 | 2,238 | 757 | 699 | 805 | 17 | 0.8 | 363 | 38 | 94 | 2,588 | 420 |
| SE Asia | 724 | 0 | 103 | 980 | 732 | 579 | 1,341 | 18 | 1.4 | 2,579 | 114 | 241 | 24 | 2,331 |
| E Asia | 1,533 | 673 | 147 | 1,473 | 1,775 | 1,545 | 2,995 | 39 | 1.0 | 601 | 10 | 37 | 16,130 | 1,769 |
| Central Asia | 86 | 2 | 7 | 311 | 665 | 477 | 2,548 | 11 | 0.2 | 6 | n/a | n/a | 1,086 | 24 |
| Australasia | 47 | 0 | 2 | 308 | 1,642 | 1,284 | 2,914 | 19 | 0.8 | 639 | 156 | 19 | 212 | 59 |
| N. Africa | 266 | 206 | 21 | 363 | 1,596 | 400 | 467 | 8 | 0.8 | 144 | 6 | 5 | 575 | 24 |
| W. Asia | 350 | 236 | 9 | 362 | 1,136 | n/a | 2,654 | 4 | 0.3 | 29 | 5 | 5 | 1,427 | 411 |
| W. Europe | 422 | 160 | 32 | 773 | 3,102 | 2,018 | 4,427 | 19 | 0.8 | 872 | 23 | 7 | 6,706 | 196 |
| C. Europe | 118 | 6 | 12 | 504 | 2,432 | 885 | 2,651 | 9 | 0.9 | 110 | 4 | 0 | 1,803 | 8 |
| E. Europe | 202 | 5 | 23 | 1,688 | 1,145 | 1,001 | 2,519 | 123 | 0.5 | 3,163 | 7 | 27 | 4,023 | 16 |
| Canada | 40 | 7 | 1 | 402 | 922 | 2,000 | 2,826 | 59 | 0.4 | 2,884 | 69 | 22 | 1,295 | 8 |
| US | 404 | 76 | 10 | 1,770 | 995 | 1,171 | 2,454 | 39 | 0.8 | 1,931 | 156 | 17 | 5,167 | 593 |
| Meso-America | 251 | 53 | 10 | 485 | 1,239 | 1,167 | 1,417 | 9 | 1.0 | 663 | 69 | 2 | 360 | 379 |
| Brazil | 224 | 0 | 16 | 490 | 1,835 | 1,776 | 2,616 | 34 | 1.1 | 6,013 | 54 | 10 | 79 | 722 |
| South America | 266 | 18 | 21 | 561 | 1,386 | 1,527 | 2,501 | 39 | 1.1 | 3,693 | 58 | 11 | 1,007 | 373 |
| Global (A1b) | 8,196 | 3,064 | 843 | 14,447 | 1,493 | 1,346 | 2,204 | 513 | 0.8 | 31,383 | 857 | 606 | 42,716 | 7,375 |
| Global (A2) | 10,387 | 4,792 | 1,083 | 2,800 | 38,876 | 2,083 | ||||||||
| Global (B1) | 8,196 | 3,064 | 843 | 910 | 40,719 | 4,696 | ||||||||
| Global (B2) | 9,021 | 3,652 | 935 | 1,150 | 40,297 | 3,380 | ||||||||
| Global (2000) | 6,122 | 1,555 | 637 | 3,100 | 30,447 | 857 | ||||||||
The vast majority of people living in water-stressed watersheds, river floodplains and flooded in coastal floods are in south and east Asia (including India, Bangladesh and China). By 2050 east Asia (predominantly China), with Europe and North America, account for the vast bulk of heating energy requirements. However, the absolute numbers hide regional variations in the proportions of people living in exposed conditions; more than 75 % of North African people would be living in water-stressed watersheds in 2050 (a slightly higher proportion than in 2000), along with two-thirds of people in west Asia (up from 35 % in 2000).
4 The regional impacts of climate change in 2050 in an A1b world
4.1 Introduction
By 2050, global average temperature under A1b emissions would be between 1.4 and 2.9 °C above the 1961-1990 mean, with an average increase across climate models of around 1.9 °C. Global average sea level would be 12 to 32 cm higher than over the period 1961-1990, with an average increase of 18 cm (note that changes in temperature under A1b are between changes under RCP6.0 and RCP8.5: IPCC 2013). However, the spatial patterns of changes in temperature, precipitation, sea level and other relevant climatic variables vary between climate models, so the projected potential impacts also vary. This section first describes the potential impacts across the world and across sectors under one example plausible climate story, and then assesses the uncertainty in impacts by region and sector.
4.2 A coherent story: Impacts under one plausible climate future
The geographic distribution of impacts under the A1b 2050 scenario: one plausible model (HadCM3). For river flood risk, white areas indicate that the grid cell floodplain population is less than 1000 people. For crop productivity, white areas indicate that the crop is not currently grown. For heating and cooling demands, white areas indicate that grid cell population is less than 10,000, light grey indicates no heating / cooling demands in either the present or the future, and magenta indicates no demand in the present but some demand in the future. For SOC and NPP, light grey denotes zero values in 2000
Regional and global impacts in 2050, under the A1b emissions and socio-economic scenario, under one plausible climate story. The socio-economic impacts are relative to the situation in 2050 in the absence of climate change (Table 2)
| Pop. exposed to increased water resources stress (millons) | Pop. with decreased water resources stress (millions) | Pop. exposed to doubled flood freq. (millions) | Pop. exposed to halved flood freq. (millions) | Cropland with decline in suitability (thousand km2) | Croplandwith increase in suitability (thousand km2) | Change in average spring wheatyield (%) | Change in average soybean yield (%) | Change in average maize yield(%) | Change in Soil Organic Carbon(SOC) content (%) | Change in regional NPP (%) | Change in forest area (%) | Change in coastal wetland (%) | Additional people flooded in coastal floods (thousands/year) | Change in heat energy demand (%) | Change in cool energy demand (%) | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| W. Africa | 26 | 16 | 18 | 3 | 552 | 0 | 40 | −15 | −32 | 2 | 22 | 6 | −7 | 29 | −33 | 35 |
| C. Africa | 5 | 0 | 5 | 1 | 68 | 4 | −30 | −44 | −42 | 1 | 13 | 2 | −13 | 4 | −88 | 37 |
| E. Africa | 79 | 12 | 4 | 1 | 102 | 14 | −17 | −40 | −14 | 3 | 34 | 6 | −12 | 6 | −85 | 93 |
| Sn Africa | 17 | 0 | 1 | 3 | 380 | 0 | −25 | −23 | −33 | 8 | 19 | 4 | −17 | 312 | −70 | 139 |
| S. Asia | 188 | 1,209 | 290 | 5 | 1,077 | 436 | 28 | 23 | −16 | −3 | 22 | 7 | −10 | 132 | −43 | 38 |
| SE Asia | 0 | 0 | 32 | 2 | 78 | 0 | 31 | −15 | −30 | 8 | 20 | 3 | −12 | 406 | −87 | 31 |
| E Asia | 0 | 636 | 80 | 0 | 130 | 419 | 7 | 1 | −16 | −3 | 27 | 5 | −22 | 222 | −28 | 97 |
| Central Asia | 3 | 0 | 0 | 0 | 289 | 20 | −3 | 99 | −31 | 8 | 9 | 12 | n/a | n/a | −24 | 127 |
| Australasia | 0 | 0 | 1 | 0 | 278 | 2 | −25 | −13 | −31 | 2 | 21 | 6 | −12 | 65 | −39 | 95 |
| N. Africa | 117 | 0 | 3 | 6 | 292 | 16 | −39 | −39 | −43 | 3 | 27 | 5 | −21 | 14 | −53 | 78 |
| W. Asia | 185 | 0 | 0 | 5 | 349 | 3 | −17 | n/a | −21 | 3 | 4 | 0 | −22 | 39 | −35 | 54 |
| W. Europe | 192 | 0 | 1 | 8 | 448 | 214 | 2 | −12 | −18 | 3 | 17 | 7 | −17 | 17 | −27 | 178 |
| C. Europe | 9 | 0 | 0 | 6 | 249 | 145 | −14 | 38 | −12 | 5 | 7 | 4 | −20 | 2 | −26 | 821 |
| E. Europe | 14 | 0 | 1 | 16 | 1,110 | 435 | −15 | 33 | −14 | −1 | 31 | 11 | −19 | 5 | −23 | 519 |
| Canada | 7 | 0 | 0 | 0 | 45 | 286 | −3 | 5 | −13 | −2 | 36 | 8 | −6 | 5 | −21 | 365 |
| US | 83 | 6 | 1 | 2 | 1,068 | 44 | −14 | −8 | −20 | 5 | 21 | 5 | −24 | 4 | −28 | 134 |
| Meso-America | 56 | 1 | 0 | 7 | 354 | 0 | −9 | −25 | −30 | 5 | −3 | 2 | −18 | 9 | −71 | 74 |
| Brazil | 28 | 0 | 6 | 6 | 166 | 0 | −25 | −30 | −32 | 14 | −8 | −1 | −9 | 11 | −93 | 69 |
| South America | 20 | 13 | 5 | 7 | 182 | 45 | −6 | −13 | −25 | 5 | 8 | 3 | −19 | 28 | −44 | 78 |
| Global | 1,025 | 1,893 | 447 | 76 | 7,215 | 2,083 | 1 | −13 | −22 | 2 | 16 | 4 | −15 | 1,309 | −30 | 73 |
Under this plausible story, approximately 1 billion people are exposed to increased water resources stress due to climate change, relative to the situation in 2050 with no climate change, and almost 450 million people are exposed to a doubling of flood frequency. In contrast, around 1.9 billion water-stressed people see an increase in runoff, and around 75 million flood-prone people are exposed to flooding half as frequently as in the absence of climate change. Approximately 1.3 million additional people are flooded in coastal floods each year. Around a half of all cropland sees a decline in suitability, but about 15 % sees an improvement. Global residential heating energy demands are reduced by 30 % (bringing them back to approximately the 2000 level) but cooling demands rise by over 70 %. The net effect is a reduction in total heating and cooling energy demands of around 15 %. There are, however, considerable regional variations in impact.
Under this story, increases in water scarcity are most apparent in the Middle East, north Africa and western Europe, whilst increases in exposure to river flooding is largest in south and east Asia. The suitability of land for cropping declines in most regions, but increases at the northern boundary of cropland and along some margins in east Asia. Spring wheat yields show a mixed pattern of change, maize yields decline everywhere except in parts of north America and eastern China, and soybean yields tend to increase in parts of south and east Asia, north America and small parts of south America, but decrease elsewhere. Increases in coastal flood risk are concentrated in Asia and eastern Africa, whilst wetland losses focus around the Mediterranean and north America. Cooling energy demands increase particularly in regions where there is currently little demand for cooling, but increase only slightly in some warm regions—because the relative change in requirements is smaller. Heating energy demands decrease most in the warmest regions.
Many regions are exposed to multiple overlaying impacts. For example, under this plausible climate story river flood risk increases across much eastern Asia, coastal flood risk increases substantially, and cooling energy demands increase by more than 70 %. At the same time, the productivity of the three example crops increases in parts of eastern Asia, but decreases across much of northern China. The suitability for agriculture appears to increase in northern and western China, although soil organic carbon contents decline (in this case because conversion of forest to arable land reduces the inputs of carbon from vegetation).
In southern Asia, crop suitability declines, productivity of maize declines but soybean productivity increases (in some parts). River flood risk increases and some coastal megacities see increased flood risk. Cooling energy demands rise by around 30–40 %, but there is little change in heating demands. Water scarcity reduces under this story across many water-scarce parts of southern Asia.
The suitability of cropland for crop cultivation declines across much of sub-Saharan Africa, primarily due to reductions in available moisture; more than 90 % of cropland in southern Africa would see a reduction in suitability for crop production. Maize yields reduce by 20–40 %. River flood risk increases substantially in parts of western Africa, and coastal flood risk increases in particular for many east African coastal cities. Across the Middle East and North Africa crop suitability declines and large populations are exposed to increased water scarcity and increased cooling energy demands; NPP also reduces in many parts of the region.
Within western and central Europe, river flood risk is little affected under this story, but around 200 million people are exposed to increased water resources stress. Crop suitability increases in the north of the region but declines elsewhere, and spring wheat productivity declines across much of central and eastern Europe. Cooling energy demands are increased very significantly—from close to zero in northern Europe—but heating energy demands fall by at least 40 %.
Under this story, the main potential impacts in North America appear to be reductions in crop suitability across much of western and central North America, but increases at the northern margins of agriculture, and mixtures of increases and decreases in crop yields. Cooling energy demands increase very significantly in the eastern parts of North America, where heating energy demands fall. Coastal wetland loss is particularly large along the west coast.
Across South America, maize and soybean yields fall and NPP decreases substantially across the Amazon basin; the suitability for cropping declines in the drier parts of eastern south America, but increases along parts of the west coast.
Global impacts in 2050, under the A1b emissions and socio-economic scenario, under six plausible stories. Compare with the global impacts in Table 3
| Pop. exposed to increased water resources stress (millions) | Pop. with decreased water resources stress (millions) | Pop. exposed to doubled flood frequency (millions) | Pop. exposed to halved flood frequency (millions) | Cropland with decline in suitability (thousand km2) | Cropland with increase in suitability (thousand km2) | Change in average spring wheat yield (%) | Change in average soybean yield (%) | Change in average maize yield(%) | Change in Soil Organic Carbon (SOC) content (%) | Change in regional NPP (%) | Change in forest area (%) | Change in coastal wetland (%) | Additional people flooded in coastal floods (thousands/year) | Change in heat energy demand (%) | Change in cool energy demand (%) | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| HadGEM1 | 1,385 | 1,385 | 202 | 94 | 7,203 | 2,150 | 9 | 0 | −12 | 4 | 23 | 5 | −14 | 1,023 | −23 | 46 |
| ECHAM5 | 1,369 | 508 | 302 | 55 | 7,631 | 1,672 | 8 | −4 | −17 | 3 | 23 | 5 | −17 | 1,568 | −31 | 65 |
| CGCM3.1 (T47) | 746 | 1,844 | 316 | 64 | 3,963 | 4,062 | 17 | −1 | −12 | 3 | 25 | 5 | −15 | 1,215 | −25 | 53 |
| CCSM3 | 639 | 1,680 | 321 | 46 | 4,362 | 2,619 | 8 | 10 | −6 | 5 | 27 | 5 | −14 | 1,003 | −20 | 37 |
| IPSL-CM4 | 2,221 | 418 | 95 | 264 | 8,882 | 1,563 | 6 | −4 | −15 | 3 | 22 | 5 | −16 | 1,503 | −30 | 62 |
| CSIRO-Mk3.0 | 1,820 | 213 | 41 | 130 | 6,722 | 2,012 | 13 | 5 | −10 | 4 | 23 | 5 | −14 | 892 | −19 | 38 |
4.3 Uncertainty in projected regional impacts
The range in regional and global impacts in 2050, under the A1b emissions and socio-economic scenario, across all climate models
| Pop. exposed to increased water resources stress (millions) | Pop. with decreased water resources stress (millions) | Pop. exposed to doubled flood freq. (millions) | Pop. exposed to halved flood freq. (millions) | Cropland with decline in suitability (thousand km2) | Croplandwith increase in suitability (thousand km2) | Change in average spring wheatyield (%) | Change in average soybean yield (%) | Change in average maize yield(%) | Change in Soil Organic Carbon(SOC) content (%) | Change in regional NPP (%) | Change in forest area (%) | Change in coastal wetland (%) | Additional people flooded in coastal floods (thousand/year) | Change in heat energy demand (%) | Change in cool energy demand (%) | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| W. Africa | 0 to 161 | 0 to 34 | 0 to 21 | 0 to 27 | 191 to 656 | 0 to 287 | 4 to 51 | −28 to−2 | −45 to−15 | −3 to 8 | 10 to 22 | 3 to 10 | −12 to−7 | 26 to 233 | −67 to−33 | 16 to 40 |
| C. Africa | 0 to 6 | 0 to 5 | 0 to 10 | 0 to 6 | 4 to 76 | 0 to 32 | −33 to−6 | −48 to−14 | −46 to−17 | 1 to 5 | 13 to 30 | 2 to 4 | −17 to−12 | 3 to 32 | −90 to−63 | 17 to 46 |
| E. Africa | 0 to 167 | 0 to 97 | 0 to 10 | 0 to 6 | 0 to 139 | 1 to 124 | −36 to−5 | −49 to−18 | −22 to 3 | 3 to 7 | 23 to 59 | 4 to 13 | −16 to−9 | 2 to 29 | −92 to−60 | 43 to 107 |
| Sn Africa | 0 to 17 | 0 to 11 | 0 to 8 | 0 to 4 | 229 to 380 | 0 to 17 | −46 to−16 | −26 to−8 | −36 to−11 | 8 to 11 | 19 to 50 | 4 to 10 | −20 to−17 | 142 to 1,909 | −71 to−43 | 57 to 148 |
| S. Asia | 52 to 1,460 | 0 to 1,397 | 9 to 290 | 1 to 161 | 215 to 2,049 | 17 to 1,621 | 12 to 41 | −17 to 23 | −28 to−1 | −6 to−1 | 5 to 30 | 4 to 9 | −13 to−9 | 93 to 716 | −52 to−29 | 20 to 45 |
| SE Asia | 0 | 0 | 1 to 71 | 0 to 40 | 0 to 159 | 0 to 2 | 18 to 33 | −32 to−9 | −42 to−15 | 7 to 9 | 15 to 23 | 2 to 3 | −15 to−11 | 349 to 643 | −87 to−62 | 15 to 35 |
| E Asia | 0 to 506 | 0 to 648 | 1 to 113 | 0 to 21 | 15 to 662 | 108 to 491 | −2 to 16 | 0 to 15 | −19 to−4 | −3 to−2 | 17 to 32 | 4 to 7 | −24 to−18 | 47 to 441 | −31 to−15 | 33 to 97 |
| Central Asia | 0 to 40 | 0 to 1 | 0 to 1 | 0 to 3 | 72 to 294 | 16 to 191 | −3 to 50 | −100 to 168 | −38 to−6 | 7 to 12 | 9 to 49 | 12 to 23 | n/a | n/a | −31 to−12 | 59 to 127 |
| Australasia | 0 | 0 | 0 to 1 | 0 to 1 | 131 to 278 | 2 to 111 | −26 to 5 | −13 to 39 | −31 to−6 | 2 to 6 | 17 to 36 | 6 to 9 | −15 to−10 | 31 to 110 | −46 to−26 | 45 to 109 |
| N. Africa | 109 to 226 | 0 to 44 | 0 to 7 | 5 to 21 | 185 to 350 | 0 to 128 | −60 to 26 | −44 to−10 | −44 to−21 | 1 to 5 | 9 to 42 | 5 to 12 | −28 to 4 | 1 to 28 | −61 to−33 | 40 to 87 |
| W. Asia | 135 to 308 | 0 to 134 | 0 to 1 | 1. to 9 | 296 to 357 | 1 to 22 | −17 to 16 | n/a | −21 to 11 | 3 to 14 | −16 to 30 | −13 to 10 | −25 to−16 | 2 to 47 | −39 to−22 | 30 to 59 |
| W. Europe | 24 to 211 | 0 to 143 | 0 to 12 | 1 to 15 | 249 to 448 | 214 to 262 | 2 to 33 | −12 to 34 | −18 to−1 | 2 to 4 | 15 to 39 | 4 to 8 | −23 to−14 | 6 to 42 | −32 to−16 | 57 to 178 |
| C. Europe | 2 to 32 | 0 to 6 | 0 to 1 | 1 to 9 | 90 to 249 | 143 to 236 | −14 to 17 | −25 to 87 | −12 to 8 | 4 to 8 | 0 to 38 | 3 to 7 | −35 to−6 | 0 to 6 | −30 to−15 | 189 to 821 |
| E. Europe | 0 to 25 | 0 to 4 | 0 to 4 | 3 to 17 | 358 to 1,144 | 388 to 893 | −15 to 22 | 10 to 87 | −14 to 7 | −1 to 1 | 17 to 36 | 7 to 11 | −23 to−12 | 1 to 58 | −27 to−11 | 116 to 519 |
| Canada | 0 to 7 | 0 to 7 | 0 | 0 to 1 | 0 to 177 | 183 to 385 | −3 to 35 | −40 to 45 | −13 to 12 | −2 to 0 | 14 to 41 | 5 to 9 | −33 to−3 | 3 to 23 | −29 to−9 | 71 to 365 |
| US | 21 to 116 | 0 to 49 | 0 to 3 | 0 to 6 | 300 to 1,545 | 25 to 270 | −14 to 7 | −10 to 28 | −20 to−1 | 4 to 6 | 9 to 37 | 3 to 5 | −28 to−20 | 3 to 17 | −35 to−13 | 47 to 137 |
| Meso-America | 0 to 112 | 0 to 53 | 0 to 6 | 0 to 10 | 188 to 415 | 0 to 2 | −30 to 4 | −34 to 7 | −43 to−3 | 2 to 11 | −9 to 30 | −1 to 6 | −25 to−18 | 8 to 37 | −71 to−34 | 28 to 77 |
| Brazil | 0 to 28 | 0 | 0 to 11 | 0 to 10 | 21 to 193 | 0 to 56 | −25 to 13 | −30 to 2 | −32 to−7 | 14 to 21 | −8 to 33 | −1 to 4 | −13 to−8 | 11 to 48 | −93 to−52 | 27 to 69 |
| South America | 0 to 31 | 0 to 18 | 0 to 12 | 0 to 12 | 52 to 405 | 20 to 185 | −13 to 15 | −21 to 0 | −31 to−8 | 5 to 9 | 8 to 30 | 3 to 6 | −22 to−18 | 23 to 142 | −48 to−23 | 31 to 79 |
| Global | 533 to 3,098 | 172 to 2,196 | 31 to 449 | 41 to 264 | 3,783 to 8,882 | 1,378 to 4,062 | 1 to 22 | −13 to 10 | −22 to−6 | 2 to 5 | 16 to 27 | 4 to 5 | −18 to−13 | 763 to 4,101 | −34 to−17 | 29 to 73 |
Uncertainty in regional impacts in 2050, under A1b emissions and socio-economic scenarios. Impacts under individual climate models are shown as open circles; the red circle shows impacts under one specific model (HadCM3)
For most impact sectors, the projected ranges are very large. In some cases—specifically the water and river flooding sectors—this is because of very large uncertainty in projected changes in regional rainfall (in south and east Asia, for example). In some other cases, the large uncertainty is because the sector in a region is particularly sensitive to change (for example where the baseline values in the absence of climate change are small—see forest and NPP change in west and central Asia). In other cases, the uncertainty range is dominated by individual anomalous regional changes. For example, the large range in estimated additional people exposed to coastal flooding is due to one particular climate model producing very considerably higher sea level rises in some regions than the others. There is least uncertainty in projected reductions in heating energy requirements and, for most regions, increases in cooling energy requirements; the greatest uncertainty here is in those regions where requirements are currently low—Europe and Canada—but the percentage changes are sensitive to small changes in temperature.
The considerable uncertainty in each region and sector needs to be interpreted carefully. It is not correct simply to add up the extremes of each range across regions and use this to characterise the global range; the global range will be smaller than the sum of the extremes because no one climate model produces the most extreme response in every region. Similarly, it is not appropriate to define the maximum impact across all sectors in a region as the sum of the maximum impacts for each sector, because again no one single climate model produces the maximum impact in all sectors. Indeed, there are some associations between impacts in different sectors between climate models. For example, models which produce the greatest increase in exposure to water resources stress tend to be those which produce the smallest increase in exposure to river flooding, and the greatest area of cropland with a decline in suitability (see Supplementary Fig. 8 for an example).
5 Impacts under different worlds and over time
Global-scale impacts of climate change in 2020, 2050 and 2080 under A1b, A2, B1 and B2 emissions and socio-economic scenarios. The grey bars represent the range across the climate models, the impacts under one specific model (HadCM3) are shown by the solid circle
6 Conclusions
This paper has presented a high-level assessment of the global and regional impacts of climate change across a range of sectors. The assessment used a harmonised set of assumptions and data sets, four scenarios of future socio-economic development and emissions, and climate scenarios constructed from 21 climate models. The distribution of impacts between regions and the relationship between different impact indicators are important, so the assessment first describes impacts under a set of discrete ‘stories’ based on different climate models, and then considers uncertainty in regional impacts separately. The paper has therefore demonstrated a method for assessing multi-dimensional, regionally-variable impacts of climate change for a global assessment.
With A1b emissions and socio-economics, one plausible climate future (based on one climate model pattern) would result in 2050 in 1 billion people being exposed to increased water resources stress, around 450 million people exposed to increased frequency of river flooding, and an additional 1.3 million people flooded each year in coastal floods. Approximately half of all cropland would see a reduction in suitability for cropping, and the productivity of three major crops—spring wheat, soybean and maize—would be reduced in most regions. Global residential cooling energy requirements would increase by over 70 % globally, but heating energy requirements would decrease so total global heating and cooling energy requirements would reduce globally. The productivity of terrestrial ecosystems would be increased, and soil organic carbon contents would generally increase, leading to improved soil productivity and increased carbon storage. However, there is strong regional variability. Under this one climate model pattern, most of the global impacts on water stress and flooding would be in south, southeast and east Asia, but spring wheat productivity increases across much of Asia. In proportional terms, impacts on water stress and crop productivity are very large in the Middle East and North Africa region, which is exposed to multiple impacts.
There is considerable uncertainty in the projected regional impacts under a given emissions and socio-economic scenario, largely due to differences in the spatial pattern of climate change simulated by different climate models; this uncertainty varies between regions and sectors. Large increases in exposure to water resources stress, for example, are associated with large reductions in crop suitability but small increases in exposure to river flooding. The full richness of relationships between impacts in different places, and in different sectors, can therefore only be understood by comparing narrative stories constructed separately from different climate model scenarios.
There are, of course, a number of caveats with the approach. The climate scenarios used here are based on SRES emissions assumptions, and not on more recent RCP forcings or the climate models reviewed in the most recent IPCC assessment (IPCC 2013). However, the spatial patterns of change in climate under the latest generation of climate model simulations are broadly similar to those used here (Knutti and Selacek 2013). The climate scenarios are constructed by pattern scaling, and whilst this allows a direct comparison between different emissions scenarios and time periods, it does assume a particular relationship between the amount of global temperature change and the spatial pattern of change in climate. The indicators used represent an ‘ensemble of opportunity’, and do not necessarily span the full range of impacts of interest; there are also alternative indicators for many of the sectors considered here. The indicators do not (with the notable exception of crop productivity) explicitly incorporate the effects of adaptation in reducing the consequences of climate change. Comparisons with other single-sector global-scale impact assessments are made difficult by the use of different impact indicators (e.g. in the water sector) and different climate model scenarios. Insofar as it is possible to make comparisons, impacts as estimated in these other assessments are within the ranges presented here, but nevertheless the impacts presented here are best interpreted as indicative only. Finally, the indicators are calculated using only one impact model per sector. It is increasingly recognised that impact model uncertainty may make a substantial contribution to total impact uncertainty in some regions (e.g. Hagemann et al. 2013), and several initiatives are currently under way (for example ISI-MIP: Warszawski et al. 2014) to systematically evaluate the effects of impact model uncertainty.
Notes
Acknowledgments
The research presented in this paper was conducted under the QUEST-GSI project, funded by the UK Natural Environment Research Council (NERC) as part of the QUEST programme (grant number NE/E001882/1). PS is a Royal Society-Wolfson Research Merit Award holder. We acknowledge the international modelling groups for providing climate change and sea level data for analysis, the Program for Climate Model Diagnosis and Intercomparison (PCMDI) for collecting and archiving the model data, the JSC/CLIVAR Working Group on Coupled Modelling (WGCM) and their Coupled Model Intercomparison Project (CMIP) and Climate Simulation Panel for organising the model data analysis activity. The IPCC Data Archive at Lawrence Livermore National Laboratory is supported by the Office of Science, US Department of Energy. We thank the reviewers for their helpful comments and suggestions.
Supplementary material
References
- Arnell NW, Gosling SN (2014) The impacts of climate change on river flood risk at the global scale. Clim Chang. doi: 10.1007/s10584-014-1084-5 Google Scholar
- Arnell NW, Lowe JA, Brown S et al (2013) A global assessment of the effects of climate policy on the impacts of climate change. Nat Clim Chang 3:512–519. doi: 10.1038/nclimate1793 CrossRefGoogle Scholar
- Arnell NW, Brown S, Gosling SN et al (2014) Global-scale climate impact functions: The relationship between forcing and impact. Clim Chang. doi: 10.1007/s10584-013-1034-7 Google Scholar
- Brown S, Nicholls RJ, Lowe J, Hinkel J (2013) Spatial variations of sea-level rise and impacts: An application of DIVA. Clim Chang. doi: 10.1007/s10584-013-0925-y Google Scholar
- Center for International Earth Science Information Network (CIESIN) (2004) Global Rural-Urban Mapping Project (GRUMP), Alpha Version: Population Grids. Socioeconomic Data and Applications Center (SEDAC), Columbia University. Available at http://sedac.ciesin.columbia.edu/gpw. (1 April 2011), Palisades: NY.
- Gosling SN, Arnell NW (2013) A global assessment of the impact of climate change on water scarcity. Clim Chang. doi: 10.1007/s10584-013-0853-x Google Scholar
- Gottschalk P et al (2012) How will organic carbon stocks in mineral soils evolve under future climate? Global projections using RothC for a range of climate change scenarios. Biogeosci 9:3151–3171CrossRefGoogle Scholar
- Gregory J, Huybrechts P (2006) Ice-sheet contributions to future sea-level change. Philos Trans R Soc Lond A 364:1709–1731CrossRefGoogle Scholar
- Hagemann S et al (2013) Climate change impact on available water resources obtained using multiple global climate and hydrology models. Earth Syst Dynamics 4:129–144CrossRefGoogle Scholar
- Harris I, Jones PD, Osborn TJ, Lister DH (2014) Updated high-resolution grids of monthly climatic observations - the CRU TS3.10 data set. Int J Climatol 34:623–642CrossRefGoogle Scholar
- Hayashi A, Akimoto K, Sano F, Mori S, Tomoda T (2010) Evaluation of global warming impacts for different levels of stabilization as a step toward determination of the long-term stabilization target. Clim Chang 98:87–112CrossRefGoogle Scholar
- Huntingford C, Booth BBB, Sitch S et al (2010) IMOGEN: an intermediate complexity model to evaluate terrestrial impacts of a changing climate. Geosci Model Dev 3:679–687CrossRefGoogle Scholar
- IPCC (2007) Climate Change 2007. In: Parry ML (ed) Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UKGoogle Scholar
- IPCC (2013) Climate Change 2013. In: Stocker T (ed) The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UKGoogle Scholar
- IPCC (2014) Climate Change 2014. In: Field C (ed) Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UKGoogle Scholar
- IPCC (Intergovernmental Panel on Climate Change) (2000) Special Report on Emissions Scenarios. Cambridge University Press, CambridgeGoogle Scholar
- Isaac M, van Vuuren DP (2009) Modeling global residential sector energy demand for heating and air conditioning in the context of climate change. Energy Pol 37:507–521CrossRefGoogle Scholar
- Knutti R, Selacek J (2013) Robustness and uncertainties in the new CMIP5 climate model projections. Nat Clim Chang 3:369–373. doi: 10.1038/nclimate1716 CrossRefGoogle Scholar
- Meehl GA et al. (2007a) The WCRP CMIP3 multimodel dataset - A new era in climate change research. Bulletin of the American Meteorological Society 88:1383-+.Google Scholar
- Meehl GA (2007) Global climate projections. In: Solomon S (ed) Climate Change 2007: The Physical Science Basis. Contribution of Working Group 1 to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, CambridgeGoogle Scholar
- Osborn T, Wallace CJ, Harris IC, Melvin TM (2014) Pattern scaling using ClimGen: monthly resolution future climate scenarios including changes in the variability of precipitation. Climatic Change, in pressGoogle Scholar
- Osborne T, Rose GA, Wheeler TR (2012) Variation in the global-scale impacts of climate change on crop productivity due to climate model uncertainty and adaptation. Agric For Meteorol 170:183–194CrossRefGoogle Scholar
- Piontek F et al (2014) Multisectoral climate impact hotspots in a warming world. Proc Natl Acad Sci 111(9):3233–3238CrossRefGoogle Scholar
- Ramankutty N, Foley JA, Norman J, McSweeney K (2002) The global distribution of cultivable lands: current patterns and sensitivity to possible climate change. Glob Ecol Biogeogr 11:377–392CrossRefGoogle Scholar
- Ramankutty N, Evan AT, Monfreda C, Foley JA (2008) Farming the planet: 1. Geographic distribution of global agricultural lands in the year 2000. Global Biogeochemical Cycles 22. doi: Gb1003 10.1029/2007gb002952
- Tebaldi C, Arblaster JM (2014) Pattern-scaling: its strengths and limitations, and an update on the latest model simulations. Clim Chang 122:459–471CrossRefGoogle Scholar
- van Vuuren DP et al (2007) Stabilizingst greenhouse gas concentrations at low levels: an assessment of reduction strategies and costs. Clim Chang 81:119–159CrossRefGoogle Scholar
- Van Vuuren DP et al (2011) The use of scenarios as the basis for combined assessment of climate change mitigation and adaptation. Glob Environ Chang 21:575–591CrossRefGoogle Scholar
- Warszawski L et al (2014) The Inter-Sectoral Impact Model Intercomparison Project (ISI-MIP): Project framework. Proc Natl Acad Sci 111(9):3228–3232CrossRefGoogle Scholar
Copyright information
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.