9: Climate Change Scenarios

Clare Goodess

Climate change scenarios provide the best-available means of exploring how human activities may change the composition of the atmosphere, how this may affect global climate, and how the resulting climate changes may impact upon the environment and human activities. They should not be viewed as predictions or forecasts of future climate, but as internally-consistent pictures of possible future climates, each dependent on a set of prior assumptions.

General circulation models (GCMs) (see Information Sheet 8) are complex, gridded, three-dimensional computer-based models of the climate system (developed from numerical weather forecasting models). They are considered to provide the best basis for the construction of climate change scenarios.

European land-sea mask for the HadCM2/HadCM3 GCMs. The HadCM2 and HadCM3 GCMs developed by the UK Met. Office Hadley Centre have a gridded resolution of 2.5° latitude by 3.75° longitude, which is typical of the current generation of models. Thus the model geography is much simpler than the real-world geography. Only five grid boxes cover the UK, for example.

A number of decisions must be taken in order to construct GCM-based climate change scenarios:

• What underlying emissions scenarios should be used to predict atmospheric greenhouse gas and aerosol concentrations?

A forcing of 1% per year increase in the equivalent CO₂ atmospheric concentration has been widely used as a 'business-as-usual' emissions scenario in GCM experiments. However, the Intergovernmental Panel on Climate Change (IPCC) Special Report on Emission Scenarios (SRES), accepted in May 2000, identifies 40 scenarios which follow four different 'storylines' and with forcing increases ranging between 0.4% and 1.2% per year. Each of the SRES scenarios is considered equally probable.


• Should the climate change scenarios be based on a single GCM (and if so, which) or should the results from several models be combined in order to reflect inter-model differences?

GCMs are run at a number of different modelling centres. The IPCC Data Distribution Centre (IPCC DDC), for example, holds GCM output from seven modelling groups based in Europe, North America, Japan and Australia. There are differences in the way physical processes and feedbacks are simulated in different GCMs. Thus the global temperature change in response to a doubling of the atmospheric CO₂ concentration (the climate sensitivity) varies from model to model (a sensitivity range of 1.5° to 4.5° C is conventionally quoted). The same GCM may give different results for the same forcing but slightly different initial conditions. Thus it may be desirable to consider results from an ensemble of runs. There are also differences in the regional estimates of climate change produced by different models (even for the same mean global warming). Identification of a single 'best' GCM is not straightforward, particularly when performance is considered across a number of different climate variables and/or regions. Climate changes from GCMs should be calculated so as to maximise signal-to-noise ratios.

(also available as PostScript, 3.2mb) (also available as PostScript, 3.2mb) Change in mean annual temperature (left-hand panel, °C) and mean summer rainfall (right-hand panel, percentage change) with respect to the 1961-90 mean for thirty-year periods centred on the 2020s, 2050s and 2080s for the UKCIP98 scenarios (Hulme and Jenkins, 1998). The background fields are interpolated from the full HadCM2 grid, while the highlighted numbers show the change for each HadCM2 grid box over the UK. The UKCIP98 scenarios were prepared for the UK Climate Impacts Programme. The four scenarios (low, medium-low, medium-high and high) span a range of future global warming rates from 0.1° to 0.3° C per decade.

• Should the original grid box GCM information be used (the current generation of GCMs typically provides information on a grid with a spatial resolution of about 300 km North-South, and which may vary in a West-East direction depending on latitude but will typically be about 400 km at 45° N or S) or is information required at a higher spatial resolution (for investigating the regional impacts of climate change on agriculture or water resources, for example)?

Raw output from GCMs is rarely used in assessments of the environmental and economic impacts of climate change because of the relatively coarse spatial scale. Furthermore, confidence in the reliability of the output tends to decrease moving from the global to the grid-box scale, and from the annual to the monthly and ultimately the daily time scale. In order to address these problems of scale and reliability, various techniques have been developed for 'downscaling' from the coarse GCM scale to the finer spatial scale required for impact assessment (which may be a particular country, or a river basin or even a single site). No one method is universally appropriate and what may seem good in one region may be poor in another climate regime.


• If scenarios with a high spatial resolution are required, what method of downscaling should be used?

Two major approaches to downscaling can be identified: model-based and empirical. The first approach involves nesting a finer-scale Regional Climate Model (with a typical spatial resolution of 50 km) within the coarse-scale GCM. This approach is considered to have good potential, but is currently subject to a number of technical problems and limitations. The second approach, empirical downscaling, requires the identification of quantitative relationships between the observed large-scale and regional climate, which are then applied to large-scale GCM output (making the assumption that these relationships will remain unchanged in a future, warmer world). It has the advantage of requiring fewer data inputs and computing resources than the model-based approach, but there is a need for further development and comparison of both approaches to downscaling. There have been, at present, very few direct comparisons of the two approaches.

The model-based approach to downscaling (also available as PostScript) (also available as PostScript) (also available as PostScript) Annual UK rainfall (mm per day) simulated by the HadCM2 GCM (left-hand panel) and the Hadley Centre RCM (centre panel), compared with observed rainfall for 1961-1990 (right-hand panel). The observed spatial pattern of annual UK rainfall is more closely reproduced by the Hadley Centre RCM (which has a spatial resolution of 50 km by 50 km) than the HadCM2 GCM (which has a spatial resolution of about 300 km). This comparison was undertaken as part of a study to investigate how the intensity of rainfall over the UK might change in the future (Jones and Reid, 2000).

The empirical approach to downscaling (also available as PostScript, 5mb) Winter rainfall scenario for the Mediterranean constructed using an empirical approach to downscaling. This scenario was constructed by Maureen Agnew for the MEDALUS project, using a Geographical Information System to interpolate output from the HadCM2 GCM to a 1 km by 1 km grid based on information such as height above sea level, distance to the sea, and latitude and longitude (Agnew and Palutikof, 2000).

Climate change scenarios are most commonly constructed for temperature variables and, with lesser confidence, for moisture-related variables. Scenarios can, however, be constructed for other variables, such as sea level change (see Information Sheet 10).

Further reading

• Agnew, M.D. and Palutikof, J.P., 2000: 'GIS-based construction of baseline climatologies for the Mediterranean using terrain variables', Climate Research, 14, 115-127.
• Giorgi, F. and Mearns, L.O., 1999: 'Introduction to special section: Regional climate modelling revisited', Journal of Geophysical Research, 104, 6335-6352.
• Hewitson, B.C. and Crane, R.G., 1996: ‘Climate downscaling: techniques and application’ Climate Research, 7, 85-95.
• Houghton, J.T., Meiro Filho, L.G., Callander, B.A., Harris, N., Kattenberg, A. and Maskell, K. (eds.), 1996: Climate Change 1995: The Science of Climate Change. Cambridge University Press, Cambridge, 572pp.
• Hulme, M. and Jenkins, G.J., 1998: Climate Change Scenarios for the United Kingdom: Scientific Report. UKCIP Technical Report No. 1, Climatic Research Unit, Norwich, 80pp.
• IPCC-TGCIA, 1999: Guidelines on the Use of Scenario Data for Climate Impact and Adaptation Assessment. Version 1, Prepared by Carter, T.R., M. Hulme and M. Lal, Intergovernmental Panel on Climate Change, Task Group on Scenarios for Climate Impact Assessment, 69pp. Available on line from the IPCC DDC web site.
• Jones, P.D. and Reid, P.A., 2000: 'Assessing future changes in extreme precipitation over Britain using Regional Climate Model integrations', International Journal of Climatology, submitted.
• Kattenberg, A., Giorgi, F., Grassl, H., Meehl, G.A., Mitchell, J.F.B., Stouffer, R.J., Tokioka, T., Weaver, A.J. and Wigley, T.M.L., 1996: 'Climate models - projections of future climate', in J.T. Houghton, L.G. Meira Filho, B.A. Callander, N. Harris, A. Kattenberg and K. Maskell (eds.), Climate Change 1995: The Science of Climate Change. Cambridge University Press, Cambridge, 285-357.
• Mearns, L.O., Bogardi, I., Giorgi, F., Matyasovszky, I. and Palecki, M., 1999: 'Comparison of climate change scenarios generated from regional climate model experiments and statistical downscaling', Journal of Geophsyical Research, 104, 6603-6621.
• von Storch, H., Zorita, E. and Cubasch, U., 1993: 'Downscaling of global climate change estimates to regional scales: an application to Iberian rainfall in wintertime', Journal of Climate, 6, 1161-1171.
• Wilby, R.L., Wigley, T.M.L., Conway, D., Jones, P.D., Hewitson, B.C., Main, J. and Wilks, D.S., 1998: 'Statistical downscaling of general circulation model output: A comparison of methods', Water Resources Research, 42, 2995-3008.

Relevant web sites

• Climate Impacts LINK project: distribution centre for output (including daily data) from the UK Met. Office Hadley Centre GCMs.
• Intergovernmental Panel on Climate Change (IPCC)
• IPCC Data Distribution Centre (IPCC DDC): distribution centre for monthly GCM output from seven international modelling groups and expert advice on scenario construction.
• IPCC Special Report on Emission Scenarios (SRES): Summary for Policymakers. The SRES scenarios have been constructed for the IPCC Third Assessment Report (due to be published in 2001).
• MAGICC and SCENGEN: Model for the Assessment of Greenhouse-gas Induced Climate Change and a global and regional SCENario GENerator.
• UKCIP98 climate scenarios report: Climate change scenarios for the United Kingdom, see Hulme and Jenkins (1998).