The rates of anthropogenic greenhouse gas and aerosol emissions that influence future climate will vary according to changes in population and economic growth, technology, energy availability and national and international policies. The Intergovernmental Panel on Climate Change developed 40 different future emissions pathways, the so-called ‘SRES scenarios’ (see section 1.3 and Appendix 1), with no evaluation of their relative probabilities of occurrence.
In order to drive detailed regional projections from such global emissions, it is necessary to use complex atmosphere–ocean global climate models (AOGCMs). These model simulations require months of supercomputer time for each scenario of forcing emissions (or atmospheric concentration). The SRES emission scenarios were not produced early enough for climate modellers to incorporate them into model projections for the IPCC Third Assessment of 2001. However, for the Fourth Assessment of 2007 (referred to as ‘AR4’), a large number of model simulations were carried out, focusing particularly on the six illustrative scenarios (A1B, A1FI, A1T, A2, B1 and B2 – see Appendix 1). The Summary for Policymakers (IPCC 2007) states that all these scenarios should be considered equally sound.
A2.2 Selection of Models
An enormous amount of data processing is required to cover the range of model and scenario information available from the AR4 modelling community (global fields of many climate variables, each for 100 years and more, for 15–20 General Circulation Models and for six emission scenarios). Initially, NIWA focused in its FRST-funded work (contracts C01X0202, C01X0701) on model simulations for the A1B emissions scenario, and this dataset is used as the basis of the present report.
Data were downloaded from the IPCC data centre, http://www-pcmdi.llnl.gov/software-portal/esg_data_portal/dapserver/ (3 April 2008) for 17 General Circulation Models that had produced projections forced by the A1B emissions scenario. The climate variables of interest are those required for downscaling over New Zealand – specifically, mean sea-level pressure (mslp), precipitation, and surface air temperature. Since we are interested in changes from the current climate, it was also necessary to download simulated data from the 20th century control run (which typically begins in the mid- to late 19th century). The 20th century data end in either 1999 or 2000 (depending on modelling institution), with the A1B simulation following on for a further 100 years to either 2099 or 2100. In some cases, the A1B simulation was extended for a further 100 years or 200 years with atmospheric concentrations of radiatively active gases fixed at the 2100 scenario level, in what were called ‘stabilisation experiments’.
Extensive validation of the control climates of the 17 General Circulation Models was carried out, comparing the period 1971–2000 (or 1970–199979) in the models with gridded observational data for 1971–2000 from the widely-used NCEP re-analysis (Kalnay et al 1996). The validation focused on the New Zealand–South Pacific region, and calculated correlations and root-mean square differences between observed and simulated climatology of: spatial pattern and seasonal variation in mslp, precipitation and temperature; position and intensity of the westerly wind maximum south of New Zealand and the high-pressure maximum north of New Zealand; the Southern Oscillation Index; and the Trenberth circulation indices Z1 and M1 (used in the downscaling procedure). The results (to be reported elsewhere) indicated that five of the 17 models performed significantly poorer than the remaining 12. For example, some of the poorer five models had the Southern Hemisphere westerly winds much further north than observed, and no clear Southern Oscillation signal in their interannual variability.
Table A2.1 lists the 12 models retained for the downscaling exercise, along with the global annual temperature changes and (downscaled) New Zealand-average annual temperature changes relative to the base period 1980–1999, under the A1B scenario. Multi-decadal variations in the rate of warming can be seen: for example, the mpi_echam5 has the least New Zealand warming by 2040 (+0.33°C), but by 2090 the model csiro_mk30 has least warming (+1.13°C). The average 100-year warming to 2080–2099 over the 12 models for the A1B scenario is 2.80°C for the globe (range 1.84–4.15°C), and 2.10°C for New Zealand (range 1.13–3.44°C).
No model projected New Zealand warming faster than the global average. The ratio of New Zealand to global warming over 100 years (1980–1999 to 2080–2099) varies between 0.56 and 0.96, with an average rate over the 12 General Circulation Models of 0.75. For the 50-year period (1980–1999 to 2030–2049), there is a larger scatter in the estimated trend, with the New Zealand-to-global warming ratio varying between 0.30 and 0.94, but an average rate over the 12 General Circulation Models of 0.73.
Table A2.1: Annual temperature changes (in °C) relative to 1980–1999 for 12 General Circulation Models forced by the SRES A1B scenario. Changes are shown for different end periods, the global and downscaled New Zealand average.
A2.2 Scaling methodology
The Summary for Policymakers (IPCC 2007) summarises the projected global warming between 1980–1999 and 2090–2099,80 as shown below in Table A2.2. For the A1B scenario, the global-average surface temperature change varies between 2.0°C and 4.3°C for the 12 models that perform well in the South Pacific region. Including the other five models increases this range slightly to 1.9–4.3°C, which corresponds fairly closely to the IPCC range of 1.7–4.4°C. The reason for the small discrepancy is that the IPCC range incorporates some expert judgement after considering a wide range of climate models that are simpler numerically than the General Circulation Models but that encompass more uncertainties such as difference rates of carbon cycling through the climate system. In simple terms, the IPCC range arises from taking the ‘best estimate’ temperature change, and subtracting 40% to get the low end, and adding 60% to get the high end of the range (see Meehl et al 2007: caption to figure 10.29 in chapter 10). This simplification was also a sensible approach given that the number of model experiments varied with the scenario (up to 23 General Circulation Models for the B1, A1B and A2 scenarios, and a lot fewer for B2, A1T and A1FI).
Table A2.2: Projected global average surface warming (in °C) from 1980–1999 to 2090–2099 for the six illustrative IPCC SRES emission scenarios.
Source: Table SPM.3 in IPCC (2007).
|Scenario||Best estimate||Likely range|
1.1 – 2.9
1.4 – 3.8
1.4 – 3.8
1.7 – 4.4
2.0 – 5.4
2.4 – 6.4
For the purposes of this Ministry for the Environment Guidance Manual, maps are presented of changes only for the A1B scenario, rather than trying to rescale to cover a range of emission scenarios (as was done in the earlier Guidance Manual, Ministry for the Environment 2004). Thus, the maps in chapter 2 (Figures 2.3–2.7) are the downscaled changes from the A1B-driven General Circulation Model projections. However, in the tables of chapter 2 (Tables 2.2–2.5), a rescaling is carried out to mimic the impact over all six illustrative scenarios, which span the full range of the 40 SRES scenarios.
Rescaling is done by taking the 17-model A1B range (ie, 1.9–4.3°C), and calculating the factors required to match this to the IPCC ‘likely ranges’ for each scenario of Table A2.2, while maintaining the same relative spacing (in global temperature change space) between the models. This scaling factor is then applied to the local change (in temperature, precipitation, etc.) from the downscaling, where only 12 models are ultimately considered. This assumption of a proportional relationship between the global temperature change and a local change is a very common one in integrated assessment modelling (Kenny et al 2001). The scaling factors vary between about 0.6 for the B1 scenario to about 1.0 for A1B and 1.4 for A1FI (but are model-dependent).
A2.3 Probability distribution of climate projections
The IPCC Fourth Assessment, 2007 does give some consideration to describing projected warming in terms of probability distributions, in addition to a simple best estimate and likely range (eg, section 10.5.4.5 in Meehl et al 2007). The distributions are estimated across the multi-model ensembles, and evaluated separately for each emission scenario. By IPCC convention, probabilities or likelihoods are not assigned to the emission scenarios themselves.
Estimated probability distributions often demonstrate a slight positive skewness (ie, a longer tail to the right or high end of projected changes). To some extent, this is expected because changes in many variables (temperature, sea level) are truncated to be non-negative at the low end of the range, at least in the global average although perhaps not regionally. Even though the high end of projected changes has a low probability, the higher risks associated with these extreme projections suggest they be given serious consideration (Kerr 2007). We also see examples of this positively skewed distribution of changes in the local downscaling results (see Appendix 3).
Kenny, GJ, Harman, JJ, Warrick, RA. 2001. Introduction: The CLIMPACTS programme and method. In: Warrick RA, Kenny GJ, Harman JJ (eds) The Effects of Climate Change and Variation in New Zealand: An assessment using the CLIMPACTS system. Chapter 1. IGCI, University of Waikato: Hamilton. 1–10.
Kerr RA. 2007. Pushing the scary side of global warming. Science 316: 1412–1415.
IPCC. 2007. Summary for Policymakers. In: Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, Tignor M, Miller HL (eds). Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press: Cambridge, United Kingdom and New York, NY, USA. http://www.ipcc.ch/pdf/assessment-report/ar4/wg1/ar4-wg1-spm.pdf (2 April 2008).
Kalnay E, Kanamitsu M, Kistler R, Collins W, Deaven D, Gandin L, Iredell M, Saha S, White G, Woollen J, Zhu Y, Chelliah M, Ebisuzaki W, Higgins W, Janowiak J, Mo KC, Ropelewski C, Wang J, Leetmaa A, Reynolds R, Jenne R, Joseph D. 1996. The NCEP/NCAR 40-year reanalysis project. Bulletin American Meteorological Society 77: 437–471.
Meehl GA, Stocker TF, Collins WD, Friedlingstein P, Gaye AT, Gregory JM, Kitoh A, Knutti R, Murphy JM, Noda A, Raper SCB, Watterson IG, Weaver AJ, Zhao ZC. 2007. Global climate projections. In: Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, Tignor M, Miller HL (eds) Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change . Cambridge University Press: Cambridge, United Kingdom and New York, NY, USA. http://www.ipcc.ch/pdf/assessment-report/ar4/wg1/ar4-wg1-spm.pdf (2 April 2008).
Ministry for the Environment. 2004. Climate Change Effects and Impacts Assessment: A Guidance Manual for Local Government in New Zealand. Wratt D, Mullan B, Salinger J, Allen S, Morgan T, Kenny G,with MfE. Ministry for the Environment Report, 153 p. http://www.mfe.govt.nz/publications/climate/effects-impacts-may04/index.html (3 April 2008).
79 Since we are comparing features of the simulated climate, and are not concerned with exact sequencing in the time series, the 1-year difference in control climate periods is unimportant. The effect of a 1-year difference in radiative forcing will likewise be trivial.
80 Note that the end period starts at the year 2090 in the IPCC table, and we use ratios for the approximately 105-year changes to rescale the A1B results to the other SRES scenarios. However, the New Zealand downscaled changes are calculated between two 20-year periods: for example, 1980–1999 to 2080–2099.