UNCERTAINTIES IN CLIMATE CHANGE PREDICTIONS

Transcription

1 OECD Workshop on the benefits of climate policy. Paris, 12/13 December 2002 UNCERTAINTIES IN CLIMATE CHANGE PREDICTIONS Geoff Jenkins Hadley Centre MetOffice,UK 1 CLIMATE CHANGE DEPENDS UPON: How man-made emissions grow population, GDP, energy use, technology... How sensitive the climate system is to emissions how much concentrations change what forcing effect this has what climate change this produces we can estimate this only with climate models 2 Page 1

2 UNCERTAINTIES IN CLIMATE PREDICTIONS EMISSIONS CONCENTRATIONS FORCING GLOBAL CLIMATE feedbacks Emissions uncertainty Science uncertainty Natural variability REGIONAL CLIMATE NATURAL VARIABILITY IMPACTS, COSTS etc 3 CO2 in SRES EMISSIONS SCENARIOS A1FI A2 B1 B2 4 Page 2

3 GLOBAL TEMPERATURE RISE Global temperature rise, degrees C IPCC A1FI emissions A2 emissions B2 emissions B1 emissions Start to diverge from mid-century HadCM ppm 850 UNCERTAINTY IN CO2 CONCENTRATIONS Hadley coupled GCM Bern High IPCC Bern Low Page 3

4 CLIMATE FORCING (to 2000) GHG sulphate direct FF black C FF org C indirect aerosol Source: IPCC TAR 7 WINTER ENGLAND & WALES PRECIPITATION observed to 1999; predicted to 2050 Change relative to average (mm) 8 Page 4

5 CHANGE IN DJF PRECIPITATION (%) 2080s - present, SRES A2, 3-member initial condition ensemble 9 GLOBAL PRECIPITATION CHANGE due to A2 emissions, predicted by 9 climate models ±70% Source: IPCC, Page 5

6 CHANGE IN GLOBAL PRECIPITATION HadCM3 under a range of emissions scenarios Change in global precipitation (mm/day) ±25% 11 % CHANGE IN WINTER PRECIPITATION 2080s, A2 12 Page 6

7 SEA LEVEL RISE REGIONAL VARIATIONS due to thermal expansion and ocean circulation changes only Source: IPCC m 13 GJJ1999 CLOUDS AND CLIMATE Low clouds reflect sunlight but trap little infra-red radiation; They act to cool climate High clouds reflect sunlight but also trap infra-red radiation; They act to warm climate Global warming may change the characteristics of clouds, and thus alter their current net cooling effect; this could exert a powerful feedback on climate change. 14 Page 7

8 100% PROBABILISTIC CLIMATE PREDICTIONS current position future position Probability Probability -40% -30% -20% -10% 0% 2050s Paris summer rainfall -40% -30% -20% -10% 0% 2050s Paris summer rainfall 15 PROBABILITY PREDICTIONS Generate X ( ) climate models by varying parametrisations within plausible limits Make simulation ( ) with each model Assess credibility of each model based on current climatology and simulation Weight prediction ( ) according to model credibility Use X runs to generate pdf of a specific change pdf can be used direct in risk analyses next: do this with other climate models 16 Page 8

9 PROBABILISTIC CLIMATE PREDICTIONS X CLIMATE MODELS ATMOSPHERE + CLOUDS LAND AND VEGETATION Probability CARBON CYCLE OCEAN CIRCULATION ETC. -40% -30% -20% -10% 0% 2050s Paris summer rainfall 17 RESOLUTION OF CLIMATE SCENARIOS 1998: GLOBAL MODEL 2002: REGIONAL MODEL 300km grid 50km grid 18 Page 9

10 SIMULATION OF A TROPICAL CYCLONE in the PRECIS regional climate model Global climate model Regional climate model Met Office/Hadley Centre 19 REGIONAL MODEL PREDICTIONS OF CHANGE IN WINTER PRECIP BY 2080s Rossby RCM Hadley RCM Both RCMs driven by Hadley Centre HadCM3 GCM 20 Page 10

11 EMISSIONS AND CONCENTRATIONS OF CO 2 from unmitigated and stabilising emission scenarios Anthropogenic CO 2 emissions (GtC/yr) CO 2 concentration (ppm) Unmitigated emissions 750 ppm stabilisation 550 ppm stabilisation 21 Source: IPCC GLOBAL AVERAGE TEMPERATURE RISE from unmitigated and stabilising emission scenarios Global temperature change ( C) Met Office/Hadley Centre Unmitigated emissions 750 ppm stabilisation 550 ppm stabilisation 22 Page 11

12 Potential change in cereal yields (%) CROP YIELD changes from the present day to the 2080s No data HadCM2 Unmitigated emissions Stabilisation of CO 2 at 750 ppm Stabilisation of CO 2 at 550 ppm University of East Anglia/ IIASA Austria /Universidad Politecnica de Madrid/Goddard Institute for Space Studies 23 ANNUAL NUMBER OF PEOPLE FLOODED with no climate change and under three emissions People flooded (millions/year) scenarios University of Middlesex HadCM s 2050s 2080s Unmitigated emissions 750 ppm stabilisation 550 ppm stabilisation No climate change 24 Page 12

13 CHANGES IN STORED CARBON Soil carbon Change in carbon content (kg C per square metre) Vegetation carbon Hadley Centre coupled climate carbon-cycle model 25 EFFECT ON CO2 CONCENTRATION OF CARBON CYCLE FEEDBACK 1000 with carbon cycle without carbon cycle Met Office / Hadley Centre CO2 concentrations (ppmv) Page 13

14 EFFECT ON LAND TEMPERATURES OF CARBON 8 CYCLE FEEDBACK 6 with carbon cycle without carbon cycle Temperature rise ( C) Met Office / Hadley Centre Impact of carbon cycle feedbacks on CO 2 concentrations (WRE550) 28 Page 14

15 Cumulative emissions consistent with WRE550 CO 2 concentration scenario 29 Impact of carbon cycle feedbacks on stabilisation level 30 Page 15

16 SUMMARY: HANDLING UNCERTAINTIES IN CLIMATE CHANGE PREDICTIONS Emissions uncertainty first half C21: uncertainty small ignore beyond: scope using range of SRES scenarios Science uncertainty larger than SRES emissions uncertainty initially: use a range of climate model results ( calibrated using simulations) soon: probability predictions Natural variability quantify via initial condition ensemble 31 Page 16