What do recent advances in quantifying climate and carbon cycle uncertainties mean for climate policy?

Joanna I House^1,5, Chris Huntingford², Wolfgang Knorr¹, Sarah E Cornell¹, Peter M Cox³, Glen R Harris⁴, Chris D Jones⁴, Jason A Lowe⁴ and I Colin Prentice¹

¹ QUEST, Department of Earth Sciences, University of Bristol, Bristol BS8 1RJ, UK
² CEH Wallingford, Maclean Building, Benson Lane, Crowmarsh Gifford, Wallingford, Oxfordshire OX10 8BB, UK
³ School of Engineering, Computer Science and Mathematics, University of Exeter, Harrison Building, North Park Road, Exeter EX4 4QF, UK
⁴ Hadley Centre for Climate Prediction and Research, Met Office, FitzRoy Road, Exeter, Devon EX1 3PB, UK

⁵ Author to whom any correspondence should be addressed

E-mail: jo.house@bristol.ac.uk

Received 4 June 2008
Accepted 10 September 2008
Published 14 October 2008

Contents

• 1. Introduction: climate policy
• 2. Carbon dioxide and long-term climate impacts
• 3. The carbon cycle, feedbacks and scientific uncertainty
• 4. Methods summary
• 5. Results
• 6. Conclusions
• Acknowledgments
• A. Detailed methods
• References

1. Introduction: climate policy

The temperature increase due to human activity since pre-industrial times has been in the order of 0.8 °C (IPCC 2007). The Intergovernmental Panel on Climate Change (2007) projected an additional global warming of 1.1–6.4 °C for the 21st century based on greenhouse gas emissions scenarios (SRES) that intentionally exclude mitigation policy (Nakićenović and Swart 2000). The United Nations Framework Convention on Climate Change commits its signatories to achieve `... stabilization of greenhouse gas concentrations in the atmosphere at a level that would prevent dangerous anthropogenic interference with the climate system'. What constitutes `dangerous' climate change is difficult to determine and highly subjective as regional impacts, rate of change and ability to cope with change are highly variable (Schellnhuber et al 2006). The EU has adopted a target of limiting global warming to 2 °C above pre-industrial levels (European Council 2007). Some argue that this may already be unachievable or undesirable due to the costs of mitigation. Nonetheless, efforts are being made to negotiate international targets to limit climate change.

The Kyoto Protocol set emissions reduction targets for industrialized countries based on what could be achieved politically, rather than what would be needed for a desired outcome (Prins and Rayner 2007). While continuing negotiations seek to strengthen and extend the scope of Kyoto, several countries and organizations have formulated aspirational global greenhouse gas emissions goals (Weaver et al 2007). The powerful G8 (Group of 8)^Note6 nations in 2007 issued a non-binding aim `to at least half global emissions of CO₂ by 2050' (G8 2007); they did not specify what the ultimate goal was in terms of concentration or climate, or what should happen after 2050.

The Stern Review (Stern 2006) is more explicit, stating that `The risks of the worst impacts of climate change can be substantially reduced if greenhouse gas levels in the atmosphere can be stabilized between 450 and 550 ppm CO₂ equivalent. Stabilization in this range would require emissions to be at least 25% below current levels by 2050, and perhaps much more. Ultimately, stabilisation—at whatever level—requires that annual emissions be brought down to more than 80% below current levels.' The emissions reductions required to achieve these levels of stabilized concentration were derived from results in the IPCC Third Assessment Report in 2001 (Prentice et al 2001). While the 2001 model results incorporated ranges due to uncertainty in carbon cycle processes, climate sensitivity to CO₂ and climate impacts on the carbon cycle, models have since been updated and in particular there has been more explicit quantification of the feedbacks between the carbon cycle and climate. The modelled effect of climate–carbon cycle feedbacks imply substantially greater impacts for a given emissions trajectory, or lower allowable emissions to meet a given concentration or temperature target (Cox et al 2000, Matthews 2006, Friedlingstein et al 2006, Jones et al 2006a, 2006b, Plattner et al 2006). This paper explores the implications for the G8 and Stern emissions targets in terms of concentration and temperature change, spanning the range of uncertainty across state-of-the-art models. It also explores the implications for allowable emissions to achieve stabilization of atmospheric CO₂ at 550 ppm.

2. Carbon dioxide and long-term climate impacts

In this analysis we focus on carbon dioxide. CO₂ currently accounts for about two thirds of the radiative forcing produced by all greenhouse gases (IPCC 2007). The contribution of other GHGs has commonly been expressed in terms of `CO₂ equivalent' concentrations and emissions. In the latter case this means emissions are scaled relative to CO₂ according to their global warming potential over a given time horizon (commonly 100 years). However, this equivalence is not meaningful at other timescales because as much as a third of all CO₂ emitted remains in the atmosphere for thousands to tens of thousands of years; while the next most radiatively important greenhouse gasses are removed over much shorter time periods (e.g. 12 years for methane, 114 years for nitrous oxide) (Archer et al 1997, Archer 2005, Denman et al 2007, Forster et al 2007). Thus in the near term, manipulation of other greenhouse gases such as methane can influence the pathway of global warming, but in the long term it is the accumulated emissions of CO₂ that count (Matthews 2006, Meinshausen et al 2006, Denman et al 2007, den Elzen et al 2007).

The long-term legacy of CO₂ emissions is compounded by inertia in the climate system. Due to the mass and thermal capacity of the oceans and ice, and the slowness of heat transport processes, it takes a long time for the atmospheric temperature to fully respond to changes in radiative forcing. If CO₂ concentrations were stabilized on a timescale of 100 or so years, temperatures would still take several centuries to stabilize, sea level rise due to thermal expansion would take centuries to millennia, and sea level rise due to ice melting would take millennia (IPCC 2001). By contrast, climate policy typically considers climate impacts and mitigation targets on the timescale of decades up to a century.

3. The carbon cycle, feedbacks and scientific uncertainty

Less than half of the total cumulative anthropogenic CO₂ emitted due to fossil fuel burning and land use change (deforestation) has remained in the atmosphere, the rest has been taken up by the land and ocean (Prentice et al 2001). The major land and ocean carbon sinks active today are responsive to the raised atmospheric CO₂ levels. The ocean sink occurs due to the partial pressure difference of CO₂ between the atmosphere and ocean. CO₂ dissolves in the ocean surface waters and ocean circulation transfers it to depth. Plants take up CO₂ during photosynthesis and convert it into biomass. Plants release CO₂ during plant (autotrophic) respiration and decay (heterotrophic respiration). Raised concentrations of CO₂ stimulate additional growth in terrestrial plants (CO₂ fertilization effect), drawing down CO₂ from the atmosphere in a negative carbon cycle feedback (Prentice et al 2001, Norby et al 2005). The magnitude and long-term persistence of the CO₂ fertilization effect is uncertain, and differs in model projections (Friedlingstein et al 2006, Plattner et al 2006, Sitch et al 2008). Note that since increasing CO₂ levels stimulate CO₂ uptake in the ocean and on land, these sinks will continue to operate as long as atmospheric CO₂ concentration is rising and the land sink does not saturate. However, these sinks must both tend to zero as a stable CO₂ concentration is approached.

The land and ocean carbon sinks are also sensitive to climate change. Warming reduces the solubility of CO₂ in the ocean, reducing ocean uptake (positive feedback). Increasing temperatures are also likely to cause vertical stratification in the ocean, reducing transport to the deep ocean (positive feedback), and affecting biological productivity (sign of feedback uncertain). On land, warming increases the rate of heterotrophic respiration (positive feedback) (Prentice et al 2001, Knorr et al 2005). Changes in temperature and precipitation will have regionally specific effects on plant growth. For example, increasing plant growth due to longer growing seasons in northern high latitudes for moderate temperature increases (negative feedback), and reducing growth due to heat stress and drought in low to mid-latitude regions (positive feedback). Many models have indicated a net positive `climate–carbon cycle feedback' such that global warming drives a reduction in net CO₂ uptake (e.g. Cox et al 2000, Friedlingstein et al 2006, Plattner et al 2006, Sitch et al 2008). While the magnitude of this feedback varied considerably between studies, some indicated a very large effect that would have major implications for projecting climate change impacts, or indeed for calculating the level of anthropogenic emissions consistent with achieving stabilization targets.

Eleven state-of-art global climate models that participated in the IPCC (2007) assessment were coupled to carbon cycle models to study the magnitude of the climate–carbon cycle feedbacks in the Coupled Climate–Carbon Cycle Model Intercomparison Project (C⁴MIP) (Friedlingstein et al 2006). These models incorporated a range of climate sensitivities^Note7, one of the key uncertainties in climate modelling (Knutti et al 2008). The models represent a range of CO₂ fertilization strengths and other differences in carbon cycle processes reflecting uncertainty in the state of knowledge. Simulations were made with the climate response incorporated, and then suppressed, to isolate the responses of the modelled ocean and land carbon sinks to CO₂ increases alone. Under a particular CO₂ emissions scenario (IPCC SRES A2) (Nakićenović and Swart 2000), projected atmospheric concentration at 2100 was greater by 20–225 ppm relative to simulations without climate–carbon cycle feedback. This range equates to an additional temperature increase of 0.1–1.5 °C (Friedlingstein et al 2006). None of the models included in the C⁴MIP analysis combined high climate sensitivity with high climate–carbon cycle feedbacks, which would imply even greater warming.

4. Methods summary

In the present analysis, a simple coupled climate–carbon cycle model HadSCCM1 (Jones et al 2006a) was calibrated so as to emulate both the carbon cycle responses to climate and the climate sensitivity for the 11 fully coupled models participating in C⁴MIP. We used global-mean outputs from these simulations to calibrate the simple model for each of the coupled models. (For more information see the appendix.)

CO₂ emissions profiles were defined that provide a smooth transition to the G8 and Stern emissions reductions targets. Results are not strongly sensitive to the detailed time course of the emissions cuts (Jones et al 2006b, Matthews 2006).

5. Results

The `G8 scenario' (figure 1) cuts current (2007) global emissions by 50% by 2050, with emissions then held constant at this target level. Atmospheric CO₂ concentration continues to increase after 2050. By 2100 the CO₂ concentration is 470 to 590 ppm and the global-mean temperature is 1.3–3.1 °C above pre-industrial levels. By 2300 CO₂ concentration has risen as high as 640–980 ppm, temperature has risen by 2.2–5.7 °C and both continue to rise rapidly thereafter.

The `Stern scenario' (figure 2) cuts emissions by 25% by 2050, with progressive cuts thereafter, down to 80%. CO<sub>2</sub> concentrations in 2100 reach 480–620 ppm with a corresponding temperature change of 1.4–3.4 °C. These 2100 levels are slightly higher than the `G8 scenario' as emission reductions happen more slowly, showing that cutting early can be as important as cutting deeply for nearer term impacts. Nevertheless, once the final target of 80% has been achieved, the CO₂ concentration increase after 2100 is slight, reaching 510–700 ppm by 2300 with temperature approaching, but not quite achieving, stabilization between 1.6 and 4.2 °C. Thus in the longer term it is the depth of cuts (or total cumulative emissions) that is more important (Matthews 2006).

When emissions are set initially to follow the G8 scenario with a 50% cut in emissions by 2050, but then continue on the same trajectory to reach an 80% cut (not shown), a warming of less than 2 °C in 2100 is shown by all models.

What mitigation action is necessary to achieve stabilization at 550 ppm (Stern's upper bound) when taking account of the climate–carbon cycle feedback and other uncertainty ranges across the C⁴MIP models? Again using the simple model to emulate all 11 C⁴MIP models, we calculated allowable emissions to achieve the IPCC WRE550 stabilization pathway (Wigley et al 1996). Our analyses show that by 2100, cuts of up to 80% are necessary. Most models actual allow a somewhat slower near-term decline in emissions than Stern proposes. The need for progressive emissions reductions in the long term however continues beyond 2100, with required cuts reaching 81–90% by 2300.

The results of the 550 ppm stabilization scenario runs are consistent with results of modelling sensitivity studies that varied the size of the climate–carbon cycle feedback, climate sensitivity and other carbon cycle processes within models (Matthews 2006, Jones et al 2006b, Plattner et al 2006). These studies found substantial variation due to both the carbon cycle (e.g. CO₂ fertilization) and climate sensitivity (which compounds climate–carbon cycle feedback uncertainty). The results that the permissible emissions to achieve stabilization must be substantially reduced when climate–carbon cycle feedbacks are included, and that to achieve stabilization at any level emissions must be substantially below present levels, are robust across a range of models and model settings. The range across model emulations is similar to the ranges obtained by changing sensitivities to processes within models (Plattner et al 2006). Using simplified models tuned to reproduce the behaviour of more comprehensive, computationally intensive, carbon cycle and climate models is a robust tool for exploring a range of emissions scenarios and climate projections spanning the range of model variability.

6. Conclusions

On the timescale of decades to a century (the more common domain of climate-based policy making), Stern's proposal of 80% emissions cuts remains an effective near-term target on the pathway to achieve stabilization of CO₂ concentrations or climate. This conclusion is robust despite the large uncertainty in the climate–carbon cycle feedback. Ultimately, however, climate stabilization at any level can only be achieved if net global CO₂ emissions decline over centuries to the level of persistent natural sinks (, or just a few % of today's emissions) (Archer et al 1997, Prentice et al 2001, Caldeira et al 2003, Archer 2005, Lenton et al 2006, Denman et al 2007, Tyrrell et al 2007). On the timescale of centuries to millennia, over which the impacts of today's CO₂ emissions are still being felt, stabilization in the presence of a non-trivial anthropogenic source of CO₂ is only possible if this source is balanced by an artificial sink (Haugen and Eide 1996, Prentice et al 2001, Weaver et al 2007). The long-term impact of today's emissions brings this planetary timescale into contemporary policy relevance.

Acknowledgments

The contribution of JIH, ICP, WK and SEC was supported by the NERC QUEST programme. The contribution of CDJ, GRH and JAL was supported by the Joint Defra and MoD Integrated Climate Programme-(Defra) GA01101, (MoD) CBC/2B/0417_Annex C5. We thank Liz Loeffler for technical editing.

A. Detailed methods

The calibration of the land component of the box model Hadsccm1 followed Jones et al (2006b) with one extension: vegetation and soil carbon turnover rates now include linear dependence on the size of their respective carbon pools. The initial values of these global-mean vegetation and soil carbon pools, as diagnosed from the C⁴MIP runs, were prescribed as part of the calibration. The half-saturation constant for land net primary productivity (NPP) as a function of ambient CO₂ concentration was estimated from the C⁴MIP runs without carbon cycle feedback. Then the climate sensitivities of NPP and soil respiration could be derived from the fully coupled simulations: equations (1)–(3) of Jones et al (2006b). Oceanic drawdown of CO₂ was modelled using the impulse–response approach of Joos et al (1996), with the depth of the mixed layer fitted to reproduce the trajectory of the ocean carbon sink in each of the C⁴MIP models (see appendix of Huntingford et al 2004). Global warming due to CO₂ increase was defined in terms of an equilibrium climate sensitivity, T_2CO2 and oceanic thermal capacity c_p. T_2CO2 was estimated by calibration against C⁴MIP model outputs, except for the UMD model where the warming constraint was not sufficient to yield a single well-defined value and a value was adopted from the published literature.

References

Archer D 2005 Fate of fossil fuel CO₂ in geologic time J. Geophys. Res. 110 C09S05 
CrossRef

Archer D, Kheshgi H and MaierReimer E 1997 Multiple timescales for neutralization of fossil fuel CO₂ Geophys. Res. Lett. 24 405–8 
CrossRef

Caldeira K, Jain A K and Hoffert M I 2003 Climate sensitivity uncertainty and the need for energy without CO₂ emission Science 299 2052–4 
CrossRefPubMed

Cox P M, Betts R A, Jones C D, Spall S A and Totterdell I J 2000 Acceleration of global warming due to carbon-cycle feedbacks in a coupled climate model Nature 408 184–7 
CrossRefPubMed

den Elzen M, Meinshausen M and van Vuuren D 2007 Multi-gas emission envelopes to meet greenhouse gas concentration targets: costs versus certainty of limiting temperature increase Glob. Environ. Change 17 260–80 
CrossRef

Denman K L et al 2007 Couplings between changes in the climate system and biogeochemistry Climate Change 2007: The Physical Science Basis (Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change) ed S Solomonet al (Cambridge: Cambridge University Press) pp 499–587 

European Council 2007 Limiting Global Climate Change to 2 Degrees Celsius—The Way Ahead for 2020 and Beyond Communication from the Commission to the Council, the European Parliament, the European Economic and Social Committee and the Committee of the Regions 

Forster P et al 2007 Changes in atmospheric constituents and radiative forcing Climate Change 2007: The Physical Science Basis (Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change) ed S Solomonet al (Cambridge: Cambridge University Press) pp 129–234 

Friedlingstein P et al 2006 Climate-carbon cycle feedback analysis: results from the C⁴MIP model intercomparison J. Clim. 19 3337–53 
CrossRef

G8 2007 Growth and responsibility in the world economy 2007 (cited 29 November 2007) http://www.g-8.de/Webs/G8/EN/G8Summit/SummitDocuments/summit-documents.html 

Haugen H A and Eide L I 1996 CO₂ capture and disposal: the realism of large scale scenarios Energy Convers. Manage. 37 1061–66 
CrossRef

Houghton R A 2003 Revised estimates of the annual net flux of carbon to the atmosphere from changes in land use and land management 1850–2000 Tellus B 55 378–90 
CrossRef

Huntingford C, Harris P P, Gedney N, Cox P M, Betts R A, Marengo J A and Gash J H C 2004 Using a GCM analogue model to investigate the potential for Amazonian forest dieback Theor. Appl. Clim. 78 177–85 
CrossRef

IPCC 2001 Climate Change 2001: Synthesis Report (Contribution of Working Groups I, II and III to the Third Assessment Report of the Intergovernmental Panel on Climate Change) ed Robert T Watsonet al (Cambridge: Cambridge University Press) pp 323–31 

IPCC 2007 Climate Change 2007: The Physical Science Basis (Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change) ed S Solomonet al (Cambridge: Cambridge University Press) p 996 

Jones C D, Cox P M and Huntingford C 2006a Impact of climate–carbon cycle feedbacks on emissions scenarios to achieve stabilisation Avoiding Dangerous Climate Change ed H J Schellnhuberet al (Cambridge: Cambridge University Press) pp 323–31 

Jones C D, Cox P M and Huntingford C 2006b Climate-carbon cycle feedbacks under stabilization: uncertainty and observational constraints Tellus B 58 603–13 
CrossRef

Joos F, Bruno M, Fink R, Siegenthaler U, Stocker T F and LeQuéré C 1996 An efficient and accurate representation of complex oceanic and biospheric models of anthropogenic carbon uptake Tellus B 48 397–417 
CrossRef

Knorr W, Prentice I C, House J I and Holland E A 2005 Long-term sensitivity of soil carbon turnover to warming Nature 433 298–301 
CrossRefPubMed

Knutti R et al 2008 A review of uncertainties in global temperature projections over the twenty-first century J. Clim. 21 2651–63 
CrossRef

Lenton T M, Williamson M S, Edwards N R, Marsh R, Price A R, Ridgwell A J, Shepherd J G and Cox S J 2006 Millennial timescale carbon cycle and climate change in an efficient Earth system model Clim. Dyn. 26 687–711 
CrossRef

Matthews H D 2006 Emissions targets for CO₂ stabilization as modified by carbon cycle feedbacks Tellus B 58 591–602 
CrossRef

Meinshausen M, Hare B, Wigley T M L, Van Vuuren D, Den Elzen M G J and Swart R 2006 Multi-gas emissions pathways to meet climate targets Clim. Change 75 151–94 
CrossRef

2000 IPCC Special Report on Emissions Scenarios ed  N Nakićenović and R Swart (Cambridge: Cambridge University Press) 

Norby R J et al 2005 Forest response to elevated CO₂ is conserved across a broad range of productivity Proc. Natl Acad. Sci. 102 18052–56 
CrossRefPubMed

Plattner G-K et al 2006 Long-term climate commitments projected with climate-carbon cycle models J. Clim. 21 2721–51 
CrossRef

Prentice I C, Farquhar G D, Fasham M J R, Goulden M L, Heimann M, Jaramillo V J, Kheshgi H S, Quéré C L, Scholes R J and Wallace D W R 2001 The carbon cycle and atmospheric carbon dioxide Climate Change 2001: The Scientific Basis. (Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change) ed J T Houghtonet al (Cambridge: Cambridge University Press) pp 183–237 

Prins G and Rayner S 2007 Time to ditch Kyoto Nature 449 973–5 
CrossRefPubMed

2006 Avoiding Dangerous Climate Change ed  H J Schellnhuber, W Cramer, N Nakićenović, T Wigley and G Yohe (Cambridge: Cambridge University Press) p 392 

Sitch S et al 2008 Evaluation of the terrestrial carbon cycle, future plant geography and climate-carbon cycle feedbacks using five dynamic global vegetation models (DGVMs) Glob. Change Biol. 14 2015–39 
CrossRef

Stern N 2006 Stern Review on the Economics of Climate Change (Cambridge: Cambridge University Press) 

Tyrrell T, Shepherd J G and Castle S 2007 The long-term legacy of fossil fuels Tellus B 59 664–72 
CrossRef

Weaver A J, Zickfeld K, Montenegro A and Eby M 2007 Long term climate implications of 2050 emission reduction targets Geophys. Res. Lett. 34 L19703 
CrossRef

Wigley T M L, Richels R and Edmonds J A 1996 Economic and environmental choices in the stabilization of atmospheric CO₂ concentrations Nature 379 240–3 
CrossRef

Notes

Note6  The Group of Eight (G8) is an international forum for the governments of Canada, France, Germany, Italy, Japan, Russia, the United Kingdom, The United States and the European Union.

Note7   Climate sensitivity refers to the change in the annual mean global surface temperature for a given change in radiative forcing, usually referenced to the equilibrium change projected for a doubling of atmospheric CO₂ concentration. It is a major uncertainty in future climate projections.