constraints from long-term observations on the future of

Constraints from long-term observations on the future of

the global carbon cycle

Martin HeimannMax-Planck-Institute for

Biogeochemistry, Jena, Germany

Anthropogenic Perturbation of the Global Carbon Cycle

Keeling et al., 2005, updated

90°N

Cs

Qn QsCn - Cs = (Qn - Qs) τ/2

90°S

Cn

0°

τ

Interhemispheric CO2 concentration gradient tracks fossil

fuel emissions

Key carbon cycle observations:

Radiocarbon (14C)

Radiocarbon as a Unique Tracer 71

The D14C depletion in the first half of the 20th century has been shown to be larger in more pollutedareas, such as Europe, than at the west coast of the United States (De Jong and Mook 1982). In par-ticular, highly populated areas show regional Suess effects up to dD14C = -100‰, especially duringwinter (Levin et al. 1989). A discussion of the use of 14C for regional studies in the context of veri-fying emission reductions (as agreed upon in the Kyoto Protocol), will be given below (RegionalSuess Effect section).

Figure 1 (a) Atmospheric CO2 increase derived from direct observations at Mauna Loa sta-tion (Keeling and Whorf 1999) and from analysis of air inclusions in Antarctic ice cores(Etheridge et al. 1996). (b) Anthropogenic CO2 emissions from fossil fuel burning, cementmanufacturing and gas flaring. (c) Temporal change of D14C in tree rings grown at the Pacificcoast: The D14C decrease is closely related to the increasing anthropogenic CO2 emissions andthe resulting increase in atmospheric CO2 mixing ratio. (Note: after 1955 the decreasing D14Ctrend ends due to the overwhelming effect of bomb 14C input into the atmosphere.)

Radiocarbon as a Unique Tracer 73

surface water (i.e. in polar regions), the rate of gas exchange becomes important for net transport ofexcess CO2 into the deep ocean (Siegenthaler 1983).

Both natural and bomb 14C have been successfully used in the past to derive the mean atmosphere/ocean gas exchange rate for major ocean basins (Stuiver 1980; Stuiver et al. 1981). The net 14Cinflux from the atmosphere to the surface ocean at a given time is proportional to the D14C gradient

Figure 2 (a) Long-term observations of D14C in atmospheric CO2 in the northern and in the south-ern hemisphere (Manning et al. 1990; Levin et al. 1992 and unpublished Heidelberg data). Shortlyafter the atmospheric test ban treaty in 1962, the 14CO2 level in the northern hemisphere was twiceas high as the natural 14CO2 level (defined as 0 in the D14C scale [Stuiver and Polach 1977]). Alsoincluded is the D14C level of DIC in surface ocean water between 17°N and 25°N derived fromannual banded coral rings (Druffel and Suess 1983; Druffel 1995) together with model calculationsfor the surface ocean by Hesshaimer (1997). (b) Temporal trend of the observed bomb 14C inven-tories of the stratosphere up to 30 km (Hesshaimer and Levin 2000), the troposphere (derived fromthe observations in [a]) the ocean (box-diffusion model estimates by Hesshaimer et al. [1994],tuned to fit the total inventory of Broecker et al. [1995]) and the biosphere (3-box compartmentmodel estimates by Hesshaimer et al. [1994]).

Levin and Hessheimer, 2001

Direct response of the global carbon cycle to the anthropogenic perturbation

CO2Atmosphere

Ocean

Landbiosphere

Emissions from burning of fossil

fuels and cement production

Changes in landuse and land

management

The Oceanic Sink forAnthropogenic CO2

Christopher L. Sabine,1* Richard A. Feely,1 Nicolas Gruber,2

Robert M. Key,3 Kitack Lee,4 John L. Bullister,1 Rik Wanninkhof,5

C. S. Wong,6 Douglas W. R. Wallace,7 Bronte Tilbrook,8

Frank J. Millero,9 Tsung-Hung Peng,5 Alexander Kozyr,10

Tsueno Ono,11 Aida F. Rios12

Using inorganic carbon measurements from an international survey effort in the1990s and a tracer-based separation technique, we estimate a global oceanicanthropogenic carbondioxide (CO2) sink for theperiod from1800 to1994of118!19petagramsof carbon. Theoceanic sink accounts for"48%of the total fossil-fueland cement-manufacturing emissions, implying that the terrestrial biosphere wasa net source of CO2 to the atmosphere of about 39! 28 petagrams of carbon forthis period. The current fraction of total anthropogenic CO2 emissions stored in theocean appears to be about one-third of the long-term potential.

Since the beginning of the industrial period inthe late 18th century, i.e., over the anthropo-cene (1), humankind has emitted large quan-tities of CO2 into the atmosphere, mainly as aresult of fossil-fuel burning, but also becauseof land-use practices, e.g., deforestation (2).Measurements and reconstructions of the at-mospheric CO2 history reveal, however, thatless than half of these emissions remain in theatmosphere (3). The anthropogenic CO2 thatdid not accumulate in the atmosphere musthave been taken up by the ocean, by the landbiosphere, or by a combination of both. The

relative roles of the ocean and land bio-sphere as sinks for anthropogenic CO2 overthe anthropocene are currently not known.Although the anthropogenic CO2 budgetfor the past two decades, i.e., the 1980s and1990s, has been investigated in detail (3),the estimates of the ocean sink have notbeen based on direct measurements ofchanges in the oceanic inventory of dis-solved inorganic carbon (DIC).

Recognizing the need to constrain the oce-anic uptake, transport, and storage of anthro-pogenic CO2 for the anthropocene and toprovide a baseline for future estimates ofoceanic CO2 uptake, two international oceanresearch programs, the World Ocean Circu-lation Experiment (WOCE) and the JointGlobal Ocean Flux Study (JGOFS), jointlyconducted a comprehensive survey of inor-ganic carbon distributions in the global oceanin the 1990s (4). After completion of the U.S.field program in 1998, a 5-year effort wasbegun to compile and rigorously quality-con-trol the U.S. and international data sets, in-

cluding a few pre-WOCE data sets in regionsthat were data limited (5). The final data setconsists of 9618 hydrographic stations col-lected on 95 cruises, which represents themost accurate and comprehensive view of theglobal ocean inorganic carbon distributionavailable (6). As individual basins were com-pleted, the ocean tracer–based #C* method(7) was used to separate the anthropogenicCO2 component from the measured DIC con-centrations (8–10). Here we synthesize theindividual ocean estimates to provide anocean data-constrained global estimate of thecumulative oceanic sink for anthropogenicCO2 for the period from "1800 to 1994 (11).Distribution and inventories of an-

thropogenic CO2 in the ocean. The objec-tively gridded individual sample estimateswere vertically integrated to produce thecolumn inventory map shown in Fig. 1.Because the global survey had limited datacoverage in the marginal basins and theArctic Ocean (north of 65°N), these areaswere excluded from the mapped regions.The cumulative oceanic anthropogenic CO2

sink in 1994, for the ocean region shown inFig. 1, is 106 ! 17 Pg C. Accounting forthe excluded regions, we estimate a globalanthropogenic CO2 sink of 118 ! 19 Pg C.The uncertainty in the total inventory isbased on uncertainties in the anthropogenicCO2 estimates and mapping errors (11).

Figure 1 shows that this anthropogenicCO2 is not evenly distributed throughout theoceans. The highest vertically integrated con-centrations are found in the North Atlantic.As a result, this ocean basin stores 23% of theglobal oceanic anthropogenic CO2, despitecovering only 15% of the global ocean area(table S1). By contrast, the Southern Oceansouth of 50°S has very low vertically inte-grated anthropogenic CO2 concentrations,containing only 9% of the global inventory.More than 40% of the global inventory isfound in the region between 50°S and 14°S

1National Oceanic and Atmospheric Administration(NOAA) Pacific Marine Environmental Laboratory, 7600Sand Point Way NE, Seattle, WA 98115, USA. 2Univer-sity of California–Los Angeles, Institute of Geophysicsand Planetary Physics and Department of Atmosphericand Oceanic Sciences, Los Angeles, CA 90095, USA.3Princeton University, Program in Atmospheric and Oce-anic Science, Forrestal Campus/Sayre Hall, Princeton, NJ08544, USA. 4Pohang University of Science and Tech-nology, San 31, Nam-gu, Hyoja-dong, Pohang 790-784,South Korea. 5NOAA Atlantic Oceanographic and Me-teorological Laboratory, 4301 Rickenbacker Causeway,Miami, FL 33149, USA. 6Institute of Ocean Sciences,Climate Chemistry Laboratory, Post Office Box 6000,Sidney, BC V8L 4B2, Canada. 7Forschungsbereich MarineBiogeochemie, Leibniz Institut fur Meereswissenschafte,an der Universitat Kiel, (IFM-GEOMAR), DusternbrookerWeg 20, D-24105 Kiel, Germany. 8Commonwealth Sci-entific and Industrial Research Organisation (CSIRO)Marine Research and Antarctic Climate and EcosystemCooperative Research Center, Hobart, Tasmania 7001,Australia. 9University of Miami, Rosenstiel School ofMarine and Atmospheric Science, Division of Marine andAtmospheric Sciences, 4600 Rickenbacker Causeway,Miami, FL 33149, USA. 10Carbon Dioxide InformationAnalysis Center, Oak Ridge National Laboratory, U.S.Department of Energy, Mail Stop 6335, Oak Ridge, TN37831–6335, USA. 11Frontier Research System forGlobal Change/Institute for Global Change Research,Sumitomo Hamamatsu-cho, Building 4F, 1-18-16Hamamatsutyo, Minato-ku, 105-0013, Japan. 12Institutode Investigaciones Marinas, Consejo Superior deInvestigationes Cientificas, c/Eduardo Cabello, 6, 36208Vigo, Spain.

*To whom correspondence should be addressed. E-mail: [email protected]

Fig. 1. Column inventory of anthropogenic CO2 in the ocean (mol m$2). High inventories are

associated with deep water formation in the North Atlantic and intermediate and mode waterformation between 30° and 50°S. Total inventory of shaded regions is 106 ! 17 Pg C.

R E S E A R C H A R T I C L E S

www.sciencemag.org SCIENCE VOL 305 16 JULY 2004 367Sabine et al., 2004

InventoriesGlobal: 118 PgCNorth Atlantic: 27 PgC

Estimated uptake rates (1980-2000):Global: 1.85 PgC yr-1

North Atlantc: 0.42 PgC yr-1

Observed anthropogenic CO2 inventory in word ocean(~ mid 1990s)

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

Q@8 R)($B+)%$"7

K>+ "1&)7 1)+B"731*1(& -./ $# ) +&1&7% (&)#%3#=>)+&# S% %" %2&O;NKRR8$ :$7"+?)7$1' +&#!"7#& %" ) #>44&7 $71+&)#& "6 %2&)%0"#!2&+$1 RKQ 1"71&7%+)%$"7 B* GT :Q@JP !!0'@ A2& 0"4&(D)# +>7 D$%2">% ) B$"("?$1)( !>0! )74 D$%2">% R)RK8 #&4$0&7%$7%&+)1%$"7@ A2& )#*0!%"%$1 )$+B"+7& 6+)1%$"7 :G8T' $# 1("#& %"%2& 5)(>& 6">74 6"+ %2& GHPJ O;NKRRG -./ :GUTV %2& #0)((4$E&+&71& 1)7 B& )%%+$B>%&4 %" %2& "7#&% "6 ) 7"7($7&)+ &E&1%4>& %" %2& 4$E&+&7% $0!>(#& #$W&# >#&4 6"+ %2& %D" 1)($B+)%$"7#'@A2& 7"7($7&)+ -./ 0"4&( D)# 12&1C&4L >#$7? 5)(>&# 6"+ %2&

12&0$1)( &=>$($B+$>0 1"7#%)7%# 1"++&#!"74$7? %" %2& !+&#&7%34)*?("B)(30&)7 %&0!&+)%>+&L )?)$7#% %2& 6>(( O;NKRR8$,# +&#!"7#&6"+ $0!>(#&# $7 D2$12 %2& !+&$74>#%+$)( )%0"#!2&+$1 RKQ D)# $731+&)#&4 B* GTL QITL GMMTL )74 8MMT :/$?@ G )74 A)B(& G'@/"+ %2& ()+?&#% $0!>(#&L %2& RKQ >!%)C& "6 %2& 7"7($7&)+ -./

0"4>(& $# ) 6&D !&+1&7% #("D&+ %2)7 $7 O;NKRR8$X %2& 7"7($7&)++&%)+4)%$"7 "6 %2& 1)+B"7 >!%)C& $# #($?2%(* "5&+&#%$0)%&4@ O"D3&5&+L #0)(( &++"+# $7 %2$# +)7?& )+& %" B& &<!&1%&4L )# %2& 7"7($7&)+)$+3#&) &<12)7?& $7 %2& 89 0"4&( $# #!)%$)((* 4&!&74&7% )74 1)77"%B& )11>+)%&(* #$0>()%&4 B* ) "7&34$0&7#$"7)( 0"4&( >#$7? "7(* )#$7?(& #&% "6 ?("B)(30&)7 12&0$#%+* )74 0$<&43()*&+ !)+)0&%&+#@A2& 4$+&1% +&4>1%$"7 $7 "1&)7$1 1)+B"7 >!%)C& B* #&) #>+6)1&

D)+0$7? %2+">?2 %&0!&+)%>+&3+&()%&4 12&0$#%+* 12)7?&# "7(* $#

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

Q@U A2& %&++&#%+$)( B$"#!2&+& 0"4>(&

; RKQ #$7C "6 +">?2(* Q \%R_*&)+ $7 %2& ?("B)( %&++&#%+$)( B$"3#!2&+& $# B&($&5&4 %" )!!+"<$0)%&(* 1"0!&7#)%& 1)+B"7 ("##&# 6+"04&6"+&#%)%$"7 )74 "%2&+ ()74 >#& 12)7?&#L 0)$7(* $7 %2& %+"!$1#@A2& )(("1)%$"7 $# )#1+$B&4 %" )11&(&+)%&4 !()7% ?+"D%2 4>& %" %2&+$#$7? RKQ 1"71&7%+)%$"7# :RKQ 6&+%$($W)%$"7' )74 7$%+)%& 6&+%$($W)3%$"7@ A2& &`1$&71* "6 %2& %&++&#%+$)( 1)+B"7 #$7C )74 %2& =>&#%$"7D2&%2&+ $% D$(( 1">7%&+)1% 6"##$(36>&( &0$##$"7# $7 %2& 6>%>+& $# 2"%(*4&B)%&4@ 9&#!$%& 7>0&+"># !)!&+# "7 %2$# %"!$1 $7 +&1&7% *&)+# :#&&&@?@ %2& +&5$&D "6 F12(&#$7?&+ GHH8V -ZRR GHHMV GHHIV A)7# &% )(@GHHMV a&&($7? )74 F2&+%W GHHQV /+$&4($7?#%&$7 &% )(@ GHHIV a&&($7?&% )(@ GHH^V F&((&+# &% )(@ GHH^V a7"++ GHHJV \)*(&+ )74 R()>##&7GHHJV Y""# )74 b+>7" GHHPV R()>##&7 &% )(@ GHHHV \)7"!"(#C$ &% )(@GHHP )74 0)7* "%2&+#'L %2& $##>& $# #%$(( 6)+ 6+"0 +&#"(5&4@ A&++&#3%+$)( B$"#!2&+& 0"4&(# "6 4$E&+&7% 1"0!(&<$%* )74 #!)%$)( +&#"(>3%$"7 2)5& B&&7 0)!!&4 "7%" RKQ >!%)C& $0!>(#& +&#!"7#& 6>71%$"70"4&(# "+ &=>$5)(&7% B"<3%*!& )7)("?>&# :N&*&+ &% )(@ GHHHVA2"0!#"7 )74 .)74&+#"7 GHHH'@c& )>?0&7%&4 ">+ "1&)7 1)+B"7 1*1(& 0"4>(& B* ) #$0!(&

6">+3B"< +&!+&#&7%)%$"7 "6 %2& %&++&#%+$)( B$"#!2&+& :F$&?&7%2)(&+)74 K&#12?&+ GHPJ'L %" )11">7% %" S+#% "+4&+ 6"+ 12)7?&# $7 %2&%&++&#%+$)( 1)+B"7 #%"+)?& >74&+ +$#$7? RKQ@ A2& ?("B)( %&++&#%+$)(7&% !+$0)+* !+"4>1%$"7 :dZZ' $# )##>0&4 %" B& !+"!"+%$"7)( %" %2&("?)+$%20 "6 %2& )%0"#!2&+$1 RKQ 1"71&7%+)%$"7L )74 %2& +&#!$+)3%"+* RKQ e>< B)1C $7%" %2& )%0"#!2&+& $# ($7&)+ $7 %2& 6">+ +&#3&+5"$+ 1"7%&7%# :Y""# &% )(@ GHH^V a$1C($?2%&+ &% )(@ GHHH'@ f$C&"%2&+ 1>++&7% %&++&#%+$)( 1)+B"7 1*1(& 0"4&(#L ">+ %&++&#%+$)( B$"3#!2&+& 0"4&( 7&?(&1%# %2& &E&1%# "6 ()743>#& 12)7?&# )74 "%2&+$7%&+6&+&71& )74 1"++&#!"74$7? ("##&# "6 B$"("?$1)( 4$5&+#$%* )74!+"4>1%$5$%*@ -% 7&?(&1%# 6>+%2&+0"+& 1"0!(&%& dZZ #)%>+)%$"7 &5&7)% 2$?2 RKQ (&5&(#L )74 $% 7&?(&1%# )7 )11&(&+)%&4 +&#!$+)%$5& +&%>+7"6 1)+B"7 %" %2& )%0"#!2&+& )# $# &<!&1%&4 $7 ) D)+0&+ 1($0)%&@ -% $#!"##$B(& %2)% %2&#& )44$%$"7)( 6&&4B)1C# D">(4 +&4>1& %2& %&++&#%+$)(RKQ >!%)C&@A2& )??+&?)%& 0"4&( "6 %2& %&++&#%+$)( 1)+B"7 1*1(& $# %>7&4 %"

0)%12 &#%$0)%&# "6 %2& %&++&#%+$)( 1)+B"7 #$7C 4>+$7? %2& GHPM#:F12$0&( &% )(@ GHHJ'@ c2&7 4+$5&7 B* &0$##$"7 #1&7)+$"#L $%#+&#!"7#& $7 %&++&#%+$)( RKQ >!%)C& D)# D&(( D$%2$7 %2& +)7?&"B%)$7&4 D$%2 1>++&7%L #!)%$)((* +&#"(5&4 0"4&(# :a$1C($?2%&+&% )(@ GHHH'@

!"#$ % d"7($7&)+ $0!>(#&3+&#!"7#& "6 %2& 7"7($7&)+ "1&)731)+B"71*1(& -./ 0"4&( :dKL !"#$% #$&'!' 1"0!)+&4 D$%2 $%# !)+&7% 89 0"4&(O;NNKR8$ :%"()%"()%*!+'%' %" $0!>(#&# $71+&)#$7? %2& !+&$74>#%+$)()%0"#!2&+$1 RKQ 1"7%&7% B* GT :#",'!('L GMMT :-'%$.-'L )74 8MMT:+$/+'!( 0.12''L +&#!&1%$5&(*@ K7(* %2& !&+%>+B)%$"7# )+& #2"D7L7"+0)($W&4 %" %2& $0!>(#& #$W&

&'()* % ;0!($%>4&# )74 %$0& 1"7#%)7%# 6"+ %2& "1&)7$1 RKQ>!%)C& -./ !"!! 1"0!>%&4 6+"0 ) (&)#%3#=>)+&# S% %" %2& O;3NKRR8$ +&#!"7#& %" ) #>44&7 GT $71+&)#& "6 $%# !+&$74>#%+$)()%0"#!2&+$1 RKQ :#&& /$?@ G'

#M #G !G g*h #Q !Q g*h #8 !8 g*h #U !U g*h

"MG M@G8Q M@8GG Q8^@I M@QI8 IH@IQ M@QMH GQ@GJ M@MHI G@QJG

GHQ O""## &% )(@X ; 7"7($7&)+ $0!>(#& +&#!"7#& 0"4&( "6 %2& 1">!(&4 1)+B"7 1*1(&31($0)%& #*#%&0 :d-RRF'

Years

~0.15 ~0.05

100kyr

Si + CaCO3 weathering

Ocean dominated uptake

Atmosphere fraction after pulse injection at t=0

time constant: ~30kyr

Archer, 2005Hoos et al., 2001

6 PgC

2000 PgC

Mid- and long-term response of carbon cycle to direct perturbation

}Effect of ocean chemistry

Stabilization of atmospheric CO2

No climate feedbacks

791

Chapter 10 Global Climate Projections

given scenario of CO2 emissions, a climate change that takes

into account the dynamic evolution of the Earth’s capacity to

absorb the CO2 perturbation.

Conversely, the climate-carbon cycle feedback will have an

impact on the estimate of the projected CO2 emissions leading

to stabilisation of atmospheric CO2 at a given level. The TAR

showed the range of future emissions for the Wigley, Richels and

Edmonds (WRE; Wigley et al., 1996) stabilisation concentration

scenarios, using different model parametrizations (including the

climate-carbon feedback, Joos et al., 2001; Kheshgi and Jain,

2003). However, the emission reduction due to this feedback

was not quantified. Similar to the C4MIP protocol, coupled and

uncoupled simulations have been recently performed in order to

specifically evaluate the impact of climate change on the future

CO2 emissions required to achieve stabilisation (Matthews,

2005; Jones et al., 2006). Figure 10.21 shows the emissions

required to achieve CO2 stabilisation for the stabilisation profiles

SP450, SP550, SP750 and SP1000 (SP450 refers to stabilisation

at a CO2 concentration of 450 ppm, etc.) as simulated by three

climate-carbon cycle models. As detailed above, the climate-

carbon cycle feedback reduces the land and ocean uptake of

CO2, leading to a reduction in the emissions compatible with

a given atmospheric CO2 stabilisation pathway. The higher

the stabilisation scenario, the larger the climate change, the

larger the impact on the carbon cycle, and hence the larger the

emission reduction relative to the case without climate-carbon

cycle feedback. For example, stabilising atmospheric CO2 at

450 ppm, which will likely result in a global equilibrium

warming of 1.4°C to 3.1°C, with a best guess of about 2.1°C,

would require a reduction of current annual greenhouse gas

emissions by 52 to 90% by 2100. Positive carbon cycle feedbacks

(i.e., reduced ocean and terrestrial carbon uptake caused by the

warming) reduce the total (cumulative) emissions over the 21st

century compatible with a stabilisation of CO2 concentration

at 450 ppm by 105 to 300 GtC relative to a hypothetical case

where the carbon cycle does not respond to temperature. The

uncertainty regarding the strength of the climate-carbon cycle

feedback highlighted in the C4MIP analysis is also evident in

Figure 10.21. For higher stabilisation scenarios such as SP550,

SP750 and SP1000, the larger warming (2.9°C, 4.3°C and 5.5°C,

respectively) requires an increasingly larger reduction (130 to

425 GtC, 160 to 500 GtC and 165 to 510 GtC, respectively) in

the cumulated compatible emissions.

The current uncertainty involving processes driving the land

and ocean carbon uptake will translate into an uncertainty in

the future emissions of CO2 required to achieve stabilisation. In

Figure 10.22, the carbon-cycle related uncertainty is addressed

using the BERN2.5CC carbon cycle EMIC (Joos et al., 2001;

Plattner et al., 2001; see Table 8.3 for model details) and the

series of S450 to SP1000 CO2 stabilisation scenarios. The range

of emission uncertainty was derived using identical assumptions

as made in the TAR, varying ocean transport parameters and

parametrizations describing the cycling of carbon through the

terrestrial biosphere. Results are thus very closely comparable,

and the small differences can be largely explained by the

different CO2 trajectories and the use of a dynamic ocean model

here compared to the TAR.

The model results confirm that for stabilisation of

atmospheric CO2, emissions need to be reduced well below year

2000 values in all scenarios. This is true for the full range of

simulations covering carbon cycle uncertainty, even including

the upper bound, which is based on rather extreme assumptions

of terrestrial carbon cycle processes.

Cumulative emissions for the period from 2000 to 2100 (to

2300) range between 596 GtC (933 GtC) for SP450, and 1,236

GtC (3,052 GtC) for SP1000. The emission uncertainty varies

SP550

SP750

SP1000

Atmo

sphe

ric C

O 2 (pp

m)Un

coup

led E

miss

ions

(PgC

yr-1

)Co

upled

Emi

ssion

s (P

gC yr

-1)

Emiss

ions R

educ

tion

(PgC

yr-1

)

1100

900

300

500

700

Year

2016

48

12

0

2016

48

12

0

8

6

0

2

4

1900 2000 2100 2200 2300

1900 2000 2100 2200 2300

1900 2000 2100 2200 2300

1900 2000 2100 2200 2300

UVic EMICHadley SMBERN2.5CC

a)

b)

c)

d)

SP450

Figure 10.21. (a) Atmospheric CO2 stabilisation scenarios SP1000 (red), SP750 (blue), SP550 (green) and SP450 (black). (b) Compatible annual emissions calculated by three models, the Hadley simple model (Jones et al., 2006; solid), the UVic EMIC (Matthews, 2005; dashed) and the BERN2.5CC EMIC (Joos et al., 2001; Plattner et al., 2001; triangles) for the three stabilisation scenarios without accounting for the impact of climate on the carbon cycle (see Table 8.3 for details of the latter two models). (c) As for (b) but with the climate impact on the carbon cycle accounted for. (d) The difference between (b) and (c) showing the impact of the climate-carbon cycle feedback on the calculation of compatible emissions.

791

Chapter 10 Global Climate Projections

given scenario of CO2 emissions, a climate change that takes

into account the dynamic evolution of the Earth’s capacity to

absorb the CO2 perturbation.

Conversely, the climate-carbon cycle feedback will have an

impact on the estimate of the projected CO2 emissions leading

to stabilisation of atmospheric CO2 at a given level. The TAR

showed the range of future emissions for the Wigley, Richels and

Edmonds (WRE; Wigley et al., 1996) stabilisation concentration

scenarios, using different model parametrizations (including the

climate-carbon feedback, Joos et al., 2001; Kheshgi and Jain,

2003). However, the emission reduction due to this feedback

was not quantified. Similar to the C4MIP protocol, coupled and

uncoupled simulations have been recently performed in order to

specifically evaluate the impact of climate change on the future

CO2 emissions required to achieve stabilisation (Matthews,

2005; Jones et al., 2006). Figure 10.21 shows the emissions

required to achieve CO2 stabilisation for the stabilisation profiles

SP450, SP550, SP750 and SP1000 (SP450 refers to stabilisation

at a CO2 concentration of 450 ppm, etc.) as simulated by three

climate-carbon cycle models. As detailed above, the climate-

carbon cycle feedback reduces the land and ocean uptake of

CO2, leading to a reduction in the emissions compatible with

a given atmospheric CO2 stabilisation pathway. The higher

the stabilisation scenario, the larger the climate change, the

larger the impact on the carbon cycle, and hence the larger the

emission reduction relative to the case without climate-carbon

cycle feedback. For example, stabilising atmospheric CO2 at

450 ppm, which will likely result in a global equilibrium

warming of 1.4°C to 3.1°C, with a best guess of about 2.1°C,

would require a reduction of current annual greenhouse gas

emissions by 52 to 90% by 2100. Positive carbon cycle feedbacks

(i.e., reduced ocean and terrestrial carbon uptake caused by the

warming) reduce the total (cumulative) emissions over the 21st

century compatible with a stabilisation of CO2 concentration

at 450 ppm by 105 to 300 GtC relative to a hypothetical case

where the carbon cycle does not respond to temperature. The

uncertainty regarding the strength of the climate-carbon cycle

feedback highlighted in the C4MIP analysis is also evident in

Figure 10.21. For higher stabilisation scenarios such as SP550,

SP750 and SP1000, the larger warming (2.9°C, 4.3°C and 5.5°C,

respectively) requires an increasingly larger reduction (130 to

425 GtC, 160 to 500 GtC and 165 to 510 GtC, respectively) in

the cumulated compatible emissions.

The current uncertainty involving processes driving the land

and ocean carbon uptake will translate into an uncertainty in

the future emissions of CO2 required to achieve stabilisation. In

Figure 10.22, the carbon-cycle related uncertainty is addressed

using the BERN2.5CC carbon cycle EMIC (Joos et al., 2001;

Plattner et al., 2001; see Table 8.3 for model details) and the

series of S450 to SP1000 CO2 stabilisation scenarios. The range

of emission uncertainty was derived using identical assumptions

as made in the TAR, varying ocean transport parameters and

parametrizations describing the cycling of carbon through the

terrestrial biosphere. Results are thus very closely comparable,

and the small differences can be largely explained by the

different CO2 trajectories and the use of a dynamic ocean model

here compared to the TAR.

The model results confirm that for stabilisation of

atmospheric CO2, emissions need to be reduced well below year

2000 values in all scenarios. This is true for the full range of

simulations covering carbon cycle uncertainty, even including

the upper bound, which is based on rather extreme assumptions

of terrestrial carbon cycle processes.

Cumulative emissions for the period from 2000 to 2100 (to

2300) range between 596 GtC (933 GtC) for SP450, and 1,236

GtC (3,052 GtC) for SP1000. The emission uncertainty varies

SP550

SP750

SP1000

Atmo

sphe

ric C

O 2 (pp

m)Un

coup

led E

miss

ions

(PgC

yr-1

)Co

upled

Emi

ssion

s (P

gC yr

-1)

Emiss

ions R

educ

tion

(PgC

yr-1

)

1100

900

300

500

700

Year

2016

48

12

0

2016

48

12

0

8

6

0

2

4

1900 2000 2100 2200 2300

1900 2000 2100 2200 2300

1900 2000 2100 2200 2300

1900 2000 2100 2200 2300

UVic EMICHadley SMBERN2.5CC

a)

b)

c)

d)

SP450

Figure 10.21. (a) Atmospheric CO2 stabilisation scenarios SP1000 (red), SP750 (blue), SP550 (green) and SP450 (black). (b) Compatible annual emissions calculated by three models, the Hadley simple model (Jones et al., 2006; solid), the UVic EMIC (Matthews, 2005; dashed) and the BERN2.5CC EMIC (Joos et al., 2001; Plattner et al., 2001; triangles) for the three stabilisation scenarios without accounting for the impact of climate on the carbon cycle (see Table 8.3 for details of the latter two models). (c) As for (b) but with the climate impact on the carbon cycle accounted for. (d) The difference between (b) and (c) showing the impact of the climate-carbon cycle feedback on the calculation of compatible emissions.

IPCC AR4, 2007 Siegenthaler and Oeschger, 1978

CO2Atmosphere

Ocean

Landbiosphere

Climate

Emissions from burning of fossil

fuels and cement production

Changes in landuse and land

management

Carbon cycle - climate system feedbacks

Basic terrestrial carbon cycle - climate system feedbacks

Basic terrestrial carbon cycle - climate system feedbacks

Air

Temperature

Soil carbon

stocks

Heterotrophic

respiration

Growing

season length

Photosynthesis

Vegetation

carbon stocks

++

CO2

concentration+

-+++

+-

+

+

520 PgC). Therefore, the difference in the atmosphericCO2 concentration between both runs, and hence the

positive climate–carbon cycle feedback is mostly driven by

the warming of 5.8 K over land areas (Fig. 1) and itsinfluence on the carbon uptake there. In the year 2100 this

amounts to 655 PgC in the uncoupled run and is reduced to

484 PgC in the coupled run.The regional distribution of the difference in net surface

carbon flux between the two simulations is illustrated in

Fig. 6 for the end of the twenty-first century. The NorthAtlantic Ocean takes up less carbon in the coupled run,

because deep water formation in the Labrador Sea and

Greenland Sea decreases and the Thermohaline Circulation(THC) slows down from 21 to 15 Sv (1 Sv = 106 m3/s).

This is partly offset by an increasing carbon flux into the

ocean in areas where sea ice melts. Globally the impact ofthe greenhouse warming on the oceanic sink is limited, as

the finite buffer capacity of the mixed-layer restricts the

carbon uptake in both runs and the decline of the THC sets

in too late to considerably reduce the carbon uptake duringthe whole coupled simulation. Interestingly, most C4MIP

climate–carbon cycle models with an explicit ocean cir-

culation show a similar sensitivity of the ocean carbonuptake to global warming (Friedlingstein et al. 2006),

which may be due to the same fundamental mechanism.

Figure 6 also shows, that the land biosphere stores morecarbon at high latitudes during the model simulation with

greenhouse warming than in the one without. By contrast,the carbon uptake of the tropical land areas is much smaller

under climate change conditions, than in the uncoupled

case. This opposite response of the cold and the warmterrestrial biosphere to global warming has also been re-

ported by Zeng et al. (2004). It is considered here in more

detail by regarding gross primary production (GPP) and netprimary production (NPP) in the latitudinal bands 30!S–30!N (tropics) and 30!N–90!N (northern temperate/boreal)

as displayed in Fig. 7.During the uncoupled simulation GPP increases by

about 50% in both regions, reflecting the strong CO2 fer-

tilization inherent to the photosynthesis scheme employed.The enhancement of NPP is even more pronounced. It

amounts to 75% in the temperate/boreal range and to 100%

in the tropics. The different growth rates of GPP and NPPcan be explained by the way elevated atmospheric CO2 acts

on vegetation. It raises GPP, but it does not significantly

affect maintenance respiration (Long et al. 2004), if apossible downregulation of photosynthetic capacity is not

considered as in our model. Therefore, the simulated

absolute increase in NPP by CO2 fertilization is generallyabout 80% of that in GPP. The growth of NPP relative to

its pre-industrial value thus depends on the pre-industrial

ratio of NPP to GPP. This is smaller for tropical than fortemperate/boreal vegetation due to high plant respiration

costs at high temperature. Consequently, the relative in-

crease in NPP is particularly pronounced in the tropics.Additionally, NPP is enhanced indirectly by a higher soil

moisture level and less water stress due to stomatal closure

under elevated atmospheric CO2. Again the tropical andsubtropical areas profit more as water stress is more fre-

quent and severe there than in the temperate/boreal zone.

Overall, these mechanisms amplify the dominant role ofthe tropics for global NPP under elevated atmospheric CO2

and result in a sizable carbon uptake of northern cold

ecosystems (179 PgC) and a vigorous carbon uptake of thetropics (445 PgC) during the simulation without global

warming.

In the coupled run the vegetation of the temperate/borealregion profits from the warmer climate—GPP and NPP

almost double during the simulated period. By contrast,

GPP in the tropics is not affected by the change in climate.This is partly a result of the only small increase in GPP

Fig. 5 Accumulation of anthropogenic carbon on land (solid line), inthe atmosphere (dashed line) and in the ocean (dotted line) during thecoupled simulation (red) and the uncoupled simulation (blue)

Fig. 6 Difference in uptake of anthropogenic C between the coupledsimulation and the uncoupled simulation in the period 2070–2100.Regions with negative (positive) values take up less (more) carbonunder global warming conditions and contribute to a positive(negative) climate–carbon cycle feedback

570 T. J. Raddatz et al.: Will the tropical land biosphere dominate the climate–carbon cycle feedback during the twenty-first century?

123

Model simulation of coupled carbon cycle - climate system

520 PgC). Therefore, the difference in the atmosphericCO2 concentration between both runs, and hence the

positive climate–carbon cycle feedback is mostly driven by

the warming of 5.8 K over land areas (Fig. 1) and itsinfluence on the carbon uptake there. In the year 2100 this

amounts to 655 PgC in the uncoupled run and is reduced to

484 PgC in the coupled run.The regional distribution of the difference in net surface

carbon flux between the two simulations is illustrated in

Fig. 6 for the end of the twenty-first century. The NorthAtlantic Ocean takes up less carbon in the coupled run,

because deep water formation in the Labrador Sea and

Greenland Sea decreases and the Thermohaline Circulation(THC) slows down from 21 to 15 Sv (1 Sv = 106 m3/s).

This is partly offset by an increasing carbon flux into the

ocean in areas where sea ice melts. Globally the impact ofthe greenhouse warming on the oceanic sink is limited, as

the finite buffer capacity of the mixed-layer restricts the

carbon uptake in both runs and the decline of the THC sets

in too late to considerably reduce the carbon uptake duringthe whole coupled simulation. Interestingly, most C4MIP

climate–carbon cycle models with an explicit ocean cir-

culation show a similar sensitivity of the ocean carbonuptake to global warming (Friedlingstein et al. 2006),

which may be due to the same fundamental mechanism.

Figure 6 also shows, that the land biosphere stores morecarbon at high latitudes during the model simulation with

greenhouse warming than in the one without. By contrast,the carbon uptake of the tropical land areas is much smaller

under climate change conditions, than in the uncoupled

case. This opposite response of the cold and the warmterrestrial biosphere to global warming has also been re-

ported by Zeng et al. (2004). It is considered here in more

detail by regarding gross primary production (GPP) and netprimary production (NPP) in the latitudinal bands 30!S–30!N (tropics) and 30!N–90!N (northern temperate/boreal)

as displayed in Fig. 7.During the uncoupled simulation GPP increases by

about 50% in both regions, reflecting the strong CO2 fer-

tilization inherent to the photosynthesis scheme employed.The enhancement of NPP is even more pronounced. It

amounts to 75% in the temperate/boreal range and to 100%

in the tropics. The different growth rates of GPP and NPPcan be explained by the way elevated atmospheric CO2 acts

on vegetation. It raises GPP, but it does not significantly

affect maintenance respiration (Long et al. 2004), if apossible downregulation of photosynthetic capacity is not

considered as in our model. Therefore, the simulated

absolute increase in NPP by CO2 fertilization is generallyabout 80% of that in GPP. The growth of NPP relative to

its pre-industrial value thus depends on the pre-industrial

ratio of NPP to GPP. This is smaller for tropical than fortemperate/boreal vegetation due to high plant respiration

costs at high temperature. Consequently, the relative in-

crease in NPP is particularly pronounced in the tropics.Additionally, NPP is enhanced indirectly by a higher soil

moisture level and less water stress due to stomatal closure

under elevated atmospheric CO2. Again the tropical andsubtropical areas profit more as water stress is more fre-

quent and severe there than in the temperate/boreal zone.

Overall, these mechanisms amplify the dominant role ofthe tropics for global NPP under elevated atmospheric CO2

and result in a sizable carbon uptake of northern cold

ecosystems (179 PgC) and a vigorous carbon uptake of thetropics (445 PgC) during the simulation without global

warming.

In the coupled run the vegetation of the temperate/borealregion profits from the warmer climate—GPP and NPP

almost double during the simulated period. By contrast,

GPP in the tropics is not affected by the change in climate.This is partly a result of the only small increase in GPP

Fig. 5 Accumulation of anthropogenic carbon on land (solid line), inthe atmosphere (dashed line) and in the ocean (dotted line) during thecoupled simulation (red) and the uncoupled simulation (blue)

Fig. 6 Difference in uptake of anthropogenic C between the coupledsimulation and the uncoupled simulation in the period 2070–2100.Regions with negative (positive) values take up less (more) carbonunder global warming conditions and contribute to a positive(negative) climate–carbon cycle feedback

570 T. J. Raddatz et al.: Will the tropical land biosphere dominate the climate–carbon cycle feedback during the twenty-first century?

123

Raddatz et al., 2006

Climate feedback: atmospheric land biosphere

carbon content

Simulated global carbon stock increases in atmosphere, ocean and land biosphere

PgC

Carbon uptake by land and ocean Difference coupled - uncoupled simulation (2070-2100)

kgC m-2

Year

C4MIP Simulations, Friedlingstein et al., 2006

11 models, historical emissions after 2000: SRES-A2 emission profile

Coupled Carbon Cycle - Climate Model Simulation

Experiments (C4MIP):

Predicted atmospheric CO2 concentration

!!!!!!!!!!!!!!!!!!!!

!!!!!!!!!!!!!!!

!!!!!!!!!!!!!

1950 2000 2050300

400

500

600

700

800

ppm

1950 2000 20500

20

40

60

80

100

120

ppm

Observations (MLO+SPO)

Decadal averages, smoothed

Difference coupled-uncoupled

simulation

C4MIP simulations:Reproduction of atmospheric CO2

increase

! ! !! ! ! !

! ! !! ! !

!! ! !

! !! !

! !! !

! !! !

! !! !

! !! !

! !! !

! !!! !

!!

1960 1970 1980 1990 2000300

320

340

360

380

400

420

! ! !! ! ! !

! ! !! ! !

!! ! !

! !! !

! !! !

! !! !

! !! !

! !! !

! !! !

! !!! !

!!

1960 1970 1980 1990 2000300

320

340

360

380

400

420

Coupled simulation

Uncoupled simulation


• 11 models, • historical CO2 emissions

from fossil fuels and land use change

ppm

ppm



Coupled Carbon Cycle - Climate Model

Simulation Experiments (C4MIP)


11 models, SRES-A2 emission profile

1950 2000 2050

0

2

4

6

8

10

PgCyr!1

1950 2000 20500

2

4

6

8

PgCyr!1

Global land uptake

Global ocean uptake

Sabine et al. 2004


Coupled Carbon Cycle - Climate Model Simulation

Experiments (C4MIP):

Climate feedback effects on global uptake by land

and ocean


11 models, SRES-A2 emission profile

Land

Ocean

1950 2000 2050

!6

!4

!2

0

PgCyr!1

1950 2000 2050

!1.0

!0.5

0.0

0.5

1.0

PgCyr!1


∆Natm

Qemiss

Decadal average + s.d.Annual estimates

Observed airborne fraction:

blue: uncoupled simulationred: coupled simulation

HADLEY

MPI

When can we expect to see a clear climate

feedback signal on global carbon cycle?

Decadal average airborne fraction simulated by

C4MIP models

Global CO2 budget over the next 100 years: Based on C4MIP results

Total uptake by land and ocean

A2-SRESEmissions


1950 2000 20500

5

10

15

20

PgCyr!1

Climate feedbacks


Carbon cycle in the 21st century:Lessons from C4MIP simulation experiments

• Ocean:

• Uncertainty due to different mixing and circulation characteristics

• Relatively small climate feedback

• Land:

• Models assume substantial “CO2 fertilization”: Effective

• Strong climate feedback

• Carbon cycle - climate feedback gain, range of C4MIP models:

• 4 - 20% (10 models),

• 31% (HadCM3LC)

β =

∆NPP

NPP0

∆C

C0

= 0.2 − 0.6

Climate feedbacks: Implications for

atmospheric CO2 stabilization

CARBON CYCLE FEEDBACKS AND STABILIZATION 607

WRE 450

1900 2000 2100 2200 2300Year

0

5

10

15

emis

sion

s (G

tC y

r)

WRE 550

1900 2000 2100 2200 2300Year

0

5

10

15

emis

sion

s (G

tC y

r)

Fig. 2. Observationally constrained permissible emissions forWRE450 and WRE550 scenarios. Only parameter combinations whichare consistent with historical emissions are chosen. The central whiteline represents the simple model results when run with the best fitparameters to HadCM3LC. The central black shading shows theuncertainty range for a climate sensitivity of 3 K (as fitted toHadCM3LC). The grey shading shows the range for climate sensitivityin the IPCC-TAR range of 1.5–4.5 K. The dashed lines show theemissions profiles from Wigley et al. (1996).

reductions in permissible emissions compared with the case ofno feedbacks.

The very long-term limit to which permissible emissions ap-proach is determined by the persistent natural sinks (Prenticeet al., 2001) such as transport of anthropogenic carbon to thedeep ocean. Since we are not perturbing ocean uptake behaviourin this study, our simulations begin to converge after about 2150.

The uncertainty bounds due to carbon cycle parameter uncer-tainty are themselves sensitive to the degree of climate change,and are much greater at higher climate sensitivities. This can beseen in Fig. 3 which shows uncertainty in the total cumulativeemissions from 2000 to 2300 as a function of climate sensitiv-ity. The solid lines correspond to simulations with the carboncycle parameters fitted to HadCM3LC (as given above), and the

WRE450

0 2 4 6 8 10Climate sensitivity (K)

0

400

800

1200

1600

2 4 6 8stabilisation temperature rise (K)

WRE550

0 2 4 6 8 10Climate sensitivity (K)

0

400

800

1200

1600

2 4 6 8 10stabilisation temperature rise (K)

Fig. 3. Sensitivity of cumulative permissible CO2 emissions from2000 to 2300 to variations in the climate sensitivity. Results are shownfor stabilization following WRE450 (top panel) and WRE550 (bottompanel) scenario. The solid lines show a strong decrease in cumulativeemissions for increasing climate sensitivity when carbon cyclefeedbacks are considered. Dot–dashed lines show values from Wigleyet al. (1996) when these feedbacks are neglected. The shaded regionsshow the range of uncertainty when the carbon cycle parameters arevaried, but constrained by the historical emissions. For reference, theglobal temperature change by 2300 is marked on a separate axis.

dot–dashed lines show the corresponding WRE values. Shadedareas show the range of cumulative emissions when the carboncycle parameters are varied, but constrained as above by com-parison with historical emissions. For low climate sensitivitythe carbon cycle parameters produce little spread because theyare responding to much lower increases in global temperature.For high climate sensitivity (and hence much greater tempera-ture rises), climate feedbacks have a stronger influence on thecarbon cycle, and thus parameter uncertainty in the latter is re-flected in larger uncertainty bounds on cumulative emissions.Figure 3 shows that for some carbon cycle parameters consistentwith data from the contemporary period and at climate sensi-tivities towards the upper end of the range of the IPCC-TAR,stabilization at 450 ppm could only be achieved by engineered

Tellus 58B (2006), 5

Stabilization target: 450ppm

Stabilization target: 550ppm

Uncertainty range from climate sensitivity1.5 - 4.5 K

Uncertainty range from climate sensitivity + carbon cycle model parameters

Emission profiles of Wigley, 1996(no feedbacks)

Standard HadCM3LC climate sensitivity: 3 K

}

}

Jones et al., 2006

C-POOLS AT RISK IN THE 21st CENTURY

Magnitude of Vulnerability [Pg C]

10 1000100

highlyunlikely

unlikely

likely

very likely

Figure 2

LAN

D


10 1000100

highlyunlikely

unlikely

likely

very likely

OC

EA

N

Permafrost

Soil carbon

Terrestrial Biomass

Wetlands & Peatlands

Solubility pump poolSoft-tissue pump

pool

Carbonate pump pool

Methane hydrates

C-POOLS AT RISK IN THE 21st CENTURY


10 1000100

highlyunlikely

unlikely

likely

very likely

Figure 2

LAN

D


10 1000100

highlyunlikely

unlikely

likely

very likely

OC

EA

NPermafrost

Soil carbon

Terrestrial Biomass

Wetlands & Peatlands

Solubility pump poolSoft-tissue pump

pool

Carbonate pump pool

Methane hydrates

PgC PgC

Gruber et al., 2006

1 PgC release ⇒

~0.25 ppm atmospheric CO2 increase (100yr time scale)

Vulnerability of carbon pools (100yr time scale)

Conclusions• Currently observed (~linear) dynamics will change

• Present records do not yet exhibit enough information for quantification or validation of non-linear dynamics

• Current models exhibit still large differences -> Indication of insufficient process knowledge

• Many vulnerable pools and biogeochemical processes not yet represented in current Earth system models (a.o. permafrost, wetlands, fire, nutrients, ozone, CH4,...)

• Effects of changes in land use and management not yet included

• Comprehensive assessment: Biogeochemical + biophysical feedbacks!

• 100yr time scale carbon cycle - climate feedbacks: positive, ~20% effect

Thank you

constraints from long-term observations on the future of

Documents