Carbon cycle processes in the ENSEMBLES simulations Richard Betts, Stephen Sitch, Colin Prentice, Pierre Friedlingstein, Chris Jones, Debbie Hemming, Peter Cox Thanks to Camilla Mathison and colleagues in RT2A and RT6.1 © Crown copyright Met Office

Model projections of global warming with IPCC SRES emissions scenarios

IPCC (2007)

Vegetation absorbs and releases carbon CO 2

• “photosynthesis” absorbs CO2 from the atmosphere, and turns it into carbon in the living vegetation

Vegetation absorbs and releases carbon CO 2

• “photosynthesis” absorbs CO2 from the atmosphere, and turns it into carbon in the living vegetation

• The plant’s metabolism releases some back to the atmosphere • “plant respiration”

• Photosynthesis – respiration = Net Primary Productivity (NPP)

CO2

Vegetation absorbs and releases carbon CO 2

• “photosynthesis” absorbs CO2 from the atmosphere, and turns it into carbon in the living vegetation

CO2

• The plant’s metabolism releases some back to the atmosphere • “plant respiration”

• Dead matter (leaves etc) falls to soil • LARGE amounts of carbon stored in the soil

“litter”

Vegetation absorbs and releases carbon CO 2

• “photosynthesis” absorbs CO2 from the atmosphere, and turns it into carbon in the living vegetation

CO2

• The plant’s metabolism releases some back to the atmosphere • “plant respiration”

• Dead matter (leaves etc) falls to soil • LARGE amounts of carbon stored in the soil • Decomposed by bacteria/microbes and released as CO2 back to the atmosphere • “soil respiration”

“litter”

CO2

Balancing the carbon What we emit…

100

Balancing the carbon What we emit…

100

Must go somewhere

= 50

atmosphere

25

25

land

ocean

Balancing the carbon What we emit…

100

Must go somewhere

=

If these go down due to climate change… 50

atmosphere

25

25

land

ocean

Balancing the carbon What we emit…

Must go somewhere This must go up

100

=

If these go down due to climate change… 50

atmosphere

25

25

land

ocean

Balancing the carbon For given emissions, the carbon cycle determines:

100

=

- How much CO2 stays in atmosphere > 50 50

atmosphere

- Hence the degree of climate change

Let’s say we have a CO2 target to stabilise climate… This must go down to balance This is our fixed target e.g. 450 ppm 100

=

If these go down, as before, due to climate change…

50

atmosphere

25

25

land

ocean

Implied CO2 emissions: E1

Vegetation and soil carbon changes

HadCM3C: A1B 2070-2099 relative to 1860-1889 Kg C m-2

Vegetation and soil carbon changes

HadCM3C: E1 2070-2099 relative to 1860-1889 Kg C m-2

Difference in vegetation and soil carbon changes HadCM3C: A1B - E1 2070-2099 relative to 1860-1889 Kg C m-2

Vegetation and soil carbon changes

HadCM3C: E1 2070-2099 relative to 1860-1889 Kg C m-2

Previous assessment of uncertainties in carbon cycle feedbacks: C4MIP A2 emissions scenario

Friedlingstein et al (2006)

ENSEMBLES work: uncertainties in vegetation responses to a specific pattern of climate change • Dynamic Global Vegetation Model (DGVM) intercomparison • 5 DGVMs • Couple to “analogue” climate model • Climate change patterns scaled with global land mean temperature • Global thermal 2-box model for global land and ocean temperature • Includes interactive CO2 and ocean carbon cycle as well as DGVMs so allows for carbon cycle feedbacks • Global box model for ocean carbon uptake / release • Driven by emissions not concentrations • Run models with and without future radiative forcing of climate: diagnose feedback

DGVMs in intercomparison DGVM

Host GCM in ENSEMBLES Stream 2 runs

No. of vegetation Reference types

HyLand

n/a

3

Friend and White (2000)

LPJ

?

9

Sitch et al (2003)

ORCHIDEE

IPSL-LOOP

10

Krinner et al (2005)

SDGVM

n/a

7

Woodward and Lomas (2004)

TRIFFID

HadCM3C

5

Cox (2001)

Seasonal patterns of temperature change per unit global land warming

Patterns from HadCM3LC GCM (Cox et al, 2000)

Sitch et al, 2008

Seasonal patterns of precipitation change per unit global land warming

Patterns from HadCM3LC GCM (Cox et al, 2000)

Sitch et al, 2008

CO2 rise with 4 SRES emissions scenarios

A1FI A2 B2 B1

Sitch et al, 2008

CO2 concentrations in 2099

1400

CO2 (ppm) in 2099

1200 1000

HyLand LPJ

800

ORCHIDEE 600

SDGVM

400

TRIFFID

200 0 A1FI

Sitch et al, 2008

A2

B2

B1

Land carbon uptake over 21st Century

Cumulative land uptake (Pg C)

600 500 HyLand

400

LPJ 300

ORCHIDEE SDGVM

200

TRIFFID

100 0 A1FI

Sitch et al, 2008

A2

B2

B1

Change in vegetation carbon 2099-1860: A1FI

Vegetation carbon change (kg C m-2) Sitch et al, 2008

Changes in tree cover 2099-1860: A1FI

Sitch et al, 2008 % change in tree cover

Change in soil carbon 2099-1860: A1FI

Soil carbon change (kg C m-2)

Sitch et al, 2008

Comparing effects of climate and CO2 on ecosystems and hydrology

© Crown copyright Met Office

CO2 effects on land surface impacts in HadCM3 / HadSM3 • Photosynthesis responds to CO2 concentration and climate • Respiration responds to climate • Transpiration (uptake and evaporation of soil moisture through plants) also affected by plant responses to CO2 as well as climate • One of the parameters perturbed in HadCM3 ensemble was “on/off” switch for the direct CO2 effect • Opportunity to examine importance of this in context of climate uncertainties

Net Primary Productivity (NPP) in the HadSM3 perturbed physics ensemble (equilibrium 2×CO2): subset with CO2 fertilization included

a) b)

Mean (kg C m-2 y-1)

Standard deviation (kg C m-2 y-1)

NPP in the HadSM3 perturbed physics ensemble (equilibrium 2×CO2): impact of CO2 fertilization c) b) c)

No CO2 fertilization

NPP (kg C m-2 y-1)

CO2 fertilization alone

CO2 rise, climate change and the hydrological cycle Affected by climate change Precipitation Surface evaporation Transpiration

Also affected directly by changing CO2 concentration

Surface runoff Infiltration

Subsurface runoff

Plant physiological responses to CO2: impacts on runoff

50

RAD RADPHYS

40

30

20

10

120

Runoff change (Kg m2 yr-1)

Percentage of ensemble members

a

100

RAD RADPHYS

RADPHYS y = 0.58x + 20.82 R2 = 0.50

80

60 40

RAD y = 0.56x + 7.39 2 R = 0.59

20

0 -20

0

20

40

60

Runoff change (Kg

80

m-2

100

yr-1)

0 -20

0

20

40

60

80

100

Precipitation change (Kg m2 yr-1)

HadCM3 perturbed physics ensemble (equilibrium 2×CO2) c Figure 1

RAD: radiative forcing only RADPYS: radiative and physiological forcing

120

EUROPE

Plant responses to CO2: impacts on runoff • Physiological responses may partly offset increased risk of drought

Runoff change (Kg m-2 yr-1)

90

RAD RADPHYS

70

50

30

RAD y = 0.48x + 11.94 R2 = 0.58

10

-10

-30 -50

-30

10

30

50

70

90

SOUTH AMERICA

150

Runoff change (Kg m-2 yr-1)

• Hydrological impacts of a given level of warming depend on climate sensitivity and on mix of GHGs (CO2 vs. non-CO2)

-10

Precipitation change (Kg m-2 yr-1)

• Unlikely to alter flood risk – depends on extreme rainfall • Hydrological impacts studies should consider plant responses to CO2

RADPHYS y = 0.59x + 16.03 R2 = 0.66

100

RAD RADPHYS

RADPHYS y = 0.31x + 13.87 R2 = 0.46

50

0

RAD y = 0.51x - 2.07 R2 = 0.63

-50

-100

-150 -500

Figure 1

-400

-300

-200

-100

0

Precipitation change (Kg m-2 yr-1)

100

200

Conclusions • Negative emissions required by E1 scenario due to emissions of CO2 from land surface (mainly soils, also Amazon forest) • Response of vegetation and soil carbon to a given pattern of climate change can vary significantly between models • HadCM3C’s own ecosystem model is not the most extreme! • Regionally, sign of response can be different between models • Vegetation carbon balance depends mainly on competition between CO2 fertilization and climate effects • CO2 effects on plant physiology also affect hydrological impacts • Hydrological drought may be partly offset by CO2 effects • Hydrological impacts affected by climate sensitivity and CO2 vs nonCO2 GHGs