Transcript Carbon-nitrogen cycle coupling regulates climate

Critical needs for new
understanding of nutrient dynamics
in Earth System Models
Peter Thornton
Oak Ridge National Laboratory
Collaborators:
Gautam Bisht, Jiafu Mao, Xiaoying Shi, Forrest Hoffman,
Keith Lindsay, Scott Doney, Keith Moore, Natalie
Mahowald, Jim Randerson, Inez Fung, Jean-Francois
Lamarque, Johannes Feddema, Yen-Huei Lee
NASA GSFC, 22 Feb 2011
Key Uncertainties
• Nutrient limitation effect on CO2 fertilization
• Nutrient – climate interactions
– Is the “nitrogen as phosphorus proxy” hypothesis
useful in the tropics?
– Nutrient dynamics in a warming Arctic
• Mechanisms and time scales for plant nutrient
dynamics:
– Competition (with microbes and other plants)
– Uptake and storage (across days and seasons)
– Deployment
Nitrogen cycle
Carbon cycle
Atm CO2
photosynthesis
Internal
(fast)
External
(slow)
denitrification
N deposition
Plant
assimilation
respiration
litterfall & mortality
Litter / CWD
Soil Mineral
N
decomposition
mineralization
Soil Organic Matter
Thornton et al., 2009
N leaching
N fixation
Land carbon cycle sensitivity to
increasing atmospheric CO2
Offline CLM-CN
Fully-coupled CCSM3.1
C-only
C-N
high Ndep
low Ndep
Effect of C-N coupling is to increase atmospheric CO2 by about 150 ppm by 2100,
compared to previous model results
Thornton et al., 2007 (left), and Thornton et al., 2009 (right)
1980s
(TAR)
1990s
(AR4)
2000-2009
(AR4)
Atmospheric
increase
3.3 ± 0.1
3.2 ± 0.1
4.1 ± 0.1
Emissions
5.4 ± 0.3
6.4 ± 0.4
7.2 ± 0.3
Net ocean-toatm
-1.8 ± 0.8
-2.2 ± 0.4
-2.2 ± 0.5
Net land-to-atm
-0.3 ± 0.9
-1.0 ± 0.6
-0.9 ± 0.6
Land use flux
1.7
(0.6 to 2.5)
1.6
(0.5 to 2.7)
n.a.
Residual land
flux
-1.9
(-3.4 to 0.2)
-2.6
(-4.3 to -0.9)
n.a.
Land partitioning:
Global C-cycle component estimates from IPCC AR4, 2007
N avail. (index)
Influence of rising CO2 on NEE and N
availability
(CO2 – control)
Single and combined effects on NEE
LULCC
All combined
N dep
CO2
Shevliakova 2009 (LM3V model result)
Interaction effects for total land C
CxN
(3-way)
N x LULCC
C x LULCC
All effects
Land components of climate-carbon cycle feedback
low Ndep
high Ndep
• Effect of C-N coupling on gamma_land is to reduce atmospheric CO2 by
about 130 ppm by 2100, compared to previous model results
• Net climate-carbon cycle feedback gain (including ocean
response) is nearly neutral or negative, compared to positive
feedback for previous models.
Thornton et al., 2009
N dep
Preind.
Trans.
Prog.
CC
CC+Ndep
Fixed
Ctrl
Ndep
All simulations with
prescribed transient fossil
fuel emissions
warmer / wetter
Rad CO2
Lower N
Higher N
cooler / drier
N availability hypothesis
Higher due to N deposition
Does climate change mimic
the effects of increased N
deposition?
Higher due to climate
change
Higher due to deposition
and climate change
Climate-carbon cycle feedback
CO2-induced climate change (warmer and wetter) leads to
increased land carbon storage
ND effect
CC effect
• Both climate change (red curve) and anthropogenic nitrogen deposition (blue
curve) result in increased land carbon storage.
• Climate change producing uptake of carbon over tropics, opposite response
compared to previous (carbon-only) results.
Thornton et al., 2009
ND effect
CC effect
GPP
• GPP response is highly
correlated with gross N
mineralization
• Relationship between
GPP and N min is similar
for effects of climate
change and direct N
fertilization
(anthropogenic N
deposition).
Gross N mineralization
Thornton et al., 2009
ND effect
CC effect
• Increased N deposition causes increase in both SOM and vegetation
carbon stocks
• Radiatively-forced climate change causes a decline in SOM and an
increase in vegetation carbon stocks.
• Consistent with the hypothesis that increased GPP under climate
change is due to transfer of nitrogen from SOM to vegetation pools.
Thornton et al., 2009
• Does warming-induced carbon uptake in
the tropics make sense if the most limiting
nutrient is P instead of N?
Photosynthesis
C-N Coupling
Schematic
Potential GPP sets N demand
Plants and microbes
compete for N on basis of
relative demand
N Immobilization
Soil
Mineral
N
N Mineralization
Plant N uptake
GPP
downregulated
by N supply
CLM-CN, GPP
Multi-site comparison
Mid-summer mean diurnal cycle
Obs
Model
Original model: no plant N storage pool
Soil mineral N
Plant allocated N
GPP
obs
model
immob.
mineralization
0
6
12
hour
18
24
12
hour
18
24
Revised model: plant N storage pool
obs
model
N from storage  (demand, storage)
Soil mineral N
Plant allocated N
Pre-allocation
plant N storage
GPP
N to storage  (demand, availability)
0
6
Implications and Conclusions
• Additional empirical constraints are
required to reduce prediction uncertainty
– warming (x CO2?) x nutrient manipulations
• Tropical forest (areal extent, C stocks, C fluxes)
• Arctic tundra and boreal forest
• Brave new models
– Introduce the known important mechanisms
• Get the wrong answer for the right reasons
• … to eventually get the right answer