Nitrogen limitation on land: How can
it occur in Earth System Models?
R. Quinn Thomas
Forest Resources and Environmental Conservation
Virginia Tech
Blacksburg, VA
What is nitrogen limitation?
Biogeochemistry
13: 87-115,199l
0 199 1 Kluwer Academic Publishers. Printed in the Netherlands.
Nitrogen limitation on land and in the sea:
How can it occur?
PETER M. VITOUSEK’
& ROBERT W. HOWARTH*
’ Department of biological Sciences, Stanford University, Stanford, CA 94305, USA
2 Section of Ecology and Systematics, Cornell University, Ithaca, NY 14853, USA
Received 1 October 1990; accepted 18 February 199 1
What is nitrogen limitation?
P
N
What is nitrogen limitation?
P
N
78%
Why is nitrogen limiting?
Why is nitrogen limiting?
• Essential element for growth
Why is nitrogen limiting?
• Essential element for growth
• Necessary for photosynthesis and carbon uptake
Why is nitrogen limiting?
• Essential element for growth
• Necessary for photosynthesis and carbon uptake
• Easily lost from ecosystems through disturbance
Why is nitrogen limiting?
• Essential element for growth
• Necessary for photosynthesis and carbon uptake
• Easily lost from ecosystems through disturbance
• Energetically costly to fix nitrogen from atmosphere
Why is nitrogen limiting?
• Essential element for growth
• Necessary for photosynthesis and carbon uptake
• Easily lost from ecosystems through disturbance
• Energetically costly to fix nitrogen from atmosphere
Boreal forests
Temperate forests
Early successional
tropical forests
Why is nitrogen limiting?
• Essential element for growth
• Necessary for photosynthesis and carbon uptake
• Easily lost from ecosystems through disturbance
• Energetically costly to fix nitrogen from atmosphere
‘Disturbed’ ecosystems
Boreal forests
Temperate forests
Early successional
tropical forests
Why Earth System Models?
• Goal: Predict the allowable emissions that avoid
exceeding a temperature target
~45%
~29%
~26%
Why Earth System Models?
• Goal: Predict the allowable emissions that avoid
exceeding a temperature target
Requires(predic,ng(the(
evolu,on(of(the(terrestrial(
~45%
carbon(sink!(
~29%
~26%
Carbon cycle feedbacks
Atmospheric CO2
+ Stabilizing Feedback
Land C sink
Carbon concentration feedback
Temperature
-
Amplifying
Feedback
+
Land C sink
Carbon climate feedback
Wide range of predictions for future carbon sink
Year
Friedlingstein et al. 2006 J of Climate
on budget, based on Eqs. (9) and (11), (0.64–0.71) is associated with their relative
Carbon
impact
effective
emissions
8. The
effect offeedbacks
the ocean carbon–
fraction
of emissions
taken up by land (0.06–0
on the overall atmospheric carbon pared to other comprehensive Earth system
oth carbon–concentration and carbon– This is related to a weaker CO2 fertilizatio
over ocean vary little between models. these models. In the absence of an explicit
cept NorESM-ME and CESM1-BGC, nitrogen cycle, the strength of the CO2 fe
ycle component dominates the overall effect in CanESM2 is ‘‘downregulated’’ bas
response of plants grown in ambient and ele
edback.
show that, because of their offsett- following Arora et al. (2009). The CO2 fertil
r values of cumulative emissions and fect in the NorESM-ME and CESM1-BGC
result even though the strength of feed- constrained by nitrogen limitation. Finally, un
erably across models. The higher air- models, which use a biogeochemical approac
umulative emissions in the CanESM2, terrestrial photosynthesis, the MIROC-ESM
SM1-BGC, and MIROC-ESM models empirical approach to model the photosynthet
FIG. 7. Contributions of the carbon–concentration and carbon–climate fe
(a) in terms of their absolute magnitudes and (b) as a fraction of cumulati
Arora et al. 2013 J of Climate
Similarly strong relations were seen for leaves (P < 0.001,
R2 = 0.66, n = 1386), stems (P < 0.001, R2 = 0.80,
n = 380) and fine roots (P < 0.001, R2 = 0.62, n = 744)
examined among all plant taxa (scaling exponents ranged
from 1.34 to 1.64; Fig. 1). Examining each plant group
separately (pooled across organs), there were strong R–N
relations for gymnosperms (P < 0.001, R2 = 0.66,
7960.001,
P. B. Reich
n = 1216), woody angiosperms (P <
R2et=al. 0.62,
n = 1069) and herbs (P < 0.001, R2 = 0.57, n = 225;
Components of the land carbon sink
scaling exponents ranged from 1.16
Fig. 2). The consistently greater than isom
of R to N in all organs and all plant grou
unit N is as a rule higher in tissues with
rates. This is likely due to generally faste
proteins, maintenance of solute gradients,
in more metabolically active tissues (Boum
a disproportionate increase in N in m
pools relative to structural N with inc
C sink =
Plant C uptake - C release by decomposition
Leaves
Respiration (nmol g–1 s–1)
10
woody angio
0.1
gymno
herb
100
1
0.3
100
3
1
Roots
10
1
0.1
0.3
1
Nitrogen (mmol g–1)
Fine roots
20-70
scaling expo
Fig. 2). The
of R to N in
unit N is as
rates. This is
proteins, mai
in more meta
a disproport
pools relativ
Figure 1 Mass-based
(nmol g)1 s)1) in relatio
Stems
Leaves
0.01
1
concentration (mmol g
Microbes0.1
5-17
10
pressed on a logarithmic
are shown for different
10
100
All
1
all organ types pooled)
labelled with different c
10
woody angio
0.1
visibility of each pan
gymno
1
herb
identical among panels
1
3
0.3
1
0.01
0.1 are the same,
1
ratios
he
compared among pan
0.1
100
Roots
100 All
include herbaceous angios
angiosperms (woody
10
3
0.01
0.1
1
gymnosperms (gymno).
10
–1
Nitrogen (mmol g )
Table 1.
1
100
Respiration (nmol g–1 s–1)
100
Similarly strong relations were seen for leaves (P < 0.001,
R2 = 0.66, n = 1386), stems (P < 0.001, R2 = 0.80,
n = 100
380) and fine roots (P < 0.001, R2 = 0.62, n = 744)
examined Stems
among all plant taxa (scaling exponents ranged
from 1.34 Tissue
to 1.64; Fig. 1). Examining
each plant group
C:N
10
separately (pooled across organs), there were strong R–N
relations Leaves
for gymnosperms (P16-60
< 0.001, R2 = 0.66,
1
n = 1216),
woody angiosperms (P < 0.001, R2 = 0.62,
n = 1069)
and herbs
(P < 0.001,
R2 = 0.57, n = 225;
Woody
tissue
212-1400
Reich et al. 1997 PNAS, Reich etal. 2008 Ecol. Letters, White et al. 2005 Earth Interactions
Discrepancy (Pg N)
e model prodistribute the
oughly equalth no change
ratios of trees
-only projecnitrogen; the
re 2.3 to 16.9
trogen fixation (12) to increase
10% (low) or 45% (high) with
bling (9). We further assume th
leaching losses are currently 3
trogen per year (13), and th
leaching would d
40
40
early with CO2 d
CO2
0 (low) to 20% (h
35
35
Combining ou
mates, 6.1 Pg o
TRIFFID
30
30
could accumulat
(see the figure). T
25
25
is less than is req
CO2-only simulat
CO2
20
20
and
four of the six C
climate
simulations (1, 2
IBIS
15
15
figure). Our low e
LPJ
VECODE
nitrogen accumu
10
10
only 1.2 Pg of n
SDGVM
High
sufficient for all s
Nitrogen
5
5
HYBRID
supply
We have focu
limits
Low
trogen, but the sit
0
0
0
200
400
600
800 1000
be worse for othe
Additional carbon stored (Pg C)
such as potassium
which
Supply and demand. (Left)Hungate
Nitrogen et
required
to support
terrestrial. 2003
Science;
Norby etphorus,
al. 2010
PNASar
tal studies show that when CO2 enrichment
increases soil C:N, decomposing microorganisms require more nitrogen. This effect
can reduce nitrogen
mineralization,
the main
optimistic
CO
2 fertilization?
source of nitrogen for plants (8, 9). It is thus
Additional nitrogen required (Pg N)
g atmospheric
890 Pg of carlate in the tere amounts
are
Overly
pected anthrointermediate
models suggest
ate change tog of carbon to
missions (1, 2).
xaggerate the
al to slow atm carbon acned by nutri, 4), through
developed in
is given in SI Materials and Methods.
Discrepancy (Pg N)
e model prodistribute the
oughly equalth no change
ratios of trees
-only projecnitrogen; the
re 2.3 to 16.9
trogen fixation (12) to increase
10% (low) or 45% (high) with
bling (9). We further assume th
leaching losses are currently 3
trogen per year (13), and th
leaching would d
40
40
early with CO2 d
CO2
0year),
(low)leaf
to 20%
mass (h
p
35
35
tent.
LAD varied
Combining
ou
and with6.1
no effect
mates,
Pg o
TRIFFID
30
30
in aCOaccumulat
2, but there
could
and LAD in eCO2
(see
the(Fig.
figure).
Ta
iment
3E),
25
25
is(Fig.
less3F).
thanCanopy
is req
CO
simulat
CO2
through
time
exce
2-only
20
20
and
NPP of
wasthe
neverthe
four
six C
climate
influences of (1,
LAD
simulations
2
IBIS
15
15
compared
with
figure).
Our
lowthe
framework, foliar [
LPJ
VECODE
nitrogen
accumu
10
10
eCO2, accounting
only
1.2 Pg
ofofn
explained
57%
SDGVM
High
sufficient
all s
additional for
15%.
Nitrogen
5
5
HYBRID
supply
We have focu
Light-saturated
limits
Low
(P < 0.05)
at
eCO
trogen,
but
the
sit
0
0
Fig. 1. Tree growth responses to elevated CO2. NPP [kilograms dry matter
0
200
400
600
800 1000
14.5
± 1.8
μmol·
be
worse
for
othe
(DM) per square meter land area per year] data are the means of three aCO2
However,
photosy
Additional
carbon
stored
(Pg C)
plots (open
symbols) and
two eCO
such
as potassium
2 plots (solid symbols) ± SEM. The number
and there was no l
at each point is the percentage increase under eCO2. Statistical information
which
−2 −1ar
0.7
·s v
Supply and demand. (Left)Hungate
Nitrogen et
required
to support
terrestrial. 2003
Science;
Norby etphorus,
al. μmol·m
2010
PNAS
tal studies show that when CO2 enrichment
increases soil C:N, decomposing microorganisms require more nitrogen. This effect
can reduce nitrogen
mineralization,
the main
optimistic
CO
2 fertilization?
source of nitrogen for plants (8, 9). It is thus
Additional nitrogen required (Pg N)
g atmospheric
890 Pg of carlate in the tere amounts
are
Overly
pected anthrointermediate
models suggest
ate change tog of carbon to
missions (1, 2).
xaggerate the
al to slow atm carbon acned by nutri, 4), through
developed in
eleased from soils in response to warming has been taken up
ed in trees in the heated area. Thus, although warming has
in a net positive feedback to the climate system, the magf the feedback has been substantially dampened by the inn storage of carbon in vegetation.
ases in vegetation carbon storage in the heated area are
ue, for the most part, to the warming-induced increase in
ogen mineralization of about 27 kg·N·ha−1·yr−1. This
nts a 45% increase relative to the nitrogen mineralization
he control area (Fig. 5).
sed the principles of ecosystem stoichiometry to explore
or not the increase in net nitrogen mineralization in the
area was large enough to support the measured increase
on storage in the trees growing there. When carbon is
n plant tissues, a small amount of nitrogen is also stored,
mass ratio of C:N specific to plant tissue type. In the
Overly negative carbon loss?
Although soil warming caused a loss o
concurrently stimulated a gain of carbon i
The forest in the study area has been r
stand-replacing hurricane in 1938, which
Fig. 1. Total annual soil COin
into
soilthe
organic
matter
loss
2 efflux
both partitioned
the control
and
heated
areas
root respiration, and fine-root
decomposition
for the
heated
and contro
study,
carbon
storage
in
the
vegetation
of
−1
areas (in Mg·C·ha ).
between 1.7 and 3.6 Mg·C·ha−1, with a
Mg·C·ha−1. For forest stands of similar
other
the the
Harvard
Barf
unit of root mass and total
rootparts
mass.ofOver
courseForest,
of the study
a biometrically
determined
carbon
stora
soil warming has resulted
in increased respiration
rates
per
unit
o
−1
tation
of
1.7
Mt·C·ha
for
the
period
19
root mass, whereas it has led to decreased fine-root mass.
The annual
rate of
carbon
storage to
i
Our estimate of the relative
contribution
of root
respiration
area
wasfalls
greater
than in
theearlie
cont
total respiration in theheated
control
area
between
two
−1
2.4
and
5.1
Mg·C·ha
with a mean
for t
estimates made in unheated areas in other ,similarly
structured
Melilloandetheated
al. 2011
Fig. 5. Net nitrogen mineralization in the control
areas.PNAS
Bars
eleased from soils in response to warming has been taken up
ed in trees in the heated area. Thus, although warming has
in a net positive feedback to the climate system, the magf the feedback has been substantially dampened by the inn storage of carbon in vegetation.
ases in vegetation carbon storage in the heated area are
ue, for the most part, to the warming-induced increase in
ogen mineralization of about 27 kg·N·ha−1·yr−1. This
nts a 45% increase relative to the nitrogen mineralization
he control area (Fig. 5).
sed the principles of ecosystem stoichiometry to explore
or not the increase in net nitrogen mineralization in the
area was large enough to support the measured increase
on storage in the trees growing there. When carbon is
n plant tissues, a small amount of nitrogen is also stored,
mass ratio of C:N specific to plant tissue type. In the
Overly negative carbon loss?
Temperature
-
Amplifying
Feedback
Soil C sink
+
-
Temperature
Although soil warming caused a loss o
Stabilizing
+ a gain of carbon i
concurrently stimulated
Feedback
The forest in the study area has been r
Plant growth + N mineralization
stand-replacing hurricane in 1938, which
Fig. 1. Total annual soil COin
into
soilthe
organic
matter
loss
2 efflux
both partitioned
the control
and
heated
areas
root respiration, and fine-root
decomposition
for the
heated
and contro
study,
carbon
storage
in
the
vegetation
of
−1
areas (in Mg·C·ha ).
between 1.7 and 3.6 Mg·C·ha−1, with a
Mg·C·ha−1. For forest stands of similar
other
the the
Harvard
Barf
unit of root mass and total
rootparts
mass.ofOver
courseForest,
of the study
a biometrically
determined
carbon
stora
soil warming has resulted
in increased respiration
rates
per
unit
o
−1
tation
of
1.7
Mt·C·ha
for
the
period
19
root mass, whereas it has led to decreased fine-root mass.
The annual
rate of
carbon
storage to
i
Our estimate of the relative
contribution
of root
respiration
area
wasfalls
greater
than in
theearlie
cont
total respiration in theheated
control
area
between
two
−1
2.4
and
5.1
Mg·C·ha
with a mean
for t
estimates made in unheated areas in other ,similarly
structured
Melilloandetheated
al. 2011
Fig. 5. Net nitrogen mineralization in the control
areas.PNAS
Bars
Carbon-nitrogen interactions impact effective
on budget, based on Eqs. (9) and (11), (0.64–0.71) is associated with their relative
emissions
8. The
effect of the ocean carbon– fraction of emissions taken up by land (0.06–0
on the overall atmospheric carbon
oth carbon–concentration and carbon–
over ocean vary little between models.
cept NorESM-ME and CESM1-BGC,
ycle component dominates the overall
edback.
show that, because of their offsettr values of cumulative emissions and
result even though the strength of feederably across models. The higher airumulative emissions in the CanESM2,
SM1-BGC, and MIROC-ESM models
pared to other comprehensive Earth system
This is related to a weaker CO2 fertilizatio
these models. In the absence of an explicit
nitrogen cycle, the strength of the CO2 fe
effect in CanESM2 is ‘‘downregulated’’ bas
response of plants grown in ambient and ele
following Arora et al. (2009). The CO2 fertil
fect in the NorESM-ME and CESM1-BGC
constrained by nitrogen limitation. Finally, un
models, which use a biogeochemical approac
terrestrial photosynthesis, the MIROC-ESM
empirical approach to model the photosynthet
FIG. 7. Contributions of the carbon–concentration and carbon–climate fe
(a) in terms of their absolute magnitudes and (b) as a fraction of cumulati
Arora et al. 2013 J of Climate
Typical result from Earth System models with
carbon and nitrogen interactions
Simulation with
carbon and nitrogen
coupled cycles
-
Simulation with
only carbon cycle
Zaehle et al. 2010 GRL
N cycle is resolved at the landunit scale
Gridcell
Landunit
Glacier
Columns
Wetland
Vegetated
Lake
Soil
Type 1
PFTs
Urban
How do we model coupled C & N cycles:
CLM-CN: Allocation and Plant N limitation
Net
photosynthesis
Potential C
available for
allocation
1st run of
allocation
scheme
Required N for C
allocation
Ratio of soil N available for
plants to N required for C
allocation for plants
maintenance
respiration
How do we model coupled C & N cycles:
CLM-CN: Allocation and Plant N limitation
Net
photosynthesis
maintenance
respiration
Fixed fraction
(loosely based
on White et al.
2000)
Potential C
available for
allocation
1st run of
allocation
scheme
Required N for C
allocation
Potential Available C
for allocation
x
Ratio of N available
to N required
Ratio of soil N available for
plants to N required for C
allocation for plants
GPP = Net photosynthesis x ratio of N available to N required
N uptake = min(N required, N available)
Fixed C:N
Phenology
Displayed
pools
Stored
leaves roots
leaves roots
Stored Live
wood
Live wood
Stored Dead
wood
Dead wood
Stored Live
coarse roots
Live coarse
roots
Stored Dead
coarse roots
Dead coarse
roots
Stored fine
roots
fine roots
Stored
growth
respiration
Growth
respiration
How do we model coupled C & N cycles:
CLM-CN: Allocation and Plant N limitation
Net
photosynthesis
maintenance
respiration
Fixed fraction
(loosely based
on White et al.
2000)
Potential C
available for
allocation
Fixed C:N
Phenology
Stored
leaves roots
leaves roots
Stored Live
wood
Live wood
Stored Dead
wood
Dead wood
Stored Live
coarse roots
Live coarse
roots
Stored Dead
coarse roots
Dead coarse
roots
Stored fine
roots
fine roots
Stored
growth
respiration
Growth
respiration
N limitation is instantaneous
N limitation directly reduces GPP
1st run of
allocation
scheme
Required N for C
allocation
Potential Available C
for allocation
x
Ratio of N available
to N required
Ratio of soil N available for
plants to N required for C
allocation for plants
GPP = Net photosynthesis x ratio of N available to N required
N uptake = min(N required, N available)
Displayed
pools
How do we model coupled C & N cycles:
Microbial example
30 C
60 C
1N
30 C
1N
4N
6C
1N
Most nitrogen is recycled
Cleveland et al. 2013 PNAS
CLM-CN 4.0: General N cycle structure
clm4cn structure
N gas loss
photosynthesis,
Fire loss
g
Litterfall and
mortality
a
SON &
litter N
b
Mineralization
norganic N
nt N uptake (d). No
re and soil water). N deposition
tion (b-c).
h
i
N fixation
Plant N
c
Fire loss
dPlant N uptake
Immobilization
Soil
Inorganic N
Harvest
f
N leaching
e
Denitrification
clm4mod structure
CLM-CN 4.0: General N cycle structure
clm4cn structure
N gas loss
photosynthesis,
Fire loss
g
Litterfall and
mortality
a
SON &
litter N
b
Mineralization
norganic N
nt N uptake (d). No
re and soil water). N deposition
tion (b-c).
h
i
N fixation
Plant N
c
Fire loss
dPlant N uptake
Immobilization
Soil
Inorganic N
Harvest
f
N leaching
e
Denitrification
clm4mod structure
al.
Tegen ettransport
al. 2004).
The greater dust
mobility
the 2004,
long-distance
of atmospheric
(Okinand
et
biological
availability
combined
N in the
atmosphere
al. 2004, Tegen
et al. of
2004).
The greater
mobility
and
implies
thatavailability
anthropogenic
P limitation
be much
biological
of combined
N in should
the atmosphere
more
(if perhaps
less persistent)
than
implieswidespread
that anthropogenic
P limitation
should be much
anthropogenic
N limitation.
more widespread
(if perhaps less persistent) than
anthropogenic N limitation.
material-based
P limitation
or
human fertilization
that can r
sustained. Where
barriersa
material-based
P soil
limitation
features of
soilssoil
(permafrost
sustained.
Where
barriers d
warming;ofmassive
placic horizo
features
soils (permafrost
they too massive
could drive
ultimate
warming;
placic
horizonP
they too could drive ultimate P
Mechanisms that govern N limitation
Steady-state and transient processes
TABLE 2. Pathways, mechanisms, and timescales of N limitation to primary production in terrestria
TABLE 2. Pathways, mechanisms, and timescales of N limitation to primary production in terrestrial
Pathway
Mechanism
Pathwaylosses
Demand-independent
losses of combined NMechanism
that organisms cannot prevent,
d
including
leaching
DON,
post- disturbance
losses, dec
Demand-independent losses
losses
of combined
N of
that
organisms
cannot prevent,
some gaseous
pathways
including
leaching
of DON, post- disturbance losses,
o
Constraints to biological N fixation
biological
N fixation
is slow or absent even when N is
d
some gaseous
pathways
limiting;Ncould
be due
to energetic
differential
Constraints to biological N fixation
biological
fixation
is slow
or absent costs,
even when
N is
dec
grazing, could
demands
for to
P, energetic
Mo, or other
limiting;
be due
costs,essential
differential
elementsdemands for P, Mo, or other essential
grazing,
Transactional
slow
release of N from complex organic into soluble
ye
elements
forms,
relative
to thecomplex
supply of
other into
resources
Transactional
slow
release
of N from
organic
soluble
yea
Sink driven
sequestration
of available
N inofanother
accumulating
d
forms, relative
to the supply
resources pool
within ecosystems
Sink driven
sequestration
of available N in an accumulating pool
dec
within ecosystems
Sources: Vitousek and Howarth (1991), Vitousek and Field (1999), and Vitousek (2004).
Sources: Vitousek and Howarth (1991), Vitousek and Field (1999), and Vitousek (2004).
Transient-only processes
Vitousek et al. 2010 Ecol. Applications
al.
Tegen ettransport
al. 2004).
The greater dust
mobility
the 2004,
long-distance
of atmospheric
(Okinand
et
biological
availability
combined
N in the
atmosphere
al. 2004, Tegen
et al. of
2004).
The greater
mobility
and
implies
thatavailability
anthropogenic
P limitation
be much
biological
of combined
N in should
the atmosphere
more
(if perhaps
less persistent)
than
implieswidespread
that anthropogenic
P limitation
should be much
anthropogenic
N limitation.
more widespread
(if perhaps less persistent) than
anthropogenic N limitation.
material-based
P limitation
or
human fertilization
that can r
sustained. Where
barriersa
material-based
P soil
limitation
features of
soilssoil
(permafrost
sustained.
Where
barriers d
warming;ofmassive
placic horizo
features
soils (permafrost
they too massive
could drive
ultimate
warming;
placic
horizonP
they too could drive ultimate P
Mechanisms that govern N limitation
Steady-state and transient processes
TABLE 2. Pathways, mechanisms, and timescales of N limitation to primary production in terrestria
TABLE 2. Pathways, mechanisms, and timescales of N limitation to primary production in terrestrial
Pathway
Mechanism
Pathwaylosses
Demand-independent
losses of combined NMechanism
that organisms cannot prevent,
d
including
leaching
DON,
post- disturbance
losses, dec
Demand-independent losses
losses
of combined
N of
that
organisms
cannot prevent,
some gaseous
pathways
including
leaching
of DON, post- disturbance losses,
o
Constraints to biological N fixation
biological
N fixation
is slow or absent even when N is
d
some gaseous
pathways
limiting;Ncould
be due
to energetic
differential
Constraints to biological N fixation
biological
fixation
is slow
or absent costs,
even when
N is
dec
grazing, could
demands
for to
P, energetic
Mo, or other
limiting;
be due
costs,essential
differential
elementsdemands for P, Mo, or other essential
grazing,
Transactional
slow
release of N from complex organic into soluble
ye
elements
forms,
relative
to thecomplex
supply of
other into
resources
Transactional
slow
release
of N from
organic
soluble
yea
Sink driven
sequestration
of available
N inofanother
accumulating
d
forms, relative
to the supply
resources pool
within ecosystems
Sink driven
sequestration
of available N in an accumulating pool
dec
within ecosystems
Sources: Vitousek and Howarth (1991), Vitousek and Field (1999), and Vitousek (2004).
Sources: Vitousek and Howarth (1991), Vitousek and Field (1999), and Vitousek (2004).
Vitousek et al. 2010 Ecol. Applications
Demand-independent losses
Symbiotic N fixation
Tissue
turnover
Demandindependent
Litter
losses
Soil organic matter
Volatilized (fire)
Free-living N fixation
Plant
Uptake
Demanddependent
losses
Inorganic N
Simple organic N
N deposition
Demandindependent
losses
Demand-independent losses
clm4cn
structure
CLM-CN
4.0
N gas loss
photosynthesis,
Fire loss
g
Litterfall and
mortality
a
SON &
litter N
b
Mineralization
norganic N
nt N uptake (d). No
re and soil water). N deposition
tion (b-c).
h
i
N fixation
Plant N
c
Fire loss
dPlant N uptake
Immobilization
Soil
Inorganic N
Harvest
f
N leaching
e
Denitrification
clm4mod structure
Demand-independent losses
clm4cn
structure
CLM-CN
4.5
N gas loss
photosynthesis,
Fire loss
g
Litterfall and
mortality
a
SON &
litter N
b
Mineralization
norganic N
nt N uptake (d). No
re and soil water). N deposition
tion (b-c).
h
i
N fixation
Plant N
c
Fire loss
dPlant N uptake
Immobilization
Soil
Inorganic N
Harvest
f
N leaching
e
Denitrification
clm4mod structure
al.
Tegen ettransport
al. 2004).
The greater dust
mobility
the 2004,
long-distance
of atmospheric
(Okinand
et
biological
availability
combined
N in the
atmosphere
al. 2004, Tegen
et al. of
2004).
The greater
mobility
and
implies
thatavailability
anthropogenic
P limitation
be much
biological
of combined
N in should
the atmosphere
more
(if perhaps
less persistent)
than
implieswidespread
that anthropogenic
P limitation
should be much
anthropogenic
N limitation.
more widespread
(if perhaps less persistent) than
anthropogenic N limitation.
material-based
P limitation
or
human fertilization
that can r
sustained. Where
barriersa
material-based
P soil
limitation
features of
soilssoil
(permafrost
sustained.
Where
barriers d
warming;ofmassive
placic horizo
features
soils (permafrost
they too massive
could drive
ultimate
warming;
placic
horizonP
they too could drive ultimate P
Mechanisms that govern N limitation
Steady-state and transient processes
TABLE 2. Pathways, mechanisms, and timescales of N limitation to primary production in terrestria
TABLE 2. Pathways, mechanisms, and timescales of N limitation to primary production in terrestrial
Pathway
Mechanism
Pathwaylosses
Demand-independent
losses of combined NMechanism
that organisms cannot prevent,
d
including
leaching
DON,
post- disturbance
losses, dec
Demand-independent losses
losses
of combined
N of
that
organisms
cannot prevent,
some gaseous
pathways
including
leaching
of DON, post- disturbance losses,
o
Constraints to biological N fixation
biological
N fixation
is slow or absent even when N is
d
some gaseous
pathways
limiting;Ncould
be due
to energetic
differential
Constraints to biological N fixation
biological
fixation
is slow
or absent costs,
even when
N is
dec
grazing, could
demands
for to
P, energetic
Mo, or other
limiting;
be due
costs,essential
differential
elementsdemands for P, Mo, or other essential
grazing,
Transactional
slow
release of N from complex organic into soluble
ye
elements
forms,
relative
to thecomplex
supply of
other into
resources
Transactional
slow
release
of N from
organic
soluble
yea
Sink driven
sequestration
of available
N inofanother
accumulating
d
forms, relative
to the supply
resources pool
within ecosystems
Sink driven
sequestration
of available N in an accumulating pool
dec
within ecosystems
Sources: Vitousek and Howarth (1991), Vitousek and Field (1999), and Vitousek (2004).
Sources: Vitousek and Howarth (1991), Vitousek and Field (1999), and Vitousek (2004).
Vitousek et al. 2010 Ecol. Applications
Constraints to Biological Fixation
Temperature,
Precipitation,
Light,
CO2,
Nitrogen
Cleveland et al. 1999 Global Biogeochemical Cycles
Constraints to Biological Fixation
Temperature,
Precipitation,
Light,
CO2,
Nitrogen
Other models use:
ET (not a strong function of N)
N demand weighted by energy
Cleveland et al. 1999 Global Biogeochemical Cycles
Constraints to Biological Fixation
0
0
640
,t
0
..
I.
o
OF TERRESTRIAL N FIXATION
acreage [Matthews, 1983].
Therefore we believe that
decreasingour value by this value (i.e., 13.6%) may provide a
subsistence
agriculture
and-15 x 106 km2 for global crop
(13.6%) of all lands have been converted for cultivation. This
value includes a contribution of-3 x 106 km 2 from extensive
1983]). Matthews[1983] suggeststhat -18 x 106 km2
thatwhile Centurydoesseemto capturebroad-scale
patternsin
N fixation (Figure 2) it also appears to significantly
underestimatetotal inputs.
These estimatesrepresentpotential nitrogen fixation (i.e.,
ecosystemareasdo not account for land-use changes), but
accounting for the decrease in natural N fixation due to
cultivation would not dramatically alter our estimate, as the
greatest reductions in area have occurred in systems
characterizedby relatively low rates of N fixation (e.g.
temperate forests and temperate grasslands; [Matthews,
N fixationinputof 25 Tg N yr4 (datanot shown),suggesting
procedurefor doing so. Thus we caution the treatment of this
value as an absolute and emphasize that we are much more
comfortablethat our range of possible global N fixation
estimatesincludesthe true value of global N fixation, than we
are that our central estimateis the true value of global N
fixation. Our data-basedestimatescompareto a modeledtotal
CLEVELAND ET AL.: GLOBAL PATFE•S
o
o
z
,
from betwe
13 Gt C to 8
of the new
remainder b
N:C ratios
Methods),
w
Cleveland et al. 1999 Global Biogeochemical Cycles
However, both our conservative and central estimates do
within individualecosystems.In addition, a greateremphasis
shouldbe placedon estimatingthe aerial coverageof N-fixing
microsites and species, in order to improve estimatesof N
fixation budgets from the watershed-to-ecosystemscale.
Finally, we would suggest that whenever possible, studies
estimateBNF inputswithin the contextof otherN inputs(e.g.,
wet and dry deposition), as well as the full suite of N loss
vectors (e.g., trace gas fluxes, organic and inorganic N
leaching). However,a recentglobal inventory of nitric oxide
(Table 13). A more rigorous evaluation of ecosystem-scale
estimates will require more studies of BNF at the local scale
Schlesinger[1991]. Despite the similarity, we believe that
our rigorous literature review provides a more documented,
constrainedestimate. Nonetheless,it is noteworthy that our
independently derived estimate of terrestrial BNF is
remarkably similar overall to this latter estimate.
Although we believe this studyprovidesa much improved
estimate of biological nitrogen fixation in terrestrial
ecosystems,perhapsthe most valuable aspectof this work is
that it points out critical gaps in our understandingof N
fixation at the local, regional, and global scale. In order to
improve our understanding, key research needs include
continuedand expandedresearchinvestigatingN fixation rates
in natural and managedecosystems,particularly in parts of
Asia, Africa, and South America, where nitrogen fixation is
likely to be important,but wheredata are scarce. For example,
several of our ecosystem-scaleestimates are basedon only
one-to-severalpublishedestimates,and data often vary widely
fall within the range of 44-200 Tg N yr-• reportedby
BNF.
previouslyreportedby Stedmanand Shetter [1983], Soderland
and Rosswall [1982], and Galloway et al., [1995]. Thus we
believe that these authorsmay have underestimatedterrestrial
ecosystems
is higherthan the valuesof 90-140 Tg N yr-•
Our central value of 195 Tg N fixed per year in terrestrial
conservative estimate of actual present-day terrestrial BNF in
natural ecosystems.
Precipitation,
Light,
CO2,
Nitrogen
o
o
•
modelsCONSTRAINT
use:
WANG AND HOULTON:Other
NITROGEN
ET (not a strong function of N)
Temperature,
N demand weighted by
theenergy
models,
03
N limitation at steady-state
N fixation < Demand-independent losses
How to classify losses
in Earth System Models?
Demand-dependent losses < N deposition
Menge 2011 Ecosystems
al.
Tegen ettransport
al. 2004).
The greater dust
mobility
the 2004,
long-distance
of atmospheric
(Okinand
et
biological
availability
combined
N in the
atmosphere
al. 2004, Tegen
et al. of
2004).
The greater
mobility
and
implies
thatavailability
anthropogenic
P limitation
be much
biological
of combined
N in should
the atmosphere
more
(if perhaps
less persistent)
than
implieswidespread
that anthropogenic
P limitation
should be much
anthropogenic
N limitation.
more widespread
(if perhaps less persistent) than
anthropogenic N limitation.
material-based
P limitation
or
human fertilization
that can r
sustained. Where
barriersa
material-based
P soil
limitation
features of
soilssoil
(permafrost
sustained.
Where
barriers d
warming;ofmassive
placic horizo
features
soils (permafrost
they too massive
could drive
ultimate
warming;
placic
horizonP
they too could drive ultimate P
Mechanisms that govern N limitation
TABLE 2. Pathways, mechanisms, and timescales of N limitation to primary production in terrestria
TABLE 2. Pathways, mechanisms, and timescales of N limitation to primary production in terrestrial
Pathway
Mechanism
Pathwaylosses
Demand-independent
losses of combined NMechanism
that organisms cannot prevent,
d
including
leaching
DON,
post- disturbance
losses, dec
Demand-independent losses
losses
of combined
N of
that
organisms
cannot prevent,
some gaseous
pathways
including
leaching
of DON, post- disturbance losses,
o
Constraints to biological N fixation
biological
N fixation
is slow or absent even when N is
d
some gaseous
pathways
limiting;Ncould
be due
to energetic
differential
Constraints to biological N fixation
biological
fixation
is slow
or absent costs,
even when
N is
dec
grazing, could
demands
for to
P, energetic
Mo, or other
limiting;
be due
costs,essential
differential
elementsdemands for P, Mo, or other essential
grazing,
Transactional
slow
release of N from complex organic into soluble
ye
elements
forms,
relative
to thecomplex
supply of
other into
resources
Transactional
slow
release
of N from
organic
soluble
yea
Sink driven
sequestration
of available
N inofanother
accumulating
d
forms, relative
to the supply
resources pool
within ecosystems
Sink driven
sequestration
of available N in an accumulating pool
dec
within ecosystems
Sources: Vitousek and Howarth (1991), Vitousek and Field (1999), and Vitousek (2004).
Sources: Vitousek and Howarth (1991), Vitousek and Field (1999), and Vitousek (2004).
Transient-only processes
Vitousek et al. 2010
Transactional N limitation
Thornton and Rosenbloom 2005 Ecological Modelling
Parton et al. 1993 Glob. Biogeochemical Cycles
Transactional N limitation
Thornton and Rosenbloom 2005 Ecological Modelling
Parton et al. 1993 Glob. Biogeochemical Cycles
al.
Tegen ettransport
al. 2004).
The greater dust
mobility
the 2004,
long-distance
of atmospheric
(Okinand
et
biological
availability
combined
N in the
atmosphere
al. 2004, Tegen
et al. of
2004).
The greater
mobility
and
implies
thatavailability
anthropogenic
P limitation
be much
biological
of combined
N in should
the atmosphere
more
(if perhaps
less persistent)
than
implieswidespread
that anthropogenic
P limitation
should be much
anthropogenic
N limitation.
more widespread
(if perhaps less persistent) than
anthropogenic N limitation.
material-based
P limitation
or
human fertilization
that can r
sustained. Where
barriersa
material-based
P soil
limitation
features of
soilssoil
(permafrost
sustained.
Where
barriers d
warming;ofmassive
placic horizo
features
soils (permafrost
they too massive
could drive
ultimate
warming;
placic
horizonP
they too could drive ultimate P
Mechanisms that govern N limitation
TABLE 2. Pathways, mechanisms, and timescales of N limitation to primary production in terrestria
TABLE 2. Pathways, mechanisms, and timescales of N limitation to primary production in terrestrial
Pathway
Mechanism
Pathwaylosses
Demand-independent
losses of combined NMechanism
that organisms cannot prevent,
d
including
leaching
DON,
post- disturbance
losses, dec
Demand-independent losses
losses
of combined
N of
that
organisms
cannot prevent,
some gaseous
pathways
including
leaching
of DON, post- disturbance losses,
o
Constraints to biological N fixation
biological
N fixation
is slow or absent even when N is
d
some gaseous
pathways
limiting;Ncould
be due
to energetic
differential
Constraints to biological N fixation
biological
fixation
is slow
or absent costs,
even when
N is
dec
grazing, could
demands
for to
P, energetic
Mo, or other
limiting;
be due
costs,essential
differential
elementsdemands for P, Mo, or other essential
grazing,
Transactional
slow
release of N from complex organic into soluble
ye
elements
forms,
relative
to thecomplex
supply of
other into
resources
Transactional
slow
release
of N from
organic
soluble
yea
Sink driven
sequestration
of available
N inofanother
accumulating
d
forms, relative
to the supply
resources pool
within ecosystems
Sink driven
sequestration
of available N in an accumulating pool
dec
within ecosystems
Sources: Vitousek and Howarth (1991), Vitousek and Field (1999), and Vitousek (2004).
Sources: Vitousek and Howarth (1991), Vitousek and Field (1999), and Vitousek (2004).
Transient-only processes
Vitousek et al. 2010
Sink-driven N limitation
CWD
2.7 years
C:N = variable
20 - 500 C:N
1.00
Litter 1
1.4 days
C:N = variable
Litter 2
14 days
C:N = variable
Litter 3
71 days
C:N = variable
0.61
0.45
0.71
Soil 1
14 days
C:N = 12
0.72
Soil 2
71 days
C:N = 12
0.54
Soil 3
1.95 years
C:N = 10
0.45
10-12 C:N
Soil 4
27.4 years
C:N =10
Sink-driven N limitation
a change in the atntration. However,
pools are cumulacumulation in an
osystem productivnce between inflow
is and outflow from
hotosynthetic stimCO2 is constant, the
ation rate declines
radually increasing
eynolds 1999). Resportional to C pool
t turnover rate (e.g.,
ne roots), it takes a
te C before the pool
brium. These pools
ols with a long resdy biomass and old
ulative change in C
cades to thousands
l size reaches a new
Progressive N limitation
Articles
N sequestered in
biomass and litter
CO2
NPP
C:N
Labile soil N
C input
to soil
N uptake
N sequestered in
in SOM
N availability
Figure 1. Two sets of feedback processes to elevated carbon dioxide (CO2) leading
to progressive nitrogen (N) limitation. In the upper pathway, the initialLuo
producet al.
2004 Bioscience
How do we measure and evaluate N limitation?
Field
Model
Add N to ecosystem
and measure response
Instantaneous metric in
models
(actual NPP/potential NPP)
Measure inorganic N
leaching relative to N
deposition
Coupled simulations with
and without N constraints
included
Measure leaf chemistry
(N:P ratios)
Measurement of N
fixation relative to DON
leaching
How do we measure and evaluate N limitation?
Field
Model
Add N to ecosystem
and measure response
Instantaneous metric in
models
(actual NPP/potential NPP)
Measure inorganic N
leaching relative to N
deposition
Coupled simulations with
and without N constraints
included
Measure leaf chemistry
(N:P ratios)
Measurement of N
fixation relative to DON
leaching
How do we measure and evaluate N limitation?
Field
Model
Add N to ecosystem
and measure response
Instantaneous metric in
models
(actual NPP/potential NPP)
Measure inorganic N
leaching relative to N
deposition
Coupled simulations with
and without N constraints
included
Measure leaf chemistry
(N:P ratios)
Measurement of N
fixation relative to DON
leaching
How do we measure and evaluate N limitation?
Field
Model
Add N to ecosystem
and measure response
Instantaneous metric in
models
(actual NPP/potential NPP)
Measure inorganic N
leaching relative to N
deposition
Coupled simulations with
and without N constraints
included
Measure leaf chemistry
(N:P ratios)
Measurement of N
fixation relative to DON
leaching
How do we measure and evaluate N limitation?
Field
Model
Add N to ecosystem
and measure response
Instantaneous metric in
models
(actual NPP/potential NPP)
Measure inorganic N
leaching relative to N
deposition
Coupled simulations with
and without N constraints
included
Measure leaf chemistry
(N:P ratios)
Measurement of N
fixation relative to DON
leaching
Model comparison to data:
Model response compared to observations
Nitrogen fertilization experiments
15N tracer studies
Plot/small catchment nitrogen budgets
Thomas et al. 2013 Global Change Biology
Nitrogen Leaching / Nitrogen Deposition
Model comparison to data:
Plot/Small Catchment Nitrogen Budgets
1.6
1.4
Temperate and boreal forests
n = 209
1.2
1.0
0.8
0.6
0.4
CLM-CN
0.0004
0.2
0
Thomas et al. 2013 Global Change Biology
Observations from NiRENA project: Goodale et al.
How do we measure and evaluate N limitation?
Field
Model
Add N to ecosystem
and measure response
Instantaneous metric in
models
(actual NPP/potential NPP)
Measure inorganic N
leaching relative to N
deposition
Coupled simulations with
and without N constraints
included
Measure leaf chemistry
(N:P ratios)
Measurement of N
fixation relative to DON
leaching
Test of nitrogen limitation in two global
biogeochemical models
CLM-CN 4.0
O-CN
(Thornton et al. 2009 Biogeosciences)
(Zaehle et al. 2011 Nature Geoscience)
Test of nitrogen limitation in two global
biogeochemical models
CLM-CN 4.0
Biogeosciences, 6, 2099–2120, 2009
www.biogeosciences.net/6/2099/2009/
© Author(s) 2009. This work is distributed under
the Creative Commons Attribution 3.0 License.
Biogeosciences
(Thornton et al. 2009 Biogeosciences)
Biogeosciences, 6, 2099–2120, 2009
www.biogeosciences.net/6/2099/2009/
© Author(s) 2009. This work is distributed under
the Creative Commons Attribution 3.0 License.
O-CN
(Zaehle et al. 2011 Nature Geoscience)
Biogeosciences
Carbon-nitrogen interactions regulate climate-carbon cycle
feedbacks: results from an atmosphere-ocean
general circulation model
P. E. Thornton1 , S. C. Doney2 , K. Lindsay3 , J. K. Moore4 , N. Mahowald5 , J. T. Randerson4 , I. Fung6 ,
J.-F. Lamarque7,8 , J. J. Feddema9 , and Y.-H. Lee3
Carbon-nitrogen interactions regulate climate-carbon cycle
1 Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6335, USA
results from an atmosphere-ocean
2feedbacks:
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole,
MA
02543-1543,
USA
general
circulation
model
3 Climate and Global Dynamics Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
LETTERS
PUBLISHED ONLINE: 31 JULY 2011 | DOI: 10.1038/NGEO1207
Carbon benefits of anthropogenic reactive
nitrogen offset by nitrous oxide emissions
Sönke Zaehle1 *, Philippe Ciais2 , Andrew D. Friend3 and Vincent Prieur2
Additions of reactive nitrogen to terrestrial ecosystems—
primarily through fertilizer application and atmospheric
deposition—have more than doubled since 1860 owing to
human activities1 . Nitrogen additions tend to increase the net
uptake of carbon by the terrestrial biosphere, but they also
4 Department of Earth System Science, University of California, Irvine, CA 92697-3100, USA
1 , S. C. Doney2 , K. Lindsay3 , J. K. Moore4 , N. Mahowald5 , J. T. Randerson4 , I. Fung6 ,
stimulate nitrous oxide release from soils2 . However, given
P.
E.
Thornton
5 Department of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY 14850, USA
that the magnitude of these effects is uncertain, and that the
7,8
9
3
Lamarque
, J.and
J. Feddema
, and Y.-H.
Lee of California, Berkeley, CA 94720-4767, USA
6J.-F.
carbon and nitrogen cycles are tightly coupled, the net climatic
Department
of Earth
Planetary Science,
University
impact of anthropogenic nitrogen inputs is unknown3 . Here
71NOAA
Earth System
Research
Laboratory,
Chemical
Sciences
Division,
325Ridge,
Broadway,
Boulder, COUSA
80305-3337, USAwe use a process-based model of the terrestrial biosphere4,5
Environmental
Sciences
Division,
Oak Ridge
National
Laboratory,
Oak
TN 37831-6335,
82Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
to evaluate the overall impact of anthropogenic nitrogen
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole,
inputs on terrestrial ecosystem carbon and nitrous oxide fluxes
9 Department of Geography, University of Kansas, Lawrence, KS 66045-7613, USA
MA 02543-1543, USA
between 1700 and 2005. We show that anthropogenic nitrogen
3 Climate and Global Dynamics Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
inputs account for about a fifth of the carbon sequestered
Received:
28 January 2009 – Published in Biogeosciences Discuss.: 26 March 2009
by terrestrial ecosystems between 1996 and 2005, and for
4 Department
of Earth
System
Science,
California,
Irvine,
CA 92697-3100,
USA
most of the increase in global nitrous oxide emissions in recent
Revised:
12 August
2009
– Accepted:
17University
September of
2009
– Published:
8 October
2009
5 Department
decades; the latter is largely due to agricultural intensification.
of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY 14850, USA
We estimate that carbon sequestration due to nitrogen
6 Department of Earth and Planetary Science, University of California, Berkeley, CA 94720-4767, USA
deposition has reduced current carbon dioxide radiative forcing
7 NOAA Earth System Research Laboratory, Chemical Sciences Division, 325 Broadway, Boulder, CO 80305-3337,byUSA
96 ± 14 mW m 2 . However, this effect has been offset by
Abstract.
Inclusion of fundamental ecological interactions
century, attributable in part to a steady decline in the ocean
the increase in radiative forcing resulting from nitrous oxide
8 Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
between carbon and nitrogen cycles in the land component of
sink fraction. Comparison to experimental studies on theemissions,
fate
which amounts to 125±20 mW m 2 .
9 Department
an
atmosphere-ocean
general University
circulation of
model
(AOGCM)
of radio-labeled
tracers in temperate forests indiof Geography,
Kansas,
Lawrence, KS
66045-7613, nitrogen
USA
The effect of the anthropogenic reactive nitrogen (hereafter Nr)
perturbation, that is the addition of ecosystem-available forms of
leads to decreased carbon uptake associated with CO2 fertilcates that the model representation of competition between
nitrogen such as ammonium and nitrate, on terrestrial ecosystems
Received:
28
January
2009
–
Published
in
Biogeosciences
Discuss.:
26
March
2009
ization, and increased carbon uptake associated with warmplants and microbes for new mineral nitrogen resources
and is
the land–atmosphere fluxes of carbon dioxide and nitrous
Revised:
August
2009 The
– Accepted:
17these
September
2009 – Published:
2009
oxide
of the tight coupling
ing
of the 12
climate
system.
balance of
two opposreasonable.8 October
Our results
suggest a weaker dependence of
nethas proved difficult to quantify because
between the carbon and nitrogen cycles1 : changes in one cycle
ing effects is to reduce the fraction of anthropogenic CO2
land carbon flux on soil moisture changes in tropical regions,
cannot be understood without considering interactions with the
predicted to be sequestered in land ecosystems. The primary
and a stronger positive growth response to warming in those
other cycle. Nevertheless, previous studies have addressed the
of anthropogenic Nr additions in isolation: nitrogen cycle
Abstract. responsible
Inclusion offorfundamental
ecological
interactions
century,
in apart
to a AOGCM
steady decline
ineffects
the ocean
mechanism
increased land
carbon storage
unregions,
thanattributable
predicted by
similar
implemented
studies have focussed on understanding and quantifying the global
between
carbon
and nitrogen
cyclesisinshown
the land
component
sink fraction.
Comparison interactions.
to experimental
on
the budget
fate through the cascading effects of Nr on terrestrial
der
radiatively
forced
climate change
to be
fertiliza- of without
land carbon-nitrogen
Westudies
expectnitrogen
that
N fluxes
and emissions6 , riverine export to coastal ecosystems7 ,
an atmosphere-ocean
model
(AOGCM) the of
radio-labeled uncertainty
nitrogen tracers
in temperate
forests
tion
of plant growth by general
increasedcirculation
mineralization
of nitrogen
between-model
in predictions
of future
at- indiand the consequences of N2 O accumulating in the atmosphere on
directly
increased
soil2 fertilormospheric
COthe
and associated
anthropogenic
leads toassociated
decreased with
carbon
uptake decomposition
associated withofCO
cates that
model representation
of competition
between
2 concentration
climate
and stratospheric ozone depletion8 . Carbon cycle studies
ganic
matter
under
a
warming
climate,
which
in
this
particclimate
change
will
be
reduced
as
additional
climate
modhave focussed
ization, and increased carbon uptake associated with warmplants and microbes for new mineral nitrogen resources
is on determining the contribution of Nr deposition
on productivity and carbon storage in natural ecosystems at
We perform a set of model simulations aimed at isolating
global and regional effects of anthropogenic Nr addition due to
fertilizer applications and atmospheric deposition from those due
to changes in land-cover, atmospheric [CO2 ], and climatic changes
over the historical period (1700–2005), using a consistent set of
model forcings (see Methods and Supplementary Information).
To calculate the overall climatic forcing of the Nr perturbation,
we propagate the net land–atmosphere exchanges of carbon
dioxide and nitrous oxide, resulting from the coupled nitrogen–
carbon model, into atmospheric concentrations using simple
box models and further information about fossil, marine, and
complementary terrestrial fluxes11,12 (Fig. 1; see Methods and
Supplementary Information).
In 1960–2005, the period for which direct atmospheric measurements are available13 , the increase in atmospheric [CO2 ] and
the interannual growth rate of atmospheric [CO2 ] are reproduced
well by the model (Fig. 1a,c), which implies that the simulated
airborne fraction of anthropogenic emissions also falls into the
observed range12 . The effect of the Nr perturbation on atmospheric CO2 levels is small ( 6 ppm). The simulated terrestrial
N2 O flux, when combined with the other terrestrial N2 O fluxes,
produces a good agreement with the observed atmospheric [N2 O]
evolution since pre-industrial times11,14–16 (Fig. 1b,d). The increase
in the terrestrial N2 O source explains more than two thirds
(33 of 45 ppb) of the atmospheric increase since 1860. Nevertheless, the average simulated [N2 O] growth rate in 1990–2005
(0.93 ± 0.12 ppb yr 1 ) is about 25% higher than the observed
trend at atmospheric stations for the same period (GAGE/AGAGE
network15,16 : 0.72 ± 0.27 ppb yr 1 , Fig. 1d).
What are the underlying causes of these terrestrial CO2 and
N2 O flux anomalies? We estimate that the Nr perturbation
has increased terrestrial C storage since pre-industrial times by
12 Pg C. This value is small compared with modelled changes
in terrestrial C storage due to deforestation (a cumulative loss
of 205 Pg C), and the enhancement of plant productivity due
to CO2 fertilization (a gain of 143 Pg C). During the period
1996–2005 we estimate that Nr deposition causes roughly 20%
(0.2 Pg C yr 1 ) of the terrestrial net C uptake (1.1 Pg C yr 1 ;
Test of nitrogen limitation in two global
biogeochemical models
CLM-CN 4.0
Biogeosciences, 6, 2099–2120, 2009
www.biogeosciences.net/6/2099/2009/
© Author(s) 2009. This work is distributed under
the Creative Commons Attribution 3.0 License.
Biogeosciences
(Thornton et al. 2009 Biogeosciences)
Biogeosciences, 6, 2099–2120, 2009
www.biogeosciences.net/6/2099/2009/
© Author(s) 2009. This work is distributed under
the Creative Commons Attribution 3.0 License.
O-CN
(Zaehle et al. 2011 Nature Geoscience)
Biogeosciences
Carbon-nitrogen interactions regulate climate-carbon cycle
feedbacks: results from an atmosphere-ocean
general circulation model
P. E. Thornton1 , S. C. Doney2 , K. Lindsay3 , J. K. Moore4 , N. Mahowald5 , J. T. Randerson4 , I. Fung6 ,
J.-F. Lamarque7,8 , J. J. Feddema9 , and Y.-H. Lee3
Carbon-nitrogen interactions regulate climate-carbon cycle
1 Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6335, USA
results from an atmosphere-ocean
2feedbacks:
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole,
MA
02543-1543,
USA
general
circulation
model
3 Climate and Global Dynamics Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
LETTERS
PUBLISHED ONLINE: 31 JULY 2011 | DOI: 10.1038/NGEO1207
Carbon benefits of anthropogenic reactive
nitrogen offset by nitrous oxide emissions
Sönke Zaehle1 *, Philippe Ciais2 , Andrew D. Friend3 and Vincent Prieur2
Additions of reactive nitrogen to terrestrial ecosystems—
primarily through fertilizer application and atmospheric
deposition—have more than doubled since 1860 owing to
human activities1 . Nitrogen additions tend to increase the net
uptake of carbon by the terrestrial biosphere, but they also
4 Department of Earth System Science, University of California, Irvine, CA 92697-3100, USA
1 , S. C. Doney2 , K. Lindsay3 , J. K. Moore4 , N. Mahowald5 , J. T. Randerson4 , I. Fung6 ,
stimulate nitrous oxide release from soils2 . However, given
P.
E.
Thornton
5 Department of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY 14850, USA
that the magnitude of these effects is uncertain, and that the
7,8
9
3
Lamarque
, J.and
J. Feddema
, and Y.-H.
Lee of California, Berkeley, CA 94720-4767, USA
6J.-F.
carbon and nitrogen cycles are tightly coupled, the net climatic
Department
of Earth
Planetary Science,
University
impact of anthropogenic nitrogen inputs is unknown3 . Here
71NOAA
Earth System
Research
Laboratory,
Chemical
Sciences
Division,
325Ridge,
Broadway,
Boulder, COUSA
80305-3337, USAwe use a process-based model of the terrestrial biosphere4,5
Environmental
Sciences
Division,
Oak Ridge
National
Laboratory,
Oak
TN 37831-6335,
82Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
to evaluate the overall impact of anthropogenic nitrogen
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole,
inputs on terrestrial ecosystem carbon and nitrous oxide fluxes
9 Department of Geography, University of Kansas, Lawrence, KS 66045-7613, USA
MA 02543-1543, USA
between 1700 and 2005. We show that anthropogenic nitrogen
3 Climate and Global Dynamics Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
inputs account for about a fifth of the carbon sequestered
Received:
28 January 2009 – Published in Biogeosciences Discuss.: 26 March 2009
by terrestrial ecosystems between 1996 and 2005, and for
4 Department
of Earth
System
Science,
California,
Irvine,
CA 92697-3100,
USA
most of the increase in global nitrous oxide emissions in recent
Revised:
12 August
2009
– Accepted:
17University
September of
2009
– Published:
8 October
2009
5 Department
decades; the latter is largely due to agricultural intensification.
of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY 14850, USA
We estimate that carbon sequestration due to nitrogen
6 Department of Earth and Planetary Science, University of California, Berkeley, CA 94720-4767, USA
deposition has reduced current carbon dioxide radiative forcing
7 NOAA Earth System Research Laboratory, Chemical Sciences Division, 325 Broadway, Boulder, CO 80305-3337,byUSA
96 ± 14 mW m 2 . However, this effect has been offset by
Abstract.
Inclusion of fundamental ecological interactions
century, attributable in part to a steady decline in the ocean
the increase in radiative forcing resulting from nitrous oxide
8 Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
between carbon and nitrogen cycles in the land component of
sink fraction. Comparison to experimental studies on theemissions,
fate
which amounts to 125±20 mW m 2 .
9 Department
an
atmosphere-ocean
general University
circulation of
model
(AOGCM)
of radio-labeled
tracers in temperate forests indiof Geography,
Kansas,
Lawrence, KS
66045-7613, nitrogen
USA
The effect of the anthropogenic reactive nitrogen (hereafter Nr)
perturbation, that is the addition of ecosystem-available forms of
leads to decreased carbon uptake associated with CO2 fertilcates that the model representation of competition between
nitrogen such as ammonium and nitrate, on terrestrial ecosystems
Received:
28
January
2009
–
Published
in
Biogeosciences
Discuss.:
26
March
2009
ization, and increased carbon uptake associated with warmplants and microbes for new mineral nitrogen resources
and is
the land–atmosphere fluxes of carbon dioxide and nitrous
Revised:
August
2009 The
– Accepted:
17these
September
2009 – Published:
2009
oxide
of the tight coupling
ing
of the 12
climate
system.
balance of
two opposreasonable.8 October
Our results
suggest a weaker dependence of
nethas proved difficult to quantify because
between the carbon and nitrogen cycles1 : changes in one cycle
ing effects is to reduce the fraction of anthropogenic CO2
land carbon flux on soil moisture changes in tropical regions,
cannot be understood without considering interactions with the
predicted to be sequestered in land ecosystems. The primary
and a stronger positive growth response to warming in those
other cycle. Nevertheless, previous studies have addressed the
of anthropogenic Nr additions in isolation: nitrogen cycle
Abstract. responsible
Inclusion offorfundamental
ecological
interactions
century,
in apart
to a AOGCM
steady decline
ineffects
the ocean
mechanism
increased land
carbon storage
unregions,
thanattributable
predicted by
similar
implemented
studies have focussed on understanding and quantifying the global
between
carbon
and nitrogen
cyclesisinshown
the land
component
sink fraction.
Comparison interactions.
to experimental
on
the budget
fate through the cascading effects of Nr on terrestrial
der
radiatively
forced
climate change
to be
fertiliza- of without
land carbon-nitrogen
Westudies
expectnitrogen
that
N fluxes
and emissions6 , riverine export to coastal ecosystems7 ,
an atmosphere-ocean
model
(AOGCM) the of
radio-labeled uncertainty
nitrogen tracers
in temperate
forests
tion
of plant growth by general
increasedcirculation
mineralization
of nitrogen
between-model
in predictions
of future
at- indiand the consequences of N2 O accumulating in the atmosphere on
directly
increased
soil2 fertilormospheric
COthe
and associated
anthropogenic
leads toassociated
decreased with
carbon
uptake decomposition
associated withofCO
cates that
model representation
of competition
between
2 concentration
climate
and stratospheric ozone depletion8 . Carbon cycle studies
ganic
matter
under
a
warming
climate,
which
in
this
particclimate
change
will
be
reduced
as
additional
climate
modhave focussed
ization, and increased carbon uptake associated with warmplants and microbes for new mineral nitrogen resources
is on determining the contribution of Nr deposition
on productivity and carbon storage in natural ecosystems at
We perform a set of model simulations aimed at isolating
global and regional effects of anthropogenic Nr addition due to
fertilizer applications and atmospheric deposition from those due
to changes in land-cover, atmospheric [CO2 ], and climatic changes
over the historical period (1700–2005), using a consistent set of
model forcings (see Methods and Supplementary Information).
To calculate the overall climatic forcing of the Nr perturbation,
we propagate the net land–atmosphere exchanges of carbon
dioxide and nitrous oxide, resulting from the coupled nitrogen–
carbon model, into atmospheric concentrations using simple
box models and further information about fossil, marine, and
complementary terrestrial fluxes11,12 (Fig. 1; see Methods and
Supplementary Information).
In 1960–2005, the period for which direct atmospheric measurements are available13 , the increase in atmospheric [CO2 ] and
the interannual growth rate of atmospheric [CO2 ] are reproduced
well by the model (Fig. 1a,c), which implies that the simulated
airborne fraction of anthropogenic emissions also falls into the
observed range12 . The effect of the Nr perturbation on atmospheric CO2 levels is small ( 6 ppm). The simulated terrestrial
N2 O flux, when combined with the other terrestrial N2 O fluxes,
produces a good agreement with the observed atmospheric [N2 O]
evolution since pre-industrial times11,14–16 (Fig. 1b,d). The increase
in the terrestrial N2 O source explains more than two thirds
(33 of 45 ppb) of the atmospheric increase since 1860. Nevertheless, the average simulated [N2 O] growth rate in 1990–2005
(0.93 ± 0.12 ppb yr 1 ) is about 25% higher than the observed
trend at atmospheric stations for the same period (GAGE/AGAGE
network15,16 : 0.72 ± 0.27 ppb yr 1 , Fig. 1d).
What are the underlying causes of these terrestrial CO2 and
N2 O flux anomalies? We estimate that the Nr perturbation
has increased terrestrial C storage since pre-industrial times by
12 Pg C. This value is small compared with modelled changes
in terrestrial C storage due to deforestation (a cumulative loss
of 205 Pg C), and the enhancement of plant productivity due
to CO2 fertilization (a gain of 143 Pg C). During the period
1996–2005 we estimate that Nr deposition causes roughly 20%
(0.2 Pg C yr 1 ) of the terrestrial net C uptake (1.1 Pg C yr 1 ;
Buffering capacity of C to changes in N
Test of nitrogen limitation in two global
biogeochemical models
CLM-CN 4.0
Biogeosciences, 6, 2099–2120, 2009
www.biogeosciences.net/6/2099/2009/
© Author(s) 2009. This work is distributed under
the Creative Commons Attribution 3.0 License.
Biogeosciences
(Thornton et al. 2009 Biogeosciences)
Biogeosciences, 6, 2099–2120, 2009
www.biogeosciences.net/6/2099/2009/
© Author(s) 2009. This work is distributed under
the Creative Commons Attribution 3.0 License.
O-CN
(Zaehle et al. 2011 Nature Geoscience)
Biogeosciences
Carbon-nitrogen interactions regulate climate-carbon cycle
feedbacks: results from an atmosphere-ocean
general circulation model
P. E. Thornton1 , S. C. Doney2 , K. Lindsay3 , J. K. Moore4 , N. Mahowald5 , J. T. Randerson4 , I. Fung6 ,
J.-F. Lamarque7,8 , J. J. Feddema9 , and Y.-H. Lee3
Carbon-nitrogen interactions regulate climate-carbon cycle
1 Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6335, USA
results from an atmosphere-ocean
2feedbacks:
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole,
MA
02543-1543,
USA
general
circulation
model
3 Climate and Global Dynamics Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
LETTERS
PUBLISHED ONLINE: 31 JULY 2011 | DOI: 10.1038/NGEO1207
Carbon benefits of anthropogenic reactive
nitrogen offset by nitrous oxide emissions
Sönke Zaehle1 *, Philippe Ciais2 , Andrew D. Friend3 and Vincent Prieur2
Additions of reactive nitrogen to terrestrial ecosystems—
primarily through fertilizer application and atmospheric
deposition—have more than doubled since 1860 owing to
human activities1 . Nitrogen additions tend to increase the net
uptake of carbon by the terrestrial biosphere, but they also
4 Department of Earth System Science, University of California, Irvine, CA 92697-3100, USA
1 , S. C. Doney2 , K. Lindsay3 , J. K. Moore4 , N. Mahowald5 , J. T. Randerson4 , I. Fung6 ,
stimulate nitrous oxide release from soils2 . However, given
P.
E.
Thornton
5 Department of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY 14850, USA
that the magnitude of these effects is uncertain, and that the
7,8
9
3
Lamarque
, J.and
J. Feddema
, and Y.-H.
Lee of California, Berkeley, CA 94720-4767, USA
6J.-F.
carbon and nitrogen cycles are tightly coupled, the net climatic
Department
of Earth
Planetary Science,
University
impact of anthropogenic nitrogen inputs is unknown3 . Here
71NOAA
Earth System
Research
Laboratory,
Chemical
Sciences
Division,
325Ridge,
Broadway,
Boulder, COUSA
80305-3337, USAwe use a process-based model of the terrestrial biosphere4,5
Environmental
Sciences
Division,
Oak Ridge
National
Laboratory,
Oak
TN 37831-6335,
82Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
to evaluate the overall impact of anthropogenic nitrogen
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole,
inputs on terrestrial ecosystem carbon and nitrous oxide fluxes
9 Department of Geography, University of Kansas, Lawrence, KS 66045-7613, USA
MA 02543-1543, USA
between 1700 and 2005. We show that anthropogenic nitrogen
3 Climate and Global Dynamics Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
inputs account for about a fifth of the carbon sequestered
Received:
28 January 2009 – Published in Biogeosciences Discuss.: 26 March 2009
by terrestrial ecosystems between 1996 and 2005, and for
4 Department
of Earth
System
Science,
California,
Irvine,
CA 92697-3100,
USA
most of the increase in global nitrous oxide emissions in recent
Revised:
12 August
2009
– Accepted:
17University
September of
2009
– Published:
8 October
2009
5 Department
decades; the latter is largely due to agricultural intensification.
of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY 14850, USA
We estimate that carbon sequestration due to nitrogen
6 Department of Earth and Planetary Science, University of California, Berkeley, CA 94720-4767, USA
deposition has reduced current carbon dioxide radiative forcing
7 NOAA Earth System Research Laboratory, Chemical Sciences Division, 325 Broadway, Boulder, CO 80305-3337,byUSA
96 ± 14 mW m 2 . However, this effect has been offset by
Abstract.
Inclusion of fundamental ecological interactions
century, attributable in part to a steady decline in the ocean
the increase in radiative forcing resulting from nitrous oxide
8 Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
between carbon and nitrogen cycles in the land component of
sink fraction. Comparison to experimental studies on theemissions,
fate
which amounts to 125±20 mW m 2 .
9 Department
an
atmosphere-ocean
general University
circulation of
model
(AOGCM)
of radio-labeled
tracers in temperate forests indiof Geography,
Kansas,
Lawrence, KS
66045-7613, nitrogen
USA
The effect of the anthropogenic reactive nitrogen (hereafter Nr)
perturbation, that is the addition of ecosystem-available forms of
leads to decreased carbon uptake associated with CO2 fertilcates that the model representation of competition between
nitrogen such as ammonium and nitrate, on terrestrial ecosystems
Received:
28
January
2009
–
Published
in
Biogeosciences
Discuss.:
26
March
2009
ization, and increased carbon uptake associated with warmplants and microbes for new mineral nitrogen resources
and is
the land–atmosphere fluxes of carbon dioxide and nitrous
Revised:
August
2009 The
– Accepted:
17these
September
2009 – Published:
2009
oxide
of the tight coupling
ing
of the 12
climate
system.
balance of
two opposreasonable.8 October
Our results
suggest a weaker dependence of
nethas proved difficult to quantify because
between the carbon and nitrogen cycles1 : changes in one cycle
ing effects is to reduce the fraction of anthropogenic CO2
land carbon flux on soil moisture changes in tropical regions,
cannot be understood without considering interactions with the
predicted to be sequestered in land ecosystems. The primary
and a stronger positive growth response to warming in those
other cycle. Nevertheless, previous studies have addressed the
of anthropogenic Nr additions in isolation: nitrogen cycle
Abstract. responsible
Inclusion offorfundamental
ecological
interactions
century,
in apart
to a AOGCM
steady decline
ineffects
the ocean
mechanism
increased land
carbon storage
unregions,
thanattributable
predicted by
similar
implemented
studies have focussed on understanding and quantifying the global
between
carbon
and nitrogen
cyclesisinshown
the land
component
sink fraction.
Comparison interactions.
to experimental
on
the budget
fate through the cascading effects of Nr on terrestrial
der
radiatively
forced
climate change
to be
fertiliza- of without
land carbon-nitrogen
Westudies
expectnitrogen
that
N fluxes
and emissions6 , riverine export to coastal ecosystems7 ,
an atmosphere-ocean
model
(AOGCM) the of
radio-labeled uncertainty
nitrogen tracers
in temperate
forests
tion
of plant growth by general
increasedcirculation
mineralization
of nitrogen
between-model
in predictions
of future
at- indiand the consequences of N2 O accumulating in the atmosphere on
directly
increased
soil2 fertilormospheric
COthe
and associated
anthropogenic
leads toassociated
decreased with
carbon
uptake decomposition
associated withofCO
cates that
model representation
of competition
between
2 concentration
climate
and stratospheric ozone depletion8 . Carbon cycle studies
ganic
matter
under
a
warming
climate,
which
in
this
particclimate
change
will
be
reduced
as
additional
climate
modhave focussed
ization, and increased carbon uptake associated with warmplants and microbes for new mineral nitrogen resources
is on determining the contribution of Nr deposition
on productivity and carbon storage in natural ecosystems at
We perform a set of model simulations aimed at isolating
global and regional effects of anthropogenic Nr addition due to
fertilizer applications and atmospheric deposition from those due
to changes in land-cover, atmospheric [CO2 ], and climatic changes
over the historical period (1700–2005), using a consistent set of
model forcings (see Methods and Supplementary Information).
To calculate the overall climatic forcing of the Nr perturbation,
we propagate the net land–atmosphere exchanges of carbon
dioxide and nitrous oxide, resulting from the coupled nitrogen–
carbon model, into atmospheric concentrations using simple
box models and further information about fossil, marine, and
complementary terrestrial fluxes11,12 (Fig. 1; see Methods and
Supplementary Information).
In 1960–2005, the period for which direct atmospheric measurements are available13 , the increase in atmospheric [CO2 ] and
the interannual growth rate of atmospheric [CO2 ] are reproduced
well by the model (Fig. 1a,c), which implies that the simulated
airborne fraction of anthropogenic emissions also falls into the
observed range12 . The effect of the Nr perturbation on atmospheric CO2 levels is small ( 6 ppm). The simulated terrestrial
N2 O flux, when combined with the other terrestrial N2 O fluxes,
produces a good agreement with the observed atmospheric [N2 O]
evolution since pre-industrial times11,14–16 (Fig. 1b,d). The increase
in the terrestrial N2 O source explains more than two thirds
(33 of 45 ppb) of the atmospheric increase since 1860. Nevertheless, the average simulated [N2 O] growth rate in 1990–2005
(0.93 ± 0.12 ppb yr 1 ) is about 25% higher than the observed
trend at atmospheric stations for the same period (GAGE/AGAGE
network15,16 : 0.72 ± 0.27 ppb yr 1 , Fig. 1d).
What are the underlying causes of these terrestrial CO2 and
N2 O flux anomalies? We estimate that the Nr perturbation
has increased terrestrial C storage since pre-industrial times by
12 Pg C. This value is small compared with modelled changes
in terrestrial C storage due to deforestation (a cumulative loss
of 205 Pg C), and the enhancement of plant productivity due
to CO2 fertilization (a gain of 143 Pg C). During the period
1996–2005 we estimate that Nr deposition causes roughly 20%
(0.2 Pg C yr 1 ) of the terrestrial net C uptake (1.1 Pg C yr 1 ;
Buffering capacity of C to changes in N
How potential primary productivity is limited by nitrogen
Test of nitrogen limitation in two global
biogeochemical models
CLM-CN 4.0
Biogeosciences, 6, 2099–2120, 2009
www.biogeosciences.net/6/2099/2009/
© Author(s) 2009. This work is distributed under
the Creative Commons Attribution 3.0 License.
Biogeosciences
(Thornton et al. 2009 Biogeosciences)
Biogeosciences, 6, 2099–2120, 2009
www.biogeosciences.net/6/2099/2009/
© Author(s) 2009. This work is distributed under
the Creative Commons Attribution 3.0 License.
O-CN
(Zaehle et al. 2011 Nature Geoscience)
Biogeosciences
Carbon-nitrogen interactions regulate climate-carbon cycle
feedbacks: results from an atmosphere-ocean
general circulation model
P. E. Thornton1 , S. C. Doney2 , K. Lindsay3 , J. K. Moore4 , N. Mahowald5 , J. T. Randerson4 , I. Fung6 ,
J.-F. Lamarque7,8 , J. J. Feddema9 , and Y.-H. Lee3
Carbon-nitrogen interactions regulate climate-carbon cycle
1 Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6335, USA
results from an atmosphere-ocean
2feedbacks:
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole,
MA
02543-1543,
USA
general
circulation
model
3 Climate and Global Dynamics Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
LETTERS
PUBLISHED ONLINE: 31 JULY 2011 | DOI: 10.1038/NGEO1207
Carbon benefits of anthropogenic reactive
nitrogen offset by nitrous oxide emissions
Sönke Zaehle1 *, Philippe Ciais2 , Andrew D. Friend3 and Vincent Prieur2
Additions of reactive nitrogen to terrestrial ecosystems—
primarily through fertilizer application and atmospheric
deposition—have more than doubled since 1860 owing to
human activities1 . Nitrogen additions tend to increase the net
uptake of carbon by the terrestrial biosphere, but they also
4 Department of Earth System Science, University of California, Irvine, CA 92697-3100, USA
1 , S. C. Doney2 , K. Lindsay3 , J. K. Moore4 , N. Mahowald5 , J. T. Randerson4 , I. Fung6 ,
stimulate nitrous oxide release from soils2 . However, given
P.
E.
Thornton
5 Department of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY 14850, USA
that the magnitude of these effects is uncertain, and that the
7,8
9
3
Lamarque
, J.and
J. Feddema
, and Y.-H.
Lee of California, Berkeley, CA 94720-4767, USA
6J.-F.
carbon and nitrogen cycles are tightly coupled, the net climatic
Department
of Earth
Planetary Science,
University
impact of anthropogenic nitrogen inputs is unknown3 . Here
71NOAA
Earth System
Research
Laboratory,
Chemical
Sciences
Division,
325Ridge,
Broadway,
Boulder, COUSA
80305-3337, USAwe use a process-based model of the terrestrial biosphere4,5
Environmental
Sciences
Division,
Oak Ridge
National
Laboratory,
Oak
TN 37831-6335,
82Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
to evaluate the overall impact of anthropogenic nitrogen
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole,
inputs on terrestrial ecosystem carbon and nitrous oxide fluxes
9 Department of Geography, University of Kansas, Lawrence, KS 66045-7613, USA
MA 02543-1543, USA
between 1700 and 2005. We show that anthropogenic nitrogen
3 Climate and Global Dynamics Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
inputs account for about a fifth of the carbon sequestered
Received:
28 January 2009 – Published in Biogeosciences Discuss.: 26 March 2009
by terrestrial ecosystems between 1996 and 2005, and for
4 Department
of Earth
System
Science,
California,
Irvine,
CA 92697-3100,
USA
most of the increase in global nitrous oxide emissions in recent
Revised:
12 August
2009
– Accepted:
17University
September of
2009
– Published:
8 October
2009
5 Department
decades; the latter is largely due to agricultural intensification.
of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY 14850, USA
We estimate that carbon sequestration due to nitrogen
6 Department of Earth and Planetary Science, University of California, Berkeley, CA 94720-4767, USA
deposition has reduced current carbon dioxide radiative forcing
7 NOAA Earth System Research Laboratory, Chemical Sciences Division, 325 Broadway, Boulder, CO 80305-3337,byUSA
96 ± 14 mW m 2 . However, this effect has been offset by
Abstract.
Inclusion of fundamental ecological interactions
century, attributable in part to a steady decline in the ocean
the increase in radiative forcing resulting from nitrous oxide
8 Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
between carbon and nitrogen cycles in the land component of
sink fraction. Comparison to experimental studies on theemissions,
fate
which amounts to 125±20 mW m 2 .
9 Department
an
atmosphere-ocean
general University
circulation of
model
(AOGCM)
of radio-labeled
tracers in temperate forests indiof Geography,
Kansas,
Lawrence, KS
66045-7613, nitrogen
USA
The effect of the anthropogenic reactive nitrogen (hereafter Nr)
perturbation, that is the addition of ecosystem-available forms of
leads to decreased carbon uptake associated with CO2 fertilcates that the model representation of competition between
nitrogen such as ammonium and nitrate, on terrestrial ecosystems
Received:
28
January
2009
–
Published
in
Biogeosciences
Discuss.:
26
March
2009
ization, and increased carbon uptake associated with warmplants and microbes for new mineral nitrogen resources
and is
the land–atmosphere fluxes of carbon dioxide and nitrous
Revised:
August
2009 The
– Accepted:
17these
September
2009 – Published:
2009
oxide
of the tight coupling
ing
of the 12
climate
system.
balance of
two opposreasonable.8 October
Our results
suggest a weaker dependence of
nethas proved difficult to quantify because
between the carbon and nitrogen cycles1 : changes in one cycle
ing effects is to reduce the fraction of anthropogenic CO2
land carbon flux on soil moisture changes in tropical regions,
cannot be understood without considering interactions with the
predicted to be sequestered in land ecosystems. The primary
and a stronger positive growth response to warming in those
other cycle. Nevertheless, previous studies have addressed the
of anthropogenic Nr additions in isolation: nitrogen cycle
Abstract. responsible
Inclusion offorfundamental
ecological
interactions
century,
in apart
to a AOGCM
steady decline
ineffects
the ocean
mechanism
increased land
carbon storage
unregions,
thanattributable
predicted by
similar
implemented
studies have focussed on understanding and quantifying the global
between
carbon
and nitrogen
cyclesisinshown
the land
component
sink fraction.
Comparison interactions.
to experimental
on
the budget
fate through the cascading effects of Nr on terrestrial
der
radiatively
forced
climate change
to be
fertiliza- of without
land carbon-nitrogen
Westudies
expectnitrogen
that
N fluxes
and emissions6 , riverine export to coastal ecosystems7 ,
an atmosphere-ocean
model
(AOGCM) the of
radio-labeled uncertainty
nitrogen tracers
in temperate
forests
tion
of plant growth by general
increasedcirculation
mineralization
of nitrogen
between-model
in predictions
of future
at- indiand the consequences of N2 O accumulating in the atmosphere on
directly
increased
soil2 fertilormospheric
COthe
and associated
anthropogenic
leads toassociated
decreased with
carbon
uptake decomposition
associated withofCO
cates that
model representation
of competition
between
2 concentration
climate
and stratospheric ozone depletion8 . Carbon cycle studies
ganic
matter
under
a
warming
climate,
which
in
this
particclimate
change
will
be
reduced
as
additional
climate
modhave focussed
ization, and increased carbon uptake associated with warmplants and microbes for new mineral nitrogen resources
is on determining the contribution of Nr deposition
on productivity and carbon storage in natural ecosystems at
We perform a set of model simulations aimed at isolating
global and regional effects of anthropogenic Nr addition due to
fertilizer applications and atmospheric deposition from those due
to changes in land-cover, atmospheric [CO2 ], and climatic changes
over the historical period (1700–2005), using a consistent set of
model forcings (see Methods and Supplementary Information).
To calculate the overall climatic forcing of the Nr perturbation,
we propagate the net land–atmosphere exchanges of carbon
dioxide and nitrous oxide, resulting from the coupled nitrogen–
carbon model, into atmospheric concentrations using simple
box models and further information about fossil, marine, and
complementary terrestrial fluxes11,12 (Fig. 1; see Methods and
Supplementary Information).
In 1960–2005, the period for which direct atmospheric measurements are available13 , the increase in atmospheric [CO2 ] and
the interannual growth rate of atmospheric [CO2 ] are reproduced
well by the model (Fig. 1a,c), which implies that the simulated
airborne fraction of anthropogenic emissions also falls into the
observed range12 . The effect of the Nr perturbation on atmospheric CO2 levels is small ( 6 ppm). The simulated terrestrial
N2 O flux, when combined with the other terrestrial N2 O fluxes,
produces a good agreement with the observed atmospheric [N2 O]
evolution since pre-industrial times11,14–16 (Fig. 1b,d). The increase
in the terrestrial N2 O source explains more than two thirds
(33 of 45 ppb) of the atmospheric increase since 1860. Nevertheless, the average simulated [N2 O] growth rate in 1990–2005
(0.93 ± 0.12 ppb yr 1 ) is about 25% higher than the observed
trend at atmospheric stations for the same period (GAGE/AGAGE
network15,16 : 0.72 ± 0.27 ppb yr 1 , Fig. 1d).
What are the underlying causes of these terrestrial CO2 and
N2 O flux anomalies? We estimate that the Nr perturbation
has increased terrestrial C storage since pre-industrial times by
12 Pg C. This value is small compared with modelled changes
in terrestrial C storage due to deforestation (a cumulative loss
of 205 Pg C), and the enhancement of plant productivity due
to CO2 fertilization (a gain of 143 Pg C). During the period
1996–2005 we estimate that Nr deposition causes roughly 20%
(0.2 Pg C yr 1 ) of the terrestrial net C uptake (1.1 Pg C yr 1 ;
Buffering capacity of C to changes in N
How potential primary productivity is limited by nitrogen
Fixed Vegetation C:N
Test of nitrogen limitation in two global
biogeochemical models
CLM-CN 4.0
Biogeosciences, 6, 2099–2120, 2009
www.biogeosciences.net/6/2099/2009/
© Author(s) 2009. This work is distributed under
the Creative Commons Attribution 3.0 License.
Biogeosciences
(Thornton et al. 2009 Biogeosciences)
Biogeosciences, 6, 2099–2120, 2009
www.biogeosciences.net/6/2099/2009/
© Author(s) 2009. This work is distributed under
the Creative Commons Attribution 3.0 License.
O-CN
(Zaehle et al. 2011 Nature Geoscience)
LETTERS
Biogeosciences
Carbon-nitrogen interactions regulate climate-carbon cycle
feedbacks: results from an atmosphere-ocean
general circulation model
P. E. Thornton1 , S. C. Doney2 , K. Lindsay3 , J. K. Moore4 , N. Mahowald5 , J. T. Randerson4 , I. Fung6 ,
J.-F. Lamarque7,8 , J. J. Feddema9 , and Y.-H. Lee3
Carbon-nitrogen interactions regulate climate-carbon cycle
1 Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6335, USA
results from an atmosphere-ocean
2feedbacks:
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole,
MA
02543-1543,
USA
general
circulation
model
3 Climate and Global Dynamics Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
PUBLISHED ONLINE: 31 JULY 2011 | DOI: 10.1038/NGEO1207
Carbon benefits of anthropogenic reactive
nitrogen offset by nitrous oxide emissions
Sönke Zaehle1 *, Philippe Ciais2 , Andrew D. Friend3 and Vincent Prieur2
Additions of reactive nitrogen to terrestrial ecosystems—
primarily through fertilizer application and atmospheric
deposition—have more than doubled since 1860 owing to
human activities1 . Nitrogen additions tend to increase the net
uptake of carbon by the terrestrial biosphere, but they also
4 Department of Earth System Science, University of California, Irvine, CA 92697-3100, USA
1 , S. C. Doney2 , K. Lindsay3 , J. K. Moore4 , N. Mahowald5 , J. T. Randerson4 , I. Fung6 ,
stimulate nitrous oxide release from soils2 . However, given
P.
E.
Thornton
5 Department of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY 14850, USA
that the magnitude of these effects is uncertain, and that the
7,8
9
3
Lamarque
, J.and
J. Feddema
, and Y.-H.
Lee of California, Berkeley, CA 94720-4767, USA
6J.-F.
carbon and nitrogen cycles are tightly coupled, the net climatic
Department
of Earth
Planetary Science,
University
impact of anthropogenic nitrogen inputs is unknown3 . Here
71NOAA
Earth System
Research
Laboratory,
Chemical
Sciences
Division,
325Ridge,
Broadway,
Boulder, COUSA
80305-3337, USAwe use a process-based model of the terrestrial biosphere4,5
Environmental
Sciences
Division,
Oak Ridge
National
Laboratory,
Oak
TN 37831-6335,
82Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
to evaluate the overall impact of anthropogenic nitrogen
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole,
inputs on terrestrial ecosystem carbon and nitrous oxide fluxes
9 Department of Geography, University of Kansas, Lawrence, KS 66045-7613, USA
MA 02543-1543, USA
between 1700 and 2005. We show that anthropogenic nitrogen
3 Climate and Global Dynamics Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
inputs account for about a fifth of the carbon sequestered
Received:
28 January 2009 – Published in Biogeosciences Discuss.: 26 March 2009
by terrestrial ecosystems between 1996 and 2005, and for
4 Department
of Earth
System
Science,
California,
Irvine,
CA 92697-3100,
USA
most of the increase in global nitrous oxide emissions in recent
Revised:
12 August
2009
– Accepted:
17University
September of
2009
– Published:
8 October
2009
5 Department
decades; the latter is largely due to agricultural intensification.
of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY 14850, USA
We estimate that carbon sequestration due to nitrogen
6 Department of Earth and Planetary Science, University of California, Berkeley, CA 94720-4767, USA
deposition has reduced current carbon dioxide radiative forcing
7 NOAA Earth System Research Laboratory, Chemical Sciences Division, 325 Broadway, Boulder, CO 80305-3337,byUSA
96 ± 14 mW m 2 . However, this effect has been offset by
Abstract.
Inclusion of fundamental ecological interactions
century, attributable in part to a steady decline in the ocean
the increase in radiative forcing resulting from nitrous oxide
8 Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
between carbon and nitrogen cycles in the land component of
sink fraction. Comparison to experimental studies on theemissions,
fate
which amounts to 125±20 mW m 2 .
9 Department
an
atmosphere-ocean
general University
circulation of
model
(AOGCM)
of radio-labeled
tracers in temperate forests indiof Geography,
Kansas,
Lawrence, KS
66045-7613, nitrogen
USA
The effect of the anthropogenic reactive nitrogen (hereafter Nr)
perturbation, that is the addition of ecosystem-available forms of
leads to decreased carbon uptake associated with CO2 fertilcates that the model representation of competition between
nitrogen such as ammonium and nitrate, on terrestrial ecosystems
Received:
28
January
2009
–
Published
in
Biogeosciences
Discuss.:
26
March
2009
ization, and increased carbon uptake associated with warmplants and microbes for new mineral nitrogen resources
and is
the land–atmosphere fluxes of carbon dioxide and nitrous
Revised:
August
2009 The
– Accepted:
17these
September
2009 – Published:
2009
oxide
of the tight coupling
ing
of the 12
climate
system.
balance of
two opposreasonable.8 October
Our results
suggest a weaker dependence of
nethas proved difficult to quantify because
between the carbon and nitrogen cycles1 : changes in one cycle
ing effects is to reduce the fraction of anthropogenic CO2
land carbon flux on soil moisture changes in tropical regions,
cannot be understood without considering interactions with the
predicted to be sequestered in land ecosystems. The primary
and a stronger positive growth response to warming in those
other cycle. Nevertheless, previous studies have addressed the
of anthropogenic Nr additions in isolation: nitrogen cycle
Abstract. responsible
Inclusion offorfundamental
ecological
interactions
century,
in apart
to a AOGCM
steady decline
ineffects
the ocean
mechanism
increased land
carbon storage
unregions,
thanattributable
predicted by
similar
implemented
studies have focussed on understanding and quantifying the global
between
carbon
and nitrogen
cyclesisinshown
the land
component
sink fraction.
Comparison interactions.
to experimental
on
the budget
fate through the cascading effects of Nr on terrestrial
der
radiatively
forced
climate change
to be
fertiliza- of without
land carbon-nitrogen
Westudies
expectnitrogen
that
N fluxes
and emissions6 , riverine export to coastal ecosystems7 ,
an atmosphere-ocean
model
(AOGCM) the of
radio-labeled uncertainty
nitrogen tracers
in temperate
forests
tion
of plant growth by general
increasedcirculation
mineralization
of nitrogen
between-model
in predictions
of future
at- indiand the consequences of N2 O accumulating in the atmosphere on
directly
increased
soil2 fertilormospheric
COthe
and associated
anthropogenic
leads toassociated
decreased with
carbon
uptake decomposition
associated withofCO
cates that
model representation
of competition
between
2 concentration
climate
and stratospheric ozone depletion8 . Carbon cycle studies
ganic
matter
under
a
warming
climate,
which
in
this
particclimate
change
will
be
reduced
as
additional
climate
modhave focussed
ization, and increased carbon uptake associated with warmplants and microbes for new mineral nitrogen resources
is on determining the contribution of Nr deposition
on productivity and carbon storage in natural ecosystems at
We perform a set of model simulations aimed at isolating
global and regional effects of anthropogenic Nr addition due to
fertilizer applications and atmospheric deposition from those due
to changes in land-cover, atmospheric [CO2 ], and climatic changes
over the historical period (1700–2005), using a consistent set of
model forcings (see Methods and Supplementary Information).
To calculate the overall climatic forcing of the Nr perturbation,
we propagate the net land–atmosphere exchanges of carbon
dioxide and nitrous oxide, resulting from the coupled nitrogen–
carbon model, into atmospheric concentrations using simple
box models and further information about fossil, marine, and
complementary terrestrial fluxes11,12 (Fig. 1; see Methods and
Supplementary Information).
In 1960–2005, the period for which direct atmospheric measurements are available13 , the increase in atmospheric [CO2 ] and
the interannual growth rate of atmospheric [CO2 ] are reproduced
well by the model (Fig. 1a,c), which implies that the simulated
airborne fraction of anthropogenic emissions also falls into the
observed range12 . The effect of the Nr perturbation on atmospheric CO2 levels is small ( 6 ppm). The simulated terrestrial
N2 O flux, when combined with the other terrestrial N2 O fluxes,
produces a good agreement with the observed atmospheric [N2 O]
evolution since pre-industrial times11,14–16 (Fig. 1b,d). The increase
in the terrestrial N2 O source explains more than two thirds
(33 of 45 ppb) of the atmospheric increase since 1860. Nevertheless, the average simulated [N2 O] growth rate in 1990–2005
(0.93 ± 0.12 ppb yr 1 ) is about 25% higher than the observed
trend at atmospheric stations for the same period (GAGE/AGAGE
network15,16 : 0.72 ± 0.27 ppb yr 1 , Fig. 1d).
What are the underlying causes of these terrestrial CO2 and
N2 O flux anomalies? We estimate that the Nr perturbation
has increased terrestrial C storage since pre-industrial times by
12 Pg C. This value is small compared with modelled changes
in terrestrial C storage due to deforestation (a cumulative loss
of 205 Pg C), and the enhancement of plant productivity due
to CO2 fertilization (a gain of 143 Pg C). During the period
1996–2005 we estimate that Nr deposition causes roughly 20%
(0.2 Pg C yr 1 ) of the terrestrial net C uptake (1.1 Pg C yr 1 ;
Buffering capacity of C to changes in N
How potential primary productivity is limited by nitrogen
Fixed Vegetation C:N
Variable Vegetation C:N
Test of nitrogen limitation in two global
biogeochemical models
CLM-CN 4.0
Biogeosciences, 6, 2099–2120, 2009
www.biogeosciences.net/6/2099/2009/
© Author(s) 2009. This work is distributed under
the Creative Commons Attribution 3.0 License.
Biogeosciences
(Thornton et al. 2009 Biogeosciences)
Biogeosciences, 6, 2099–2120, 2009
www.biogeosciences.net/6/2099/2009/
© Author(s) 2009. This work is distributed under
the Creative Commons Attribution 3.0 License.
O-CN
(Zaehle et al. 2011 Nature Geoscience)
LETTERS
Biogeosciences
Carbon-nitrogen interactions regulate climate-carbon cycle
feedbacks: results from an atmosphere-ocean
general circulation model
P. E. Thornton1 , S. C. Doney2 , K. Lindsay3 , J. K. Moore4 , N. Mahowald5 , J. T. Randerson4 , I. Fung6 ,
J.-F. Lamarque7,8 , J. J. Feddema9 , and Y.-H. Lee3
Carbon-nitrogen interactions regulate climate-carbon cycle
1 Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6335, USA
results from an atmosphere-ocean
2feedbacks:
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole,
MA
02543-1543,
USA
general
circulation
model
3 Climate and Global Dynamics Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
PUBLISHED ONLINE: 31 JULY 2011 | DOI: 10.1038/NGEO1207
Carbon benefits of anthropogenic reactive
nitrogen offset by nitrous oxide emissions
Sönke Zaehle1 *, Philippe Ciais2 , Andrew D. Friend3 and Vincent Prieur2
Additions of reactive nitrogen to terrestrial ecosystems—
primarily through fertilizer application and atmospheric
deposition—have more than doubled since 1860 owing to
human activities1 . Nitrogen additions tend to increase the net
uptake of carbon by the terrestrial biosphere, but they also
4 Department of Earth System Science, University of California, Irvine, CA 92697-3100, USA
1 , S. C. Doney2 , K. Lindsay3 , J. K. Moore4 , N. Mahowald5 , J. T. Randerson4 , I. Fung6 ,
stimulate nitrous oxide release from soils2 . However, given
P.
E.
Thornton
5 Department of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY 14850, USA
that the magnitude of these effects is uncertain, and that the
7,8
9
3
Lamarque
, J.and
J. Feddema
, and Y.-H.
Lee of California, Berkeley, CA 94720-4767, USA
6J.-F.
carbon and nitrogen cycles are tightly coupled, the net climatic
Department
of Earth
Planetary Science,
University
impact of anthropogenic nitrogen inputs is unknown3 . Here
71NOAA
Earth System
Research
Laboratory,
Chemical
Sciences
Division,
325Ridge,
Broadway,
Boulder, COUSA
80305-3337, USAwe use a process-based model of the terrestrial biosphere4,5
Environmental
Sciences
Division,
Oak Ridge
National
Laboratory,
Oak
TN 37831-6335,
82Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
to evaluate the overall impact of anthropogenic nitrogen
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole,
inputs on terrestrial ecosystem carbon and nitrous oxide fluxes
9 Department of Geography, University of Kansas, Lawrence, KS 66045-7613, USA
MA 02543-1543, USA
between 1700 and 2005. We show that anthropogenic nitrogen
3 Climate and Global Dynamics Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
inputs account for about a fifth of the carbon sequestered
Received:
28 January 2009 – Published in Biogeosciences Discuss.: 26 March 2009
by terrestrial ecosystems between 1996 and 2005, and for
4 Department
of Earth
System
Science,
California,
Irvine,
CA 92697-3100,
USA
most of the increase in global nitrous oxide emissions in recent
Revised:
12 August
2009
– Accepted:
17University
September of
2009
– Published:
8 October
2009
5 Department
decades; the latter is largely due to agricultural intensification.
of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY 14850, USA
We estimate that carbon sequestration due to nitrogen
6 Department of Earth and Planetary Science, University of California, Berkeley, CA 94720-4767, USA
deposition has reduced current carbon dioxide radiative forcing
7 NOAA Earth System Research Laboratory, Chemical Sciences Division, 325 Broadway, Boulder, CO 80305-3337,byUSA
96 ± 14 mW m 2 . However, this effect has been offset by
Abstract.
Inclusion of fundamental ecological interactions
century, attributable in part to a steady decline in the ocean
the increase in radiative forcing resulting from nitrous oxide
8 Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
between carbon and nitrogen cycles in the land component of
sink fraction. Comparison to experimental studies on theemissions,
fate
which amounts to 125±20 mW m 2 .
9 Department
an
atmosphere-ocean
general University
circulation of
model
(AOGCM)
of radio-labeled
tracers in temperate forests indiof Geography,
Kansas,
Lawrence, KS
66045-7613, nitrogen
USA
The effect of the anthropogenic reactive nitrogen (hereafter Nr)
perturbation, that is the addition of ecosystem-available forms of
leads to decreased carbon uptake associated with CO2 fertilcates that the model representation of competition between
nitrogen such as ammonium and nitrate, on terrestrial ecosystems
Received:
28
January
2009
–
Published
in
Biogeosciences
Discuss.:
26
March
2009
ization, and increased carbon uptake associated with warmplants and microbes for new mineral nitrogen resources
and is
the land–atmosphere fluxes of carbon dioxide and nitrous
Revised:
August
2009 The
– Accepted:
17these
September
2009 – Published:
2009
oxide
of the tight coupling
ing
of the 12
climate
system.
balance of
two opposreasonable.8 October
Our results
suggest a weaker dependence of
nethas proved difficult to quantify because
between the carbon and nitrogen cycles1 : changes in one cycle
ing effects is to reduce the fraction of anthropogenic CO2
land carbon flux on soil moisture changes in tropical regions,
cannot be understood without considering interactions with the
predicted to be sequestered in land ecosystems. The primary
and a stronger positive growth response to warming in those
other cycle. Nevertheless, previous studies have addressed the
of anthropogenic Nr additions in isolation: nitrogen cycle
Abstract. responsible
Inclusion offorfundamental
ecological
interactions
century,
in apart
to a AOGCM
steady decline
ineffects
the ocean
mechanism
increased land
carbon storage
unregions,
thanattributable
predicted by
similar
implemented
studies have focussed on understanding and quantifying the global
between
carbon
and nitrogen
cyclesisinshown
the land
component
sink fraction.
Comparison interactions.
to experimental
on
the budget
fate through the cascading effects of Nr on terrestrial
der
radiatively
forced
climate change
to be
fertiliza- of without
land carbon-nitrogen
Westudies
expectnitrogen
that
N fluxes
and emissions6 , riverine export to coastal ecosystems7 ,
an atmosphere-ocean
model
(AOGCM) the of
radio-labeled uncertainty
nitrogen tracers
in temperate
forests
tion
of plant growth by general
increasedcirculation
mineralization
of nitrogen
between-model
in predictions
of future
at- indiand the consequences of N2 O accumulating in the atmosphere on
directly
increased
soil2 fertilormospheric
COthe
and associated
anthropogenic
leads toassociated
decreased with
carbon
uptake decomposition
associated withofCO
cates that
model representation
of competition
between
2 concentration
climate
and stratospheric ozone depletion8 . Carbon cycle studies
ganic
matter
under
a
warming
climate,
which
in
this
particclimate
change
will
be
reduced
as
additional
climate
modhave focussed
ization, and increased carbon uptake associated with warmplants and microbes for new mineral nitrogen resources
is on determining the contribution of Nr deposition
on productivity and carbon storage in natural ecosystems at
We perform a set of model simulations aimed at isolating
global and regional effects of anthropogenic Nr addition due to
fertilizer applications and atmospheric deposition from those due
to changes in land-cover, atmospheric [CO2 ], and climatic changes
over the historical period (1700–2005), using a consistent set of
model forcings (see Methods and Supplementary Information).
To calculate the overall climatic forcing of the Nr perturbation,
we propagate the net land–atmosphere exchanges of carbon
dioxide and nitrous oxide, resulting from the coupled nitrogen–
carbon model, into atmospheric concentrations using simple
box models and further information about fossil, marine, and
complementary terrestrial fluxes11,12 (Fig. 1; see Methods and
Supplementary Information).
In 1960–2005, the period for which direct atmospheric measurements are available13 , the increase in atmospheric [CO2 ] and
the interannual growth rate of atmospheric [CO2 ] are reproduced
well by the model (Fig. 1a,c), which implies that the simulated
airborne fraction of anthropogenic emissions also falls into the
observed range12 . The effect of the Nr perturbation on atmospheric CO2 levels is small ( 6 ppm). The simulated terrestrial
N2 O flux, when combined with the other terrestrial N2 O fluxes,
produces a good agreement with the observed atmospheric [N2 O]
evolution since pre-industrial times11,14–16 (Fig. 1b,d). The increase
in the terrestrial N2 O source explains more than two thirds
(33 of 45 ppb) of the atmospheric increase since 1860. Nevertheless, the average simulated [N2 O] growth rate in 1990–2005
(0.93 ± 0.12 ppb yr 1 ) is about 25% higher than the observed
trend at atmospheric stations for the same period (GAGE/AGAGE
network15,16 : 0.72 ± 0.27 ppb yr 1 , Fig. 1d).
What are the underlying causes of these terrestrial CO2 and
N2 O flux anomalies? We estimate that the Nr perturbation
has increased terrestrial C storage since pre-industrial times by
12 Pg C. This value is small compared with modelled changes
in terrestrial C storage due to deforestation (a cumulative loss
of 205 Pg C), and the enhancement of plant productivity due
to CO2 fertilization (a gain of 143 Pg C). During the period
1996–2005 we estimate that Nr deposition causes roughly 20%
(0.2 Pg C yr 1 ) of the terrestrial net C uptake (1.1 Pg C yr 1 ;
Buffering capacity of C to changes in N
How potential primary productivity is limited by nitrogen
Fixed Vegetation C:N
Fixed Soil Organic Matter C:N
Variable Vegetation C:N
Test of nitrogen limitation in two global
biogeochemical models
CLM-CN 4.0
Biogeosciences, 6, 2099–2120, 2009
www.biogeosciences.net/6/2099/2009/
© Author(s) 2009. This work is distributed under
the Creative Commons Attribution 3.0 License.
Biogeosciences
(Thornton et al. 2009 Biogeosciences)
Biogeosciences, 6, 2099–2120, 2009
www.biogeosciences.net/6/2099/2009/
© Author(s) 2009. This work is distributed under
the Creative Commons Attribution 3.0 License.
O-CN
(Zaehle et al. 2011 Nature Geoscience)
LETTERS
Biogeosciences
Carbon-nitrogen interactions regulate climate-carbon cycle
feedbacks: results from an atmosphere-ocean
general circulation model
P. E. Thornton1 , S. C. Doney2 , K. Lindsay3 , J. K. Moore4 , N. Mahowald5 , J. T. Randerson4 , I. Fung6 ,
J.-F. Lamarque7,8 , J. J. Feddema9 , and Y.-H. Lee3
Carbon-nitrogen interactions regulate climate-carbon cycle
1 Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6335, USA
results from an atmosphere-ocean
2feedbacks:
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole,
MA
02543-1543,
USA
general
circulation
model
3 Climate and Global Dynamics Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
PUBLISHED ONLINE: 31 JULY 2011 | DOI: 10.1038/NGEO1207
Carbon benefits of anthropogenic reactive
nitrogen offset by nitrous oxide emissions
Sönke Zaehle1 *, Philippe Ciais2 , Andrew D. Friend3 and Vincent Prieur2
Additions of reactive nitrogen to terrestrial ecosystems—
primarily through fertilizer application and atmospheric
deposition—have more than doubled since 1860 owing to
human activities1 . Nitrogen additions tend to increase the net
uptake of carbon by the terrestrial biosphere, but they also
4 Department of Earth System Science, University of California, Irvine, CA 92697-3100, USA
1 , S. C. Doney2 , K. Lindsay3 , J. K. Moore4 , N. Mahowald5 , J. T. Randerson4 , I. Fung6 ,
stimulate nitrous oxide release from soils2 . However, given
P.
E.
Thornton
5 Department of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY 14850, USA
that the magnitude of these effects is uncertain, and that the
7,8
9
3
Lamarque
, J.and
J. Feddema
, and Y.-H.
Lee of California, Berkeley, CA 94720-4767, USA
6J.-F.
carbon and nitrogen cycles are tightly coupled, the net climatic
Department
of Earth
Planetary Science,
University
impact of anthropogenic nitrogen inputs is unknown3 . Here
71NOAA
Earth System
Research
Laboratory,
Chemical
Sciences
Division,
325Ridge,
Broadway,
Boulder, COUSA
80305-3337, USAwe use a process-based model of the terrestrial biosphere4,5
Environmental
Sciences
Division,
Oak Ridge
National
Laboratory,
Oak
TN 37831-6335,
82Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
to evaluate the overall impact of anthropogenic nitrogen
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole,
inputs on terrestrial ecosystem carbon and nitrous oxide fluxes
9 Department of Geography, University of Kansas, Lawrence, KS 66045-7613, USA
MA 02543-1543, USA
between 1700 and 2005. We show that anthropogenic nitrogen
3 Climate and Global Dynamics Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
inputs account for about a fifth of the carbon sequestered
Received:
28 January 2009 – Published in Biogeosciences Discuss.: 26 March 2009
by terrestrial ecosystems between 1996 and 2005, and for
4 Department
of Earth
System
Science,
California,
Irvine,
CA 92697-3100,
USA
most of the increase in global nitrous oxide emissions in recent
Revised:
12 August
2009
– Accepted:
17University
September of
2009
– Published:
8 October
2009
5 Department
decades; the latter is largely due to agricultural intensification.
of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY 14850, USA
We estimate that carbon sequestration due to nitrogen
6 Department of Earth and Planetary Science, University of California, Berkeley, CA 94720-4767, USA
deposition has reduced current carbon dioxide radiative forcing
7 NOAA Earth System Research Laboratory, Chemical Sciences Division, 325 Broadway, Boulder, CO 80305-3337,byUSA
96 ± 14 mW m 2 . However, this effect has been offset by
Abstract.
Inclusion of fundamental ecological interactions
century, attributable in part to a steady decline in the ocean
the increase in radiative forcing resulting from nitrous oxide
8 Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
between carbon and nitrogen cycles in the land component of
sink fraction. Comparison to experimental studies on theemissions,
fate
which amounts to 125±20 mW m 2 .
9 Department
an
atmosphere-ocean
general University
circulation of
model
(AOGCM)
of radio-labeled
tracers in temperate forests indiof Geography,
Kansas,
Lawrence, KS
66045-7613, nitrogen
USA
The effect of the anthropogenic reactive nitrogen (hereafter Nr)
perturbation, that is the addition of ecosystem-available forms of
leads to decreased carbon uptake associated with CO2 fertilcates that the model representation of competition between
nitrogen such as ammonium and nitrate, on terrestrial ecosystems
Received:
28
January
2009
–
Published
in
Biogeosciences
Discuss.:
26
March
2009
ization, and increased carbon uptake associated with warmplants and microbes for new mineral nitrogen resources
and is
the land–atmosphere fluxes of carbon dioxide and nitrous
Revised:
August
2009 The
– Accepted:
17these
September
2009 – Published:
2009
oxide
of the tight coupling
ing
of the 12
climate
system.
balance of
two opposreasonable.8 October
Our results
suggest a weaker dependence of
nethas proved difficult to quantify because
between the carbon and nitrogen cycles1 : changes in one cycle
ing effects is to reduce the fraction of anthropogenic CO2
land carbon flux on soil moisture changes in tropical regions,
cannot be understood without considering interactions with the
predicted to be sequestered in land ecosystems. The primary
and a stronger positive growth response to warming in those
other cycle. Nevertheless, previous studies have addressed the
of anthropogenic Nr additions in isolation: nitrogen cycle
Abstract. responsible
Inclusion offorfundamental
ecological
interactions
century,
in apart
to a AOGCM
steady decline
ineffects
the ocean
mechanism
increased land
carbon storage
unregions,
thanattributable
predicted by
similar
implemented
studies have focussed on understanding and quantifying the global
between
carbon
and nitrogen
cyclesisinshown
the land
component
sink fraction.
Comparison interactions.
to experimental
on
the budget
fate through the cascading effects of Nr on terrestrial
der
radiatively
forced
climate change
to be
fertiliza- of without
land carbon-nitrogen
Westudies
expectnitrogen
that
N fluxes
and emissions6 , riverine export to coastal ecosystems7 ,
an atmosphere-ocean
model
(AOGCM) the of
radio-labeled uncertainty
nitrogen tracers
in temperate
forests
tion
of plant growth by general
increasedcirculation
mineralization
of nitrogen
between-model
in predictions
of future
at- indiand the consequences of N2 O accumulating in the atmosphere on
directly
increased
soil2 fertilormospheric
COthe
and associated
anthropogenic
leads toassociated
decreased with
carbon
uptake decomposition
associated withofCO
cates that
model representation
of competition
between
2 concentration
climate
and stratospheric ozone depletion8 . Carbon cycle studies
ganic
matter
under
a
warming
climate,
which
in
this
particclimate
change
will
be
reduced
as
additional
climate
modhave focussed
ization, and increased carbon uptake associated with warmplants and microbes for new mineral nitrogen resources
is on determining the contribution of Nr deposition
on productivity and carbon storage in natural ecosystems at
We perform a set of model simulations aimed at isolating
global and regional effects of anthropogenic Nr addition due to
fertilizer applications and atmospheric deposition from those due
to changes in land-cover, atmospheric [CO2 ], and climatic changes
over the historical period (1700–2005), using a consistent set of
model forcings (see Methods and Supplementary Information).
To calculate the overall climatic forcing of the Nr perturbation,
we propagate the net land–atmosphere exchanges of carbon
dioxide and nitrous oxide, resulting from the coupled nitrogen–
carbon model, into atmospheric concentrations using simple
box models and further information about fossil, marine, and
complementary terrestrial fluxes11,12 (Fig. 1; see Methods and
Supplementary Information).
In 1960–2005, the period for which direct atmospheric measurements are available13 , the increase in atmospheric [CO2 ] and
the interannual growth rate of atmospheric [CO2 ] are reproduced
well by the model (Fig. 1a,c), which implies that the simulated
airborne fraction of anthropogenic emissions also falls into the
observed range12 . The effect of the Nr perturbation on atmospheric CO2 levels is small ( 6 ppm). The simulated terrestrial
N2 O flux, when combined with the other terrestrial N2 O fluxes,
produces a good agreement with the observed atmospheric [N2 O]
evolution since pre-industrial times11,14–16 (Fig. 1b,d). The increase
in the terrestrial N2 O source explains more than two thirds
(33 of 45 ppb) of the atmospheric increase since 1860. Nevertheless, the average simulated [N2 O] growth rate in 1990–2005
(0.93 ± 0.12 ppb yr 1 ) is about 25% higher than the observed
trend at atmospheric stations for the same period (GAGE/AGAGE
network15,16 : 0.72 ± 0.27 ppb yr 1 , Fig. 1d).
What are the underlying causes of these terrestrial CO2 and
N2 O flux anomalies? We estimate that the Nr perturbation
has increased terrestrial C storage since pre-industrial times by
12 Pg C. This value is small compared with modelled changes
in terrestrial C storage due to deforestation (a cumulative loss
of 205 Pg C), and the enhancement of plant productivity due
to CO2 fertilization (a gain of 143 Pg C). During the period
1996–2005 we estimate that Nr deposition causes roughly 20%
(0.2 Pg C yr 1 ) of the terrestrial net C uptake (1.1 Pg C yr 1 ;
Buffering capacity of C to changes in N
How potential primary productivity is limited by nitrogen
Fixed Vegetation C:N
Fixed Soil Organic Matter C:N
Variable Vegetation C:N
Variable Soil Organic Matter C:N
Test of nitrogen limitation in two global
biogeochemical models
CLM-CN 4.0
Biogeosciences, 6, 2099–2120, 2009
www.biogeosciences.net/6/2099/2009/
© Author(s) 2009. This work is distributed under
the Creative Commons Attribution 3.0 License.
Biogeosciences
(Thornton et al. 2009 Biogeosciences)
Biogeosciences, 6, 2099–2120, 2009
www.biogeosciences.net/6/2099/2009/
© Author(s) 2009. This work is distributed under
the Creative Commons Attribution 3.0 License.
O-CN
(Zaehle et al. 2011 Nature Geoscience)
LETTERS
Biogeosciences
Carbon-nitrogen interactions regulate climate-carbon cycle
feedbacks: results from an atmosphere-ocean
general circulation model
P. E. Thornton1 , S. C. Doney2 , K. Lindsay3 , J. K. Moore4 , N. Mahowald5 , J. T. Randerson4 , I. Fung6 ,
J.-F. Lamarque7,8 , J. J. Feddema9 , and Y.-H. Lee3
Carbon-nitrogen interactions regulate climate-carbon cycle
1 Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6335, USA
results from an atmosphere-ocean
2feedbacks:
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole,
MA
02543-1543,
USA
general
circulation
model
3 Climate and Global Dynamics Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
PUBLISHED ONLINE: 31 JULY 2011 | DOI: 10.1038/NGEO1207
Carbon benefits of anthropogenic reactive
nitrogen offset by nitrous oxide emissions
Sönke Zaehle1 *, Philippe Ciais2 , Andrew D. Friend3 and Vincent Prieur2
Additions of reactive nitrogen to terrestrial ecosystems—
primarily through fertilizer application and atmospheric
deposition—have more than doubled since 1860 owing to
human activities1 . Nitrogen additions tend to increase the net
uptake of carbon by the terrestrial biosphere, but they also
4 Department of Earth System Science, University of California, Irvine, CA 92697-3100, USA
1 , S. C. Doney2 , K. Lindsay3 , J. K. Moore4 , N. Mahowald5 , J. T. Randerson4 , I. Fung6 ,
stimulate nitrous oxide release from soils2 . However, given
P.
E.
Thornton
5 Department of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY 14850, USA
that the magnitude of these effects is uncertain, and that the
7,8
9
3
Lamarque
, J.and
J. Feddema
, and Y.-H.
Lee of California, Berkeley, CA 94720-4767, USA
6J.-F.
carbon and nitrogen cycles are tightly coupled, the net climatic
Department
of Earth
Planetary Science,
University
impact of anthropogenic nitrogen inputs is unknown3 . Here
71NOAA
Earth System
Research
Laboratory,
Chemical
Sciences
Division,
325Ridge,
Broadway,
Boulder, COUSA
80305-3337, USAwe use a process-based model of the terrestrial biosphere4,5
Environmental
Sciences
Division,
Oak Ridge
National
Laboratory,
Oak
TN 37831-6335,
82Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
to evaluate the overall impact of anthropogenic nitrogen
Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole,
inputs on terrestrial ecosystem carbon and nitrous oxide fluxes
9 Department of Geography, University of Kansas, Lawrence, KS 66045-7613, USA
MA 02543-1543, USA
between 1700 and 2005. We show that anthropogenic nitrogen
3 Climate and Global Dynamics Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
inputs account for about a fifth of the carbon sequestered
Received:
28 January 2009 – Published in Biogeosciences Discuss.: 26 March 2009
by terrestrial ecosystems between 1996 and 2005, and for
4 Department
of Earth
System
Science,
California,
Irvine,
CA 92697-3100,
USA
most of the increase in global nitrous oxide emissions in recent
Revised:
12 August
2009
– Accepted:
17University
September of
2009
– Published:
8 October
2009
5 Department
decades; the latter is largely due to agricultural intensification.
of Earth and Atmospheric Sciences, Cornell University, Ithaca, NY 14850, USA
We estimate that carbon sequestration due to nitrogen
6 Department of Earth and Planetary Science, University of California, Berkeley, CA 94720-4767, USA
deposition has reduced current carbon dioxide radiative forcing
7 NOAA Earth System Research Laboratory, Chemical Sciences Division, 325 Broadway, Boulder, CO 80305-3337,byUSA
96 ± 14 mW m 2 . However, this effect has been offset by
Abstract.
Inclusion of fundamental ecological interactions
century, attributable in part to a steady decline in the ocean
the increase in radiative forcing resulting from nitrous oxide
8 Atmospheric Chemistry Division, National Center for Atmospheric Research, Boulder, CO 80307-3000, USA
between carbon and nitrogen cycles in the land component of
sink fraction. Comparison to experimental studies on theemissions,
fate
which amounts to 125±20 mW m 2 .
9 Department
an
atmosphere-ocean
general University
circulation of
model
(AOGCM)
of radio-labeled
tracers in temperate forests indiof Geography,
Kansas,
Lawrence, KS
66045-7613, nitrogen
USA
The effect of the anthropogenic reactive nitrogen (hereafter Nr)
perturbation, that is the addition of ecosystem-available forms of
leads to decreased carbon uptake associated with CO2 fertilcates that the model representation of competition between
nitrogen such as ammonium and nitrate, on terrestrial ecosystems
Received:
28
January
2009
–
Published
in
Biogeosciences
Discuss.:
26
March
2009
ization, and increased carbon uptake associated with warmplants and microbes for new mineral nitrogen resources
and is
the land–atmosphere fluxes of carbon dioxide and nitrous
Revised:
August
2009 The
– Accepted:
17these
September
2009 – Published:
2009
oxide
of the tight coupling
ing
of the 12
climate
system.
balance of
two opposreasonable.8 October
Our results
suggest a weaker dependence of
nethas proved difficult to quantify because
between the carbon and nitrogen cycles1 : changes in one cycle
ing effects is to reduce the fraction of anthropogenic CO2
land carbon flux on soil moisture changes in tropical regions,
cannot be understood without considering interactions with the
predicted to be sequestered in land ecosystems. The primary
and a stronger positive growth response to warming in those
other cycle. Nevertheless, previous studies have addressed the
of anthropogenic Nr additions in isolation: nitrogen cycle
Abstract. responsible
Inclusion offorfundamental
ecological
interactions
century,
in apart
to a AOGCM
steady decline
ineffects
the ocean
mechanism
increased land
carbon storage
unregions,
thanattributable
predicted by
similar
implemented
studies have focussed on understanding and quantifying the global
between
carbon
and nitrogen
cyclesisinshown
the land
component
sink fraction.
Comparison interactions.
to experimental
on
the budget
fate through the cascading effects of Nr on terrestrial
der
radiatively
forced
climate change
to be
fertiliza- of without
land carbon-nitrogen
Westudies
expectnitrogen
that
N fluxes
and emissions6 , riverine export to coastal ecosystems7 ,
an atmosphere-ocean
model
(AOGCM) the of
radio-labeled uncertainty
nitrogen tracers
in temperate
forests
tion
of plant growth by general
increasedcirculation
mineralization
of nitrogen
between-model
in predictions
of future
at- indiand the consequences of N2 O accumulating in the atmosphere on
directly
increased
soil2 fertilormospheric
COthe
and associated
anthropogenic
leads toassociated
decreased with
carbon
uptake decomposition
associated withofCO
cates that
model representation
of competition
between
2 concentration
climate
and stratospheric ozone depletion8 . Carbon cycle studies
ganic
matter
under
a
warming
climate,
which
in
this
particclimate
change
will
be
reduced
as
additional
climate
modhave focussed
ization, and increased carbon uptake associated with warmplants and microbes for new mineral nitrogen resources
is on determining the contribution of Nr deposition
on productivity and carbon storage in natural ecosystems at
We perform a set of model simulations aimed at isolating
global and regional effects of anthropogenic Nr addition due to
fertilizer applications and atmospheric deposition from those due
to changes in land-cover, atmospheric [CO2 ], and climatic changes
over the historical period (1700–2005), using a consistent set of
model forcings (see Methods and Supplementary Information).
To calculate the overall climatic forcing of the Nr perturbation,
we propagate the net land–atmosphere exchanges of carbon
dioxide and nitrous oxide, resulting from the coupled nitrogen–
carbon model, into atmospheric concentrations using simple
box models and further information about fossil, marine, and
complementary terrestrial fluxes11,12 (Fig. 1; see Methods and
Supplementary Information).
In 1960–2005, the period for which direct atmospheric measurements are available13 , the increase in atmospheric [CO2 ] and
the interannual growth rate of atmospheric [CO2 ] are reproduced
well by the model (Fig. 1a,c), which implies that the simulated
airborne fraction of anthropogenic emissions also falls into the
observed range12 . The effect of the Nr perturbation on atmospheric CO2 levels is small ( 6 ppm). The simulated terrestrial
N2 O flux, when combined with the other terrestrial N2 O fluxes,
produces a good agreement with the observed atmospheric [N2 O]
evolution since pre-industrial times11,14–16 (Fig. 1b,d). The increase
in the terrestrial N2 O source explains more than two thirds
(33 of 45 ppb) of the atmospheric increase since 1860. Nevertheless, the average simulated [N2 O] growth rate in 1990–2005
(0.93 ± 0.12 ppb yr 1 ) is about 25% higher than the observed
trend at atmospheric stations for the same period (GAGE/AGAGE
network15,16 : 0.72 ± 0.27 ppb yr 1 , Fig. 1d).
What are the underlying causes of these terrestrial CO2 and
N2 O flux anomalies? We estimate that the Nr perturbation
has increased terrestrial C storage since pre-industrial times by
12 Pg C. This value is small compared with modelled changes
in terrestrial C storage due to deforestation (a cumulative loss
of 205 Pg C), and the enhancement of plant productivity due
to CO2 fertilization (a gain of 143 Pg C). During the period
1996–2005 we estimate that Nr deposition causes roughly 20%
(0.2 Pg C yr 1 ) of the terrestrial net C uptake (1.1 Pg C yr 1 ;
Buffering capacity of C to changes in N
How potential primary productivity is limited by nitrogen
Fixed Vegetation C:N
Fixed Soil Organic Matter C:N
Variable Vegetation C:N
Variable Soil Organic Matter C:N
Differing mechanisms governing N loss
Global nitrogen fertilization experiment
• 25 year simulations (1985-2009)
• Nitrogen applied globally at five levels continuously
• Low application to parallel plausible changes in nitrogen deposition
(0.5 g N m-2 yr-1)
• Higher applications to parallel field experimental additions of
nitrogen fertilizer to terrestrial ecosystems
(2.0, 4.0, 10.0 g N m-2 yr-1)
• High application to test nitrogen saturation (30.0 g N m-2 yr-1)
• Same climate inputs and land-use history
Global nitrogen fertilization response:
High addition (30.0 g N m-2 yr-1)
CLM-CN ∆Net Primary Productivity
-200
-100
0
100
200
300
400
500
g C m-2 yr-1
Thomas et al. 2013 Global Change Biology
Global nitrogen fertilization response:
High addition (30.0 g N m-2 yr-1)
O-CN ∆Net Primary Productivity
-200
-100
0
100
200
300
400
500
g C m-2 yr-1
Thomas et al. 2013 Global Change Biology
separately (pooled across organs), there were strong R–N
relations for gymnosperms (P < 0.001, R2 = 0.66,
7960.001,
P. B. Reich
n = 1216), woody angiosperms (P <
R2et=al. 0.62,
n = 1069) and herbs (P < 0.001, R2 = 0.57, n = 225;
Global nitrogen fertilization response:
High addition (30.0 g N m-2 yr-1)
100
proteins, maintenance of
in more metabolically act
a disproportionate incre
pools relative to struct
Similarly strong relations were seen for leaves (P < 0.001,
R2 = 0.66, n = 1386), stems (P < 0.001, R2 = 0.80,
n = 100
380) and fine roots (P < 0.001, R2 = 0.62, n = 744)
examined Stems
among all plant taxa (scaling exponents ranged
from 1.34 to 1.64; Fig. 1). Examining each plant group
10
separately (pooled across organs), there were strong R–N
relations for gymnosperms (P < 0.001, R2 = 0.66,
1
n = 1216),
woody angiosperms (P < 0.001, R2 = 0.62,
n = 1069) and herbs (P < 0.001, R2 = 0.57, n = 225;
Leaves
O-CN ∆Net Primary Productivity
woody angio
0.1
gymno
herb
100
1
0.3
3
1
Figure
100
Leaves
0.01
0.1
100
Roots
10
1
0.3
0.1
1
Nitrogen (mmol g–1)
3
10
100
All
1
10
woody angio
gymno
herb
1
1
0.3
0.1
100
10
3
1
0.01
100
0
–1)
-100
10
Herbaceous
angiosperms
100
0.1
0.01
Roots
100
0.1
Nitrogen (mmol
1
g–1)
10
1
All
(nmo
conce
presse
are sh
all or
labelle
visibil
identi
0.
ratios
comp
includ
angio
gymn
Table
0.1
1
-200
1
Stems
10
Respiration (nmol g–1 s–1)
Respiration (nmol g–1 s–1)
10
Woody
0.3
0.1
angiosperms
3
1
0.01
0.1
100
200
300 Nitrogen400
500
(mmol g–1)
Nitroge
g C m-2 yr-1
10
Thomas et al. 2013 Global Change Biology
500
500
400
400
Temperate
Tropical latitudes: CLM−CN
Tropical latitudes: O−CN
Mid−latitudes: CLM−CN
Mid−latitudes: O−CN
High latitudes: CLM−CN
High latitudes: O−CN
Tropical
Fertilzation NPP − Control NPP (gC/m2/yr)
Fertilization NPP - Control NPP (g C m-2 yr-1)
CLM-CN more responsive to nitrogen than O-CN
CLM-CN
300
300
200
200
Boreal
100
100
0
Boreal
Temperate
Tropical
0
-100 0
0
−100
O-CN
5
5
10
15
20
10 Fertilization15
rate (gN/m2/yr) 20
25
25
Nitrogen fertilization rate (g N m-2 yr-1)
30
30
Thomas et al. 2013 Global Change Biology
Model comparison to data:
Model response compared to observations
Nitrogen fertilization experiments
15N tracer studies
Plot/small catchment nitrogen budgets
Thomas et al. 2013 Global Change Biology
Model comparison to data:
NPP response to N fertilization
120
n = 39
% ANPP response
100
Observed
CLM-CN
O-CN
80
n = 32
60
40
20
0
Grasslands
Temperate/boreal forests
Thomas et al. 2013 Global Change Biology
allocation, and stoichiometry).
e: Soil inorganic N leaching
f: Excess denitrification proportional the soil inorganic N
that exceeds immobilization (c) and plant N uptake (d). No
direct environmental controls (temperature and soil water).
g: N gas loss proportional to net N mineralization (b-c).
h: N fixation
i: N deposition
c
d
What controls the C cycle response to N
f
h
Soil BGD
additions? Alternative versions of the CLM-CN
i
Model
15: clm4mod
clm4cn
structure
CLM-CN
4.0
structure
g
h
i
c
d
Soil
Inorganic N
b
f
e
intoN
Plant
aInsights
SON &
litter N
h
i
clm4mod structure
Discussion Paper
b
organic N
N uptake (d). No
and soil water).
on (b-c).
Plant N
l
|
otosynthesis,
SON &
litter N
Inorganic N
e 2013
10, 1635–1683,
CLM-CN
structure
clm4mod4.5
structure
a: Plant turnover of N
b: Gross N mineralization
c: Gross N immobilization
d: Plant N uptake (Michaelis-Menten)
e: Soil inorganic N leaching
f: Excess denitrification proportional the soil inorganic N
that exceeds immobilization (c) and plant N uptake (d). No
direct environmental controls (temperature and soil water).
g: Not shown; only in clm4cn and models 2 - 9
h: N fixation
i: N deposition
j: Nitrification. Proportional to NH4+. Limited by
temperature and soil water.
NH4
k: N2O loss. Proportional to nitrification (j) and NH4+.
l: Dissolved organic N leaching. Proportional to the turnover of
lignin litter pool.
a
Discussion Paper
b
cNH4
mechanisms
Labile N
governing
C and N
cNO3
interactions
d
d
NH4+
eNH4
NH4
R. Q. Thomas etiNO
al.3
j
NO3f
k
eNO3
Title Page
Removed N gas loss that is 1% of net mineralization
Abstract
lFig.
Plant
Nclm4mod model
a clm4cn
• Denitrification
on
environmental
conditions
3. Overview
of based
the
and
structure.
SON
&
Conclusions
• Soil
litterNH
N 4+ and NO3- pools
Labile N
cNO3use efficiency (Bonan et al. 2011,2013)
• Reduced light
rganic N
Tables
N uptake (d). No
c
d
NH
NO3
•
4 layers dNH4
Vertical
soil
and soil water).
b
• Modified the timing of N fixation
J
iNO3 in high latitudes
|
•
j
1680
Discussion Pape
h
NO3
Introduction
References
Figures
I
NPP response to N fertilization:
CLM-CN 4.0 vs. CLM-CN 4.5
300
CLM-CN
4.0
NPP Nitrogen Fertilization 2000−1900
0
NPP Nitrogen Fertilization 2000−1900
100
-100
300
300
50
CLM-CN
4.5
200
200
100
0
100
0
0
−50
−100
−100
Challenges associated with including coupled C
and N cycles
Challenges associated with including coupled C
and N cycles
• Model requires 1000s of years for spin-up
Challenges associated with including coupled C
and N cycles
• Model requires 1000s of years for spin-up
• Data on key inputs (fixation) and losses (denitrification) are lacking.
Challenges associated with including coupled C
and N cycles
• Model requires 1000s of years for spin-up
• Data on key inputs (fixation) and losses (denitrification) are lacking.
• Degree of stoichiometric flexibility is unknown (buffering).
Challenges associated with including coupled C
and N cycles
• Model requires 1000s of years for spin-up
• Data on key inputs (fixation) and losses (denitrification) are lacking.
• Degree of stoichiometric flexibility is unknown (buffering).
• What governs the potential productivity?
Challenges associated with including coupled C
and N cycles
• Model requires 1000s of years for spin-up
• Data on key inputs (fixation) and losses (denitrification) are lacking.
• Degree of stoichiometric flexibility is unknown (buffering).
• What governs the potential productivity?
• What is the appropriate time scale for simulating the coupling between the C
and N scales?
Challenges associated with including coupled C
and N cycles
• Model requires 1000s of years for spin-up
• Data on key inputs (fixation) and losses (denitrification) are lacking.
• Degree of stoichiometric flexibility is unknown (buffering).
• What governs the potential productivity?
• What is the appropriate time scale for simulating the coupling between the C
and N scales?
• What is the competition between plants and microbes for N?
Challenges associated with including coupled C
and N cycles
• Model requires 1000s of years for spin-up
• Data on key inputs (fixation) and losses (denitrification) are lacking.
• Degree of stoichiometric flexibility is unknown (buffering).
• What governs the potential productivity?
• What is the appropriate time scale for simulating the coupling between the C
and N scales?
• What is the competition between plants and microbes for N?
• Role of disturbance in N cycling difficult to simulate.
Challenges associated with including coupled C
and N cycles
• Model requires 1000s of years for spin-up
• Data on key inputs (fixation) and losses (denitrification) are lacking.
• Degree of stoichiometric flexibility is unknown (buffering).
• What governs the potential productivity?
• What is the appropriate time scale for simulating the coupling between the C
and N scales?
• What is the competition between plants and microbes for N?
• Role of disturbance in N cycling difficult to simulate.
• Involves coarsely representing fine-scale non-linear processes.
Many other pathways that N influences climate:
What happens when adding N?
cooling
Wet Deposition (kg N ha-1 yr-1)
11.1
2.5
Human creation of
reactive nitrogen (Nr)
Nr Creation Rate (Tg N yr-1)
200
160
120
80
40
1960
1970
1980
1990
Pinder et al. 2012 PNAS
2000
2010
warming
Many other pathways that N influences climate:
What happens when adding N?
cooling
Wet Deposition (kg N ha-1 yr-1)
11.1
2.5
Human creation of
reactive nitrogen (Nr)
Nr Creation Rate (Tg N yr-1)
200
160
120
80
40
1960
1970
1980
1990
Pinder et al. 2012 PNAS
2000
2010
warming
Many other pathways that N influences climate:
What happens when adding N?
cooling
Wet Deposition (kg N ha-1 yr-1)
11.1
2.5
Human creation of
reactive nitrogen (Nr)
Nr Creation Rate (Tg N yr-1)
200
160
120
80
40
1960
1970
1980
1990
Pinder et al. 2012 PNAS
2000
2010
warming
Questions?
R. Quinn Thomas
Forest Resources and Environmental Conservation
Virginia Tech
Email: rqthomas@vt.edu
Transactional N limitation
CWD
2.7 years
C:N = variable
1.00
Litter 1
1.4 days
C:N = variable
Litter 2
14 days
C:N = variable
Litter 3
71 days
C:N = variable
0.61
0.45
0.71
Soil 1
14 days
C:N = 12
0.72
Soil 2
71 days
C:N = 12
0.54
Soil 3
1.95 years
C:N = 10
0.45
Thornton and Rosenbloom 2005 Ecological Modelling
Soil 4
27.4 years
C:N =10
Transactional N limitation
Parton et al.' Modeling GrasslandBiomes
SURFACE
(1-Fm)
787
ROOT
LITTER
•
Fm
=.85
-.018
*L/N
(1
-Fm
)••Fm
•k,
STRUCTURAL
i
STRUCTURAL
METABOLIC
IC
METABOLIC
I
i ¸
C
LIGNIN I CELLULOSE
LITTER
C
LIGNIN I CELLULOSE
CO
•6
A
Lc
C•
K2
•
Lc
CO2
.55
Lc
.3
.c I "c;
ø""
CsA=(1 - Csp-.55).
x
SOIL
MICROBE
C
A CO2
,
CO2
.3
,
Tm=(1-.75' T)
CAS= (1 - CAp-CAL-Ft)
K3
A
j
cø2
Ft =,85-.68'T
SLOW
C
L/N = Ligninnitrogenratio
CAL=
(H2030/18)*(.01+ .04' Ts)
CO2
A = Abioticdecomposition
factor
T = Siltplusclaycontent(fraction)
Ts= Sandcontent(fraction)
Tc= Claycontent(fraction)
.55
Ls = Faction
ofstructural
C thatislignin
Csp=.003
' 're
CAp=
.003+ .032' C
LEACHED
Lc = Exp( -3' Ls )
H203o=H20 leachedbelow30 m(cmm'1)
KL= Maximum
decayrate(y-l)
Fm= Fractionof litterthatis metabolic
PASSIVE
C
C
:
Ki = 3.9,4.8,7.3,6.0,14.8,18.5,.2,.0045y -1
Parton et al. 1993 Glob. Biogeochemical Cycles
Fig. 1. Flow diagramfor theCenturysoilorganicC model.
Transactional N limitation
Parton et al.' Modeling GrasslandBiomes
SURFACE
(1-Fm)
787
ROOT
LITTER
•
Fm
=.85
-.018
*L/N
(1
-Fm
)••Fm
•k,
STRUCTURAL
i
STRUCTURAL
METABOLIC
IC
METABOLIC
I
i ¸
C
LIGNIN I CELLULOSE
LITTER
C
LIGNIN I CELLULOSE
CO
•6
A
Lc
C•
K2
•
Lc
CO2
.55
Lc
.3
.c I "c;
ø""
CsA=(1 - Csp-.55).
x
SOIL
MICROBE
C
A CO2
,
CO2
.3
,
Tm=(1-.75' T)
CAS= (1 - CAp-CAL-Ft)
K3
A
5 yrs.
j
cø2
Ft =,85-.68'T
SLOW
C
L/N = Ligninnitrogenratio
CAL=
(H2030/18)*(.01+ .04' Ts)
CO2
A = Abioticdecomposition
factor
T = Siltplusclaycontent(fraction)
Ts= Sandcontent(fraction)
Tc= Claycontent(fraction)
.55
Ls = Faction
ofstructural
C thatislignin
Csp=.003
Lc = Exp( -3' Ls )
H203o=H20 leachedbelow30 m(cmm'1)
KL= Maximum
decayrate(y-l)
Fm= Fractionof litterthatis metabolic
' 're
CAp=
.003+ .032' C
LEACHED
222 yrs.
PASSIVE
C
C
:
Ki = 3.9,4.8,7.3,6.0,14.8,18.5,.2,.0045y -1
Parton et al. 1993 Glob. Biogeochemical Cycles
Fig. 1. Flow diagramfor theCenturysoilorganicC model.
Carbon response to N addition:
Nitrogen deposition vs. nitrogen fertilization
7
N input (g N m-2 yr-1)
6
5
N deposition
N fertilization
4
3
Wet Deposition (kg N ha-1 yr-1)
2
11.1
1
0
1800
2.5
1850
1900
1950
Year
2000
2050
Carbon response to N addition:
Nitrogen deposition vs. nitrogen fertilization
7
N input (g N m-2 yr-1)
6
5
N deposition
N fertilization
4
3
2
1
0
1800
1850
1900
1950
Year
2000
2050
allocation, and stoichiometry).
e: Soil inorganic N leaching
f: Excess denitrification proportional the soil inorganic N
that exceeds immobilization (c) and plant N uptake (d). No
direct environmental controls (temperature and soil water).
g: N gas loss proportional to net N mineralization (b-c).
h: N fixation
i: N deposition
c
d
i
Model clm4cn
15: clm4mod
structure
g
h
i
Plant N
c
SON &
litter N
d
Soil
Inorganic N
b
f
h
e
i
clm4mod structure
Discussion Paper
b
organic N
N uptake (d). No
and soil water).
on (b-c).
l
|
otosynthesis,
SON &
litter N
Inorganic N
e 2013
10, 1635–1683,
clm4mod structure
a: Plant turnover of N
b: Gross N mineralization
c: Gross N immobilization
d: Plant N uptake (Michaelis-Menten)
e: Soil inorganic N leaching
f: Excess denitrification proportional the soil inorganic N
that exceeds immobilization (c) and plant N uptake (d). No
direct environmental controls (temperature and soil water).
g: Not shown; only in clm4cn and models 2 - 9
h: N fixation
i: N deposition
j: Nitrification. Proportional to NH4+. Limited by
temperature and soil water.
NH4
k: N2O loss. Proportional to nitrification (j) and NH4+.
l: Dissolved organic N leaching. Proportional to the turnover of
lignin litter pool.
a
Discussion Paper
What controls the C cycle response to N
f
h
Soil
BGD
additions? Alternative versions of the CLM-CN
b
cNH4
Plantinto
N
a Insights
mechanisms
Labile N
governing C and N
cNO3
interactions
d
d
NH4+
eNH4
NH4
j
R. Q. Thomas iet
NOal.
3
k
NO3-
|
j
3
1680
Discussion Pape
h
eNO3
f
Title Page
Michaelis-Menten plant N uptake
Abstract
l • Overview
N
Reduced
in
extra-tropical
forests
Fig. 3.
of N
theafixation
clm4cn Plant
andmature
clm4mod
model structure.
SON
&
Conclusions
•litter
Removed
N gas loss
that
is
1%
of
net
mineralization
N
Labile N
• Denitrification
cNO3 based on environmental conditionsTables
rganic N
+
N uptake (d). No
• Soil cNH
d
d
34 pools
NH4 4 and NO
NH
NO3
and soil water).
• bReduced light use efficiency (Bonan et al. 2011,2013)
J
iNO
•
NO3
Introduction
References
Figures
I
Model comparison to data:
Model response compared to observations
5 sites, 6 fertilization experiments
(4 in Michigan, 1 in Massachusetts)
10+ years of observations
% ANPP response to N fertilization
Model comparison to data:
NPP response to N fertilization
180
160
140
N fertilization
N limitation relieved
120
100
80
60
40
20
0
Observed(
CLM,CN((
Observations
CLM-CN 4.0
(Magill et al. 2004;
Pregitizer et al. 2008)
CLM,CN(Mod(
CLM-CN Mod.
Thomas et al. 2013 Biogeosciences
proportion of nitrogen added to ecosystem
Model comparison to data:
15N Tracer studies
0.6
0.5
0.4
0.3
0.2
0.1
0
Observa(ons+
CLM/CN+
CLM/CN+mod+
Observations CLM-CN 4.0 CLM-CN Mod.
Thomas et al. 2013 Biogeosciences
Model comparison to data:
C increment response to N deposition
70
dCincrement/dNdeposition
60
50
40
30
20
10
0
Observations
Observed
(Thomas et al. 2010)
CLM-CN
4.0
CLM-CN
CLM-CN Mod.
CLM-CN
Mod
Thomas et al. 2013 Biogeosciences
Thomas et al. 2010 Nature Geoscience
70
Response to N
depostion
60
30
Discussion Paper
dCACI / dNdeposition
40
N loss switched from a
f(N mineralization) to f(NO3)
|
Change from
previous modified
model
50
Michaelis-Menten plant N uptake
20
10
0
-10
:c
lm
Discussion Pa
el
1
|
4c
M n
od
el
M 2
od
el
M 3
od
el
M 4
od
el
M 5
od
el
M 6
od
el
M 7
od
el
M 8
od
el
9
M
od
el
M 10
od
el
M 11
od
el
M 12
od
el
M
od
M 13
o
el
15 del
:c
1
lm 4
4m
od
-20
M
od
Discussion Paper
Implications of alternative approaches to
modeling N cycling
Thomas et al. 2013 Biogeosciences