The sensitivity of terrestrial carbon fluxes to
temperature and precipitation: a preliminary
analysis with ISI-MIP2a
Hanqin Tian, Susan Pan and Jia Yang
International Center for Climate and Global
Change Research
Auburn University, USA
The Carbon fluxes: terminology
GPP – Gross Primary Production
Ra – Plant Respiration
NPP - Net Primary Production (GPP-Ra)
Rh - Heterotrophic Respiration
NEP - Net Ecosystem Production (NPP-Rh)
Sensitivity analysis: Multiple Regression Approach
𝑌𝑌𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 = γ 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑋𝑋𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 + γ𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑋𝑋𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 + ԑ
𝑌𝑌𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 : detrended anomaly of annual carbon flux (g C m-2 yr-1),
including GPP, NPP, Rh, NEP
𝑋𝑋𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 : detrended anomaly of annual temperature (°C)
𝑋𝑋𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 : detrended anomaly of annual precipitation (mm yr-1)
γ 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 : carbon flux sensitivity to interannual variations in temperature
(g C m-2 yr-1 per °C)
γ𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 : carbon flux sensitivity to interannual variations in precipitation
(g C m-2 yr-1 per mm yr-1)
ԑ: residue error
Global average temperature during
1971-2010 based on GSWP3
Global average precipitation during
1971-2010 based on GSWP3
standard deviation of detrended
temperature anomaly during 19712010 based on GSWP3
standard deviation of detrended
precipitation anomaly during 19712010 based on GSWP3
0,4
Precipitation anomaly
(mm yr-1)
Temperature anomaly
(°C)
Detrended anomaly of climate and carbon fluxes
0,2
0
-0,2
-0,4
1981
1991
0
-25
-50
30
20
15
10
0
-15
1971
1981
1991
2001
1971
1981
1991
2001
1971
1981
1991
2001
0
-10
-20
-30
1971
1981
1991
2001
10
20
5
10
NEP anomaly
(g C m-2 yr-1)
Rh anomaly
(g C m-2 yr-1)
25
2001
NPP anomaly
(g C m-2 yr-1)
GPP anomaly
(g C m-2 yr-1)
1971
50
0
-5
0
-10
-20
-10
1971
1981
1991
2001
Result 1:
The sensitivity of carbon fluxes to
temperature variation
The sensitivity of GPP to temperature variation
γ
CARAIB
ORCHIDEE
DLEM
LPJml
VEGAS
VISIT
ensemble
MTE
GPP sensitivity (g C m-2 yr-1 per °C) to inter-annual variation in temperature
Global average: -10.5 ± 16.9 g C m-2 yr-1 per °C
Tair
Sensitivity of GPP to temperature variation
200
GPP sensitivity to temperature
Largest divergence in tropics among models
100
CARAIB
DLEM
LPJml
0
ORCHIDEE
VEGAS
VISIT
ensemble
-100
MTE
-200
90
60
30
0
Latitude
-30
-60
-90
MTE GPP and model ensemble show that GPP is positively correlated with
temperature in mid- and high latitudes, but negatively related to temperature in
topical region (30°N – 35°S).
Sensitivity of GPP to temperature variation
Tair
MAP (mm yr-1)
γ
MAT (°C)
GPP is positively correlated with temperature in low MAT areas, but
negatively correlated with temperature in high MAT areas.
Sensitivity of NPP to temperature variation
γ
NPP sensitivity (g C m-2 yr-1 per °C) to inter-annual variation in temperature
Global average: -16.5 ± 9.1 g C m-2 yr-1 per °C
Tair
Sensitivity of NPP to temperature variation
150
NPP sensitivity to temperature
100
50
CARAIB
DLEM
0
LPJml
ORCHIDEE
-50
VEGAS
VISIT
ensemble
-100
-150
90
60
30
0
Latitude
-30
-60
-90
NPP is positively correlated with temperature in mid- and high latitudes, but
negatively related to temperature in topical region (30°N – 35°S).
Sensitivity of NPP to temperature variation
MAP (mm yr-1)
γ
MAT (°C)
Tair
Sensitivity of Rh to temperature variation
γ
CARAIB
ORCHIDEE
DLEM
LPJml
VEGAS
VISIT
ensemble
Rh sensitivity (g C m-2 yr-1 per °C) to inter-annual variation in temperature
Global average: 11.4 ± 5 g C m-2 yr-1 per °C
Tair
Sensitivity of Rh to temperature variation
100
50
Rh sensitivity to temperature
CARAIB
DLEM
LPJml
0
ORCHIDEE
VEGAS
VISIT
-50
ensemble
-100
90
60
30
0
Latitude
-30
-60
-90
Rh is positively related to temperature almost across all the
latitudinal bands, but shows large divergent among models
Sensitivity of Rh to temperature variation
MAP (mm yr-1)
γ
MAT (°C)
Tair
Sensitivity of NEP to temperature variation
γ
CARAIB
ORCHIDEE
DLEM
LPJml
VEGAS
VISIT
ensemble
NEP sensitivity (g C m-2 yr-1 per °C) to inter-annual variation in temperature
Global average: -27.9 ± 11.7 g C m-2 yr-1 per °C
Tair
Sensitivity of NEP to temperature variation
NEP sensitivity to temperature
200
100
CARAIB
DLEM
0
LPJml
ORCHIDEE
VEGAS
-100
VISIT
ensemble
-200
90
60
30
0
Latitude
-30
-60
NEP is negatively related to temperature in almost all the latitudinal
bands. The most negative sensitivity is in tropical region.
-90
Sensitivity of NEP to temperature variation
γ
MAP (mm yr-1)
Most negative temperature
sensitivity is in high MAT and
high MAP areas.
MAT (°C)
Tair
Sensitivity of carbon fluxes to temperature variation
Sensitivity of carbon fluxes to temperature
100
50
GPP
NPP
0
Rh
NEP
-50
-100
boreal
tropical temperate temperate temperate boreal
tropical
evergreen deciduous broadleaf needleleaf deciduous evergreen deciduous
forest
forest
forest
evergreen evergreen
forest
forest
forest
forest
mixed
forest
grassland shrubland
tundra
cropland
Result 2:
The sensitivity of carbon fluxes to
precipitation variation
Sensitivity of GPP to precipitation variation
γ
CARAIB
ORCHIDEE
DLEM
LPJml
VEGAS
VISIT
ensemble
MTE
prec
GPP sensitivity (g C m-2 yr-1 per mm yr-1) to inter-annual variation in precipitation
Global average: 0.3 ± 0.11 g C m-2 yr-1 per mm yr-1.
Sensitivity of GPP to precipitation variation
GPP sensitivity to precipitation
1,5
0,75
CARAIB
DLEM
LPJml
0
ORCHIDEE
VEGAS
VISIT
-0,75
ensemble
MTE
-1,5
90
60
30
0
Latitude
-30
-60
-90
MTE GPP and model ensemble show that GPP is positively correlated with
precipitation in almost all the latitudinal bands.
Sensitivity of GPP to precipitation variation
prec
MAP (mm yr-1)
γ
MAT (°C)
Highest sensitivity is in the low and
medium MAP areas.
Sensitivity of NPP to precipitation variation
γ
prec
NPP sensitivity (g C m-2 yr-1 per mm yr-1) to inter-annual variation in precipitation
Global average: 0.2 ± 0.08 g C m-2 yr-1 per mm yr-1.
Sensitivity of NPP to precipitation variation
1,5
NPP sensitivity to precipitation
0,75
CARAIB
DLEM
LPJml
0
ORCHIDEE
VEGAS
VISIT
-0,75
ensemble
-1,5
90
60
30
0
Latitude
-30
-60
-90
Sensitivity of NPP to precipitation variation
MAP (mm yr-1)
γ
MAT (°C)
prec
Sensitivity of Rh to precipitation variation
γ
CARAIB
ORCHIDEE
DLEM
LPJml
VEGAS
VISIT
prec
ensemble
Rh sensitivity (g C m-2 yr-1 per mm yr-1) to inter-annual variation in precipitation
Global average: 0.09 ± 0.03 g C m-2 yr-1 per mm yr-1
Sensitivity of Rh to precipitation variation
1
Rh sensitivity to precipitation
0,5
CARAIB
DLEM
LPJml
0
ORCHIDEE
VEGAS
VISIT
-0,5
ensemble
-1
90
60
30
0
Latitude
-30
-60
-90
Rh is positively correlated with precipitation in almost all the latitudinal bands.
Sensitivity of Rh to precipitation variation
MAP (mm yr-1)
γ
MAT (°C)
prec
Sensitivity of NEP to precipitation variation
γ
CARAIB
ORCHIDEE
DLEM
LPJml
VEGAS
VISIT
prec
ensemble
NEP sensitivity (g C m-2 yr-1 per mm yr-1) to inter-annual variation in precipitation
Global average: 0.11 ± 0.1 g C m-2 yr-1 per mm yr-1
Sensitivity of NEP to precipitation variation
0,6
NEP sensitivity to precipitation
0,3
CARAIB
DLEM
0
LPJml
ORCHIDEE
VEGAS
-0,3
VISIT
ensemble
-0,6
90
60
30
0
Latitude
-30
-60
-90
NEP is positively correlated with precipitation in almost all the latitudinal bands.
Sensitivity of NEP to precipitation variation
prec
MAP (mm yr-1)
γ
MAT (°C)
Similar to GPP sensitivity, highest NEP sensitivity to precipitation is in
the low and medium MAP areas as well.
Sensitivity of carbon fluxes to precipitation variation
Sensitivity of carbon fluxes to precipitation
0,5
GPP
0,25
NPP
Rh
NEP
0
boreal
boreal
tropical temperate temperate temperate
tropical
evergreen deciduous broadleaf needleleaf deciduous evergreen deciduous
forest
forest
forest
evergreen evergreen
forest
forest
forest
forest
mixed
forest
grassland
shrubland
tundra
cropland
Auburn News
Auburn professor confirms new threshold for climate
change mitigation
To curb the rate of climate change and maintain ecosystem services, the increase
in surface air temperature must remain below 1.5 degrees Celsius, according to a
recently published paper led by Shufen (Susan) Pan, …..
The Paris Climate Agreement: Why 2 Degrees C?
The COP 21 UN Climate change conference
195 nations
The 1.5 degree Celsius limit is a significantly safer defense line
against the worst impacts of a changing climate.
Projection: Impacts of temperature and precipitation changes
on Global NPP (2010-2099)
A2 Scenario
Terrestrial NPP increase
until 2050s under high
emission scenario;
however, NPP would
level off or decline after
2050 when temperature
increases by more than
1.5oC.
Pan, S. et al. 2014
36
ET Projection: ET/Precipitation: Water Scarcity
Increase terrestrial ET affecting global water balance especially in the Sahel
region, US great plain region, Australia and central Asia (ET/Prec > 1).
Pan, S. et al. 2015
37
Conclusions
1. Global GPP, NPP and NEP show negative correlation with
temperature, but Rh shows positive correlation with
temperature. Tropical region is the most sensitive area, also
most divergent among models.
2. Precipitation enhanced all the carbon fluxes (GPP, NPP and Rh)
and carbon sink (NEP). But there is large divergence among
models.
3. As indicated by model ensemble results, 1 degree temperature
increase would reduce carbon sink by -27.9 g C m-2 yr-1, while 1
mm yr-1 precipitation increase would enhance carbon sink by
0.11 g C m-2 yr-1.
4. Models could overestimate GPP sensitivity to precipitation
across most latitudinal bands, and GPP sensitivity to
temperature in tropical region.