Transcript What are the Most Important Factors for Climate

Controls of carbon budgets in terrestrial
ecosystems
Does carbon storage in terrestrial ecosystems really
depend on temperature?
What factors do we need to consider when we assess
vegetation—atmosphere coupling in climate change
scenarios?
•How well do coupled climate—vegetation models simulate the response
of forest CO2 fluxes and carbon sequestration to variations in
temperature?
•What controls the long-term storage of organic carbon in boreal soils ?
[current content: 200—400 ppm of CO2 ; all soils, > 700 ppm ]
Steven C. Wofsy, Harvard University
Presented at: ICDC7 Boulder, CO, 29 Sep 2005
“Land carbon storage depends on the balance between the
input of carbon as Net Primary Productivity (NPP), and the loss
of carbon as heterotrophic (soil) respiration (Rh).”
•NPP modeled as dependent on light, nutrients, precipitation
•Rh modeled as a function of T. Is this right?
Daily Mean Ecosystem Respiration (mmole/m2/s) vs T (C)
Photosynthesis vs sunlight
Harvard Forest Canopy Photosynthesis vs Sunlight
0
-10
-15
-20
-30
40
-25
60
__ -34.6*PAR/(552 + PAR)
20
Exp fit
0
R
80
Canopy Photosynthesis
100
-5
120
Harvard Forest
-20
-10
0
10
T oC
20
30
0
500
1000
Sunlight (mE/m2/s)
1500
Modeling long-term Net Ecosystem Carbon
Production (NEP)(1st-order):
“tree”
solar
Atmosphere
CO2, H2O
Leaves (P, R6)
CO2
Org C
H2O,
nutrients
Boles (R5)
Org C
Roots (R4)
Soils R1
R2
R3
Do climate—ecosystem models capture the
temperature dependence of carbon storage?
Or the long-term trends in C storage?
Example 1: a typical mid-latitude forest (Harvard Forest,
Central New England; agricultural use 1750-1850).
60-80 year old mixed deciduous
forest, with CO2 fluxes to/from
the atmosphere measured every
half hour, for 15 years (19912005) (S. Wofsy, J. W. Munger, B. Daube,
M. Goulden, C. Barford, S. Urbanski, many
others),
plus a soil warming
manipulation (J. Melillo, K.
Nadelhofer, P. Studler).
Model (IBIS) vs. Observed C balance
at a transition deciduous forest.
Errors in both
seasons are due
to T driving
respiration
emission
uptake
phenology
R vs T: fundamental to climate change—the mean is right, but the model
is never right in any month and the CO2 Flux–T feedback is wrong!
Modeled and observed respiration at Harvard Forest
“Overall, warming treatment did not significantly change soil respiration
either with or without clipping.” (Luo et al., 2001)
Annual change (%) in soil respiration due to 5o soil warming over
a thirteen-year period at Harvard Forest [Melillo, Nadelhofer,
Studler, et al. –Marine Biological Laboratory, Woods Hole, MA].
Does this
extra N
increase
enhance
tree
growth?
Annual change (%) in N mineralization due to soil warming over a
thirteen-year period at Harvard Forest [Melillo, Nadelhofer,
Studler, et al. –Marine Biological Laboratory, Woods Hole, MA].
0
•Accel. G, R
•Accel. NEE
•Higher LUE
-1
-2
NEE
-3
-4
NEE (tonC/ha/yr)
Harvard Forest
AmeriFlux Data: 1991--2004
AGWB
incr. rate
net uptake
1992
1994
1996
1998
2000
2002
2004
1992
1994
1996
1998
2000
2002
2004
-16
-18
1994
1996
1998
2000
2002
2000
2002
NEElight sat’d
-20
12
emission
1992
-22
14
R
-1x GEE
10
R (tonC/ha/yr)
16
uptake
mean NEE PAR 1200-1500 (mmole/sq. m/yr)
Year
More efficient
1992
1994
1996
1998
Year
Long-term changes atYear
Harvard Forest J. W. Munger, S. Urbanski, S. C. Wofsy et al.
20
Example 2:
Climate and C
at a Boreal
Forest (NOBS)
flux site,
Thompson, MB
-15
45% cover
January
60
-20
Jan
30
-25
-30
40
T (C)
50
Thompson
LynnLake
1975
1980
1985
1990
1995
2000
2000
Decade
April
Snow cover
1900
Temperature
30
1920
1940
1900
(rapid warming)
1980
2000
2000
Year
July
10
14
15
20
T (C)
16
25
1960
Regional Composite T
Regional Smoothed
NOBST
18
1970
1970
-35
20
Monthly mean Snow (cm)
50 cm
PEAT
January
Jul
12
5
Monthly mean Snow (cm)
30
1970
1975
1980
1985
1990
1995
2000
1900
1920
1940
1960
Decade
Year
1980
2000
50
88
44
a. NEE
48
5
0
-6
-4
-29
-50
-27
-40
-60
Uptake | emission
g C m-2 yr-1
100
2
°C
1
0
-1
b. T anomaly
-2
cm
5
0
c. CMI.3 anomaly
-5
Temperatures
warm up,
Boreal Forest: Comparison of (a) NEE, (b) temperature ecosystem
anomaly, (c) 3-year lagged climate moisture index
switches:
anomaly (10 years of AmeriFlux Data)
sourcesink
1995
1997
1999
2001
2003
Temper1ature, °C


6
R, measured
R, modelled
Water table depth
-15
-20
1 Sep
-25
1 Aug
-30
1 Jul
-2
4
0
2
-10
0
-20
-30
Jun. 1
Jul. 1
Aug. 1
Sep. 1
Water table depth, cm
R (gC mg-2C mday-1)
8
1 Jun
Daily respiration,
Water table depth, cm
0
T
-10
20
10
10
20
Water Table
0
Water Table Depth (cm)
or T (c)
30
WT
15cm
25cm
35cm
30
NEE: C balance at a bog: Jun—Sep
Summer
2002
Summer
2003
Summer
2004
50
88
44
a. NEE
48
5
0
-6
-4
-29
-50
-27
-40
-60
Uptake | emission
g C m-2 yr-1
100
2
°C
1
0
-1
b. T anomaly
-2
cm
5
0
c. CMI.3 anomaly
-5
1995
1997
1999
2001
2003
Water balance &
Boreal Forest: Comparison of (a) NEE, (b) temperature (temperature)
explain the
anomaly, (c) 3-year lagged climate moisture index
transition:
anomaly (10 years of AmeriFlux Data)
sourcesink
Correlation: {DT, D soil moisture index}
CCSM1-Carbon Control Simulation
DJF
JJA
Positive correlation  warmer-wetter; or cooler-drier
Negative correlation  warmer-drier; or cooler-wetter
slide courtesy Inez Fung [I. Fung, S. Doney, et al.]]
Norway House
Gillam
NOBS
0
-2
-4
high T, high ppt
Low T, low ppt
-6
15
10
5
0
-5
-10
-15
1970
1980
1990
Precipitation
anom aly, cm
Mean tem perature, °C
Thompson
Flin Flon
Lynn Lake
2
2000
Recent climate variations in central Canada have been
[cold:dry] and [warm:wet] …
-20
-10
lat
0
10
Holdridge Life Zones & potential vegetation: Mean T, Precip, and E/P control
vegetation cover: warmer-drier leads to strong degradation in the tropics.
Data courtesy of D. Skole
-80
-70
P
-60
-50
lon
-40
1.5
E/P
6
T
24
Holdridge life zones (Holdridge 1967)
Summary
•How well do coupled climate—vegetation models simulate the
response of forest CO2 fluxes and carbon sequestration to warming?
Many models over-estimate the sensitivity of ecosystem C storage
to T.
• What controls the long-term storage of organic carbon in boreal
soils ? [current content: 200—400 ppm of CO2 ; all soils, > 700 ppm
equivalent]. Hydrological balance in concert with T dominate; if
changed climate has warmer wetter covariance, boreal lands may
actually increase C storage in a warmer world.
What did we learn?
Coupled climate vegetation modeling
must carefully and critically examine
predictions for climate change and
ecosystem response in terms of the key
parameters (•regional T •Precipitation
•Human impacts (ignition, agriculture,
air pollution) and their •covariances .
Consider advancing beyond the traditional focus
on global mean T.
These are all 1st order factors.