estimates of atmospheric potential oxygen (apo) fluxes ...€¦ · atmosphere co2 o2 apo co2 o2 apo...
TRANSCRIPT
Estimates of Atmospheric Potential Oxygen (APO) FluxesBased on O2/N2 and CO2 Concentration Measurements:
What Can They Tell Us About The Global Carbon Cycle?
C. Rodenbeck1, C. Le Quere1,2, R.F. Keeling3, Y. Tohjima4, N. Cassar5, A. Manning1, and M. Heimann1
1Max Planck Institute for Biogeochemistry Jena (Germany)2Univ. of East Anglia and British Antarctic Survey (UK)
3Scripps Institution of Oceanography, University of California, San Diego, California (USA)4National Institute for Environmental Studies, Tsukuba (Japan)
5Department of Geosciences, Princeton University, Princeton NJ (USA)
Many thanks to:
Computing centers: DKRZ, GWDG
CONCEPT
What is Atmospheric Potential Oxygen?
Atmosphere
CO2 CO2 CO2
↑ l l
F B O
Fossil fuel burning Terrestrial biosphere Ocean
1
CONCEPT
What is Atmospheric Potential Oxygen?
Atmosphere
CO2 O2 CO2 O2 CO2
↑ ↓ l l l
F -1.4 F B -1.1 B O
Fossil fuel burning Terrestrial biosphere Ocean
2
CONCEPT
What is Atmospheric Potential Oxygen?
Atmosphere
CO2 O2 CO2 O2 CO2 O2
↑ ↓ l l l l
F -1.4 F B -1.1 B O Z
Fossil fuel burning Terrestrial biosphere Ocean
Decoupling CO2,O2: • Carbonate chemistry (buffer effect)• Physical transport (dilution)•Warming / cooling (solubility changes)
3
CONCEPT
What is Atmospheric Potential Oxygen?
Atmosphere
CO2 O2 CO2 O2 CO2 O2
↑ ↓ l l l l
F -1.4 F B -1.1 B O Z
Fossil fuel burning Terrestrial biosphere Ocean
APO = O2 + 1.1·CO2
(Stephens et al., 1998)
4
CONCEPT
What is Atmospheric Potential Oxygen?
Atmosphere
CO2 O2 APO CO2 O2 APO CO2 O2 APO
↑ ↓ l l l l l l
F -1.4 F -0.3 F B -1.1 B -/- O Z “≈Z”
Fossil fuel burning Terrestrial biosphere Ocean
APO = O2 + 1.1·CO2
(Stephens et al., 1998)
5
CONCEPT
What is Atmospheric Potential Oxygen?
Atmosphere
CO2 O2 APO CO2 O2 APO CO2 O2 APO
↑ ↓ l l l l l l
F -1.4 F -0.3 F B -1.1 B -/- O Z “≈Z”
Fossil fuel burning Terrestrial biosphere Ocean
APO = O2 + 1.1·CO2
Information on: • Marine biological activity• Mixing/stratification, upwelling• Gas exchange→ Impact also on carbon cycle
6
CONCEPT
What is Atmospheric Potential Oxygen?
Atmosphere
CO2 O2 APO CO2 O2 APO CO2 O2 APO
↑ ↓ l l l l l l
F -1.4 F -0.3 F B -1.1 B -/- O Z “≈Z”
Fossil fuel burning Terrestrial biosphere Ocean
APO = O2 + 1.1·CO2
Information on: • Marine biological activity• Mixing/stratification, upwelling• Gas exchange→ Impact also on carbon cycle
Quantification: • Based on atmospheric measurements• Method: ‘Atmospheric inversion’
7
METHOD
Inversion principle
&%'$ &%
'$Il l l l l
Causality: Fluxes, Transport −→ Concentration gradientKnowledge:
∑Fluxes ←− Concentration gradient, Transport
cmeas ←→ cmod = Af + c0
Inversion = Multidimensional linear regression
J(f) = (cmeas − cmod)TQc(cmeas − cmod) −→ min
8
METHOD
Inversion principle
&%'$ &%
'$Il l l l l
Causality: Fluxes, Transport −→ Concentration gradientKnowledge:
∑Fluxes ←− Concentration gradient, Transport
cmeas ←→ cmod = Af + c0
Inversion = Multidimensional linear regression
‘A-priori information’: f = ffix + Fp
• Spatio-temporal weighting• Spatial correlations• Temporal correlations
No a-priori phase information
9
METHOD
Oxygen (APO) Data
Data availability: Sampling locations:
1993 1995 1997 1999 2001 2003 2005 2007 2009year (A.D.)
fxALTs SIOfxBRWp PUfxCBAs SIOfxCOIn NIESfxLJOs SIOdxTHDsfxHATn NIESfxKUMs SIOfxMLOs SIOfxSMOs SIOfxAMSp PUfxCGOs SIOfxMQAp PUfxPSAs SIOfxSYOp PUfxSPOs SIO
# meas./month: 1- 2 3- 8 9- 20 21-
OCEAN TOTAL
1993 1995 1997 1999 2001 2003year (A.D.)
-300-200
-100
0
100
200300
AP
O F
lux
(Tm
olO
2/yea
r)
Inversion runs based on:5 sites7 sites
16 sites(at least latitudinal coverage)
10
FIRST RESULTS
Do We Have Enough Information?
Inverting synthetic data(at actual locations and times):
Retrieval of“Known thruth”
5 sites7 sites
16 sites
→ Seasonality ‘seen’
NH Ocean
-2000
-1000
0
1000
2000
AP
O F
lux
(Tm
olO
2/yea
r)
Tropical Ocean
-2000
-1000
0
1000
2000
SH Ocean
1997 1998 1999 2000 2001 2002 2003-2000
-1000
0
1000
2000
OCEAN TOTAL
1997 1998 1999 2000 2001 2002 2003year (A.D.)
-2000
-1000
0
1000
2000
11
FIRST RESULTS
Do We Have Enough Information?
Inverting synthetic data(at actual locations and times):
Retrieval of“Known thruth”
5 sites7 sites
16 sites
→ Main interannual patterns ‘seen’
NH Ocean
-200
-150
-100
-50
AP
O F
lux
(Tm
olO
2/yea
r)
Tropical Ocean
100
150
200
250
SH Ocean
1997 1998 1999 2000 2001 2002 2003-150
-100
-50
0
OCEAN TOTAL
1997 1998 1999 2000 2001 2002 2003year (A.D.)
-100
-50
0
50
100
12
FIRST RESULTS
Do We Have Enough Information?
Inverting synthetic data(at actual locations andtimes):
Retrieval of“Known thruth”
5 sites7 sites
16 sites
→ Southern ocean:split incompletelyresolved
NH Pacific
-100
-75
-50
-25
0
AP
O F
lux
(Tm
olO
2/yea
r)
Tropical Pacific
25
50
75
100
125
150
SH Pacific
1997 1998 1999 2000 2001 2002 2003-150
-125
-100
-75
-50
NH Atlantic
-100
-75
-50
-25
0
Tropical Atlantic
50
75
100
125
150
SH Atlantic
1997 1998 1999 2000 2001 2002 2003year (A.D.)
-25
0
25
50
75
OCEAN TOTAL
1997 1998 1999 2000 2001 2002 2003-100
-75
-50
-25
0
25
5075
13
FIRST RESULTS
Do We Have Enough Information?
Inverting synthetic data(at actual locations and times):
Retrieval of“Known thruth”
5 sites7 sites
16 sites
→ Finer regional splitpoorly resolved
N Pacific
-100
-75
-50
-25
0
N Pacific Gyre
-50
-25
0
25
50
Tropical Pacific
25
50
75
100
125
AP
O F
lux
(Tm
olO
2/yea
r)
S Pacific Gyre
-50
-25
0
25
50
Southern Pacific
1997 1998 1999 2000 2001 2002 2003-150
-125
-100
-75
-50
North polar ocean
-50
-25
0
25
50
N Atlantic
-50
-25
0
25
50
N Atlantic Gyre
-100
-75
-50
-25
0
Tropical Atlantic
0
25
50
75
100
S Atlantic Gyre
-25
0
25
50
75
Southern Atlantic
1997 1998 1999 2000 2001 2002 2003year (A.D.)
-25
0
25
50
75
OCEAN TOTAL
-100-75
-50
-25
0
25
5075
Indian Tropics
-25
0
25
50
75
S Indian Gyre
0
25
50
75
100
Southern Indian
1997 1998 1999 2000 2001 2002 2003-100
-75
-50
-25
0
14
FIRST RESULTS
Robustness?
Varying a-priori assumptions:
• (Standard)• Uncertainty• Spatial correlations• IAV spatial weighting• Temporal correlations
→ Main patterns stable
NH Ocean
-200
-100
0
100
AP
O F
lux
(Tm
olO
2/yea
r)
Tropical Ocean
0
100
200
300
400
SH Ocean
1997 1998 1999 2000 2001 2002 2003-200
-100
0
100
OCEAN TOTAL
1997 1998 1999 2000 2001 2002 2003year (A.D.)
-200
-100
0
100
200
300
15
FIRST RESULTS
Comparison to other data?
Seasonality:
• Phase agreement with∆pO2-based fluxes
[Garcia & Keeling (2001)]
• Similar amplitudes(but larger in NorthernExtratropics)
NH Ocean
-3000
-2000
-1000
0
1000
2000
3000
AP
O F
lux
(Tm
olO
2/yea
r)
Tropical Ocean
-3000
-2000
-1000
0
1000
2000
3000
SH Ocean
1997 1998 1999 2000 2001 2002 2003-3000
-2000
-1000
0
1000
2000
3000
OCEAN TOTAL
1997 1998 1999 2000 2001 2002 2003year (A.D.)
-3000
-2000
-1000
0
1000
2000
3000
16
FIRST RESULTS
First Discussion
• Tropics: El Nino signal(Signal from differences)
NH Ocean
-300
-200
-100
0
100
200
300
AP
O F
lux
(Tm
olO
2/yea
r)
Tropical Ocean
-300
-200
-100
0
100
200
300
SH Ocean
1993 1995 1997 1999 2001 2003-300
-200
-100
0
100
200
300
OCEAN TOTAL
1993 1995 1997 1999 2001 2003year (A.D.)
-300
-200
-100
0
100
200
300
17
FIRST RESULTS
First Discussion
• Tropics: El Nino signal
• Important interann. variab.also in Extratropics
NH Ocean
-300
-200
-100
0
100
200
300
AP
O F
lux
(Tm
olO
2/yea
r)
Tropical Ocean
-300
-200
-100
0
100
200
300
SH Ocean
1993 1995 1997 1999 2001 2003-300
-200
-100
0
100
200
300
OCEAN TOTAL
1993 1995 1997 1999 2001 2003year (A.D.)
-300
-200
-100
0
100
200
300
18
FIRST RESULTS
First Discussion
• Tropics: El Nino signal
• Important interann. variab.also in Extratropics
• Compare to ocean models:
– Larger amplitudes– Little phase agreement
NH Ocean
-300
-200
-100
0
100
200
300
AP
O F
lux
(Tm
olO
2/yea
r)
Tropical Ocean
-300
-200
-100
0
100
200
300
SH Ocean
1993 1995 1997 1999 2001 2003-300
-200
-100
0
100
200
300
OCEAN TOTAL
1993 1995 1997 1999 2001 2003year (A.D.)
-300
-200
-100
0
100
200
300
19
FIRST RESULTS
First Discussion
• Tropics: El Nino signal
• Important interann. variab.also in Extratropics
• Compare to ocean models:
– Larger amplitudes– Little phase agreement
• Relation to heat flux?(change in heat content[Levitus et al., GRL 32 (2005)]assuming 6.1 nmolO2/J)
– Cannot use heat fluxto calculate CO2 budgeton short (≈5 year) time scale
NH Ocean
-300
-200
-100
0
100
200
300
AP
O F
lux
(Tm
olO
2/yea
r)
Tropical Ocean
-300
-200
-100
0
100
200
300
SH Ocean
1993 1995 1997 1999 2001 2003-300
-200
-100
0
100
200
300
OCEAN TOTAL
1993 1995 1997 1999 2001 2003year (A.D.)
-300
-200
-100
0
100
200
300
20
Conclusion – Outlook• Interannual atmospheric inversion of APO:
At least 3 latitudinal bands resolved
• Further research questions:
– Quantifying climate impacts on oxygen signal→ O2-based oceanic carbon budgets?
– Climatic trends→ increasing ocean stratification?
– Quantifying air-sea gas exchange– Improving CO2 inversions by oxygen constraint