Measurements and Models of Oceanic O2 and CO2 Fluxes
Mark Battle (Bowdoin College)Sara Mikaloff Fletcher (UCLA)Michael Bender (Princeton)
Ralph Keeling (SIO) Nicolas Gruber (UCLA)
Pieter Tans (NOAA/CMDL)Melissa B. Hendricks (Princeton) David T. Ho (Princeton/Columbia)
Carrie Simonds (Bowdoin College)
Robert Mika (Princeton)Andrew Manning (SIO)
Bill Paplawsky(SIO)
AGUFall 2004
OS11C-08
Funding from:NSF
NOAA GCRPBP-Amoco
AGU poster Fall
2003 A52B-0793
On the agenda:
• What is APO?• Historical context• Our dataset• From sparse data to meridional
gradients• Modeling• Data-model comparison
Atmospheric Potential Oxygen
APO O2 + 1.1 CO2
APO changes solely due to oceanicprocesses*
*This is almost true
1.4
1.1
1.4
1.1
Atmospheric Potential Oxygen
APO O2 + 1.1 CO2
APO responds to oceanic O2 & CO2 fluxes
Land biota doesn’t change APO
Fossil fuels change APO a little
In the beginning…
Stephens et al., 1998
Models don’t getinterpolar gradientright (physics?)
Equatorial datawould be nice.
The next chapter…
Gruber et al., 2001
Eliminate BGCModel.
Results seem Independent ofOcean physics
Equatorial dataWould be nice.
We have equatorial data!
Sampling locations used in this work
Ships of opportunity
NOAA ship Ka’imimoana
Uneven spatio-temporal data density
Uneven spatio-temporal data density
Annual-mean gradients from sparse data?
2-D interpolation (latitude and time):
Create gradients at specific times through year
Average the gradients over climatological year
3 examples of gradients…
A weighted average of all gradients
Annual-mean gradients from sparse data?
2-D interpolation (latitude and time):
Create gradients at specific times through year
Average the gradients over climatological year
Seasonal Cycles:
Sine fits to data at each sampling latitudeAnnual means from sine fits
Annual means from seasonal cycles
New data deserve a new model
•Aseasonal O2: Ocean inversion (Gruber 2001)•Aseasonal N2: Heat inversion (Gloor 2001)•Oceanic CO2: pCO2 (Takahashi 1999)•Seasonal O2 and N2: Ocean heat fluxes (Garcia and Keeling 2001)•FF CO2 and O2: CDIAC (Marland 2000)•Atmospheric Transport: TM3.8•Winds: NCEP 1995 – 2000 (repeated and averaged)
Data-model comparison: 2-D interpolation
Data-model comparison: seasonal cycles
Is this different from old models?
The models really are different!
What has changed?
• Atmospheric Transport:– Was GCTM– Is now TM3
• Seasonal O2 & N2:
– Was Najjar & Keeling/Esbensen & Kushnir
– Is now Garcia and Keeling
In summary…
• Data coverage much greater• Equatorial “bulge” exists• Interpolar gradient smaller• Newest model gives much better overall
agreement with data• Past disagreements primarily due to
atmospheric transport• Disagreement persists at SYO, SAB and
CBA• Watch for a publication soon.
UCLA model annual mean APO
Merging PU & SIO datasets
Problems at Cold Bay
Time slices: Data
Time slices: Model
Data quality at Sable Island