Cindy Nevison University of Colorado Mati Kahru, Ralph Keeling, Manfredi Manizza, B. Greg Mitchell

Evaluating satellite ocean color-derived export production in the Southern Ocean

using atmospheric O2/N2 dataCindy Nevison

University of Colorado

Mati Kahru, Ralph Keeling, Manfredi Manizza, B. Greg MitchellScripps Institution of Oceanography

Matt Charette, Kanchan Maiti Woods Hole Oceanographic Institution

Acknowledgements:NASA NNX08AB48G, Michael Bender, Nicolas Cassar, Ray Langenfelds, N2O

and CFC-12 data from AGAGE, NOAA CCGG and CSIRO

Outline

1) Export Production from Ocean Color

2) Atmospheric O2/N2 Data

3) Decomposition of O2/N2 seasonal cycle into Component Signals (land, thermal, ventilation, production)

4) Atmospheric Transport Model Uncertainty – APO Transcom Experiment

Improved Chl-a Algorithm for Southern Ocean

B.G. Mitchell et al.

NPP estimated from Ocean ColorModified Behrenfeld and Falkowski = f(Surface T, Chl-a, PAR)

Palmer LTER data using SPGANT algorithm (Kahru and Mitchell)

Export Production = ef*NPPexport fraction (ef) = f(NPP, SST)

Southern Ocean in situ data compiled from the literature (Maiti et al.)

Export Production from Ocean Color

Atmospheric O2/N2 Monitoring Stations

SIO (R. Keeling), Princeton (M. Bender, N.

Cassar)

Global South of 50S 18 Pg C/yr 1.6 Pg C/yr (Unmodified Laws model for ef)12 Pg C/yr 1.0 Pg C/yr (Laws scaled to Schlitzer, 2002)

Atmospheric O2/N2: Decreasing Trend and Seasonal Cycles

Bender et al., 2005

per m

eg

Atmospheric Potential Oxygen: Barrow, AlaskaRemove Land Signal from O2/N2 using CO2 data

O2/N2 APO

CO X1.1 /NO APO 2

o222

What causes Seasonal Cycles in APO?How are they linked to carbon export production?

Adapted from Keeling et al., 1993Corg 1 to sediment

Base of Seasonal Thermocline

Euphotic Zone

Deep Ocean

Main Thermocline

Corg 20

Corg 2

Corg 10

Production 100Consumption 80

Atmosphere O2 Flux 15

O2 Flux 5

O2 Flux 3

O2 Flux 1

Wintertime VentilationO2 Flux 8

Thermocline VentilationO2 Flux 4

Deep Water VentilationO2 Flux 3

Consumption 10

Consumption 8

Consumption 1

O2 Flux 1

Export Production

APO Seasonal Cycle at Cape Grim, TasmaniaCompare to MATCH:NCEP transport model run with satellite-derived export flux

assume 1.4 mole O2 released per mole C exported

APOobserved = APOthermal + APOnet_biological

= APOthermal + APOdeep_ventilation+ APOsurface_production

Thermally Corrected APO Seasonal Cycle at CGO

Thermal cycle estimated based on Ar/N2 data or QST/Cp (need ATM)

APO Seasonal Cycle at Cape Grim, TasmaniaCompare to MATCH:NCEP transport model run with satellite-derived export flux

APOobserved = APOthermal + APOnet_biological

= APOthermal + APOsurface_production + APOdeep_ventilation

Estimation of APO ventilationsignal based on atmospheric N2O data

AGAGE data from Cape Grim, Tasmania

Correlated Changes in Atmospheric N2O and APO during Coastal Upwelling Events at Trinidad Head

Lueker et al., 2003

APO

N2O

N2O is produced during subsurface remineralization, but not affected by euphotic zone primary production

Grey Band is Upwelling Event

N2O and O2 are Anticorrelated in Ocean Depth Profiles

Southern Ocean data from Brian Popp and Tom Trull

N2O/AOU ratio is well defined

TheN2O/AOU molar ratio in the subsurface ocean is reflected in observed N2O/APO atmospheric variations

during coastal upwelling events

Lueker et al., 2003

Decomposition of APO and N2O Seasonal Cycles at Cape Grim, Tasmania

N2O APOthermal + stratospheric + ventilation thermal + ventilation + production

Compare Decomposed APO Seasonal Cycle at Cape Grim to MIT Ocean Model Results

M. Manizza

N2O Seasonal Cycles in Individual Years at Macquarie Island (CSIRO Data)

Separating Stratospheric and Oceanic N2O Signals at Macquarie Island

Decomposition of APO and N2O Seasonal Cycles at Macquarie Island

N2O APO thermal + ventilation thermal + ventilation + production

Decomposition of APO Seasonal Cycle at Palmer Station, Antarctica

30 per meg * 0.2095 * 10-6 = O2/N (moles O2/moles dry air)But what is N? Some fraction of 1.77 x 1020 total moles dry air

Atmospheric Transport Model (ATM) Uncertainty

Evaluation of oceanic O2 fluxes based on observed APO seasonal cycles or latitudinal gradient, e.g., Naegler et al. [2007]

Conclusion: These evaluations are difficult because shortcomings in ATMs and problems of data representativity cannot be distinguished from uncertainties in the O2 fluxes.

APO Transcom Experiment (Blaine, 2005)

9 atmospheric transport models forced with Garcia and Keeling [2001] monthly oceanic O2 and N2 flux anomalies. Below: APO seasonal cycles at Cape Grim, Tasmania

APO Transcom Seasonal Amplitude vs. Vertical Gradient at four Southern Hemisphere Stations

Summer Winter

Red dotted line is observed APO amplitude

APO Transcom Vertical Profiles near Cape Grim, Tasmania vs. CSIRO Aircraft Data

Summer Winter

Data courtesy of Ray Langenfelds

Conclusions1) Satellite ocean color data: uncertainties compounded in the sequence from Chl-a to NPP to Export Production.

2) Partitioning of net biological APO signal into surface production and deep ventilation: Atmospheric N 2O-based method, MIT ocean model and satellite ocean color data converge on similar answer.

3) Can now compare surface production signal in APO directly to export production from ocean color, but some major uncertainties:

1) Need for atmospheric transport model as intermediary 2) N2O cycle decomposition – thermal and stratospheric signals

Supplementary Slides

Southern Ocean CO2 Sinkand Biological Pump

Takahashi et al., 2009 CO2 Flux Export Production from Ocean Color Satellite 14º-50ºS sink = 1.05 Pg C/yr

Southern Ocean as Source and Sink of CO2

.

Note scale differenceLovenduski et al., 2007

Components of net biological signal in APO: Three Possible Views observed – thermal = net biological = production + ventilation

MIT Ocean Model Results from M. Manizza

APO at Cape Grim, TasmaniaRemove Land Signal from O2/N2 using CO2 data

O2/N2 APO

CO X1.1 /NO APO 2

o222