Oceanic Constraints on the Size of the Terrestrial

Oceanic Constraints on the Size of the Terrestrial CO 2 Fertilization Sink • A joint atmosphere-ocean inversion New air-sea fluxes No “regularization” • A large CO 2 fertilization is sink not needed • Why do we get this result? Is it reliable? Andy Jacobson, Jorge L. Sarmiento, Manuel Gloor, Niki Gruber, and Sara Mikaloff Fletcher Atmospheric transport estimates: Trans. Com 3 modelers, Atmospheric observations: NOAA CMDL Globalview-CO 2 Program, GLODAP Ocean Carbon Survey: Robert M. Key, Richard Feely, Chris Sabine

Flux Comparison Fluxes for tropical land regions and southern land regions are correlated cannot be reliably estimated separately. Positive flux: source to atmosphere Land-air flux Sea-air flux mean ± 1 s. d. 80% CI

How much of the answer comes from data? Transport model: MATCH (R. Law)

How much of the answer comes from data? Transport model: MATCH (R. Law)

How much of the answer comes from data? Transport model: MATCH (R. Law)

How much of the answer comes from data? Transport model: MATCH (R. Law)

How much of the answer comes from data? Transport model: MATCH (R. Law)

Joint Inverse: Ocean Flux Constraints Key atmospheric inversions Trans. Com 3 L 2 control inversion result (Gurney et al. 2004) Takahashi (1999) Takahashi (2002) MOM 3 Forward sims ocean inversions

Joint Inverse: Ocean Flux Constraints Key atmospheric inversions Trans. Com 3 L 2 control inversion result (Gurney et al. 2004) Takahashi (1999) Takahashi (2002) MOM 3 Forward sims ocean inversions

Joint Inverse: Ocean Flux Constraints Key atmospheric inversions Trans. Com 3 L 2 control inversion result (Gurney et al. 2004) Takahashi (1999) Takahashi (2002) MOM 3 Forward sims ocean inversions

Joint Inverse: Ocean Flux Constraints Key atmospheric inversions Trans. Com 3 L 2 control inversion result (Gurney et al. 2004) Takahashi (1999) Takahashi (2002) MOM 3 Forward sims ocean inversions

Oceanic Constraints on Terrestrial Fluxes Independent Fluxes Reject zero at 95% confidence level for T+SL flux. 77% probability that the T+SL flux is greater than 1. 0 Pg. C/yr (vs. 49% previous best guess).

Implications for a CO 2 Fertilization Sink Residual ~0 ± 1 Pg. C/yr Small residual flux is not compatible with the existence of a large fertilization sink in the tropics.

Air-Sea Flux Comparison Contemporary Fluxes 1992 -6

Air-Sea Flux Comparison Contemporary Fluxes 1992 -6

Anthropogenic and Preindustrial Flux Estimates

Anthropogenic and Preindustrial Flux Estimates

Anthropogenic and Preindustrial Flux Estimates

C* = DIC - Cbio = Cgasex + Cant + const

Can Forward Models be so Wrong? Ocean uptake Ocean outgassing Zonally-averaged air-sea fluxes [Murnane et al. 1999]. In the Southern Ocean, a modest net flux is the result of large, offsetting contributions from the solubility pump and from the biological pump.

Can p. CO 2 Fluxes be so Wrong? The Southern Ocean is undersampled, especially in the winter. Summertime p. CO 2 is subject to biological drawdown, but wintertime levels may be higher. Figure courtesy of Colm Sweeney

Conclusions • Atmospheric data alone cannot reliably resolve regional fluxes • Ocean interior data constrain air/sea fluxes very strongly • Large tropical & southern land (T&SL) source detected • Magnitude of T&SL source similar to deforestation estimates. Hence, a large CO 2 fertilization sink is not needed to close carbon budget.

Partitioning of Sink: Land vs. Ocean -1. 1 ± 0. 2 Pg. C/yr Estimates for 1992 -96 except where noted; By convention a negative flux represents uptake from the atmosphere; Outgassing of riverine carbon (0. 45 Pg. C/yr) attributed to ocean. -1. 7 ± 0. 2 Pg. C/yr

Construction of the Joint Inverse Key atmospheric inversions Trans. Com 3 control inversion result Takahashi (1999) Takahashi (2002) MOM 3 Forward Sims ocean inversions joint inversions

Oceanic Constraints on Terrestrial Fluxes • Anticorrelated Fluxes • Can we reject zero?

Joint Inversion Constraints 67 484 observations in ocean, 76 on land

IPCC Terrestrial Models: CO 2 Flux to Atmosphere (Pg. C yr-1) Predicted Response to Rising Atmospheric CO 2 Concentrations and Climate Change 1980 s Fertilization Sink: 0. 3 to -3. 8 Pg. C/yr (reported) 4 2 0 -2 -4 -6 Hybrid LPJ -8 IBIS SDGVM -10 -12 1850 VECODE 1992 -96 Fertilization Sink: about -1 to -3. 5 Pg. C/yr (rough estimates from graph) Triffid CO 2 Effect Only (model range) 1900 1950 2000 2050 2100