Regulation of atmospheric CO 2 by deep‐sea sediments in an Earth system model

Using the full ensemble results to provide probabilistic estimates of model variables or parameters would not be a valid approach because of large uncertainties in model structure.

“because of large uncertainties in model structure, meaning that the model is known a priori to be a far from perfect model, using the full ensemble results to provide probabilistic estimates of model variables or parameters would not be a valid approach, as discussed by Ridgwell et al. [2007].”

• cites Ridgwell et al. [2007] sourceSupported

Adding 2-D sediment data to the EnKF resulted in less than 1% increase in the mean square difference between data and model fields for alkalinity and phosphate compared to observations.

“The improvement in wt% CaCO 3 afforded by the addition of 2-D sediment data in the EnKF resulted in only a very slight degradation in the quality of alkalinity and phosphate fields compared to observations, less than 1% increase in the mean square difference between the data and the model fields compared to that presented by Ridgwell et al. [2007].”

• cites Ridgwell et al. [2007] sourceSupported

Previous analysis of the role of deep-sea sediments in the long-term fate of fossil fuel CO2 employed off-line tracer-transport marine carbon cycle models in which the ocean circulation was prescribed.

“Previous analysis of the role of deep-sea sediments in the long-term fate of fossil fuel CO 2 employed off-line tracer-transport marine carbon cycle models in which the ocean circulation was prescribed [e.g., Archer et al., 1997Archer et al., , 1998;;Sundquist, 1990].”

• cites Archer et al., 1997 sourceNot checked — no full text of the source
• cites Archer et al., 1998 sourcePartly supported

  The Model Description (page unavailable) says the authors coupled an ocean-circulation and water-column chemistry model to a deep-sea sediment model, supporting the citation as an example of the stated research. It does not specify that the model was off-line or that ocean circulation was prescribed; searches for those terms did not establish that methodological detail. (LLM, conf=0.82)

  Experimental — accuracy not yet measured
• cites Sundquist, 1990 sourceNot checked — no full text of the source

Because the GENIE-1 model contains a climate component, the authors are able to extend previous analysis by explicitly assessing the impact of major feedbacks between CO2 and climate on carbonate compensation.

“Because the GENIE-1 model contains a climate component we are able to extend the analysis made in formative earlier studies [e.g., Archer et al. , 1997Archer et al. , , 1998;;Sundquist, 1990] by now explicitly assessing the impact of major feedbacks between CO 2 and climate on carbonate compensation.”

• cites Archer et al., 1997 sourceNot checked — no full text of the source
• cites Archer et al., 1998 sourceSupported
• cites Sundquist, 1990 sourceNot checked — no full text of the source

Important feedback may exist between CO2 and marine carbonate production.

“Important feedback may also exist between CO 2 and marine carbonate production.”

• cites Orr et al., 2005 sourceSupported
• cites Caldeira and Wickett, 2003 sourceNot checked — no full text of the source
• cites Zondervan et al., 2001 sourceSupported

With a fixed ocean circulation and climatology, it is difficult to account for the importance of feedbacks with climate except in a highly parameterized manner.

“With a fixed ocean circulation and climatology it is difficult to account for the importance of feedbacks with climate except in a highly parameterized manner [Archer et al., 2004;Archer, 2005].”

• cites Archer et al., 2004 sourcePartly supported

  Section 4.1.1 describes Hamocc2 as an offline model in which temperature changes affect CO₂ solubility but not ocean circulation; it imposes a uniform temperature response to atmospheric CO₂. The Fig. 3 caption describes that response as “hard wired” into the model. These passages support the claim’s description of parameterized climate feedback under fixed circulation, but do not establish the broader assertion that accounting for such feedbacks is generally difficult except in that manner. (LLM, conf=0.78)

  Experimental — accuracy not yet measured
• cites Archer, 2005 sourceNot checked — no full text of the source

Most (~70-80%) of the CaCO3 rain flux reaching the sediment surface dissolves and is returned to the ocean.

“most ($70À80%) of the CaCO 3 rain flux reaching the sediment surface dissolves and is returned to the ocean anyway”

Excess of carbonate weathering on land over deep-sea burial results in a further net transformation of 14% of CO2 emissions (261 ppm) into HCO3- on a timescale of ~8.3 ka.

“Excess of carbonate weathering on land over deep-sea burial results in a further net transformation of 14% of CO 2 emissions (261 ppm) into HCO 3 À (''terrestrial CaCO 3 neutralization'') on a timescale of $8.3 ka.”

The calibrated GENIE-1 model predicts a mean surface sediment CaCO3 content of 32.5 wt%, close to the observed value of 34.8 wt%.

“The resulting calibrated model predicts a mean surface sediment content (32.5 wt%) close to the observed value (34.8 wt%)”

• Also stated in: § 4. Present-Day Global Carbonate Cycling · plain

  “Global mean surface sediment CaCO 3 content is now 32.5 wt% and much closer to the observational-based estimate of 34.8 wt%”

An additional 12% of CO2 emissions (223 ppm) are sequestered as bicarbonate ions by reaction with deep-sea carbonates on a timescale of ~1.7 ka.

“An additional 12% of CO 2 emissions (223 ppm) are sequestered as bicarbonate ions (HCO 3 À ) by reaction with deep-sea carbonates (''seafloor CaCO 3 neutralization'') on a timescale of $1.7 ka.”

In the absence of any interaction between ocean and sediments, an equilibrium partitioning of CO2 is reached within ~1000 years of emissions ceasing, with 34% (645 ppm) remaining in the atmosphere out of a total fossil fuel burn of 4173 PgC.

“In the absence of any interaction between ocean and sediments, an equilibrium partitioning of CO 2 is reached within $1000 years of emissions ceasing, with 34% (645 ppm) remaining in the atmosphere out of a total fossil fuel burn of 4173 PgC.”

In an open system where ocean carbonate chemistry adjusts through interaction with deep-sea sediments, the long-term (>1000 year) evolution of atmospheric CO2 is markedly different from the closed system.

“The long-term (>1000 year) evolution of atmospheric CO 2 (Figure 5b) is markedly different in an open system in which the carbonate chemistry of the ocean adjusts through interaction with carbonates deposited to deep-sea sediments.”

In the open system, the timescale of CO2 drawdown is much longer than in the closed system.

“The timescale of CO 2 drawdown is much longer”

Erosion and progressive net dissolution of previously carbonate-rich sediments occurs in the open system experiment.

“Erosion and progressive net dissolution of previously carbonate rich sediments occurs (Figures 5d and 6)”

Alternative configurations of SEDGEM (uniform detrital flux, GCM-simulated aeolian deposition, or ocean model grid for hydrostatic pressure) led to a poorer simulation of the wt% distribution of CaCO3 in the sediments.

“All combinations of these alternative assumptions led to a poorer simulation of the wt% distribution of CaCO 3 in the sediments (not shown).”

In the open system, eventual (assumed, year 50,000) anthropogenic CO2 storage by the ocean represents 92.4% of the total release.

“the ''eventual'' (assumed, year 50,000) anthropogenic CO 2 storage by the ocean now represents 92.4% of the total release.”

• AssumptionThe authors state this as an assumption: Year 50,000 is reasonably close to steady state.

Model-predicted global export production of CaCO3 is 1.26 PgC yr-1, close to the preferred estimate of 1.14 PgC yr-1 diagnosed by Jin et al. [2006].

“Model-predicted global export production of CaCO 3 is 1.26 PgC yr À1 is very close to the ''preferred estimate'' of 1.14 PgC yr À1 diagnosed by Jin et al. [2006] from global nutrient and alkalinity distributions.”

• cites Jin et al. [2006] sourceNot checked — no full text of the source

The model calibrated against ocean PO4 and ALK and sedimentary wt% CaCO3 predicts a distribution of carbonates markedly closer to sediment observations than the previous calibration with PO4 and ALK only.

“The distribution of carbonates in the surface sediments predicted by the GENIE-1 model calibrated against ocean PO 4 and ALK and sedimentary wt% CaCO 3 is shown in Figure 3, and is markedly closer to the sediment observations than with the previous calibration with PO 4 and ALK only.”

From year ~3000-3500 onward, CO2 sequestration by seafloor CaCO3 neutralization proceeds with a well-defined timescale of ~1.7 ka.

“from the year $3000À3500 onward, CO 2 sequestration by seafloor CaCO 3 neutralization proceeds with a well-defined timescale of $1.7 ka (Figure 7b).”

Because the rate of burial of CaCO3 is reduced, there is a net accumulation of the products of continental weathering in the ocean.

“because the rate of burial of CaCO 3 is reduced there is also a net accumulation of the products of continental weathering in the ocean.”

In the closed system configuration, the ocean is able to sequester a maximum of 66.3% of the total atmospheric release of 4173 PgC, storing it mainly in the form of bicarbonate ions.

“In this case, the ocean is able to sequester a maximum of 66.3% of the total atmospheric release of 4173 PgC, storing it mainly in the form of bicarbonate ions (HCO 3 À ) (Figure 5c).”

The model's difference between ambient [CO3 2-] and that required for thermodynamic saturation with respect to calcite is generally ~10-20 mmol kg-1 higher than estimated from marine geochemical observations.

“That the difference between ambient [CO 3 2À ] and that required for thermodynamic saturation with respect to calcite (DCO 3 2À ) in the model is generally $10-20 mmol kg À1 higher than we estimate from marine geochemical observations”

Once emissions have ceased, it takes approximately 1000 years for fossil fuel CO2 to become fully mixed throughout the ocean.

“once emissions have ceased, it takes $1000 years for fossil fuel CO 2 to become fully mixed throughout the ocean.”

Across the four experiments, year 50,000 CO2 is virtually independent of the model configuration.

“It is notable that across the four experiments, year 50,000 CO 2 is virtually independent of the model configuration”

The estimated timescale of terrestrial CaCO3 neutralization is 8.3 ka.

“Our estimated timescale of 8.3 ka (Figure 7c)”

The effect of subgridding the SEDGEM sediments module at 72x72 and consequential subtle redistribution of CaCO3 preservation has relatively little effect on the trajectory of atmospheric CO2 compared to using the same parameter set but with SEDGEM at 36x36.

“The results (Table 1) demonstrate that the effect of subgridding the SEDGEM sediments module at 72 Â 72 and consequential subtle redistribution of CaCO 3 preservation (compare Figure 3d with Figure 4e) has relatively little affect on the trajectory of atmospheric CO 2 compared to using the same parameter set but with SEDGEM at 36 Â 36.”

Even the 41 ppm year 10,000 difference between PAC_8 and PA_8 predictions is small compared to the overall 365 ppm importance of ocean-sediment interactions.

“However, even the 41 ppm year 10,000 difference between ''PAC_8'' and ''PA_8'' predictions is small compared to overall 365 ppm importance of oceansediment interactions (Figure 5).”

The inclusion of a deep-sea sediment module in GENIE-1 has allowed innovative use of available global sedimentary data via an ensemble Kalman filter.

“The inclusion of a deep-sea sediment module in GENIE-1 has allowed us to make innovative use of available global sedimentary data via an ensemble Kalman filter.”

The calibrated model successfully captures most of the observed features of the global distribution of wt% CaCO3 in marine sediments.

“The calibrated model successfully captures most of the observed features of the global distribution of wt% CaCO 3 in marine sediments”

The results demonstrate that sediment composition data provides an improved constraint on modern marine carbon cycling compared to assimilation of ocean geochemical observations alone.

“our results demonstrate that it provides an improved constraint on the modern marine carbon cycling compared to assimilation of ocean geochemical observations alone.”

The calibrated representation of atmosphere-ocean-sediment carbon cycling within a 3-D ocean circulation based climate model achieves >1000 years per (2.4 GHz) CPU hour on a desktop PC.

“Our calibrated representation of atmosphere-oceansediment carbon cycling within a 3-D ocean circulation based climate model achieves >1000 years per (2.4 GHz) CPU hour on a desktop PC.”

The adjustment of ocean chemistry to rectify imbalances between weathering and carbonate burial is characterized by a much longer timescale of 8.3 ka.

“In contrast, the adjustment of ocean chemistry to rectify imbalances between weathering and carbonate burial is characterized by a much longer timescale of 8.3 ka.”

The 0.121 PgC yr-1 estimate for the deep-sea carbonate sink is in line with previous model and data-based estimates.

“our 0.121 PgC yr À1 estimate for the deep-sea carbonate sink is in line with previous model and data-based estimates.”

Decreased particulate organic carbon export due to surface warming and increased ocean stratification weakens seafloor CaCO3 neutralization.

“We have also carried out the first explicit analysis of the coupled atmosphere-ocean-sediment plus climate system and found that decreased particulate organic carbon export due to surface warming and increased ocean stratification weakens seafloor CaCO 3 neutralization.”

Deep-sea sediments are fundamental to how atmospheric CO2 concentration is regulated on timescales of thousands to tens of thousands of years.

“The sediments of the deep-sea are fundamental to how the concentration of CO 2 in the atmosphere is regulated on timescales of thousands to tens of thousands of years (see Ridgwell and Zeebe [2005] for a review).”

• cites Ridgwell and Zeebe [2005] sourceNot checked — no full text of the source

By inducing a change in the dissolution flux of solutes out of the sediments, oceanic concentrations of carbonate ions and silicic acid are buffered against perturbation.

“By inducing a change in the dissolution flux of solutes out of the sediments, the oceanic concentrations of carbonate ions (CO 3 2À ) and silicic acid (H 4 SiO 4 ) are buffered against perturbation.”

• AssumptionRests on an assumption: The rate of dissolution of biogenic minerals deposited to marine sediments, particularly CaCO3 and opal, is responsive to changes in the overlying ocean geochemistry.

The reduction in sequestration of fossil fuel CO2 resulting from enhanced seafloor CaCO3 preservation represents a new positive feedback on millennial-scale climate change.

“The resulting reduction in the sequestration of fossil fuel CO 2 represents a new positive feedback on millennial-scale climate change.”

Sediment property data provides a means of validating models of modern ocean carbon cycling.

“Sediment property data thus provides a means of validating models of modern ocean carbon cycling.”

• AssumptionRests on an assumption: Observed distributions of surface sedimentary properties, particularly CaCO3 and opal, reflect ecological patterns among marine plankton at the ocean surface as well as the depth of the seafloor and geochemistry of the abyssal ocean.

The preservation of CaCO3 deposited to deep-sea sediments represents the loss of dissolved inorganic carbon and Ca2+ and thus alkalinity from the ocean.

“The preservation of CaCO 3 deposited to deep-sea sediments represents the loss of dissolved inorganic carbon (DIC) and Ca 2+ and thus alkalinity (ALK) from the ocean,”

Because two moles of HCO3- are associated with each mole of CaCO3 weathered or precipitated, the global sedimentary sink of carbonate must be equal to half the HCO3- supply flux at steady state, some 12-20 Tmol CaCO3 yr-1 (=0.14-0.23 PgC yr-1).

“Because two moles of HCO 3 À are associated with each mole of CaCO 3 weathered or precipitated, the global sedimentary sink of carbonate must be equal to half the HCO 3 À supply flux at steady state, some 12À20 Tmol CaCO 3 yr À1 (= 0.14 À 0.23 PgC yr À1 ).”

Current estimates of modern shallow water CaCO3 burial place it approximately equal to the deep ocean.

“Although current estimates of modern shallow water CaCO 3 burial place it approximately equal to the deep ocean [Milliman and Droxler, 1996;Vecsei, 2004], there is substantial uncertainty in this.”

• cites Milliman and Droxler, 1996Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Vecsei, 2004 sourceNot checked — no full text of the source

The dynamical response of a closed system can be very different from that of an open system where there is a continual throughput.

“The dynamical response of a closed system can be very different from that of an ''open'' system where there is a continual throughput (i.e., both sources and sinks of a dissolved species to the ocean) [Hotinski et al., 2000].”

• cites Hotinski et al., 2000 sourceNot checked — no full text of the source

Dissolution driven by organic carbon metabolism in the sediments is slightly overrepresented in the model.

“suggests that dissolution driven by organic carbon metabolism in the sediments is slightly overrepresented.”

There is substantial uncertainty in current estimates of modern shallow water CaCO3 burial.

“there is substantial uncertainty in this.”

• cites Milliman and Droxler, 1996Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Vecsei, 2004 sourceNot checked — no full text of the source

Assimilation of sediment composition data has provided constraints on the operation of marine carbon cycling that ocean geochemical observations alone could not provide, at least in the context of the resolution and parameterizations of the model.

“Thus assimilation of sediment composition data has provided constraints on the operation of marine carbon cycling that ocean geochemical observations alone could not provide, at least in the context of the resolution and parameterizations of our model.”

The observational-based estimate of global mean surface sediment CaCO3 content (34.8 wt%) is probably an overestimate because data gaps are predominantly located in carbonate-poor regions.

“although this data-based mean is probably something of an overestimate because the data gaps are predominantly located in carbonate poor regions.”

Carbonate compensation, the buffering of ocean chemistry and atmospheric CO2 by changes in CaCO3 preservation in deep-sea sediments, is a key regulatory mechanism in the modern global carbon cycle on timescales of 5-10 ka.

“The overall buffering of ocean chemistry and atmospheric CO 2 (Figure 5b) by changes induced in the preservation of CaCO 3 in deep-sea sediments is known as ''carbonate compensation'' [Broecker and Peng, 1987] and represents a key regulatory mechanism in the modern global carbon cycle on timescales of 5À10 ka.”

• cites Broecker and Peng, 1987 sourceNot checked — no full text of the source

The two effects of sediment erosion/dissolution and reduced CaCO3 burial are responsible for the additional atmospheric CO2 drawdown and sequestration by converting CO2 to HCO3- compared to ocean invasion alone.

“These two effects are responsible for the additional atmospheric CO 2 drawdown and sequestration by converting CO 2 !HCO 3 À compared to by ocean invasion alone.”

At the global scale, the large-scale circulation of the ocean and water mass geochemistry dominates the CaCO3 content of deep-sea sediments.

“In contrast, at the global scale, the large-scale circulation of the ocean and water mass geochemistry dominates [Ridgwell and Zeebe, 2005].”

• cites Ridgwell and Zeebe, 2005 sourceNot checked — no full text of the source

On a scale smaller than ocean basins (<1000 km), the dominant control on the CaCO3 content of deep-sea sediments is hydrostatic pressure.

“On a scale rather smaller than ocean basins (<1000 km), the dominant control on the CaCO 3 content of deep-sea sediments is hydrostatic pressure [Archer, 1996].”

• cites Archer, 1996 sourceNot checked — no full text of the source

The consequence of this reduction in sedimentary CaCO3 erosion is a weaker buffering of atmospheric CO2.

“The consequence of this reduction in sedimen-tary CaCO 3 erosion is a weaker buffering of atmospheric CO 2 .”

The uniform final CO2 value represents the titration of fossil fuel CO2 with CaCO3 to the point at which the pre-industrial saturation state of the ocean is restored.

“The uniform final CO 2 value represents the titration of fossil fuel CO 2 with CaCO 3 to the point at which the pre-industrial saturation state of the ocean is restored.”

This is the first time that sediment composition data has been used in inverse modeling.

“This is the first time that sediment composition data has been used in inverse modeling”

The main reason for the differences in timescale estimates is the isolation of the component of ocean chemistry and atmospheric CO2 changes due to seafloor CaCO3 neutralization.

“However, we believe that the main reason for the differences in timescale estimates is due to our isolation of the component of ocean chemistry and atmospheric CO 2 changes due to seafloor CaCO 3 neutralization.”

The effect of CO2-calcification feedback has not previously been considered in long-term (multimillennial scale) analysis of the fate of fossil fuel CO2.

“the effect of CO 2 -calcification feedback has not previously been considered in long-term (multimillennial scale) analysis of the fate of fossil fuel CO 2 .”

The subsequent fate of this residual CO2 would be drawdown by silicate weathering feedback.

“The subsequent fate of this residual CO 2 would be drawdown by silicate weathering feedback [Ridgwell and Zeebe, 2005].”

• cites Ridgwell and Zeebe, 2005 sourceNot checked — no full text of the source

The analysis has not included a possible dependence of carbonate rain fluxes on the efficiency of organic matter transport into the deep ocean, which would reduce the predicted impact of CO2-calcification feedback on the sediments.

“One caveat to this analysis it that we have not included a possible dependence of carbonate rain fluxes on the efficiency of organic matter transport into the deep ocean [Armstrong et al., 2002;Klaas and Archer, 2002], which would reduce our predicted impact of CO 2calcification feedback on the sediments [Ridgwell, 2003].”

• cites Armstrong et al., 2002Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Klaas and Archer, 2002 sourceNot checked — no full text of the source
• cites Ridgwell, 2003 sourceNot checked — no full text of the source

This represents a new positive feedback on millennial-scale future climate change.

“This represents a new positive feedback on millennial-scale future climate change.”