Sequestration by the biological carbon pump: Do we really know what we are talking about?
André Visser
Abstract
Marine scientists have an alarming habit of attributing carbon sequestration to processes that really only relate to the turnover of long-lived natural carbon pools. This conflicts with the general understanding that carbon sequestration implies a potential offset to anthropogenic emissions. Without a critical re-examination of the terminology used, we run the risk of seriously damaging an informed political debate on climate action from carbon trading to marine CO2 removal strategies, as well as undermining our credibility. There is little evidence to suggest that the biological carbon pump (BCP) has directly offset any anthropogenic carbon emissions. Yet, we as a community talk persistently about how the BCP sequesters carbon. We often use this terminology to promote our work to a broader audience—funding agencies, nongovernmental organizations, government agencies, and politicians—and while our interpretation may be nuanced, to the rest of the world, carbon sequestration is synonymous with an offset of carbon emissions. Overstating or misrepresenting the role of the BCP in future climate regulation can have serious consequences for policies and implementation of climate mitigation actions. At the heart of this miscommunication lies the rather cavalier manner in which we use the term sequestration—it may mean one thing in ocean sciences, but something entirely different elsewhere. Here I propose a distinction that I hope will clarify how we report this fundamental BCP metric lest we confuse not only climate policy but also ourselves. It is well understood that the BCP (Fig. 1a in Box 1) is very nearly in equilibrium (Volk and Hoffert 1985; Lévy et al. 2013; Hain et al. 2014; DeVries 2022). Primary production in the surface ocean produces organic carbon, a fraction of which is exported to depth where it is respired by microbes and metazoans into dissolved inorganic carbon (DIC), is transported by the meridional overturning circulation to eventually be outgassed back to the atmosphere a few hundred years later. The BCP (Volk and Hoffert 1985), that is, the combined process of production, export, remineralization, circulation, and outgassing, holds in place a mass of about 2000 PgC (Fig. 2 in Box 1) as respired DIC below the surface mixed layer (Sarmiento and Gruber 2006; Boyd et al. 2019; Carter et al. 2021; DeVries 2022)—a reservoir large in comparison to the mass of carbon in the atmosphere (both preindustrial and current levels) but small compared to the total amount of DIC stored in the ocean. The biological carbon pump (BCP) is imbedded within the global meridional overturning circulation (MOC) (Fig. 1a) into which CO2 (i.e., dissolved inorganic carbon [DIC] via the oceans' carbonate buffering system) is injected by the oxygenation of organic matter (green circles) either by microbial remineralization of sinking detrital material, the respiration by resident and vertically migrating metazoans, or the physical subduction of organic carbon (orange arrows). This also includes organic matter that reaches the seabed and is subsequently respired by benthic communities. The MOC that transports these dissolved compounds is driven in part by deep-water formation but is also critically dependent on mixing (Wunsch and Ferrari 2004)—a large fraction of which is thought to occur in the deep oceans close to the seabed. Upwelling and mixing close the circulation and transport DIC back to the surface ocean where it is outgassed to the atmosphere. The time scale for the overturning circulation is about 1200 yr (Koeve et al. 2015). The BCP together with the solubility carbon pump (SCP) leads to a vertical gradient in DIC (Fig. 1b). While primarily driven by temperature and CO2 solubility in deep-water formations, the BCP enhances the SCP by reducing the saturation of surface waters. Current estimates of ocean carbon reservoirs (Fig. 2) indicate a total mass of DIC of C DIC ≈ 37,200 ± 200 PgC (Keppler et al. 2020). Subsets of the total DIC pool can be variously attributed to (1) C ant ≈ 185 ± 20 PgC as uptake from anthropogenic sources since the industrial revolution (25 PgC in the surface ocean and 160 PgC at depth), (2) DIC produced by the respiration of organic matter totaling a mass of C res ≈ 2000 ± 300 PgC, and (3) that originating from remineralized biogenic carbonates C bcarb ≈ 830 ± 90 PgC. In addition (4) a reservoir of about C DOC ≈ 700 ± 20 PgC is stored as dissolved organic carbon (DOC) and (5) about C sed ≈ 2200 ± 20 PgC stored as various forms of carbon in the upper layers of marine sediments (Atwood et al. 2020). Burial in marine sediments is often thought of as an important carbon sink. Discounting the biological activity of benthic communities (both epi- and endo-) which is accounted for in the BCP reservoir estimates, carbon burial is quite slow. Indeed, very little of the organic C that leaves the surface ocean makes it to the seafloor, and only a small fraction is incorporated into sediments (a fraction of a Pg yr−1). Further, C burial in sediments generally operates on a much longer time scale (100,000 yr) than the MOC, and a balanced accounting would need to also consider similar long time scale processes such as weathering, calcium carbonate dynamics, seafloor subduction, and volcanism. One other reservoir of note is (6) C bio ≈ 5 ± 1 PgC representing the living biomass in the oceans (Bar-On and Milo 2019). This includes everything from bacteria to whales. The total mass of biogenic carbon stored in the ocean ( C res + C bcarb + C DOC ) is in excess of 3500 PgC and far exceeds the living biomass in the oceans. In this light, the biogenic carbon reservoirs in the ocean can be seen as legacy carbon laid down by generations of marine biota spanning back hundreds of years. On a technical note, the original estimates for (2) and (3) C res * ≈ 1300 ± 230 PgC and C bcarb * ≈ 540 ± 60 PgC, respectively, come from inverse modeling of biogeochemical observations (Carter et al. 2021) and are predicated on first passage time (Primeau 2005) that is, the mean time taken from injection of DIC at some depth in the ocean to its first transit to the surface mixed layer. As pointed out (DeVries 2022), exchange of CO2 between the ocean mixed layer and the atmosphere is sluggish, and a considerable fraction of DIC is subducted before equilibrium can be established. That is, residence times of DIC is generally longer (by a factor of about 1.6) than first passage times, and are reflected in the values of C res and C bcarb reported here and are consistent with the estimate C res ≈ 1800 PgC accounting for atmosphere ocean disequilibrium (Nowicki et al. 2024). Over geologic time, the size of these reservoirs wax and wane. Indeed, the BCP appears to have been responsible in large part for the partitioning of carbon between the ocean and atmosphere over the glacial—interglacial cycles of the last 400,000 yr (Sigman and Boyle 2000; Hain et al. 2014). More recently, over much of the Holocene (the past 10,000 yr or so), the BCP reservoirs have been in near equilibrium with fluxes in balancing fluxes out, as witnessed by the near uniform atmospheric CO2 levels during that period (DeVries 2022). It was only since the 1950s with the great acceleration of the Anthropocene (Steffen et al. 2011) that we have entered a transient phase as Earth's carbon cycle adjusts to human-induced perturbations. Fossil fuel emissions are perhaps the most evident, but there are other perturbations relevant for oceanic carbon cycles, notably the development of industrial-scale fisheries and an increased global supply of fixed nitrogen through the Haber-Bosch process. Superimposed on this are climate-induced responses of the ocean system—everything from changing community structures to respiration rates to ocean circulation patterns. While all of these topics deserve attention, with regard to climate and the ocean carbon cycle, the overriding metric has to be the expansion or contraction of oceanic carbon reservoirs. Simply put, are the bubbles in Fig. 2 in Box 1 getting bigger or smaller? Carbon reservoirs can be defined in terms of their province and provenance (where it is and how it got there). Mathematically speaking, they are subsets of all the carbon in the Earth System. Definitions need not be unique; a reservoir can be a subset of another, and reservoirs can intersect. It is no accident that Fig. 2 in Box 1 resembles a Venn diagram used to illustrate logical relationships in set theory. Seen in this light, sequestered carbon is simply all carbon in the Earth System that is not in the atmosphere, a definition that meshes with that of the IPCC. Any given reservoir, irrespective of the details of its dynamics, can be defined in terms of residence time weighted influx and efflux functions, q θ t and p θ t . Specifically q θ t dθ is the influx of material at time t that will remain in the reservoir for a period in the range θ θ + dθ . Similarly, p θ t dθ is the efflux of material at time t that has had a residence time in the reservoir in the range θ θ + dθ . These functions, although in general different, are not independent. Conservation of mass within the reservoir means that the efflux of material at time t that had residence time θ must be equal to the influx with the same residence time that entered the reservoir at time t − θ . That is p θ t = q θ t − θ . By symmetry, this also means that q θ t = p θ t + θ . These are general conditions for any mass conserving system and can be arrived at via rigorous mathematical approach in terms of the McKendrick-von Foerster equation (Calabrese and Porporato 2015). These conditions are not dependent on the internal dynamic of the reservoir, whether it is composed of multiple interconnected compartments, and/or the dynamics are time dependent (Rasmussen et al. 2016; Sierra et al. 2017). Such facets will change the form of q and p but not their dependence. The total influx at time t is Q t = ∫ 0 ∞ q θ t dt likewise P t = ∫ 0 ∞ p θ t dt is the total efflux at time t . Noting that p θ t contains all the information on residence times prior to t , while q θ t describes residence times after t , the total mass of the reservoir can also be derived as C t = ∫ 0 ∞ p θ t θ dθ = ∫ 0 ∞ q θ t − θ θ dθ . Finally, the mean residence time can be formulated as T t = ∫ 0 ∞ p θ t θ / P t dθ , or more succinctly as C t = P t T t . This relationship is quite general and describes in a rigorous sense how the mass of a reservoir changes in terms of total efflux P t and mean residence time T t . With respect to the main ocean carbon reservoirs, we can infer that up until recently (specifically, time scale of disturbance 75 yr that is less than their residence times 200–800 yr) they were in near equilibrium. That is P ≈ Q from which Eq. (1) follows. Since the industrial revolution, anthropogenic emissions (fossil fuels plus net land use) have released 470 PgC of which about 280 PgC has accumulated in the atmosphere (Friedlingstein et al. 2020), about 160 PgC has been absorbed by the oceans below the mixed layer depth (Sabine et al. 2004; DeVries 2014; Davila et al. 2022) while about 25 PgC remains in the surface ocean (Davila et al. 2022). Within the bounds of uncertainty, the oceanic uptake can be entirely attributed to physics and chemistry—the response of the surface ocean to increased pCO2 in the atmosphere, carbonate chemistry and the meridional overturning circulation (Sabine et al. 2004; Gruber et al. 2019; Davila et al. 2022). While the BCP plays an indirect role in this by maintaining undersaturated surface waters and so strengthening the solubility pump, its direct role in offsetting anthropogenic carbon (i.e., increasing C res reservoir) appears minimal. The only estimate I can find of this direct role, is about 0.1 ± 0.03 PgC yr−1 (Koeve et al. 2020) as a mean accumulation rate since the 1970s as evidenced by an increase in global oxygen utilization (Schmidtko et al. 2017). Oxygen utilization is the biogeochemical flipside of organic matter respiration and has been used previously to quantify the BCP (Anderson and Sarmiento 1995; Koeve et al. 2020; Wilson et al. 2022). It has been pointed out that oxygen utilization estimates are a more direct and reliable means of measuring variations in C res and hence the role of the BCP in climate regulation (Frenger et al. 2024) than our current reliance on particulate organic carbon (POC) export and flux attenuation estimates (Boyd et al. 2019). The oxygen utilization derived estimate of BCP sequestration rate reported by (Koeve et al. 2020) is small. It corresponds to about 1% of our current emissions. It is also wildly different from other reported “sequestration” rates (Guidi et al. 2015; Ricour et al. 2023; Berzaghi et al. 2025) which are orders of magnitude greater. These cannot be equated to sequestration rates in the sense of an offset—simple mass balance estimates with atmospheric CO2 levels or ocean oxygen content will reveal why—but are rather estimates of the carbon flux into the ocean passing some specified depth or time horizon (e.g., carbon exported below the permanent thermocline or below 2000-m depth, carbon with a residence time longer than 100 or 50 yr). These estimates only quantify the turnover fluxes into a specified carbon reservoir (some subset of C res ) and say nothing about whether these reservoirs are expanding or contracting. As pointed out by (Frenger et al. 2024) this is like trying to guess your bank balance by only looking at deposits with no consideration of withdrawals. A statement attributed to the IPPC in 2007 defined sequestered carbon as carbon stored out of contact with the atmosphere for more than 100 yr (Passow and Carlson 2012). This is a perfectly valid way to define a reservoir with relevance for the partitioning of carbon in the Earth System and its implications for climate. The act of sequestration though is a different matter. The long-standing practice in marine science is to define sequestration as a process that removes carbon from contact with the atmosphere for more than 100 yr (Guidi et al. 2015; Baker et al. 2022; Ricour et al. 2023). It seems logical given the IPCC definition above. Unfortunately, this definition in and of itself says nothing about the re-partitioning of carbon in the Earth System. A flux into a reservoir defined as “carbon that will remain in the ocean for more than 100 yr” does not preclude that this flux is simply replacing an equal amount of carbon that has already been in that reservoir for more than 100 yr and is now outgassing back to the atmosphere. The same argument can be applied to any geologically relevant time or depth horizon. In the same set of documents as that cited above, the IPCC also defines carbon sequestration as “the process of increasing the carbon content of a reservoir/pool other than the atmosphere” (IPCC AR4 2007 Glossary to WG2, https://www.ipcc.ch/site/assets/uploads/2018/02/ar4-wg2-app-1.pdf). The meaning here is clear and unambiguous; sequestration is synonymous with and offset of emissions. When speaking of the biological pump and carbon sequestration, we need to be much more precise. Strictly speaking, we should only use the term sequestration when we can demonstrate a net transfer into a relevant oceanic reservoir with potential offset of carbon emissions. This is in keeping with common usage and its original definition by the IPCC. Any flux measurements or estimates that cannot fulfill this criterion should come with a disclaimer to reflect that they are gross, rather than net flux estimates. Long storage times are intuitively important for carbon capture and storage. For instance, newly created carbon reservoirs—whether established through artificial means (various proposed marine CO2 removal schemes) or through global change (e.g., emerging novel ecosystems associated with sea-ice retreat in the Arctic)—are precisely where sequestration (i.e., offset) based on a relevant time horizon (e.g., 100 yr) makes sense. However, such time scales become unhelpful for long-lived natural carbon reservoirs. For instance, it is somewhat ironic that the surface ocean—an ocean province that has actually taken up a significant amount of anthropogenic emissions (≈ 25 Pg C; Davila et al. 2022) is generally discounted as a site for sequestration because of its rapid turnover time (< 1 yr). More importantly, as reservoirs mature, leakage becomes more important, and as their age approaches their residence time, the flux into the reservoir changes from sequestration (i.e., offset) to turnover (no offset): perhaps an important issue for gauging marine CO2 removal schemes? Any natural carbon reservoir that has existed longer than its associated residence time will be near equilibrium with flux in approximately balancing flux out. How fast it turns over is of diminishing importance for net sequestration. If we persist in current practices, we should distinguish between sequestration sensu turnover (Guidi et al. 2015; Ricour et al. 2023; Berzaghi et al. 2025) and sequestration sensu offset (Koeve et al. 2020). This distinction impacts directly on the potential valorization of BCP services (Berzaghi et al. 2025) where it could be surmised that turnover would rank a lot less valuable than offset. Understanding the oceans' role in climate really boils down to describing and predicting the expansion or contraction of the major carbon reservoirs of marine systems. Of particular interest are any systematic changes in processes that impact both the turnover flux and the residence time. As a case in point, predicting the expected response of the biological pump to climate change has focused largely on export flux Q exp , where best estimates indicate a ~ 10% reduction on a global scale (Cabré et al. 2015). However, the process that leads to this reduction—an increase in ocean in a more meridional overturning circulation (Frenger et al. 2024) meaning that down that is, longer residence indicate that increased residence times out over export flux so that the net is an increase in the size of the respired DIC reservoir in the range PgC by the of the (Koeve et al. 2020; Wilson et al. 2022), that is, the BCP may more carbon climate change a with a in increased oxygen The oceans significant reservoirs of biogenic carbon that have been laid down by generations of biota spanning back of years. It appears that during the past 10,000 these reservoirs were in near equilibrium with fluxes in balancing fluxes out, a that up until about 75 yr changes have in the Earth System since there are of legacy biogenic carbon in the oceans ( C res + C bcarb + C DOC ~ 3500 that are and to the atmosphere. Without and ecosystems to the the oceans will become net of biogenic carbon to the a already in with various for to fisheries over the past yr et al. and the of the of in the et al. 2022). Conservation and are for maintaining legacy carbon reservoirs as well as and but whether they offset fuel emissions remains When marine carbon and whether the BCP or other carbon is not only unhelpful but to an informed political is a term that a lot of in this and should be it to a net mass balance for a We should also be of the anthropogenic on natural cycles biogenic carbon. The of marine living when a in legacy carbon emissions that are understood but should be accounted This work was and by This work was by the and the and and were created or during this is not to this