New Global Carbon Data Revives the "Missing Sink" Problem

Once again, it is about our misunderstanding of the biosphere

On March 21, 2025, Science published a study “Recent gains in global terrestrial carbon stocks are mostly stored in nonliving pools” by Bar-On et al. showing that between 1992 and 2019, plant biomass on land has remained nearly constant, with a stock change of only about 1 GtC. At the same time, a mass conservation analysis—accounting for atmospheric carbon accumulation, fossil fuel combustion, and CO₂ dissolution in the ocean—indicates that about 35 GtC should have been sequestered somewhere. And, for more than a decade, scientists believed this missing carbon was stored in trees.

This crucial result matters for several reasons. The most immediate is that dynamic vegetation models—the components of global climate models that represent the biosphere—profoundly misrepresent ecosystem responses to carbon perturbations. A deeper and more important issue is that our understanding of the biosphere remains primitive, shaped by concepts (like that of the limiting nutrients) that are inapplicable to natural ecosystems. We have an unprecedented wealth of data at our disposal, yet we remain clueless about the laws of the biosphere that these data could reveal—if our civilization’s collective mind weren’t hardwired to ignore them.

To set the stage, it’s worth noting that for a long time, scientists were reluctant to acknowledge the biosphere’s role in absorbing about a third of accumulating atmospheric carbon. This uncertainty was known as the "missing sink" problem—where does all this carbon go?

Let me quote our 2020 paper:

Modern civilization consumes energy from two main sources – fossil fuels and biomass. Fossil fuel burning emits to the atmosphere about 9 GtC/yr, to which at least 2 GtC/yr is added from the destruction of the organic stores of the biosphere (soil degradation and deforestation) (data are given to the accuracy of 0.5 GtC/yr (Friedlingstein et al., 2019)). About one-half of these emissions, 5 GtC/yr, accumulates in the atmosphere in the form of the greenhouse gas CO₂, such that its atmospheric concentration is on the rise.

As of today, the atmospheric CO₂ concentration has grown by one third as compared to its pre-industrial era and got out of equilibrium with the concentration of carbon dioxide dissolved in the ocean. Tending to restore the equilibrium, the ocean consumes about 2 GtC/yr from the atmosphere in inorganic form. The fate of the remaining 4 GtC/yr in the carbon balance (the so-called “missing sink”) had long remained enigmatic, even though it might seem natural to immediately assign a role to the biosphere. The global biota synthesizes organic matter at a rate of about 100 GtC/yr and, in a steady state, decomposes it at an equal rate. This exceeds the rate of anthropogenic carbon emissions by one order of magnitude.

Global carbon budget in 2019 according to Friedlingstein et al. 2019. Note that soils represent a major reservoir of organic carbon, surpassing atmospheric carbon stock by several times. Additionally, the ocean contains nearly as much dissolved organic carbon as the total amount of CO₂ in the atmosphere. The "missing sink" we are discussing corresponds to the difference of +1.7 GtC/year—arising from the -1.5 GtC/year source due to land-use change and the +3.2 GtC/year “land uptake.” This missing sink accounts for roughly one-third of the carbon accumulating in the atmosphere (4.9 GtC/year). 1 GtC (gigaton of carbon) = one billion metric tons of carbon. The gross “missing sink” of +3.2 GtC/year accounts for roughly one-third of anthropogenic carbon emissions (9.5 GtC/year).

The scientific community has long been reluctant to recognize and admit the existence of a biotic carbon sink. That was due to the prevailing concept that the biota does not regulate concentrations of life-important substances but, instead, is limited by them. As far as the concentrations of nitrogen and phosphorus – the presumed limiting biogens – do not change, the ecologists believed that the biosphere productivity could not grow in response to the increasing concentration of atmospheric CO₂ (Popkin, 2015).

In other words, what is now widely known as CO₂ fertilization—often cited as the explanation for the biotic carbon sink on land (despite failing to actually explain it, as discussed below)—was not the scientists’ first choice. In fact, they resisted this idea for quite some time, sometimes passionately. For example, Prof. Stephen Pacala, then director of the Princeton Environmental Institute, called the missing sink “the most vexing problem in global change science”, arguing that any additional plant biomass is almost immediately decomposed, preventing the emergence of a net sink even if photosynthesis rates increase with rising atmospheric CO₂.

It was almost at the level of Anton Chekhov’s “This cannot be, because this can never be.”

However, the law of mass conservation was hard to dispute with, and the notion that "we don’t know where the carbon goes" posed a serious challenge to the credibility of climate science as a whole. This could not continue. Eventually, a decision was taken; the missing carbon must be stored in trees. Dynamic vegetation models were subsequently developed, reinforcing the idea that, indeed, the excessive carbon is stored in tree biomass.

Let me quote our paper again:

However, forest tree biomass cannot grow infinitely (Hubau et al., 2020). The biotic sink will persist if, after the tree death, the wood is not fully decomposed but is partially deposited in soil in the form of long-lived organic substances. Recent studies revealed that the turnover time of organic compounds in soil is mostly determined not by their chemical composition but by functioning of the entire ecological community, plants and soil biota included (Schmidt et al., 2011; Gross, Harrison, 2019; Kuznetsova et al., 2019).

In other words, it is only possible to assert that biomass is stored in trees for a limited period of time, while the uncertainty in global biomass measurements obscures the actual dynamics. However, as time progresses, the amount of carbon that must have been stored in the "missing sink" continues to increase. As a result, any mismatch between the actual stock and the expected storage becomes more apparent as time goes. The study by Bar-On et al. uses twenty-seven years of data and demonstrates that it is now possible to say with certainty that tree biomass has not increased at the rate predicted by post hoc models of CO₂ fertilization (again, the latter previously considered impossible).

The strength of their approach is that they constrain the net sink by plant biomass dynamics, rather than evaluating separately the gross sources and sinks, as conventionally done in the global carbon budget analyses (see the figure above).

One might get an incorrect impression from the abstract of Bar-On et al. (see also “Why it is important to read scientific papers beyond their abstracts, especially when it is about the role of CO₂ in climate”) suggesting that human factors such as wood harvest or even landfill waste could account for a major portion of the reincarnated missing sink. It would be more accurate to mention in the abstract that these “gains from anthropogenic activity” potentially contribute to only about one-tenth of the missing sink. Mentioning these minor potential contributors makes the abstract read less controversial, as if the explanations are in place already. The actual message of the paper is, in my opinion, “the missing sink is back” (and we have no clues).

Indeed, the results of Bar-On et al. expose the lack of the ecological and climate community’s understanding of what is happening in the biosphere. It’s not that the data is lacking (I’ve previously discussed in How much wild nature do we need? that most data anyway comes from disturbed ecosystems that don’t function the same way as undisturbed ecosystems, the latter maintaining environmental homeostasis). Rather, it reflects a conceptual cluelessness, a hardwired neglect of the natural ecosystems’ capacity to stabilize the environment.

This conceptual gap is also related to what can be called Odum's paradox. Eugene Odum, the prominent American ecologist, believed that ecological succession culminates in ecosystem's maximum control of the environment (Odum, 1969). However, if, as was simultaneously believed, the ecosystem operates on the basis of closed matter cycles, its environmental impact (and, hence, environmental control) would be zero by definition. (That was one of the reasons why ecologists resisted the idea of a biotic carbon sink.)

Slide from from Makarieva & Nefiodov (2021)

The concept of biotic regulation introduced the idea of non-random openness in matter cycles to compensate for environmental disturbances (Gorshkov, 1995). In the absence of disturbances, the matter cycles in the ecosystem are closed. When disturbances occur, the ecosystem opens its cycles to compensate for them. We explained:

If we accept the concept of the biotic regulation of the environment as the main principle of life organization, then the existence of a biotic sink of atmospheric carbon becomes self-evident. The natural biota should react to an environmental disturbance – anthropogenic CO₂ emissions – according to the Le Chatelier’s principle (Gorshkov, 1995). It should counteract CO₂ accumulation and remove the excessive carbon from the atmosphere by transforming it into an inert organic form. To explain the [gross] missing sink of 4 GtC/yr it is sufficient that global photosynthesis exceeds global decomposition of organic matter by four per cent. Carbon is the main element used by life. Even if СО₂ were not a greenhouse gas, the biota should have been compensating for the modern deviation of its atmospheric concentration from the optimal value. … the biotic regulation concept predicted that, at constant nitrogen and phosphorus, the stabilizing biotic response should take the form of an excessive synthesis of carbohydrates that do not contain either nitrogen or phosphorus (Gorshkov, 1986, p. 86).

So, where do we stand now? Clearly, CO₂ fertilization cannot explain the missing sink on land. Such a sink must be a community-level response: heterotrophs must refrain from consuming the excess organic matter synthesized by plants under warming conditions. Only then can a net sink emerge.

Furthermore, a net sink could be ensured by heterotrophs alone, even if photosynthesis rates remain constant. If they reduce consumption—i.e., slow the decomposition of plant-synthesized matter—a net sink will form! Why isn’t this possibility considered? Because, as humans, we tend to judge by our own behaviors and struggle to imagine systems maintained by natural selection, where some elements “sacrifice” their own needs (i.e., reduce consumption) for the “common good” (environmental stability). However, if we accept that such systems are possible, we open our minds to considering a far richer repertoire of life’s compensatory responses to atmospheric carbon accumulation.

The soil carbon stock was not evaluated by Bar-On et al. As we noted,

Lack of an adequate account of soil carbon dynamics in undisturbed forests and other intact ecosystems results in a potential underestimate of the negative impact on the carbon cycle associated with ecosystem transition from the undisturbed state to a disturbed one, in particular, by a perturbation of the water regime (Kittler et al., 2017; Sheil et al., 2019; Mayer et al., 2020). If the undisturbed Russian forests ensure a soil sink of atmospheric carbon, then turning a given area of undisturbed forest into a disturbed, exploited plot with cultivated trees that have a higher net primary productivity, will lead to an increase in carbon emissions. The stabilizing impact (carbon absorption by the intact forest) will be replaced by a destabilizing one (loss of soil carbon in the disturbed ecosystem).

Analyses of chronosequences in natural undisturbed forests reveal that such forests, at constant plant biomass, are able to ensure a soil sink for atmospheric carbon at a rate of about 1% of primary productivity or about 5 gC m^–2 yr^–1 in a boreal ecosystem (Wardle et al. 2012). Meanwhile modern increase of atmospheric CO₂ appears to enhance the stabilizing response of undisturbed ecosystems by at least one order of magnitude, up to 50 gC m^–2 yr^–1 and more (Zhou et al., 2006; Kittler et al., 2017). Long-term year-round monitoring is required to quantify this response reliably (Kittler et al., 2017).

In other words, if the primary carbon reservoir is in the soils beneath boreal and temperate forests, their role in global climate stability is far greater than currently acknowledged.

Beyond forests, there is another neglected aspect of the missing sink problem. Victor Gorshkov argued that the ocean—where primary producers (phytoplankton) are too small to be easily “harvested” (unlike trees), making them less disturbed than terrestrial biota—should be the dominant biospheric compartment stabilizing the carbon cycle. In a talk, I discussed how this possibility was dismissed because ocean primary productivity was assumed to be limited by factors other than carbon (haven’t we heard this before?). But if CO₂ fertilization fails to explain a biotic carbon sink on land, why should it be necessary to explain one in the ocean?

Slide from from Makarieva & Nefiodov (2021). Note that the oceanic sink in the conventional global carbon budget, 2.5 GtC/year in the above figure of Friedlingstein et al. 2019, represents an inorganic carbon sink.

Given the current uncertainties, it remains entirely possible that the ocean plays a considerable role as a biotic carbon sink in the form of dissolved organic carbon—one of the major reservoirs of organic carbon alongside soils. As some marine scientists attempted to warn some time ago, “an entire revision of the carbon cycle will be necessary.”

In summary, the global carbon budget is in serious turmoil, even if it is not loudly announced. With CO₂ fertilization failing as an explanation, even landfill waste is now being considered as a potential factor in resolving the missing carbon puzzle. Hopefully, the necessary revision will take place at a higher conceptual level—one that fully accounts for the stabilizing role of natural ecosystems in regulating the environment and climate—before we disrupt natural ecosystems beyond the threshold of self-recovery.