A Major Climate Unknown: Natural Ecosystems

We do not even recognize the existence of a parallel universe: natural ecosystems. Yet understanding their role in climate stability is essential for any viable ecological strategy for our species.

Please allow me a philosophical prelude. The more factual narrative begins in the next section of this post.

I first met Victor Gorshkov, the founder of the biotic regulation concept, when I was twenty. Once, in early March, as we were walking, a bird crossed the road. I asked what bird it was. “You don’t know??? It’s a black thrush—an unbeaten singer, according to Kaigorodov.” After a pause, he added with quiet pride: “I learnt about all our birds as a schoolboy.”

In that moment, I lived through an extraordinary experience. The enormous value Victor, a prominent theoretical physicist, placed on knowing the birds resonated with me in an unexpected way. Suddenly, a new dimension of the world was unfolding in front of me—a mostly city-grown kid with only infrequent exposure to real wilderness, and rarely in spring. There were birds. Birds sing. Birds sing in spring.

Victor was also deeply interested in classical music and played the piano. With a portrait of Beethoven before him as he played, he often drew parallels between classical music and birdsong—always concluding that the birds were the superior artists, the ones who had inspired the great composers. For example, he told me, and I later verified, that the opening of Tchaikovsky’s Sleeping Beauty Waltz follows the song of the goldcrest.

The goldcrest (or “kinglet” in Russian), one of the smallest European birds, prefers tall spruce trees and can endure the boreal winter without migrating to warmer lands. I took this photo on a tiny island in the White Sea, where it was sitting unusually low and calm on a pine tree. Possibly it was a little king in disgrace.

My future scientific work became intertwined with my emotional and intellectual awakening, each reinforcing the other, as I entered the world of birds and the natural world more broadly. It felt almost too late to begin; we are most receptive to other living beings in our earliest years. But with Victor as my guide, and with some critical legacy to build upon, I managed, it seems, to catch the last train.

Robin singing — my own very unprofessional recording

In the spring of 2019, when Victor was already very ill, the following lines painfully rushed out from within, revealing how deep the connection had become:

Mute the birds; I cannot bear to hear them,
while he is locked in the hospital ward,
holding the line against death.
It hurts so much. Stop the spring!

In the spring of 2021, during Covid, as my stressed body hovered alarmingly in a kind of limbo, uncertain whether to recover or to begin shutting down, I would wake early and hear a black thrush singing in a small patch of trees near our house in the very urban city of St. Petersburg. That song, rising before the city awoke with its grudging noises, felt like a tender yet uncompromising call back toward life. (The breathing exercises advised by friends facilitated the recovery.)

The above is a feeble attempt to sketch the process of my growing into a new world of which I had been unaware. Indeed there are universes on the Earth that we are unnecessarily blind to—sometimes to our own misery. While our civilization has become highly specialized and sophisticated as a whole (to the point that Nate Hagens warns we will soon pass through a Great Simplification), this global sophistication has come at the expense of the more primitive environments and ways of living we each actually inhabit and hold. This leaves us blind to the larger picture of the Living Earth.

The Unknown Universe Where Nature Rules

In my appearance on Nate Hagens’ podcast, I made the following point:

We are calibrating climate models using the most disturbed ecosystems. It is as if we went to a hospital, observed only severely ill patients, and then defined an average human being based on their limited capacity to function. Such observations would reflect acute illness rather than normal physiology. This is not how a healthy organism operates, and it is not how intact forest ecosystems function. As a result, our understanding of forests is deeply distorted, and we largely ignore the very forests that play the strongest role in stabilizing the climate.

Let us now look at where these least disturbed and most important forest ecosystems are actually located.

https://glad.earthengine.app/view/intact-forests

An Intact Forest Landscape (IFL) is a seamless mosaic of forest and naturally treeless ecosystems within the zone of current forest extent, which exhibit no remotely detected signs of human activity or habitat fragmentation and is large enough to maintain all native biological diversity, including viable populations of wide-ranging species. Source: https://intactforests.org/

These forests covered about 11 million square kilometers in 2020, or roughly one fifth of global tree cover. One might argue that this is not a large area and ask why these ecosystems matter. First, by studying these ecosystems and their climate-stabilizing effects, we can better quantify the extent to which current environmental problems stem from the loss of natural ecosystems, and how much could be regained by allowing them to recover. Second, because any regulating system, as long as it remains functional, amplifies its response as perturbations increase, these least disturbed ecosystems should respond more effectively to ongoing global climate change than ecosystems already heavily altered by human activity. In other words, their climate-stabilizing value per unit area is disproportionately high.

These crucial and breathtakingly beautiful ecosystems have remained relatively intact largely because they are far from human activity. But what is far away is easily ignored. Out of sight, out of mind, or, as the Russian expression literally puts it, “away from the eyes, out of the heart.” As a result, scientific generalizations about how the living world is organized are routinely made while the least disturbed forests are effectively excluded.

For example, it is widely debated in the ecological and environmental literature whether a greater number of species makes an ecosystem more resilient. From the biotic regulation perspective, this is an ill-posed question, but that is a topic for another post. To systematically investigate the relationship between resilience and diversity, Lipoma et al. (2024) analyzed literature data from 26 studies encompassing 69 sites. As the map below illustrates, the least disturbed forest ecosystems made a negligible contribution to the analysis and, consequently, to its conclusions. (Notably, the authors found no correlation between resilience and diversity, which is hardly surprising. How a machine functions is not determined by the number of its components; it is the design that matters.)

Fig. 1 from Lipoma, L., Kambach, S., Díaz, S., Sabatini, F. M., Damasceno, G., Kattge, J., ... & Bruelheide, H. (2024). No general support for functional diversity enhancing resilience across terrestrial plant communities. Global Ecology and Biogeography, 33(10), e13895. https://doi.org/10.1111/geb.13895

We have discussed on several occasions that evidence already exists for greater resilience in less disturbed ecosystems (see, in particular, “The Rabbit–Duck Illusion in Climate Messaging: An Example from Wildfire Policy”). What is missing is not evidence, but systematic study. These patterns are simply not investigated in the comprehensive and deliberate way they should be.

From ecosystem resilience, let us now turn to our lack of knowledge about the climate impacts of the least disturbed ecosystems.

Natural Ecosystems as a Major Climate Unknown: The Carbon Cycle

Readers of Biotic Regulation and Biotic Pump may be aware that the global carbon cycle, or more precisely the scientific community’s understanding of it, is in an awkward state. Until recently, the standard narrative held that roughly one third of anthropogenic carbon emissions are absorbed by the terrestrial biosphere, primarily as woody biomass. Yet recent analyses show that, contrary to the predictions of vegetation models, global wood biomass has not increased over recent decades. Once again, this leaves a basic and uncomfortable question unanswered: where has the carbon gone? (For details, see “Nature is trying to fix our mess—it’s time to recognize its power”).

The next reservoir to consider after woody biomass is soil, which stores more carbon than the atmosphere and woody biomass combined. Measuring soil organic carbon, especially its long-term trends, is extremely demanding and therefore relatively rare. Nevertheless, a few such studies do exist. Recently, a group of scientists attempted to infer a global picture from existing experimental data, see
Jia, R., Fricke, E., Malhotra, A., Bar-On, Y. M., Deng, J., Piñeiro, G., ... & Terrer, C. (2025). A large global soil carbon sink informed by repeated soil samplings. BioRxiv, 2025-04. https://doi.org/10.1101/2025.04.25.650716

(To appreciate the extent of the current confusion in carbon cycle research, note that the conclusions of Jia et al. contrast sharply with those of another recent study in which some of the Jia et al. co-authors also participated.)

The figure above from Jia et al. (2025) shows sources and sinks of soil organic carbon, expressed in Tg C per year, for the period 1992 to 2020. One teragram equals one million metric tons of carbon. For comparison, anthropogenic carbon emissions amount to roughly ten thousand teragrams of carbon per year. Read in this context, the figure suggests that old forests in South America and Siberia together remove and store in soils about one tenth of global emissions. By contrast, young forests act mainly as sources of soil carbon, while grasslands contribute little.

These are very strong and potentially far-reaching claims. Importantly, the authors make a rare effort to distinguish, even if only crudely, between forest types, specifically old versus young forests. Such distinctions are seldom made, despite being essential. We would not say, “I looked out of the window and saw a human in the street.” We would say a girl with a ball, an old man with a cane, or a woman carrying heavy bags. Yet the term forest is routinely used in a broad and indiscriminate way, as though it referred to a single, well-defined entity, even though it encompasses systems that differ far more profoundly from one another than the full range of human conditions ever could.

But before considering the wider implications of these results, it is worth examining the geographical distribution of the repeated soil sampling on which they are based. The corresponding figure is not shown in the main text but appears only in the Appendix. It is striking.

That figure immediately reveals a near-universal lack of soil carbon data for the least disturbed forest landscapes. The large Siberian sink is inferred largely by extrapolating data from other regions, while for old forests in South America there is effectively a single data point. In other words, we are debating how to mitigate the accumulation of carbon in the atmosphere without any solid empirical foundation for the ecosystems that may matter most.

This reflects a deeper problem in modern Earth system science. Long-term soil carbon studies cluster around universities, simply because such measurements are convenient. Researchers sample nearby sites, publish the results, and secure further funding, while vast and potentially critical regions remain largely unmeasured. At present, there is little intellectual or financial incentive to change this situation. We are effectively searching under the lamppost and have no clue why we should not.

Natural Ecosystems as a Major Climate Unknown: The Water Cycle

One might point to remote sensing data, but such data are only as good as their ground calibration. Unsurprisingly, this calibration is minimal precisely where ecosystems are least disturbed. Let us consider transpiration, a key biotic component of the water cycle.

Take a tree growing in your yard. It grows, and it transpires, releasing large amounts of water vapor, on the order of hundreds of water molecules for every carbon dioxide molecule fixed during photosynthesis. Measuring this is trivial if the tree grows in a container. One simply tracks the added water while preventing soil evaporation. In open soil, however, the problem becomes far more complex. Soil moisture declines not only because of transpiration, but also because of drainage, and rainfall intermittently replenishes it. Separating these effects is not straightforward.

In principle, transpiration can be estimated from the atmosphere. Water vapor released by the canopy is transported upward by turbulence, and measuring this turbulent flux should allow transpiration to be inferred. This is the idea behind eddy covariance flux towers, which use high-frequency wind measurements to estimate fluxes of water vapor and carbon dioxide between vegetation and the atmosphere.

Here, too, fundamental problems arise. Horizontal transport of water vapor and carbon dioxide can far exceed vertical fluxes, making tower calibration critically important. Even when calibrated, flux towers sometimes produce paradoxical results, such as forests appearing to transpire less than grasslands, despite direct catchment-scale water budget measurements showing the opposite (see Teuling, 2018 “A Forest Evapotranspiration Paradox Investigated Using Lysimeter Data”). Despite these limitations, flux towers remain an important source of ecosystem data.

But as the two maps below make clear, the same pattern repeats. Flux towers are overwhelmingly located in accessible, human-modified landscapes. The least disturbed ecosystems, which are most relevant for understanding how transpiration and climate regulation actually work, are once again largely ignored.

The distribution of eddy covariance sites in 2000 (above) and 2025 (below), according to Xia et al. 2025. Vegetation types: evergreen needleleaf forests (ENF), evergreen broadleaf forest (EBF), deciduous needleleaf forest (DNF), deciduous broadleaf forest (DBF), mixed forest (MF), closed shrubland (CSH), open shrubland (OSH), woody savannas (WSA), savannas (SAV), grassland (GRA), wetland (WET) and cropland (CRO).

Xiao, J., Baldocchi, D., Ichii, K., Li, F., & Papale, D. (2025). Insights into terrestrial carbon and water cycling from the global eddy covariance network. Nature Reviews Earth & Environment, 1-20. https://doi.org/10.1038/s43017-025-00743-1

Summary and outlook

What we have discussed is not a collection of isolated data gaps, but a systematic methodological failure. Modern Earth system science relies heavily on measurements from ecosystems that are already degraded or operating near their limits, and then treats these observations as representative of how the living world functions. These data are scaled up, embedded in models, and used to inform policy, even though they describe failure modes rather than functioning regulation. In other words, by calibrating our science and policies on degraded ecosystems, we have mistaken breakdown for normality. As long as the least disturbed ecosystems remain largely unexplored, we can neither fully understand the climate change that has already occurred and is ongoing, nor devise a credible strategy for exiting the crisis.

Our blindness to vast, intact parts of the living world is itself a consequence of ongoing environmental destruction. As natural ecosystems disappear from everyday experience, they also disappear from observation and from scientific and policy attention. This loss of visibility feeds back into decision-making, reinforcing policies that further marginalize the remaining intact ecosystems. A recent attempt to map the planet’s so-called critical natural assets, published in a leading journal, illustrates this dangerous dynamic: most of the least disturbed forest ecosystems were excluded, in part because they lie far from human settlements and economic activity. Destruction thus produces blindness, and blindness legitimizes further destruction.

Map showing critical natural assets according to Chaplin-Kramer, R., Neugarten, R. A., Sharp, R. P., Collins, P. M., Polasky, S., Hole, D., ... & Watson, R. A. (2023). Mapping the planet’s critical natural assets. Nature Ecology & Evolution, 7(1), 51-61. https://doi.org/10.1038/s41559-022-01934-5

How can this cycle be broken? How can we embrace a natural world we have never seen? One way forward is to build an intellectual connection to what ultimately matters to us emotionally. By understanding why these ecosystems are important, we can come to see them as part of the world of ideals that shapes what we value and seek to preserve. After all, ideal concepts have long inspired individuals and even entire nations, from the Golden Fleece to the Mill of Sampo, a mythical source of abundance. In this sense, untouched natural ecosystems could emerge as a defining ideal of a new civilization, shaping its values, aspirations, and limits.

The starting point, then, is intellectual appreciation. The main thing we are missing, as our perception of the living world has been shaped by the dominance of artificial systems, is that natural ecosystems have function. They are not simply there by chance. Maintaining a stable environment is difficult, and so far we have been unable to achieve this even locally. We have polluted nearly everything that can be polluted, and continue to do so.

Yet avoiding pollution is not the core problem that life has solved. The deeper challenge is that a life-compatible environment is inherently unstable. Keeping it in dynamic equilibrium requires continuous effort and ingenuity, and this is precisely the function that natural ecosystems perform.

It is not the number of species that matters. Under different conditions, the same function can be carried out by many species or by only a few, and the latter case is not necessarily simpler. As with a story written in one volume or many, it is not the number of parts but the content that counts.

Natural ecosystems that remain capable of self-recovery are our most valuable asset as a species. Their ability to self-perpetuate is not unlimited and depends critically on the area they occupy. Once reduced below a minimal threshold, an ecosystem can collapse due to boundary effects and may never recover. Total area therefore matters crucially, both for resilience and for the efficiency of climate regulation. We must move beyond our current self-destructive horizon and act decisively to prevent further destruction of natural ecosystems, cooperating internationally wherever possible.

This post develops the ideas expressed in “How much wild nature do we need?”. I very much welcome your feedback.

A spring symphony. The voices of bog (recording from 1998)