Planetary Boundaries and their Limitations

This week, I am continuing with the recent theme of sustainability, particularly as it relates to the scale of human society. Recent posts were on peak oil—a forecast decline in oil availability that will cause severe economic dislocation—and on The Population Bomb, which forecast imminent famines and other problems as a result of overpopulation. Both of these topics fit into the general framework established in 1798 by Thomas Malthus. Another Malthusian fear is of internal limits to growth, the idea that wealthier societies devote an increasing share of their resources to consumption and leisure, ultimately cannibalizing the basis of their wealth.

Since the Great Recession of 2007-2009 and the decline of the peak oil movement, the world’s predominant Malthusian fear has shifted from food and resource availability to ecological disruption that is ultimately caused by excessive human pressures on the environment. Although climate change is the center of these fears, I don’t think any framework better encapsulates the anxiety than the planetary boundaries framework, which is the subject of today’s post.

The concept of planetary boundaries was introduced by Rockström et al. (2009), updated by Steffen et al. (2015), and updated again by Richardson et al. (2023). I will focus on the latter paper. In this post, I will describe what planetary boundaries are, how the boundaries are quantified, and whether we should expect the planetary boundaries framework to age better than the other ideas mentioned above.

What are Planetary Boundaries?

According to Schlebusch et al. (2017), anatomically modern humans appeared around 350,000 to 260,000 years ago, but insofar as civilization can be defined and its origins agreed upon, the appearance of civilization is dated at or after the Near East agricultural revolution. Thus civilization is contained in the Holocene epoch, a relatively stable climactic era with conditions favorable to agriculture as we know it.

Over geologic time scales, Earth’s climate has not been stable, and most of Earth’s history features climates that would not be conducive to the thriving of human civilization. The causes and effects of climate change (here, the term is used in a broad sense, and not in the narrow sense of global warming induced by greenhouse gases) are poorly understood.

Environmental processes driven by human activities, such as greenhouse gas emissions, land use change for agriculture, freshwater withdrawals, and so forth have created climate conditions that in several respects are outside the norms of the Holocene epoch. These conditions create a risk that Earth’s climate will change, perhaps irrevocably, in a way that is highly damaging to human wellbeing.

Following on earlier work, Richardson et al. (2023) posit nine planetary boundaries. Each boundary is a quantified metric (control value) with a reference value, called the “safe operating space”, set in relation to preindustrial Holocene norms. Invoking the precautionary principle, the paper sets strict limits as to what constitutes humanity’s safe operating space. Transgressing those boundaries creates a risk of harmful ecosystem disruption, and the greater the transgression, the greater the risk.

The planetary boundaries in the three major assessments, as portrayed by the Stockholm Resilience Center. The apparent worsening of impacts in the 2023 assessment is largely due to a more complete assessment and changes in methodology, as discussed below.

The paper emphasizes interconnectedness, or holism, a critical strand of thought in ecology, as documented by Levins and Lewontin (1994), and it discusses several important interactions between the nine areas assessed. Nevertheless, to present a comprehensible picture, the authors separate the nine issues and present quantified boundaries for each, and so I will treat them separately as well.

The Nine Boundaries

We will now briefly examine each of the nine boundaries: what the boundaries are and the outlook for that area.

Biosphere Integrity

Those who are old enough will remember a time when biodiversity was the signature issue of the environmental movement. Protection of endangered species, such as the giant panda that is the logo of the World Wildlife Foundation, was of paramount importance, and extinction was the gravest environmental crime. Nowadays, climate change has displaced biodiversity as the signature issue, but biosphere integrity is the first planetary boundary assessed by Richardson et al. (2023).

The paper assesses two metrics associated with biosphere integrity. One, genetic diversity, is the extinction rate, with the boundary set at 10 extinctions per million species-years (E/MSY) (in other words, in a given year, 1/100,000 of all species go extinct). Cellabos et al. (2015) (Paul Ehrlich, whose book I discussed last week, is a coauthor on this paper) assess the background extinction rate at 2 E/MSY for mammals and less than that for other vertebrates, and today’s extinction rate is up to 200 E/MSY. The rather polemical Cowie, Bouchet, and Fontaine (2022) cite background extinction rates on the order of 0.1-1 E/MSY and, with admittedly limited data, extrapolate to find contemporary extinction rates on the order of hundreds of E/MSY.

The idea that Earth is in the midst of its sixth mass extinction since the Cambrian Explosion, popularized by, for example, The Sixth Extinction: An Unnatural History by Elizabeth Kolbert, may be exaggerated. It is nevertheless clear that extinction rates today are well above preindustrial norms. It should also be kept in mind that an anthropogenic role in extinction predates industrialization; Barnosky et al. (2014) assess that human hunting was one of, though not the only, contributor to the late Pleistocene extinctions (the Pleistocene was the geological epoch predating the Holocene, and it encompasses all existence of Homo sapiens prior to agriculture).

The other biodiversity metric is functional diversity, and this is estimated by Net Primary Productivity, replacing the less empirically tractable Biodiversity Intactness Index of the previous version. NPP is estimated at 55.9 billion tons of carbon per year (GtC/y) as the Holocene norm. Current NPP is estimated at 65.8 GtC/y. Richardson et al. (2023) does not set the NPP boundary at a fixed NPP level, whose relevance they dispute because it is artificially inflated by carbon fertilization, but rather that the portion of NPP appropriated by human activity should be no more than 10% of total NPP. Citing Krausmann et al. (2013), the authors estimate that human appropriated NPP was 23.5% in 2020, well above the limit.

The paper cites Gerten et al. (2020), which shows that the Earth could feed 10.2 billion people within planetary boundaries. The abstract of that paper states,

Key prerequisites are spatially redistributed cropland, improved water–nutrient management, food waste reduction and dietary changes.

One wonders how many more people could be fed if biotechnology, greenhouses, and precision agriculture were included in the list of solutions.

Climate Change

Contemporary global warming is caused primarily by an elevated concentrated of greenhouse gases, principally carbon dioxide, in the atmosphere. Richardson et al. (2023) take the boundary as a concentration of 350 parts per million (ppm) of CO₂ in the atmosphere, compared to a preindustrial value of around 250 ppm and a 2023 value of 419.3 ppm. The boundary for radiative forcing is 1 watt per square meter, and the International Panel on Climate Change assesses forcing in 2022 at 2.91 W/m² relative to the 1750 value.

According to the EI Statistical Review of World Energy, world CO₂ emissions from energy, the primary source, are continuing to increase as of 2023 and were about 35 billion tons/yr. However, emissions in the Organization for Economic Co-operation and Development—rich countries generally—peaked at just under 13.8 billion tons/yr in 2007 and have since declined by 19%. This decline has been more than offset by the rest of the world, and a decline of 19% is insufficient; the decline should be 100%, and then some to remove excess CO₂ that is already in the atmosphere. Nevertheless, the decline is a sign that the greenhouse gas problem is tractable. The main solutions are low carbon energy sources, such as nuclear, solar, and wind; carbon removal; and energy efficiency.

Novel Entities

“Novel entities” are a catch-all for various anthropogenic materials without natural counterparts: microplastics, endocrine disruptors (e.g. bisphenol A), persistent organic pollutants, nuclear waste and weapons, and genetically modified organisms, for instance. Given the novel nature of these entities, following Persson et al. (2022), Richardson et al. (2023) set the boundary at zero unless the entities are “thoroughly tested” under the EU Registration, Evaluation, Authorisation and Restriction of Chemicals.

I cannot comment on whether REACH is the appropriate standard of testing. However, with some entities such as nuclear waste and genetically modified crops, there is no evidence of a systematic environmental hazard after decades of research.

Stratospheric Ozone Depletion

The level of stratospheric ozone is typically given in Dobson units. A Dobson unit is the measure of concentration of a trace gas in the atmosphere such that, if the gas was pure, it would form a layer of 10 micrometers around the Earth at standard temperature and pressure. The boundary value of stratospheric ozone is 276 Dobson units, 5% below preindustrial levels. Current concentration is 284 Du.

As I discussed in a post last year, the ozone layer is recovering following the Montreal Protocol, which banned ozone-depleting chlorofluorocarbons (CFCs), and it is one track to recover globally by 2066. However, the effect of an expanded spaceflight industry, if that comes to pass, is unclear.

Freshwater Change

Richardson et al. (2023) consider blue water and green water as two boundary variables. Blue water is that which is stored in lakes, rivers, and aquifers. Green water is stored in soil and used by plants. The metrics are a bit complex. They divide the world into 30 arc-minute grid cells, and for each cell, consider a deviation to be an amount of water that is outside of the 95% variability range, either on the side of being too wet or too dry. The boundaries are that 10% of cells should experience blue water deviations, and 11% should experience green water deviations. Actual deviations, according to Porkka et al. (2024), are 18.2% and 15.8% respectively.

Note that Richardson et al. (2023) shows a substantial deterioration of the world’s freshwater status compared to the previous version of planetary boundaries, Steffen et al. (2015). This is primarily because of a change in boundary conditions, not a change in actual conditions.

I discussed desalination a couple years ago, one of the more obvious—though far from the only—way to reduce humanity’s freshwater needs. Dieter et al. (2018) assess water withdrawals in the United States every five years from 1950 to 2015. The peak rate of withdrawal over that time was in 1980, and by 2015, withdrawals had decreased by about a quarter from the peak. Water challenges are very tractable, though I am not sufficiently aware of the ecological issues around the water cycle.

Atmospheric Aerosol Loading

Aerosols are fine particles that are suspended in air, and they have a wide range of categories, as well as a wide range of natural and anthropogenic sources. The distinction is blurred, though, when one considers that “natural” aerosols from wildfires or dust from dry land may have indirect anthropogenic causes.

Aerosols have effects on rainfall and monsoon patterns that are not fully understood, but it is generally understood that aerosols enhance cloud formation and lead to more frequent but less heavy rainfall. Aerosols should also generally have a cooling effect on the climate, which is why intentional stratospheric aerosol injection as a method to counteract global warming is of interest. Wang et al. (2021) find that human sources constitute about 30% of aerosols over the Arctic Ocean, of which Miller et al. (2023) find that shipping is the largest source.

Richardson et al. (2023) consider the metric of aerosol optical depth, or AOD, which is the fraction of direct sunlight that is blocked from reaching the Earth’s surface, by absorption or scattering, as a result of aerosols. The boundary is set at 0.25, or a quarter of the light blocked. Vogel et al. (2022) find that the global AOD mean is 0.14, though some regions exceed the boundary value. Additionally, since differences in AOD drive monsoon patterns, an interhemispheric difference boundary of 0.1 is set. Zanis et al. (2020) find the actual value to be 0.076±0.006.

As I noted last week, wealthy country have made great strides in reducing conventional air pollution, such as smog.

Ocean Acidification

As the oceans absorb excessive carbon dioxide from the atmosphere, they form carbonic acid, lowering the pH level and becoming more acidic, as explained here. A typical metric for ocean acidification is the saturation rate of aragonite, a dissoluble form of calcium carbonate. The preindustrial value of this metric was Ω_arag ≈ 3.44, and the boundary is set at 80% of this value. Jiang et al. (2015) estimate this value at 2.8, which is 81% of the preindustrial value, or just within the boundary conditions.

The most obvious solution for ocean acidification is to reduce CO₂ emissions. There are geoengineering approaches, such as adding lime to the ocean, though Renforth, Jenkins, and Kruger (2013) find that liming on a large scale would be at the edge of economic feasibility.

Land System Change

Land use required by human activities, especially agriculture, is, in my judgment, the most pressing environmental issue today. It is more pressing than even climate change. There are many metrics that Richardson et al. (2023) could have chosen for a land use boundary; the metric they chose is forest cover. In particular, the metrics are the amount of tropical, temperate, and boreal forests in the world compared to the potential amount throughout the Holocene. The boundaries for these three biomes are 85%, 50%, and 85% respectively, or a 75% weighted average. The actual value today is 60%.

According to the Food and Agriculture Organization of the United Nations, net deforestation in the world is continuing, though has slowed from 7.8 million acres per year in the 1990s to 5.2 million in the 2000s to 4.7 million in the 2010s. There are 1.11 billion hectares of primary forest—native trees without clear sign of human activity—in the world, with an 81 million hectare loss from 1990 to 2020 (the numbers don’t seem to add up because the first statistic considers all forests, and the second is primary forest).

Hannah Ritchie at Our World in Data documents that the world has probably passed a peak in agricultural land, which is the predominant human land use. There is hope that net deforestation will soon come to an end, and then we will measure reforestation instead of deforestation.

Biogeochemical Flows

“Biogeochemical” is a complex word that refers to the flows of various elements. The two considered in Richardson et al. (2023) are nitrogen and phosphorus, two elements that are critical in artificial fertilizers and thus in feeding the world. The fixation of large amounts of atmospheric nitrogen via the Haber-Bosch process and mining of large quantities of phosphorus has altered the environmental balance of these elements.

The global boundary for phosphorus flow to the ocean is set at 11 million tons per year, with an observed value of 22 million tons per year. For application to soils, the boundary is 6.2 million tons per year and actual value is 17.5 million tons per year. Finally, the boundary for artificially fixed nitrogen is 62 million tons per year, and the actual value is 112 million tons per year.

How Will Planetary Boundaries Age?

I have several criticisms of Richardson et al. (2023) and the planetary boundaries framework, but I want to start with some positive points. I appreciate how the paper brings together several major environmental topics. It is far from comprehensive, and indeed it couldn’t be, but I think it is a good selection of topics. I also appreciate how it uses a common framework to assess boundaries, which helps us get a unified understanding of what may seem to be disparate topics.

But there are several serious problems. The paper suffers from scientism, which statements like this (bold added by me):

The planetary boundaries framework formulates limits to the impact of the anthroposphere on Earth system by identifying a scientifically based safe operating space for humanity that can safeguard both Earth’s interglacial state and its resilience.

In reality, the control variables and the boundary values appear to be quite arbitrary. There is no explanation as to how they are chosen, other than to approximate Holocene conditions.

From The Breakthrough Institute, Blomqvist, Nordhaus, and Shellenberger (2012) wrote a detailed critique of the original 2009 paper (Rockström et al. (2009)). Applying their analysis to the 2023 paper, only three of the nine boundaries—climate change, stratospheric ozone depletion, and ocean acidification—can be said to be truly planetary boundaries. The rest are regional impacts, in that, for instance, aerosol loading and freshwater imbalances in one part of the world have little bearing on another part. Therefore, global aggregations of impacts for these measurements have little relevance, and there is no global tipping point.

Speaking of tipping points, the authors respond to the polemical Montoya, Donahue, and Pimm (2018) by asserting that their framework does not engender tipping points. However, to date little harm to human civilization has been observed from ecosystem disruption. The very idea that major harm will result from significant deviation from Holocene environmental norms, especially to invoke the precautionary principle, implies sharply nonlinear impacts resulting from cumulative ecosystem disruption. Semantics aside, that is a tipping point by any reasonable definition, and I will call it such.

With that out of the way, the paper does not expound on what specific harms will actually result from transgressing boundaries. This omission should not be waved away with the precautionary principle, nor should the precautionary principle be applied as though it were Pascal’s wager, as Johnson (2012) cautions.

Finally, there is the philosophical aspect. The language of boundaries is not overtly antihumanist, but it strongly implies—and this is made explicit in other papers by the authors—that the resolution to transgression of planetary boundaries is for human civilization to “know its place” and contract its footprint by mechanisms that for today I will leave to the reader’s imagination. Instead, throughout this article, I have tried to focus on many of the technological solutions to the same challenges that focus on expanding, rather than suppressing, human agency.

Will planetary boundaries age well? I’m sure that it will age better than The Population Bomb, for the simple reason that it does not make specific predictions that will be obviously falsified. Many environmental problems, such as smog, lead, and ozone depletion, have been greatly mitigated by innovation, and none by contraction of civilization. I am optimistic that 50 years from now, the news on most or all of these planetary boundaries will be good, and that this will be achieved through innovation and without severe harm to human civilization.

Quick Hits

In last week’s post, I included the infamous quote from the preface of The Population Bomb,

The battle to feed all of humanity is over. In the 1970s and 1980s hundreds of millions of people will starve to death in spite of any crash programs embarked upon now.

A reader pointed out that in the first edition of 1968, the “and 1980s” clause was absent. Already by 1971, it was clear that the Green Revolution was increasing yields in such a way as to negate the most pessimistic forecasts, and Ehrlich was already hedging his bets.