What a Chemist Keeps Wanting to Ask About Climate Policy

I am a theoretical chemist by training. I spend most of my time running calculations on molecules, working out how electrons arrange themselves around nuclei, figuring out why one reaction happens and another does not. It is quiet, exacting work, and the thing that keeps me honest is that the numbers do not care what I want them to be. You cannot sweet-talk a bond length. You cannot lobby against an ionisation energy. If the maths says the molecule is unstable, the molecule is unstable. You learn, fairly early, that your job is not to negotiate with nature. It is to understand the constraints and work within them.

That habit of thinking is hard to switch off. And over the past few years, as I have spent some time reading about how we govern the planet, I have kept running into something that bothers me. We know, with increasing precision, what the Earth system can tolerate. We have numbers. We have thresholds. We have error bars. And yet the targets we actually set (the climate pledges, the emissions reductions, the biodiversity commitments) are almost never pegged to those numbers. They are pegged to what is politically convenient.

I keep wondering: what would it look like if we treated the planet the way I have to treat a molecule?

The gap between what we aim for and what actually matters

Here is what I mean. Most of the environmental targets you hear about in the news are relative. Reduce emissions by 30% compared to 2005 levels. Cut energy intensity by 20% per unit of GDP. These are not meaningless. A relative target can be a perfectly serviceable policy tool. But they have a subtle problem when they are the only thing we aim for: they are measured against a baseline that we chose, and unless someone has done the work to connect them back to a limit the planet actually set, they can drift away from the number that matters.

It reminds me of something that happens in computational chemistry. You can set up a simulation, tweak the parameters until the output looks reasonable, and pat yourself on the back. But “looks reasonable” is not the same as “matches reality.” You can optimise forever and still be wrong if you are optimising against the wrong reference.

There is a serious scientific effort to define the right reference for the planet. It is called the Planetary Boundaries framework, and it was first proposed in 2009 by Johan Rockström, Will Steffen, and colleagues [1]. The idea is straightforward, even if the execution is not. It asks: for each of the major processes that keep the Earth system stable (the carbon cycle, the water cycle, the nitrogen cycle, the ozone layer, biodiversity, and so on), where is the threshold beyond which things start to go wrong in ways that are dangerous and potentially irreversible?

The framework was substantially updated in 2023 by Katherine Richardson, Rockström, and a large team of co-authors [2], and again in 2025 by the Planetary Boundaries Science Lab at PIK, the Potsdam Institute for Climate Impact Research [3].

Figure 1. The state of the nine planetary boundaries as of early 2025, prior to the September Planetary Health Check. Green indicates safe operating space. Yellow through red indicates increasing risk. Six of the nine boundaries were already in the red at this point; ocean acidification, the seventh, was confirmed as transgressed later that year. Image: Azote for Stockholm Resilience Centre, based on analysis in Sakschewski and Caesar et al. 2025 (CC BY-NC-ND 3.0).

Now, these boundaries are not physical constants in the way that, say, the speed of light is a physical constant. They are modelled thresholds, informed by observation, with uncertainty ranges and precautionary judgements built into them. I want to be honest about that. But they are also not arbitrary. They are the best estimate we have of where the guardrails actually sit. And by that measure, the 2023 paper found that we have already crossed six of the nine. Not approached them. Crossed them. Every boundary that was already overstepped in the previous assessment had moved further into the danger zone.

Then, in September 2025, PIK released its annual Planetary Health Check and added a seventh to the list [3].

The seventh boundary: what is happening to the oceans

When Richardson and colleagues published their 2023 update, ocean acidification was flagged as “close to being breached.” Two years later, the PIK assessment confirmed that it had crossed the line, making it the seventh of nine boundaries to be transgressed.

I find this one particularly unsettling, partly because the chemistry is so clean. CO₂ dissolves in seawater. It forms carbonic acid. The acid dissociates and releases hydrogen ions, which lower the pH. It is first-year chemistry, really. There is nothing ambiguous about the mechanism.

Since the start of the industrial era, the ocean’s surface pH has dropped by roughly 0.1 units. That might not sound like much, but pH is a logarithmic scale. A drop of 0.1 corresponds to a 26% increase in hydrogen ion concentration; the commonly cited 30–40% range reflects regional variation and measurement uncertainty. Either way, the water is becoming meaningfully more corrosive to the calcium carbonate that corals, molluscs, and certain plankton use to build their shells and skeletons.

Coral bleaching caused by warming and acidifying oceans, an early visible sign of ocean stress — Figure 2. Coral bleaching on a tropical reef. Bleaching is driven primarily by marine heat stress (elevated sea surface temperatures), though ocean acidification compounds the damage by weakening the carbonate structures that corals depend on for growth and repair. Cold-water corals, tropical reefs, and Arctic marine ecosystems are all at particular risk. Image: NOAA.

The oceans have been doing us an enormous, quiet favour for a very long time. They have absorbed roughly a quarter of all the CO₂ we have emitted since industrialisation. That absorption is part of why surface temperatures on land have not risen even faster. But the favour has not been free. The cost has been accumulating in the water column, and it is now showing up as a boundary crossed.

Johan Rockström, director of PIK, put it bluntly in the 2025 report: “More than three-quarters of the Earth’s support systems are not in the safe zone.” Of the nine boundaries, only two (stratospheric ozone depletion and atmospheric aerosol loading) remain within safe limits.

The one time it actually worked

I do not want this whole essay to feel like a slow-motion catastrophe, because there is one story in here that genuinely gives me hope, and it is worth telling properly.

In 1974, two chemists named Mario Molina and Sherwood Rowland published a paper showing that chlorofluorocarbons (CFCs), the compounds used in refrigerators, air conditioners, and aerosol sprays, were drifting up into the stratosphere and systematically destroying the ozone layer. The chemistry was elegant. Ultraviolet light breaks a chlorine atom off the CFC molecule. That chlorine atom destroys an ozone molecule. Then it goes on to destroy another one. And another. A single chlorine atom can take out tens of thousands of ozone molecules before it is finally deactivated.

The science was clear. And, to the lasting credit of the international community, the world actually responded. The Montreal Protocol was signed in 1987. Countries phased out CFCs. The ozone layer has been slowly recovering ever since.

NASA satellite composite showing the Antarctic ozone hole in 1979 versus 2019 — Figure 3. The Antarctic ozone hole as observed by satellite in (a) 1979, (b) 1987, (c) 1996, and (d) 2019. The progressive recovery since the mid-1990s is one of the clearest success stories in the history of global environmental policy. Images: NASA.

As Richardson and colleagues note in their 2023 paper: “The boundary for ozone depletion was exceeded in the 1990s but — thanks to global initiatives, catalyzed by the Montreal Protocol — this boundary is no longer transgressed.”

As a chemist, I find this story deeply satisfying. There was a specific mechanism. There was a measurable quantity (stratospheric ozone concentration) that everyone could track. There was a clear causal chain running from the molecule to the damage to the policy response. The science was unambiguous enough that it could not be easily argued away.

What else made the ozone story tractable? CFCs were produced by a relatively small number of companies. Viable chemical substitutes (HFCs and others) already existed or were in development. The economic disruption of the phaseout was narrow and manageable. The political problem was, in other words, unusually well-suited to a coordinated international response.

The harder question, the one I keep turning over, is whether that combination (clear science plus a tractable political economy) can be replicated for problems like carbon emissions, nitrogen pollution, or biodiversity loss, where the sources are diffuse, the economic stakes are vast, the substitutes are incomplete, and the politics are orders of magnitude more complicated.

The allocation problem, or: why knowing the total does not solve anything

This is where things get genuinely difficult, and where I find myself reaching for analogies from my own work.

In quantum chemistry, when you have a system of many interacting electrons, you cannot solve for each one in isolation. They are all coupled. The behaviour of each electron depends on the positions and momenta of all the others. You have to make approximations, and the choice of approximation always encodes assumptions about what you think matters and what you are willing to ignore. There is no neutral vantage point.

Dividing up a global environmental budget is structurally the same kind of problem. Take nitrogen, for example. The science tells us roughly how much reactive nitrogen the biosphere can absorb each year before ecosystems start to degrade. That is the budget. But the budget belongs to everyone. You have roughly two hundred countries, thousands of corporations, billions of people who all have some claim on it. And there is no formula, no first-principles derivation, that tells you how to split it.

Do you divide it equally per person? That sounds fair in the abstract, but it would require the wealthiest countries to make reductions so steep that no political system in history has managed anything close. Do you let countries keep roughly what they currently use? That locks in the historical advantage of those who industrialised first, emitted the most, and got rich doing so. Do you weight it by capacity to change? By GDP? By vulnerability?

Multiple competing principles can be defended. None of them is neutral. And some are much harder to defend than others: grandfathering current emissions, for instance, is difficult to justify when the largest emitters are also the wealthiest. But the central point stands. The science can tell us the size of the budget. It cannot tell us whose name goes on each line item. That is not a failure of the science. It is just a different kind of question, one that requires political philosophy, moral reasoning, and negotiation rather than measurement.

The PIK framework explicitly acknowledges this. The boundaries themselves are Earth system science. What we do about them (who must act, how fast, at whose expense) is a separate conversation, and it is one that science alone cannot settle.

Why delay is not just expensive: it can be irreversible

There is a concept I keep coming back to from materials science, one that I think deserves more attention in the policy conversation: hysteresis. The basic idea is that some systems, once pushed past a certain point, do not simply return to their original state when you release the pressure. The path back is not the reverse of the path forward.

If you take a ferromagnet and apply a strong enough external field, then remove the field, the material retains some of its magnetisation. Something in its internal structure has changed. To get it back to where it started, you would need to actively push it in the opposite direction.

Many parts of the Earth system behave like this. The Amazon rainforest is a particularly vivid example. The forest generates a significant portion of its own rainfall through transpiration: trees release water vapour, which forms clouds, which produce rain, which sustains the trees. It is a self-maintaining feedback loop. But if you remove enough of the forest, the loop starts to break. Rainfall drops. The remaining trees dry out. Fires become more frequent. The system can tip into a drier, more open state, not because of anything new happening from outside, but because the internal feedback that was keeping it alive has collapsed.

Aerial view of Amazon deforestation and intact forest, illustrating the fragmentation that can push the system toward a tipping point — Figure 4. Fragmentation of the Amazon basin. Left panel: estimated extent of original forest cover prior to European colonisation. Right panel: accumulated total forest loss from that original extent to the present (2022). Once enough forest is removed, the remaining patches lose the ability to generate their own rainfall, creating a self-reinforcing feedback loop toward further degradation. Image and Data: ACA/MAAP.

Getting back from that state is not simply a matter of replanting. The atmospheric and hydrological conditions that allowed the original forest to exist may no longer be present. The system has, in a real sense, moved to a new equilibrium, and returning to the old one would require far more effort than preventing the shift would have.

This matters enormously for how we think about policy timing. The usual framing is that delay is expensive: the longer you wait, the more it costs to fix the problem. That is true, but it understates the issue. The deeper problem is that delay, if it pushes you past certain thresholds, changes the nature of the problem. The window for one class of solutions closes, and you are left facing a different, harder problem that may not have a clean solution at all.

The 2025 PIK report makes this point directly. The seven transgressed boundaries are not just beyond safe limits. They are moving further into the risk zone each year. It is not only the position that matters; it is the direction.

A small experiment in Amsterdam

I want to end with something that I find cautiously hopeful, even though it is small.

In 2020, the city of Amsterdam adopted a framework for thinking about its economy called Doughnut Economics. The concept, developed by the economist Kate Raworth, is elegant: there are two boundaries that matter for any society. There is an ecological ceiling (the planetary limits we should not exceed) and a social floor (the minimum conditions that every person deserves: food, shelter, education, healthcare, political voice). The goal is to live in the space between them.

Amsterdam's City Portrait using the Doughnut Economics model — Figure 5. Amsterdam's City Doughnut, 2020. The framework maps both local and global social dimensions (inner ring) against local and global ecological footprints (outer ring). It is, to my knowledge, one of the first attempts by any city to operationalise this kind of thinking in actual governance. Image: Doughnut Economics Action Lab.

What interests me about this is not that Amsterdam has solved the problem. It has not. But it took an abstract scientific and philosophical framework and started using it to make concrete decisions about construction materials, food procurement, and circularity targets. The ecological ceiling stopped being something that existed only in academic papers and started showing up in municipal planning meetings.

It is a proof of concept, specifically that abstract ecological and social thresholds can be translated into planning tools at municipal scale. Whether that approach translates to the level of nations, or to the global commons, is a much harder question that Amsterdam has not answered. But proof of concept is not nothing. You have to show that something is possible at any scale before you can argue about whether it scales.

Where the science stops and the rest of it begins

The thing I keep returning to is this: the science can tell us where the limits are. It can tell us, with reasonable confidence, which thresholds matter and what happens when we cross them. That in itself is extraordinary. The fact that we can quantify, even approximately, the safe operating space of an entire planet is a remarkable achievement of modern Earth system science.

The analogy that Richardson and colleagues use in their paper is a good one. Think of planetary boundaries like blood pressure. A reading above 120/80 mm Hg does not mean you are going to have a heart attack tomorrow. But it means the risk is elevated. And no sensible doctor would tell you to wait for the heart attack before doing something about it. You try to bring the number down.

The same logic applies to the planet. We do not need to wait for a system collapse to take the readings seriously.

But here is the part where physics runs out of answers. Knowing the total size of the safe operating space does not tell you how to divide it. And that second question (who bears the cost of staying within the limits, what we owe to people in countries we will never visit, what obligations we have to people not yet born) is one that no equation can resolve. It requires something that the periodic table does not contain: a shared understanding of what is fair.

I do not think that makes the science less important. If anything, it makes it more important. The limits it identifies should function like constitutional constraints: things you design around, not things you negotiate away. But the design itself, the politics of it, the ethics of it, the compromises it will require, needs something that no amount of computation can supply.

That conversation, I suspect, is the most important one we are not quite having yet.

References

Rockström, J., Steffen, W., Noone, K. et al. (2009). A safe operating space for humanity. Nature, 461, 472–475. doi.org/10.1038/461472a
Richardson, K., Steffen, W., Lucht, W. et al. (2023). Earth beyond six of nine planetary boundaries. Science Advances, 9(37), eadh2458. doi.org/10.1126/sciadv.adh2458
Planetary Boundaries Science (PBScience) (2025). Planetary Health Check 2025. Potsdam Institute for Climate Impact Research (PIK), Potsdam, Germany. planetaryhealthcheck.org
Fanning, A.L., O’Neill, D.W., Hickel, J. & Roux, N. (2022). The social shortfall and ecological overshoot of nations. Nature Sustainability, 5, 26–36. doi.org/10.1038/s41893-021-00799-z
Lenton, T.M., Rockström, J., Gaffney, O. et al. (2019). Climate tipping points: too risky to bet against. Nature, 575, 592–595. doi.org/10.1038/d41586-019-03595-0
Armstrong McKay, D.I., Staal, A., Abrams, J.F. et al. (2022). Exceeding 1.5°C global warming could trigger multiple climate tipping points. Science, 377(6611), eabn7950. doi.org/10.1126/science.abn7950