Written by Mudit Joshi
Almost every headline about carbon dioxide is about harm. More CO₂ warms the climate. As the ocean absorbs it, seawater turns more acidic, which is a problem for corals and shell-builders. The 2025 Planetary Health Check counts seven of nine planetary boundaries as crossed, with ocean acidification the newest. The proposed safe level for climate, 350 ppm of CO₂, was passed in the late 1980s. This May, the Mauna Loa observatory recorded a peak of about 432 ppm.
For land plants, though, CO₂ is not a pollutant. It is food. Every gram of grain is built from carbon taken out of the air. That makes plants the one large group of organisms for which the CO₂ level matters directly, in both directions. The low end of that scale is where farming began.
Why cereals care about CO₂
The enzyme that fixes carbon in plants, Rubisco, is old and imprecise. It sometimes grabs oxygen instead of CO₂, and every such mistake costs the plant energy and carbon (photorespiration). The lower the CO₂, the more often it errs. Wheat, barley and rice use this pathway directly (the “C3” pathway), so they respond strongly to CO₂. Sorghum, millets and maize use the C4 pathway. They pump CO₂ into special cells around Rubisco, which keeps it close to saturated even when the air holds little CO₂. C4 photosynthesis is itself a product of a low-CO₂ world: it evolved independently dozens of times as CO₂ declined over the past 30 million years.
There is a second, less familiar benefit: plants can also save water. Leaves take in CO₂ through tiny pores called stomata, but water escapes through those same openings. When CO₂ levels are higher, plants do not need to open their stomata as widely. In open-field experiments at around 550 ppm CO₂, this reduced water loss from crops by about 10%. In drylands, where water is often the main constraint on growth, that saving can be just as important as the extra carbon available for photosynthesis. It is also the main way that C4 crops benefit from higher CO₂.
Farming at the low end of the scale
Air bubbles in Antarctic ice show that during the last Ice Age, CO₂ stood at around 180–190 ppm. Between about 18,000 and 11,700 years ago it rose to roughly 260 ppm. It then stayed between 260 and 280 ppm until industrialisation (Figure 1). In 1995 the plant ecologist Rowan Sage asked whether glacial CO₂ was simply too low for farming to pay off. Peter Richerson, Robert Boyd and Robert Bettinger developed the idea in a 2001 paper with a provocative title: was agriculture “impossible during the Pleistocene but mandatory during the Holocene?” They pointed to low CO₂ combined with an unstable glacial climate.
The physiology has since been tested. At the University of Sheffield, wild einkorn and wild barley, ancestors of the first Near Eastern crops, were grown at 180 and at 270 ppm. At the higher level they produced about 50% more grain. Wild C4 millets gained only 10–15%. Put the other way round, at glacial CO₂ a stand of wild wheat gives roughly a third less grain.
But pots are not fields, and history is less tidy than the hypothesis. Wheat and barley were cultivated in south-west Asia within about a thousand years after the Holocene began, soon after the CO₂ rise. Africa’s own cereals, sorghum and pearl millet, are C4 crops. They were domesticated much later, roughly 5,500–4,500 years ago on the southern margins of the Sahara, when CO₂ had already been near its pre-industrial level for thousands of years. There, the more obvious driver is water: the “Green Sahara” humid period and its end. If CO₂ mattered to Africa’s first farmers, it would have mattered mainly through the water channel, not through Rubisco.
The same physiology, running forward
Since the industrial revolution, we have pushed CO₂ from 280 to about 430 ppm. That is twice the size of the whole rise at the end of the Ice Age, and it took less than two centuries instead of several thousand years. Plants have noticed. Satellite records show greening over 25–50% of the world’s vegetated land between 1982 and 2009, and model experiments attribute about 70% of that trend to CO₂ fertilisation. Thirty years of free-air CO₂ enrichment (FACE) experiments show that adding about 200 ppm raised C3 crop yields by around 18% on average, and by much less when nitrogen was short. C4 maize and sorghum gained nothing when well watered and about 20% under drought (Figure 2).
How much of today’s harvest is already owed to extra CO₂ is still disputed. Statistical studies of US yields suggest effects two to six times larger than FACE experiments show. Nor is the gain free. Wheat grown at 546–586 ppm contained about 9% less zinc, 5% less iron and 6% less protein. These are nutrients that many diets already lack.
So the same physiology that may have shaped the first fields now sits at the centre of a present-day debate, at the opposite end of the scale. The low end is not only history. Field experiments cover today’s CO₂ and above, so the past is the natural experiment for the range below, where crop models have very little to test themselves against.
Testing an old idea with new tools
This is where my own work sits. I’m Mudit Joshi, a doctoral researcher in the AGRI-DRY network. I use a process-based vegetation and crop model, LPJmL, to reconstruct how rainfed cereals could have fared across Africa and south-west Asia since the end of the last Ice Age, and to separate the parts played by climate, CO₂ and nutrients. Ideas like Sage’s were proposed when the data to test them barely existed. Today, ice-core and pollen records, climate simulations and process-based models make it possible to test them: confirm them, refine them, or retire them. If you work on any part of this, whether palaeoclimate, archaeobotany, crop physiology or modelling, I would be glad to hear from you.
References
1. Sage, R. F. (1995). Was low atmospheric CO₂ during the Pleistocene a limiting factor for the origin of agriculture? Global Change Biology 1: 93–106.
2. Richerson, P. J., Boyd, R., & Bettinger, R. L. (2001). Was agriculture impossible during the Pleistocene but mandatory during the Holocene? A climate change hypothesis. American Antiquity 66: 387–411. https://doi.org/10.2307/2694241
3. Cunniff, J. et al. (2017). Yield responses of wild C3 and C4 crop progenitors to subambient CO₂: a test for the role of CO₂ limitation in the origin of agriculture. Global Change Biology 23: 380–393. https://doi.org/10.1111/gcb.13473
4. Ainsworth, E. A., & Long, S. P. (2021). 30 years of free-air carbon dioxide enrichment (FACE): What have we learned about future crop productivity and its potential for adaptation? Global Change Biology 27: 27–49. https://doi.org/10.1111/gcb.15375
5. Myers, S. S. et al. (2014). Increasing CO₂ threatens human nutrition. Nature 510: 139–142. https://doi.org/10.1038/nature13179
🌱 Keep following AGRI-DRY on LinkedIn and Instagram for all the exciting science coming up!