📄 agriculture

Stronger-than-expected CO2 fertilization of soybean inferred from paired C3/C4 yield records

By leveraging the contrasting photosynthetic responses of rotated US soybean (C3) and maize (C4) crops to isolate shared agronomic trends, this study reveals that rising atmospheric CO2 has driven a 43% increase in soybean yields since 1979—approximately three times higher than Free-Air CO2 Enrichment estimates—highlighting the critical role of co-evolving cultivars and management in amplifying the fertilization effect.

Original authors: Yi Yin, Qingyu Wang, Manasvin Anand, Xinlei Liu

Published 2026-07-16

📖 3 min read☕ Coffee break read

CC BY 4.0

Original authors: Yi Yin, Qingyu Wang, Manasvin Anand, Xinlei Liu

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the Earth's atmosphere as a giant, invisible greenhouse, but instead of just trapping heat, it's also filling up with a specific ingredient that plants need to eat: carbon dioxide (CO2). For decades, scientists have been trying to figure out how much this extra "food" is actually helping crops grow versus how much is just the result of farmers getting better at their jobs (using better seeds, more fertilizer, and smarter machinery) or how the weather is changing. It's like trying to taste a single spice in a complex stew; if the whole dish gets better, is it because of the new spice, or because the chef finally learned the perfect recipe? This question is crucial because if we don't know exactly how much CO2 helps crops, we can't accurately predict if we'll have enough food to feed everyone as the climate continues to shift.

Now, picture a clever detective story played out in the cornfields of the United States. The paper you're about to read uses a natural experiment involving two very different types of plants: soybeans and corn (maize). Think of soybeans as "open-mouthed eaters" (C3 plants) that get a huge energy boost when there's more CO2 in the air. Corn, on the other hand, is a "specialized eater" (C4 plant) that already has a built-in turbocharger to concentrate CO2, so extra CO2 in the air doesn't really make it grow faster—unless it's thirsty. The researchers noticed that in many counties, these two crops are grown in rotation on the exact same land, facing the same weather and the same farming improvements. This setup is like a perfect control group: if corn and soybeans both get better at the same time, it's probably because the farmers got better. But if soybeans suddenly get much better than corn, that extra boost must be the CO2 doing its work.

The authors used 45 years of data to separate these signals. They found that the CO2 boost for soybeans is actually much stronger than what scientists had guessed from small-scale field experiments. While those experiments suggested a modest gain, this real-world analysis suggests that for every 1 part per million (ppm) increase in CO2, soybean yields jump by about 0.32%. That's roughly three times higher than the previous estimates! The study suggests this happens because, over decades, farmers have been planting denser crops and breeding better varieties that work with the rising CO2, creating a "co-evolution" that small experiments missed.

However, there's a catch: this superpower only works well when the plants have plenty of water. The paper shows that when drought hits, the gap between how much soybeans and corn benefit from CO2 shrinks. It's like the turbocharger on the corn finally kicks in when it's dry, while the soybeans get a bit less of a boost. The researchers estimate that since 1979, about 43% of the massive growth in US soybean yields is thanks to this CO2 fertilization, while 46% comes from agronomic progress (better farming), and the rest is due to climate factors like temperature and rain.

In short, the paper argues that the "free lunch" of CO2 helping crops is real and larger than we thought, but it's a delicate deal that depends heavily on how well we manage our farms and how wet our soil stays. If we keep improving our farming techniques and managing water wisely, that CO2 boost could be a huge help for food security. But if we ignore the water needs or stop improving our methods, that extra boost might not be enough to save the harvest.

Technical Summary: Stronger-than-expected CO2 fertilization of soybean inferred from paired C3/C4 yield records

Problem Statement
Projecting future agricultural productivity and food security under climate change requires disentangling the contributions of rising atmospheric CO2, shifting climate patterns, and agronomic improvements to observed crop yield growth. A persistent obstacle in this field is the collinearity of these drivers: atmospheric CO2 has risen monotonically alongside agronomic advancements (e.g., breeding, fertilizer use, mechanization) and heterogeneous climate changes. Consequently, isolating the specific contribution of CO2 fertilization from general agronomic progress has remained unresolved. Traditional Free-Air CO2 Enrichment (FACE) experiments provide controlled estimates but often fail to capture the long-term, system-scale evolution of commercial agriculture, leading to potential underestimations of realized CO2 fertilization effects (CFE).

Methodology
This study exploits a "paired natural experiment" embedded in U.S. agriculture, leveraging the contrasting photosynthetic pathways of soybean (C3) and maize (C4).

The Natural Experiment: Soybean and maize are rotated across the same counties, sharing agronomic trends and environmental conditions. However, C3 soybeans are sensitive to rising CO2 (stimulating photosynthesis), while C4 maize is largely insensitive to CO2 under well-watered conditions due to its carbon-concentrating mechanism.
Agronomy Proxy: The authors assume that under well-watered conditions, the long-term yield trend of maize serves as a proxy for shared agronomic progress. By removing the climate variability from yield records, they demonstrate a strong contemporaneous correlation between maize and soybean yield growth rates in 705 counties, validating the shared non-climate trajectory.
Statistical Modeling: The authors trained two classes of models (Ridge linear regression and Neural Networks) on 45 years (1979–2023) of county-level data.
- For maize, a "Climate + Time" model was used to extract the background agronomic trend.
- For soybean, a "Soybean Full Model" was constructed using climate variables, the maize-derived agronomy proxy, and CO2 concentration. This structure isolates the CO2 signal by controlling for the shared agronomic trend.
Process-Based Validation: To verify empirical findings against theory, a coupled leaf-level biochemical-stomatal model (Farquhar–von Caemmerer–Berry for C3; von Caemmerer for C4) was parameterized using SoyFACE measurements. This model simulated net photosynthesis ( $A$ ) and intercellular CO2 ( $C_i$ ) responses under varying soil moisture and CO2 levels.
Attribution: A counterfactual approach was applied to the trained neural network to attribute yield growth to specific drivers (agronomy, CO2, precipitation, temperature) by fixing each driver at baseline levels while allowing others to vary.

Key Results

Magnitude of CO2 Fertilization: The study estimates a soybean CO2 fertilization effect of approximately 0.32% yield ppm⁻¹ under well-watered conditions. This is roughly three times higher than typical FACE-derived estimates (~0.09% yield ppm⁻¹).
Drivers of Yield Growth: Since 1979, U.S. soybean yields have increased by 75%. The attribution analysis reveals:
- Agronomic Progress: Accounts for 46% of the gain.
- CO2 Fertilization: Accounts for 43% of the gain.
- Climate: Accounts for the remaining ~11% (primarily interannual variability with limited net long-term contribution).
Soil Moisture Modulation: The differential CO2 response between C3 and C4 crops narrows under drought.
- Under well-watered conditions, the differential ( $\Delta$ CFE) is stable at ~0.32% yield ppm⁻¹.
- Under moderate drought, $\Delta$ CFE declines to ~0.21% yield ppm⁻¹.
- This narrowing occurs because drought induces stomatal closure in maize, pulling its operating $C_i$ off the saturation plateau of its A/ $C_i$ curve, thereby granting maize a previously negligible CO2 sensitivity. This pattern is consistent across spatial attribution, the 2012 drought event, and leaf-level theory.
Spatial Heterogeneity: The highest integrated CFE values are observed in the core Corn Belt and Mississippi River valley, where high-input agriculture supports the translation of physiological gains into yield. In contrast, drier or lower-input regions show CFE values closer to FACE estimates.

Significance and Claims
The paper claims that realized CO2 fertilization in modern U.S. agriculture is larger than experimental estimates suggest. The authors attribute the gap between their observational estimate (0.32% ppm⁻¹) and FACE estimates to several factors:

Curve Curvature: FACE experiments often measure sensitivity across a large step-change in CO2, averaging over the flatter, upper portion of the concave A/ $C_i$ curve. In contrast, the marginal sensitivity at current ambient CO2 levels (~420 ppm) sits on a steeper part of the curve.
System-Scale Amplification: Commercial agriculture involves continuous co-evolution of cultivars and management (e.g., increased planting density, improved sink capacity) that amplifies per-leaf gains into per-acre yield benefits, a dynamic difficult to capture in short-term, fixed-cultivar FACE experiments.
Operational CO2: Canopy-level CO2 is often drawn down by photosynthesis, placing the true operating point at lower effective CO2 concentrations where the A/ $C_i$ slope is steeper.

The study concludes that CFE is not a fixed physiological constant but depends on the agricultural system. Process-based crop models calibrated primarily to FACE data may misrepresent how CFE evolves as agriculture intensifies. The authors assert that incorporating the dependence of CFE on soil moisture and agronomic state is essential for accurate food-security planning under continued atmospheric and climate change.

More like this