Climate-smart agriculture offers a pathway to boost China’s carbon efficiency

Agriculture sits at the center of one of the most difficult equations in climate science: the world must produce more food even as it produces fewer greenhouse gas emissions. A new study from China offers one of the most detailed answers yet to how that balance can actually be achieved on the ground, combining a decade of provincial data with sophisticated statistical modeling to identify the concrete pathways by which climate-smart agriculture can lift the country’s agricultural carbon emission efficiency.

The research, published in the journal Air Quality, Atmosphere & Health by a team at Fujian Agriculture and Forestry University led by Jiadong Zhang, Tao Xu, Shengquan Wang, Shaoxiong Wu and Lingxin Bao, examines agricultural carbon emission efficiency—often abbreviated ACEE—across all 31 Chinese provinces from 2010 to 2022. ACEE is a measure that captures how effectively a region converts agricultural inputs into grain output relative to the carbon it emits in the process. A high ACEE score means a province is producing more food per unit of agricultural carbon, integrating the twin objectives of grain production growth and multi-source emission reductions into a single quantitative framework.

The concept of climate-smart agriculture, or CSA, was developed by the Food and Agriculture Organization of the United Nations as a paradigm that pursues three goals simultaneously: sustainably increasing agricultural productivity, adapting and building resilience to climate change, and reducing or removing greenhouse gas emissions wherever possible. In practice, CSA encompasses technologies such as water-saving irrigation, straw-return—the practice of working crop residues back into the soil rather than burning them—and no-tillage planting, which minimizes soil disturbance and the carbon losses associated with it. While these practices have been widely adopted in parts of the developing world and are increasingly embedded in agricultural policy in Europe and North America, their implementation in China has been uneven, largely because local levels of agricultural sustainability vary so dramatically across the country’s vast and ecologically diverse territory.

To map that unevenness, the researchers first constructed a comprehensive indicator system for ACEE, drawing on emission accounting methods consistent with the Intergovernmental Panel on Climate Change guidelines for national greenhouse gas inventories. Agricultural emissions in China arise from multiple sources, including nitrogen fertilizer application, rice paddies, livestock, soil management and the energy consumed by farm machinery. Rather than treating these as a monolithic total, the team’s framework integrates both the desired output—grain production—and the undesired outputs of various emission streams, reflecting the reality that a province cannot simply cut emissions by producing less food.

The efficiency calculations were performed using a technique known as super-efficiency slacks-based measurement, or super-efficiency SBM, an advanced form of data envelopment analysis. Conventional efficiency analysis struggles to rank decision-making units that all sit on the “efficient frontier”—the boundary representing the best achievable performance. The super-efficiency variant solves this by allowing efficient units to exceed a score of one, effectively ranking them against a frontier from which they have been temporarily removed. This matters in a national comparison, because without it, many provinces would simply tie at maximum efficiency and the analysis could not distinguish, say, a moderately efficient grain belt from an exceptional one.

To track how the distribution of ACEE has evolved over the twelve-year study window, the team then applied kernel density estimation, a non-parametric statistical method that reconstructs the underlying probability distribution of efficiency scores from observed data without imposing assumptions about its shape. This allowed the researchers to detect subtle shifts in the ” geography” of Chinese agricultural carbon performance that simple provincial averages would obscure. Their findings are striking: the national average ACEE remained broadly stable over the period, but the spatial distribution exhibited an asymmetric pattern the authors describe as “high-value contraction” and “low-value stability.” In other words, provinces at the top of the efficiency distribution appear to have become more tightly clustered—converging on a shared high-efficiency profile—while lower-performing provinces held their positions without marked improvement. Within China’s three major regions, internal disparities in ACEE remained evident, with varying degrees of polarization, suggesting that the gap between leaders and laggards has not closed and, in some places, may have widened.

Having quantified where efficiency is high and low, the study’s central contribution lies in explaining why. Guided by an analytical framework built around climate-smart agriculture, the researchers examined explanatory factors across three dimensions: CSA technology, policy support and the social environment. For this they turned to the Geodetector model, a spatial analysis tool designed to measure how much of the spatial variation in a variable can be explained by a stratifying factor. Geodetector works by comparing the within-stratum variance of the outcome variable to its total variance; the resulting q-statistic ranges from zero to one and expresses the explanatory power of each factor. Unlike conventional regression, Geodetector makes no assumption about linearity and is robust to multicollinearity, which makes it well suited to disentangling the effects of interrelated social, technological and environmental variables.

The Geodetector results pointed clearly to technology. The adoption levels of three CSA technologies—water-saving irrigation, straw-return and no-tillage planting—showed relatively strong explanatory power for the spatial disparities in ACEE. Provinces where these practices had penetrated more deeply tended to be provinces where agricultural carbon efficiency was higher, even after accounting for other conditions. But the single most important finding of the spatial analysis may be about interaction rather than individual factors: the explanatory power of factor combinations significantly exceeded their independent contributions. This is a classic signature of synergistic causation, in which technologies or conditions that are only moderately powerful on their own become highly consequential when deployed together. A water-saving irrigation system paired with supportive policy instruments and a favorable social environment, for instance, delivers efficiency gains that no single component could achieve alone.

To translate that insight into actionable strategy, the researchers integrated dynamic qualitative comparative analysis—QCA—into their framework. QCA is a set-theoretic method rooted in the work of Charles Ragin that treats cases, in this study provinces, as configurations of conditions rather than as independent data points. Rather than asking whether factor X has an average effect on outcome Y across all cases, QCA asks which combinations of conditions are sufficient, or necessary, to produce the outcome. The dynamic extension of the method allows these configurations to be examined across time, capturing how the recipe for high efficiency may change as regions develop. Configurational methods are increasingly favored in sustainability research precisely because they embrace what scholars call causal complexity: multiple, different routes to the same outcome, with conditions substituting for one another in some configurations and complementing one another in others.

The QCA analysis identified four differentiated configuration pathways that enhance ACEE under the CSA framework. The team labeled these pathways as those driven by “policy and environment,” by “technology and policy,” by “technology, policy and environment” jointly, and by “technology” alone. Each represents a distinct recipe that a province can follow. A policy-and-environment pathway suggests that in some regions, strong governmental support combined with favorable social and natural conditions can deliver high efficiency even without leading-edge technology adoption. A technology-driven pathway indicates that in other regions, the diffusion of CSA practices itself is sufficient to propel efficiency gains. The combined pathways, meanwhile, confirm the Geodetector’s finding that the most reliable route to high performance is the deliberate stacking of technological, institutional and social conditions.

The policy implications are significant, both for China and for the wider world. China is simultaneously the world’s largest agricultural producer and a major agricultural emitter, and its stated “dual carbon” goals—peaking carbon emissions before 2030 and achieving carbon neutrality before 2060—cannot be met without transforming the farm sector. The study suggests that a one-size-fits-all national CSA mandate would be a mistake. Provinces should instead be matched to the pathway that fits their existing endowments: regions with strong fiscal and institutional capacity might lead with policy and environmental measures, while agronomically advanced regions could accelerate technology-led transitions. The finding that factor interactions outperform individual factors also cautions against fragmented, siloed interventions—subsidizing a single technology in isolation is unlikely to replicate the gains seen where technology is embedded in supportive governance and social context.

The research also carries a note of urgency. The “high-value contraction, low-value stability” pattern implies that the provinces best positioned to improve may be plateauing at high efficiency while the laggards remain stuck, a dynamic that could entrench regional inequality in agricultural sustainability. Because ACEE integrates food production with emission performance, stagnation among low-efficiency provinces threatens both climate objectives and food security, the very trade-off CSA is designed to resolve.

The work was supported by the Natural Science Foundation of Fujian Province and the Special Fund for Science and Technology Innovation of Fujian Agriculture and Forestry University. The corresponding author is Lingxin Bao of the College of Computer and Information Sciences at Fujian Agriculture and Forestry University. While the methodology is grounded in Chinese data, the framework—linking an integrated efficiency indicator, spatial diagnostics, and configurational pathway analysis—offers a transferable template for any nation wrestling with how to feed a growing population on a warming, carbon-constrained planet. As climate pressures intensify, the study’s core message is clear: the future of low-carbon agriculture will be won not by single silver-bullet technologies, but by smartly assembled combinations of technology, policy and social conditions tailored to each region’s circumstances.