Key points are not available for this paper at this time.
ABSTRACT Biochar application in croplands has shown significant potential for enhancing crop yields and mitigating greenhouse gas (GHG) emissions. However, results from regional field trials vary widely, indicating that the effectiveness of biochar application is influenced by multiple factors, including biochar properties, environmental factors, and management practices, thereby limiting its large‐scale implementation. Here, we developed extreme gradient boosting (XGBoost) models based on global datasets to predict staple crop yield and soil GHG emission responses to biochar application, with interpretation via Shapley additive explanations. Through global‐scale iterative simulations, we identified optimal biochar application strategies accounting for geographical variability and management practices. The mitigation potential of biochar systems was further quantified using life cycle assessment. Our results showed that XGBoost models achieved strong predictive performance, with soil properties and climatic conditions identified as the dominant factors shaping biochar application effects. Biochar was found to enhance yields and reduce emissions across most global croplands simultaneously, though the magnitude and optimal strategy varied regionally: yield improvements were more pronounced in tropical areas, whereas emission reductions dominated in temperate zones. Under optimized strategies, sustainable global biochar implementation could achieve an annual net emission reduction of 1.08 Pg CO 2 ‐equivalent (offsetting 17% of global agricultural GHG emissions) while boosting staple food production by ~15 million tons. These findings highlight the considerable role of biochar in climate change mitigation and food security, emphasizing the need for region‐specific application strategies tailored to local soil and climate contexts. Such approach could facilitate its advancement from localized trials to global environmental governance.
Lu et al. (Sun,) studied this question.