Dynamic Ex-Post Baselines for Enhancing the Credibility of Improved Forest Management Carbon Offsets

Forest carbon offsets are pivotal to the global climate strategy, yet their claimed benefits are often undermined by flawed baselines, compromising their environmental integrity. Here, we assess five Improved Forest Management projects in China by constructing baselines using ex-post counterfactual methods based on satellite observations from 2001 to 2020. Applying methods including synthetic control, statistical matching, and machine learning, our counterfactuals, which accurately tracked pre-intervention carbon dynamics, reveal systematic underestimation of carbon sequestration in the reported baseline scenarios. We find that project forests would have continued to accumulate carbon stock naturally, contrary to the projects' static and artificially low baseline projections. Using our empirically derived baselines, we find no evidence that management activities are additional across our study samples. These biased baselines contribute to systematic over-crediting: ex-ante projected emission reductions exceed our estimates by a factor of 3.7 (uncertainty range: 1.7-5.7), while ex-post issued carbon credits exceed our estimates by a factor of 1.7 (uncertainty range: 0.8-2.5). These findings expose a fundamental limitation of current offset protocols, which fail to account for real and dynamic forest trends, leading to inflated climate benefits. We provide a scientifically grounded framework for robust, independent monitoring and dynamic baseline setting, which is crucial for ensuring the credibility of nature-based solutions.