Matters Arising: Critical Methodological Flaws in Qin et al. (2025) "Mangrove sediment carbon burial offset by methane emissions from mangrove tree stems"

Damien Maher, Luke Jeffrey, Judith A. Rosentreter, Scott Johnston, Vincent Gauci, Sunitha Pangala

The authors report global mangrove stem CH₄ emissions of 730.60 Gg yr⁻¹, offsetting 16.9% of carbon burial. However, their analysis suffers from critical methodological flaws involving the failure to remove extreme statistical outliers, with 14.2% of chamber measurements and 15.5% of site observations identified as outliers by standard criteria. The analysis demonstrates inappropriate handling of severely skewed data, where the mean is 13.3 times the median for chamber data, and training of machine learning models on outlier-contaminated data without quality control. The machine learning model is then applied globally, which results in an overestimation of global mangrove stem CH₄ emission estimates. The study lacks sensitivity analysis to assess robustness of estimates and relies on inflated R² metrics that reward overfitting to measurement errors. Our reanalysis using standard statistical methods and proper data quality control yields global emissions of 22.6 – 205.8 Gg yr⁻¹, representing a 3.5 to 32.4 fold overestimation in the published study. This range of estimates is more representative of previously published estimates of site-specific emission rates. The carbon burial offset should therefore be 0.5 to 4.7%, not 16.9% as claimed.