Pitfalls of forest damage detection using medium-resolution imagery: A response to the paper of Karpov et al. (2024)

A recent paper by Karpov et al. (2024) discusses forest disturbances on Kunashir Island due to windstorms and bark beetle outbreaks. The study uses medium-resolution satellite data, which we argue is conceptually flawed for analyzing windthrows and spruce dieback in mixed forests of the study area. The results significantly underestimate disturbed areas by over 50 %. Here we highlight the methodological shortcomings of the Karpov et al. (2024) study and caution against using medium-resolution data for forest disturbance detection in mixed forests. Such underestimations can provide an extremely imprecise baseline for disturbance impacts under climate change, potentially misguiding projections of impacts on forest ecology and management. •Landsat and Sentinel-2 data may underestimate windthrow and forest pest areas•Underestimation rate of disturbances in Kunashir island is about 50 %•Very high-resolution imagery and neural networks provide more accurate assessment•Use medium-resolution data cautiously for local disturbance estimations in mixed forests