Contributing to the disturbance ecology of Scots pine in Northwest Germany: Analyzing practitioners’ reports on damage severities

• in Northwest German forests, damage records amassed 600,000 observations since 2014. • foresters’ crowd sensing serves to explain forest damages. • globally, in stands with Pinus sylvestris share, storms cause the most disturbances. • locally, damages of non-wind and biotic factors are more severe and catch up. Complex and interdependent disturbance factors drive the decline of Scots pine ( Pinus sylvestris ) under climate change. However, the magnitude of low- to high-severity disturbances in forests remains difficult to quantify. Forest managers require information to adapt staff as well as machinery deployment strategies, which vary based on whether entire stands collapse or single trees die. Despite its demand, comprehensive data to assess disturbance regimes remains scarce. One decade after the introduction of the governmental hub ‘Forest Health Notification Portal’ (FHNP) of four Northwest German states, we take an exploratory approach to assess practitioners’ reports via three research questions: i. Which disturbance factors lethally affect pine forests? ii. What are the causal influences behind forest mortality? iii. Can a robust distribution of damage severities be derived from FHNP entries? Spatially linking damage reports, management plans, and climate covariates created a unique dataset, which was analyzed via a beta distribution regression model. Our findings align with national logging statistics and related studies: i. as a trigger for damage, biotic factors approach storms in both frequency and severity. ii . Damages are influenced by admixture competition, and by the ecotonal status of pine, which attenuates under positive climatic water balances. While non-wind disturbances cause more severe stand-scale damage, storm remains the dominant overall driver of timber loss. iii . Lethal damage severities average 0.5 ha but have a median of just 0.12 ha, reflecting a left-skewed density function modeled with a Bayesian R² of 38 %. The results suggest that meta-level limitations of the dataset are negligible.