Birch, please: The limits of threshold-based disturbance interpretation in Birch pollen records

Betula pollen is widely used in palaeoecology as an indicator of disturbances and early successional stages linked to woodland regeneration and landscape openness, but its interpretation remains largely qualitative. This study explores whether Betula pollen accumulation rates (PAR) can be converted into a quantitative threshold to differentiate various disturbance states. Using a modern experimental dataset from the Forchtenberg Project, which includes independently documented treatment histories since 1998 and a continuous record of pollen deposition from 2005 onwards, various thresholding methods, including percentile-based, receiver operating characteristic (ROC)-derived, and model-based approaches, were systematically tested. All methods exhibited weak, unstable performance, with Betula deposition failing to reliably distinguish disturbance from non-disturbance states. This outcome is consistent even under controlled experimental conditions, suggesting that the threshold failure reflects the ecological behaviour of Betula rather than limitations of data quality or analytical methods. To address this behaviour, the analysis extended beyond simple threshold classification to include temporal structure. Grouping observations by years since treatment revealed a consistent pulse–collapse–recovery trajectory, and a set of phase-sensitive interpretive bands was developed to describe Betula deposition as low, intermediate, or high within a graded framework. These bands are not proposed as universal thresholds but as context-dependent reference zones for fossil interpretation. Comparison with fossil lake records shows similar episodic and heterogeneous patterns of Betula influx, supporting the conclusion that threshold behaviour is not a universal feature of the proxy.