Applying integrated population models to guide conservation planning: Western Capercaillie in Scotland

Abstract Many conservation interventions prove ineffective because they lack a rigorous evidence-base. Multifactorial drivers and data limitations often hinder anticipatory planning and are difficult to tackle with standard methods. Predictive modelling approaches that integrate diverse data sources with biological hypotheses can bridge this gap by clarifying the drivers of decline and evaluating management options under uncertainty. We developed a Bayesian integrated population model (IPM) applicable to lekking species, combining incomplete data from different life stages and seasons while accounting for observation error. IPMs are widely recognised for reducing bias relative to single-data-stream analyses, yet their application in conservation planning remains limited. We applied the model to the threatened Western Capercaillie population in Scotland, which has been declining since at least the 1980s despite decades of conservation efforts, a typical case requiring urgent, evidence-based management. By fitting the model to 30 years of biased and incomplete monitoring data, we investigated associations between demographic processes and their potential drivers, including weather variables, fence collision mortality and a proxy of predation pressure. We used model predictions to evaluate the joint effectiveness of proposed management actions aimed at improving vital rates, including fence management and diversionary feeding. Data integration improved population estimate precision by 17-49% relative to standalone national survey estimates, confirming that the decline continued from 1990-2023. Reproductive rate was related to the pattern of April warming, negatively affected by pre-breeding precipitation and positively by vole abundance, the latter consistent with the alternative prey hypothesis. Model predictions indicated that combining diversionary feeding with fence removal produced the most favourable conservation outcomes, but growth remained limited and uncertainty included possible continued decline. 5. This modelling approach forms a key component of Scotland’s Capercaillie Emergency Plan 2025-2030, guiding management by assessing population-level responses and predicting conservation intervention outcomes. In future, the model could support formal adaptive management through iterative evaluation of interventions as new data become available. Beyond this case study, our integrated approach offers a transferable framework for managing multifactorial declines in other threatened species under uncertainty.