Beyond expert opinion: A scalable, data-driven model for habitat quality assessment

Habitat quality (HQ) modelling plays a critical role in biodiversity conservation and ecosystem service planning. However, widely used HQ models, such as Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST), rely on expert-derived coefficients for land-cover suitability, threat sensitivity, and spatial decay parameters, introducing subjectivity and limiting reproducibility. Here, we present a data-driven approach to parameterize the InVEST HQ model by integrating ensemble species distribution models (SDMs) with land-cover and anthropogenic threat datasets across South Korea. Using occurrence records of 29 mammal and 336 vascular plant species, we derived land-cover–specific habitat values, threat weights, maximum influence distances, and decay functions via regression and spatial analyses. The resulting HQ maps were validated against endangered species occurrence (Naemorhedus caudatus, Martes flavigula, Prionailurus bengalensis) and vegetation conservation grades, showing improved ecological relevance compared to conventional expert-based HQ maps. This approach enhances objectivity, replicability, and taxon-specific differentiation in HQ modelling, offering a scalable method that can be generalized to other regions with species occurrence data. By bridging SDMs and HQ modelling, this framework advances transparent and data-driven conservation planning, providing a methodological contribution to environmental modelling and decision-support tools.