A farm-level framework for integrating ecosystem service supply, scarcity, and feasibility: The case of crop pollination

Biodiversity decline in agricultural landscapes is well documented, yet decision-support approaches rarely connect ecologically plausible ecosystem service (ES) diagnostics with farm-level constraints and implementable management decisions. This article introduces a codebook-based, farm-level framework integrating ES supply, multidimensional scarcity, and implementation feasibility into a transparent decision-support logic. It addresses three persistent limitations: the disconnect between spatial or biophysical ES supply assessments and farm-level decision-making; the reduction of relevance to monetary valuation, which obscures a farm's dependence on the ES and its exposure to risk; and the insufficient integration of feasibility constraints such as labor, costs, and coordination. Derived from a thorough literature review, the framework combines qualitative, rule-based scorecards for ES supply potential, a multidimensional scarcity concept capturing ecological, functional, spatial, and institutional constraints, and a structured cost module reflecting farm-operational feasibility, all governed by a transparent codebook that ensures traceability and auditability under realistic data conditions. The framework is operationalized for crop pollination and tested, as a proof of concept, on four southwestern German farms through single-farm interviews, without additional biophysical data or specialist expertise, the conditions most farms and advisory services in agri-environmental schemes actually face. Results show that supply, scarcity, and feasibility do not necessarily align, and that implementation barriers can outweigh ecological scarcity. By linking ecological diagnostics to farm-level decision logic, the framework offers a transferable approach to biodiversity-related decision-making, with the scarcity dimension requiring re-specification for services with strong public-good characteristics. It presents a conceptual framework and a first tentative application, not a statistically validated tool.