Multi-objective optimization for timber harvest management incorporating wildlife habitat goals

Abstract

Context: The forestry industry provides important goods, services and economic benefits, but timber harvest can adversely impact ecosystem services, including wildlife habitat. Timber harvest planning can integrate wildlife habitat quality through multi-objective optimization for timber harvest and wildlife habitat suitability. Objectives: Our objective was to develop a method to find optimal solutions for timber harvest and wildlife habitat suitability individually and concurrently, then apply the method to Roosevelt elk (Cervus elaphus roosevelti) on Afognak Island, Alaska. Methods: We developed three seasonal habitat suitability models using elk locations and landscape variables including historical timber harvest on Afognak Island, Alaska. We used threshold-accepting optimization over a 50-year planning horizon to maximize timber harvest yield and habitat suitability in each season, then used multi-objective goal-deviation optimization to simultaneously maximize timber harvest volume and seasonal habitat suitability. Results: The optimal solution for timber yield decreased seasonal average habitat suitability by 5.7%. Elk habitat suitability and corresponding optimal solutions varied seasonally; elk generally selected open landcovers and early- to mid-successional timber stands over late-successional and mature stands. Therefore, in the optimal solutions, stands were harvested before they reached maximum volume and few stands were harvested in early planning periods, resulting in a seasonal average loss of 17.5% yield. Multi-objective optimization decreased seasonal average suitability by 3.9% and yield by 1.4% compared to single-objective optimization. Conclusions: Our multi-objective optimization approach that incorporates data-driven habitat suitability models using open-source software can enable managers to achieve desired quantity and quality of wildlife habitat while providing for resource extraction.