Simulation and optimization techniques for incorporating ecological objectives into forest harvest scheduling.

摘要

Including ecological objectives within strategic planning on forested lands is important because timber harvesting can reduce the value of these objectives. Harvesting is a valuable source of economic revenue, but changes the age-class structure of the forest, often significantly reducing the amount of late-seral stands. Late-seral stands help meet a wide variety of objectives, such as biodiversity, water quality, and, recreation. From an economic perspective, proper management of late-seral stands often is necessary to acquire “social license” for continued harvesting operations.; Most forested lands in North America are managed under some form of multiple-use, and there are many decision support tools available that integrate timber harvesting and seral objectives. However, due to the conflicting requirements of harvesting and some ecological objectives, there is growing evidence that some form of zoning may be a more efficient land-use method than multiple-use. I investigated questions of how best to define, distribute and maintain objectives requiring intact forest stands, focusing on the creation and use of decision support systems for zoning.; I first demonstrate a decision support system for landscape-level zoning that uses site attributes to create large (>5,000 ha) static zones. The Zone Allocation Model (ZAM) uses the Simulated Annealing algorithm to allocate areas into zones defined around the intensity of harvesting: Old Growth zone, Habitat zone, and Timber zone. Important ecological criteria, such as ecological representation, size, and shape of “reserves” in the Old Growth zone, are optimized relative to criteria that influence economic returns, such as site productivity and ownership, in the Timber zone. On a 1.2 million hectare landbase from coastal British Columbia, the ZAM model found solutions within 1.7% of the calculated optimum level. I then demonstrate a decision support system for small-scale zoning that uses stand attributes to reserve seral patches. The Saltus model uses simulation algorithms to create dynamic zones that move around on the landbase as disturbance creates the need to replace previously reserved stands. Saltus is demonstrated on a 139,966 hectare landbase from coastal British Columbia, as well as on computer generated landbases. The small-scale zoning method is shown to separate reserve and harvesting objectives, increasing operational flexibility. (Abstract shortened by UMI.)