Linking Temporal-Optimization and Spatial-Simulation Models for Forest Planning

Abstract

Increasingly, resource management agencies and researchers have turned their analysis and modeling efforts towards spatial and temporal information. This is driven by the need to address wildlife concerns, landscape issues, and social/economic questions. Historically, the USDA Forest Service has used optimization models (i.e., FORPLAN and Spectrum) for timber harvest scheduling in national forest planning. Spatial details often were based on geographic strata, model constraints and mapping. Recently, more spatial analyses have been used in forest plan revisions. But many spatial analyses are limited in their flexibility to address concerns about management standards and guidelines and relationships to non-agency lands. To address these limitations, we are linking two extant models, Spectrum and HARVEST, for forest plan revision on the Chequamegon-Nicolet National Forest in northern Wisconsin. Spectrum provides an initial optimal solution given standard linear programming inputs. Resulting vegetative treatments are filtered through an interface that distributes selected treatments to HARVEST’S raster-based management areas over time. HARVEST’S parameters adjust cut sizes, buffers and adjacency considerations. HARVEST is used to simulate implementation of Spectrum schedules in a spatial context. Spatial outputs such as area of closed forest and patchiness are calculated and displayed for the planning horizon. Unsatisfactory spatial patterns can be adjusted through subsequent