Determining the Impact of Moose Browsing on the Vegetation and Natural Fire Regime at Isle Royale National Park

Principal Investigators:
Ron Moen, Department of Forest Resources, 612-624-9796  

Project Personnel:
Jerrilyn L. Thompson, Department of Forest Resources and CPSP

Project Duration:
October 2000 – September 2002

Funding:
Isle Royale National Park, NPS

Summary:
Herbivory by ungulates is a management issue in many national parks. The southern boreal / northern hardwoods vegetation of Isle Royale National Park has been dramatically affected by moose herbivory. There has been very little fire activity since the 1970s.Several thousand years would be required to burn most of the island under the current rate of fire spread, much slower than the natural fire return interval in southern boreal / northern hardwood ecosystems. The study assesses the effect of moose herbivory and forest composition on fire spread, vegetation, and species diversity of Isle Royale under current and predicted climatic conditions. It will link park vegetation, soils, and topographic data with a forest growth model to predict the effect of moose herbivory on future vegetation patterns under different fire intensity scenarios. The cumulative effects of both moose herbivory and fire patterns will be predicted for the entire island. This will provide an evaluation of the effects of moose browsing on the natural fire interval and the possible effects of a prescribed burn program on vegetation patterns in the park.

The Linkages forest growth model has been modified to simulate Isle Royale forests and the components necessary to simulate fire have been added to develop an expanded model called “FireLink.” FireLink simulates fuel accumulation, calculates fuel loading, imports stand-specific initial conditions and implements moose browsing. Sensitivity analysis simulations indicate that selective browsing on deciduous species by moose affects vegetation more than soil type. The strength of the predicted effect depends on the species composition of each forest type. The implementation of moose browsing and selective regeneration results in forest composition and characteristics similar to the existing forests of Isle Royale. We compare predictions of forest composition and fuel accumulation to measurements made on northern hardwood and boreal forest types. The figures illustrate predicted aboveground biomass and 1-hour time lag fuel load for the aspen-birch-conifer forest type if moose are absent, and when moose are present at medium and high density. The effect of selective herbivory on deciduous species by moose results in a lower biomass, and a lower fuel load, particularly in the first 100 years of the simulation. Later, as the forest matures, moose browsing has less of an effect because canopy trees are out of reach of browsing moose.

The current phase of the modeling work is to implement FireLink on a spatially-explicit representation of Isle Royale that uses the vegetation map for Isle Royale prepared under the USGS/BRD-NPS Vegetation Mapping Program. These data will be used to model current and future vegetation patterns on Isle Royale and to assess the effect of fires on vegetation, species diversity, and the moose population under current and future climatic conditions.