Ph.D. Dissertation Abstract by Don McKenzie (1998)

Climatic changes anticipated for the next century are expected to alter the effects of fire on large-scale vegetation patterns. It is unlikely that future interactions between fire and vegetation can be predicted from knowledge of current and historic patterns. Thus, there is a need for simulation models that will produce realistic large-scale projections. Three topics were addressed in this paper: 1) the difficulties in applying current fine-scale models across coarse scales, 2) qualitative modeling at continental scales, and 3) semi-qualitative modeling at regional scales. A review of extrapolation problems revealed that a variety of methods have been developed by modelers; each has its advantages and disadvantages. A continental-scale model of vegetation changes expected from increased fire frequency suggested the large-scale patterns would be more homogeneous as a result of new dominant vegetation in fire-sensitive ecosystems. A regional-scale model that predicted fire frequency from environmental variables and vegetation types produced GIS coverages of mean fire return intervals at 1 km resolution for the Interior Columbia River Basin, and demonstrated a semi-qualitiative method that can be used in the absence of fully quantitative data.