Statistical models of landscape pattern metrics, with applications to regional scale dynamic forest simulations

Forest managers in Canada need to model landscape pattern or spatial configuration over large (100,000 km(2)) regions. This presents a scaling problem, as landscape configuration is measured at a high spatial resolution, but a low spatial resolution is indicated for regional simulation. We present a statistical solution to this scaling problem by showing how a wide range of landscape pattern metrics can be modelled from low resolution data. Our study area comprises about 75,000 km(2) of boreal mixedwood forest in northeast Alberta, Canada. Within this area we gridded a sample of 84 digital forest cover maps, each about 9500 ha in size, to a resolution of 1 ha and used FRAGSTATS to compute a suite of landscape pattern metrics for each map. We then used multivariate dimension reduction techniques and canonical correlation analysis to model the relationship between landscape pattern metrics and simpler stand table metrics that are easily obtained from non-spatial forest inventories. These analyses were performed on four habitat types common in boreal mixedwood forests: young deciduous, old deciduous, white spruce, and mixedwood types. Using only three landscape variables obtained directly from stand attribute tables ( total habitat area, and the mean and standard deviation of habitat patch size), our statistical models explained more than 73% of the joint variation in five landscape pattern metrics ( representing patch shape, forest interior habitat, and patch isolation). By PCA, these five indices captured much of the total variability in the rich set of landscape pattern metrics that FRAGSTATS can generate. The predictor variables and strengths of association were highly consistent across habitat classes. We illustrate the potential use of such statistical relationships by simulating the regional, cumulative effects of wildfire and forest management on the spatial arrangement of forest patches, using non-spatial stand attribute tables.