CANCELLED - The spatial history of wildfire, eastern spruce budworm, and forest harvest in Ontario

Background/Question/Methods

A chief guiding principle behind forest management planning in Canada and elsewhere is the emulation of natural disturbance. This policy and management approach posits that by emulating the spatial pattern of natural disturbances, forest harvest can maintain biodiversity and ecological structure and function on the landscape. However, the extent to which spatial patterns of forest harvest mirror those of natural disturbances is not well understood, particularly for disturbances other than wildfire. Biotic disturbances, like widespread defoliation and mortality due to insects are rarely considered, despite their influence on both wildfire through the creation of dead fuel and on harvest through salvage operations. Analyses of historical patterns of disturbance have often been hampered by uncertainties in either spatial extent, temporal scope, or uncertainty in underlying data, such as that from remote sensing classification challenges. Here, we overcome these challenges with a probabilistic approach that integrates several alternative methodologies to disturbance detection. We quantify the uncertainty in disturbance detection and propagate that uncertainty in subsequent analyses.

Results/Conclusions

Here, we present an analysis of 50 years of major disturbance throughout the managed forests of Ontario, Canada. This novel database brings together numerous data sources to create probabilistic maps of forest harvest, wildfire, and eastern spruce budworm disturbance with continuous “wall to wall” coverage throughout the managed boreal and northern temperate forests of Ontario, Canada. We take a landscape ecological approach to compare the spatiotemporal patterns of individual disturbance types and cumulative disturbance on the landscape. We assess the similarity in area disturbed, landscape fragmentation, patch configuration, and heterogeneity due to different disturbance types, and ultimately the cumulative landscape disturbance. Forest harvest has been much more consistent in spatial and temporal patterns than natural disturbance, which is more variable both across ecoregions and across years. Harvest more closely resembles patterns of wildfire than spruce budworm disturbance. However, these patterns varied among forest ecoregions. Our results depended strongly on data source (remote sensing classification methodology), highlighting the substantial uncertainties in historical remote sensing products and their interpretations.