Canadian fractional landcover maps from 1985 onward

Abstract

Canada’s northern ecosystems are undergoing rapid environmental change driven by climate warming, shifting disturbance regimes, and vegetation transitions. However, national-scale datasets describing non-treed and sparsely treed environments remain limited. This study presents a pan-Canadian dataset of annual fractional landcover at 30-m spatial resolution for the period 1985–2025, extending the Spatialized Canadian National Forest Inventory (SCANFI 1 ) framework to include both treed and non-treed surface components. The dataset integrates imagery from Landsat, Sentinel-2 and Very High-Resolution data to estimate the annual fractional cover of water, rock, soil, burn scars, lichen, herbaceous vegetation, low shrubs, tall shrubs, broadleaf trees, and coniferous trees. Fractional estimates were produced using machine learning techniques and harmonized within a vertical surface projection framework to ensure that all pixel-level components sum to 100%. The resulting dataset provides temporally consistent information for monitoring vegetation dynamics, post-disturbance recovery, and land-surface composition across Canada.