Broad ecosystem based classifications are increasingly applied as a context to consider, understand, and manage biodiversity. The need for more spatially explicit, repeatable, transferable, transparent, and defensible environmental regionalization has become apparent. Increased computing power, sophisticated analysis software, and the availability of spatially explicit descriptions of the environment, principally derived from Earth observation data, have facilitated the development of statistical ecosystem regionalizations. These regionalizations are desired to produce environmentally unique ecoregions to provide the basis for stratification for ongoing biodiversity monitoring efforts. Using a suite of indicators of the physical environment, available energy such as vegetation production, and habitat suitability all derived from remote sensing technology at 1 km spatial resolution, we undertook an environmental regionalization using a two-stage multivariate classification of terrestrial Canada. A relatively large number of classes were initially derived (100) and a hierarchical clustering approach was then applied to derive a 40 level classification. These clusters where then used to assess which clusters were the most dissimilar to the majority thus providing indication of the most unique environmental domains across Canada. Secondly, a 14 class stratification was then produced to emulate the current ecozone stratification commonly used in Canada. Results indicated that a number of unique clusters exits across Canada, specifically the forest/urban-industrial/cropland mosaic in the southern portion of Ontario, the mixed wood forests in south–central Ontario and western Quebec, the foothills of south western Alberta, regions of the southern Arctic and the northern Boreal shield (particularly the areas south of Hudson Bay and Labrador). A resemblance between the 14 class stratification and the ecozone classification for Canada is evident; locations of within and between ecozone heterogeneity are also indicated. A critical key benefit of utilising ecoregions quantitatively using key indicators, such as those derived from remote sensing observations, is the capacity to establish, and quantify, how well particular networks of sites, or plot locations, represent the overall environment. As such, the incorporation of these types of methods, and remotely derived indicators, into biodiversity assessment is an important area of ongoing research.