Corroboration of biogeoclimatic ecosystem classification climate zonation by spatially modelled climate data

Abstract

The biogeoclimatic ecosystem classification (BEC) method for distinguishing areas of reasonably homogeneous macroclimate has been used in British Columbia for over 20 years. Because of the paucity of actual long-term climate data, the method used other means to map climate. We tested how well the BEC climate units could be discriminated from one another using spatially modelled climate data. We tested the ability of climate data to distinguish three units for each of four climatically different zones at two levels of the climatic classification using discriminant analysis. For each analysis, 60 points were randomly selected from within the boundaries of the mapped unit and climate data were generated by ClimateBC. Even at the finest level of the mapping, over 70% of the randomly selected points were correctly classified according to the mapped unit based on selected climate variables. A large proportion of the misclassified points were within 1 km horizontal distance or 100 m elevation of the boundary and are typically climatically transitional areas. We recommend that the BEC climate unit should form the basic unit for examining climate change at multiple scales from the provincial scale to the scale of watersheds or basins, and that further analysis be conducted to both improve biogeoclimatic unit mapping and climate models.