Novel climates: Trajectories of climate change beyond the boundaries of British Columbia's forest management knowledge system

The non-stationary climates of the 21st century are compelling forest managers to seek non-local species, provenances, and silvicultural regimes that are better suited to the anticipated future climates of their operating areas. Ideally, forest managers can source this information from climate analogs within their jurisdictions, but the emergence of unfamiliar climates is a distinct possibility with particular challenges. Here, we present an assessment of the emergence of mid-21st-century climates with no analog in the 20th-century climates of British Columbia (BC), and the extent to which these novel climates are described by climate analogs elsewhere in North America. We use a recently developed linear method of novel climate detection in parallel with Random Forest classification to evaluate the robustness of novel climate inferences. Our results suggest that a majority of the province's area will remain free of novel climates over this time period, and therefore that BC's ecological knowledge system, the Biogeoclimatic Ecosystem Classification, can remain the dominant source of climate analogs for mid-21st-century forest management planning horizons. Nevertheless, we detected a robust pattern of climates that are novel to BC in mid-21st-century climate projections at low elevations in the coastal, southern interior, and northeastern regions of the province. There appears to be potential to inform forest management in some of these novel climates with analogs from adjacent states and provinces. We demonstrate that extrapolations into novel climates typically understate the magnitude of climate change and modeling uncertainty, creating a false impression of robust predictions in locations where model performance is poorest. By identifying portions of their landscapes that are prone to emergence of novel climates, forest managers can avoid management errors and prioritize the search for analogs beyond the boundaries of their knowledge systems.