Planning for a variable future: Environmental stochasticity, species richness, and protected areas in Canada

Priority areas are typically identified based on mean conditions, while ignoring variance around the mean (here termed ‘stochasticity’). This is problematic, as high environmental stochasticity can increase extinction risk and reduce the effectiveness of protected areas. Identifying habitats that are both productive and stable should be essential in the design of protected areas. Using more than four decades of daily Normalized Difference Vegetation Index data, we generated spatially-explicit estimates of both the mean and variance in environmental productivity across Canada between 1981 and 2025. From these models, we found that environmental stochasticity shows strong spatial structure and has been steadily increasing over the past four decades. There was also no clear relationship between stochasticity and protection status, suggesting that Canada’s network of protected areas are not well-buffered against this increasing stochasticity. Additionally, we identified a strong negative relationship between environmental stochasticity and species richness, which suggests that the increasing stochasticity risks driving widespread species declines. Promisingly, we identified 2,709,580 km\textsuperscript{2} of currently unprotected land that may minimise the impact(s) of growing stochasticity. While the establishment of protected areas is a complicated process that requires local stakeholder engagement, this work provides a framework for incorporating environmental stochasticity into conservation planning.