Soil effects on plant distributions and potential migration in Eastern North America

Abstract Many plant species are predicted to migrate poleward in response to climate change. Species distribution models (SDMs) have been widely used to quantify future suitable habitats and potential migration distances, but SDM studies typically neglect soil properties, despite their importance for plant fitness. In this study, we built three SDMs – one with only climate predictors, one with only soil predictors, and one with both climate and soil predictors – for each of 1870 plant species in Eastern North America (ENA), in order to investigate the relative importance of soil properties in determining plant distributions and potential migration under climate change. While climate variables were the most important predictors (mean relative importance = 0.573), soil properties also had a substantial influence on continental-scale plant distributions (mean relative importance = 0.369). Given spatial correlations between climate and soil variables, considerable variation in distributions cannot be uniquely attributed to one or the other. Under future climate scenarios, models including soil predicted much smaller northward shifts in distributions than climate-only models, strongly suggesting that high-latitude soils are likely to impede ongoing plant migration. Our findings highlight the necessity of incorporating soil properties into models and predictions for plant distributions and migration under environmental change.