LB 28-244 - Do we predict right? How landscape legacies and population lags affect species distribution models.

Populations and entire communities respond often with delays to environmental changes. Species Distribution Models employ contemporaneous species and landscape data neglecting how organisms could be in a transient state between past and present conditions due to past landscapes legacies. Here, we demonstrate how the relationships between present organisms and habitat suitability can shrink or inflate in its magnitude depending on legacies from historical environmental change. Bridging temporal simulations of population dynamics with Species Distribution Models, we explore how their inference and subsequent spatial predictions are affected by altered landscape variance, change synchrony and species intrinsic growth rate. Slower growing populations, suffering from a larger delay in their response to past environmental changes, manifest inflation or shrinkage in their estimated parameters. Environmental homogenization, where overall landscape variance decreased in time, resulted in inflated slope coefficients whilst environmental heterogenization, where variance increased in time, shrank slopes estimates. Less synchrony in change magnitudes, across space or time, also produced shrinkage. Shrinkage or inflation in estimated parameter magnitudes, results in erroneous predictions. Inflated inference drives a false picture of biotic homogenization, making species seem rarer and concentrated in specific hotspots across space. Viceversa, predictions suffering from shrank estimates portray biotic homogenization, making species appear more widespread and common. Legacy effects from altered landscapes continue to influence present populations and we highlight how these processes if unacknowledged can lead to biodiversity conservation and management actions being misinformed and target at the incorrect spatial scale.