Novel genomic offset metrics account for local adaptation in climate suitability forecasts and inform assisted migration

Abstract

Local adaptation is increasingly being integrated into macroecological models, offering an evolutionary perspective that has been largely missing from climate change biogeography. Genomic offsets, which quantify the disruption of existing genotype-environment associations under environmental change, are an informative landscape genomic tool that allows for the incorporation of intra-specific adaptive differentiation into forecasts of climate suitability and thus management planning. Gradient forest (GF), a method originally developed to model community turnover along environmental gradients, is now the most commonly used approach in genomic offset estimation. However, major hurdles in the application of GF-derived offsets are (1) an inability to interpret the absolute magnitude of genomic offsets in a biologically meaningful way and (2) uncertainty in how genomic offsets compare to established species-level approaches like Ecological Niche Models (ENMs). We used both ENMs and novel, genomic offset metrics derived from GF modeling of genomic variation along climatic gradients to assess the climate change vulnerability of red spruce (Picea rubens), a cool-temperate tree species endemic to eastern North America. We show how genomic offsets can be standardized relative to contemporary genomic variation across the landscape to better represent their biological significance and facilitate comparisons among studies. In three common gardens, we found a significant negative relationship between standardized genomic offsets and red spruce growth and higher explanatory power for standardized offsets than (raw) climate transfer distances. We also derived new threshold-based metrics that we term Donor and Recipient Importance and which quantify transferability of propagules between donor and recipient localities while minimizing disruption of genotype-environment associations. This approach leverages landscape genomic information to account for local adaptation when predicting climate suitability. ENMs and our novel genomic offset metrics largely agreed in forecasting drastic northward range shifts. Combining several offset-based metrics, we show that the projected northward shift of suitable climate mainly applies to populations located in the center and northern parts of the current range, whereas southern populations might be able to persist in situ owing to specific local climate adaptations. The novel metrics thus yield refined, region-specific prognoses for local persistence and show how management could be improved by considering assisted migration.