Evaluating Genomic Offset Predictions in a Forest Tree with High Population Genetic Structure

Genomic offset models are increasingly popular tools for identifying populations at risk of maladaptation under climate change. These models estimate the extent of genetic change required for populations to remain adapted under future climate change scenarios but face strong limitations and still lack broad empirical testing. Using 9,817 single-nucleotide polymorphisms (SNPs) genotyped in 454 trees from 34 populations of maritime pine, a species with a marked population genetic structure, we found substantial variability across genomic offset predictions from different methods, SNP sets, and general circulation models. Using five common gardens, we mostly found positive associations between genomic offset predictions and mortality, as expected. However, contrary to our expectations, we observed very few negative monotonic associations between genomic offset predictions and height. Higher mortality rates were also observed in national forest inventory plots with high genomic offset, but only for some methods and SNP sets. The differing genomic offset patterns produced by the best-validated methods across the maritime pine range hindered drawing definitive conclusions for the species. Our study demonstrates the imperative of employing different methods and validating genomic offset predictions with independent data sources before using them as reliable metrics to inform conservation or management.