Restoring temperate grasslands often necessitates the introduction of large quantities of seeds, a process that is regulated by seed transfer zones in many countries. These zones are commonly delineated based on abiotic factors. Consequently, it remains uncertain to what extent existing seed zones represent, and thereby protect or erode the spatial distribution of genetic variation. Empirical data on the spatial genetic structure of grassland species are therefore essential to address this knowledge gap. Moreover, as seed zones are increasingly expected to provide genotypes pre-adapted to climate change, such data can also inform predictions of maladaptation and support the identification of suitable donor populations. Here, we focus on Galium album, a widespread perennial grassland species, which we sampled systematically across Germany, with an average of one population per 25 km-square area. Based on 8,348 SNP loci, we analysed the population genetic structure using Bayesian clustering. We identified four spatially coherent genetic clusters, which explained 2.43 % of genomic variation, but showed little congruence with current seed zones. Yet, seed zones still capture a significant component of spatial genetic structure (1.92 %), which is also reflected in a significant isolation by distance among zones. Seed transfer practices are increasingly challenged by climate change, shifting the adaptive requirements for populations. We performed a genotype–environment association analysis using redundancy analysis, and estimated the genomic offset, i.e. the genomic change required to maintain the current genotype-environment relationship under climate change. The genomic offset was generally moderate across Germany, even under a pessimistic climate scenario projected into the more distant future (SSP5-8.5, 2081-2100). For one of the few locations where the temporal genomic offset slightly exceeded a previously proposed threshold, we identified suitable donor regions harbouring potentially pre-adapted genotypes for targeted assisted migration, both within the same and in adjacent zones.

Introduction of large amounts of seeds is essential for restoration of temperate grasslands and is often regulated by seed transfer zones. These zones are commonly derived from abiotic parameters only. In order to evaluate seed zones as a means for the protection of genetic diversity and to avoid potential detrimental effects such as maladaptation and homogenisation of seed transfer within zones, empirical data on spatial genetic structure is paramount. Here, we focussed on Galium album, a widespread perennial grassland species, which we thoroughly sampled on average once per 25 km x 25 km in all of Germany. Based on 9,403 SNP loci, we described the genetic structure using Bayesian clustering. We identified four spatially coherent genetic clusters that were rarely congruent with the current seed zones. Therefore, current zone-based seed transfer potentially distorts and homogenises some spatial differentiation. Yet, we found significant isolation-by-distance among seed zones, showing that they still reflect a substantial part of spatial genetic differentiation. Seed transfer practice is challenged by climate change that shifts adaptive requirements for populations. We conducted redundancy analysis (RDA)-based genotype-environment association (GEA) analysis, and assessed necessary genomic turnover to maintain current levels of adaptation (temporal genomic offset). The resulting patterns suggest that climate change might impose a risk for regional adaptation in parts of Southern and Central Germany. We found that targeted assisted migration across seed zone borders might in some cases mitigate the most adverse GEA disruptions in seed zones that do not harbour suitable donor material themselves.