High landscape-scale forest cover favours cold-adapted plant communities in agriculture-forest mosaics

Aim The ongoing climate warming is expected to reshuffle understorey plant community composition by increasing the occurrence of warm-adapted species at the expense of cold-adapted species. This process has been evidenced before by a warming community temperature index (CTI) over time. However, data indicate that the local tree canopy can partly explain an observed lag between understorey plant CTI and climate warming rates, though landscape-scale forest cover effects have not yet been investigated. Here, we test the hypothesis that the amount of forest cover in the landscape lowers local CTI.Location Temperate forests in France.Time period 2005-2019.Major taxa studied Forest vascular plants.Methods We compared 2,012 pairs of neighbouring French forest inventory plots with contrasting percentages of forest cover within a 1-km radius area (landscape forest cover). We computed the difference in the CTI of the understorey communities for each pair and tested the contributions of the landscape-scale forest cover, local canopy cover, and soil conditions to the differences in CTI.Results Plots located in highly forested areas (> 80% in the 1-km area) had an average CTI 0.26 ? lower (0.81? SD) than plots in sparsely forested areas (< 30% in the 1-km area). Fifty percent of this difference was explained by landscape-scale forest cover. Bioindicated soil conditions such as pH and available nutrients, which correlated with cold-adapted species preferences, explained the remaining 50%.Main conclusions Highly forested landscapes allow cold-adapted species to survive in given macroclimatic conditions. These landscapes meet cold-adapted species' soil requirements and may cool the regional climate. Further microclimatic studies are needed to confirm the cooling capacity of landscape-scale forest cover.