Comparative seed germination traits in alpine and subalpine grasslands: higher elevations are associated with warmer germination temperatures

Abstract

Seed germination traits in alpine grasslands are poorly understood, despite the sensitivity of these communities to climate change. We hypothesise that germination traits predict species occurrence along the alpine-subalpine elevation gradient. Phylogenetic comparative analyses were performed using fresh seeds of 22 species from alpine and subalpine grasslands (1600-2400 m) of the Cantabrian Mountains, Spain (43° N, 5° W). Laboratory experiments were conducted to characterise germinability, optimum germination temperature and effect of cold and warm stratification on dormancy breaking. Variability in these traits was reduced by phylogenetic principal component analysis (phyl.PCA). Phylogenetic generalised least squares regression (PGLS) was used to fit a model in which species average elevation was predicted from their position on the PCA axes. Most subalpine species germinated in snow-like conditions, whereas most alpine species needed accumulation of warm temperatures. Phylogenetic signal was low. PCA1 ordered species according to overall germinability, whilst PCA2 ordered them according to preference for warm or cold germination. PCA2 significantly predicted species occurrence in the alpine-subalpine gradient, as higher elevation species tended to have warmer germination preferences. Our results show that germination traits in high-mountain grasslands are closely linked to the alpine-subalpine gradient. Alpine species, especially those from stripped and wind-edge communities, prefer warmer germination niches, suggesting that summer emergence prevents frost damage during seedling establishment. In contrast, alpine snowfield and subalpine grassland plants have cold germination niches, indicating that winter emergence may occur under snow to avoid drought stress.

Keywords: Altitude; functional traits; germination temperature; high-mountain flora; phylogenetic comparative methods; phylogenetic generalised least squares; physiological seed dormancy.