ABSTRACT Questions Rare species play a key role in ecosystems that far exceeds their numbers, especially in the ecologically sensitive grassland systems of Gannan Prefecture, China. However, with the intensification of human activities, the rate of species extinction far exceeds expectations, and rare species are likely to be the first to disappear. However, little is known about the impact of rare‐species loss on ecosystem functions. Therefore, it is crucial to study how ecosystems respond to the loss of rare species. Location The alpine and subalpine meadows of Gannan, situated on the eastern edge of China's Qinghai–Tibet Plateau. Methods Rare species were defined by integrating three complementary dimensions of rarity (local abundance, geographic range size and habitat breadth) into a composite rarity index calculated separately for each study area. The index ranged from 0 to 1, and species with values between 0.5 and 1 were classified as rare. We then analysed the distribution of rare‐species traits in functional space across 81 subplots and used Monte Carlo simulations to evaluate how rare‐species loss altered community functional redundancy (FRe) and stability. Results Compared with common species, rare species exhibit low abundance, yet they occupy larger functional niches and possess higher FRe. In addition, the synergistic loss of rare and common species triggers non‐linear synergistic effects. When rare species suffer a 30% loss and common species a 25% loss in synergistic loss, the probability of community collapse increases significantly, with a risk 2.3 times higher than that of a single‐species loss pattern. Conclusions Our research indicates that the loss of rare species reduces FRe and stability within communities, with these effects being driven primarily by the richness effect. Although rare species are not abundant, they make disproportionately important contributions to maintaining FRe. In the grasslands of Gannan, China, conservation efforts should therefore prioritise rare species and their unique functions.
Yuan et al. (Sun,) studied this question.