Water as a multifaceted environmental filter of tundra vegetation

Abstract

The hydrological cycle of tundra has intensified due to accelerated environmental changes. Climatic changes affect tundra vegetation by altering water conditions (1). Plant-available water mediates climate change impacts, namely against rising temperatures and changing snow dynamics (2). Vegetation is limited by water resources, but water forms also major stress and disturbance. However, climate change impact studies often cover water inadequately in cold regions, which are assumed to be energy-limited ecosystems (3). Thus, we used statistical modelling methods to test if the inclusion of different water factors improved species distribution, species richness, and community composition models of vascular plants, mosses, and lichens in tundra. We collected occurrence data of 289 species on 378 1 m² plots in European mountain tundra. We also recorded eight environmental predictors comprising of direct and resource predictors of the species. The predictors consisted of three water-related factors – namely water resources (soil moisture level), water stress (soil moisture change), and water disturbance (fluvial accumulation and erosion) – and five other factors, namely temperature (growing degree day), nutrients (soil pH), light (radiation), cryogenic processes (solifluction and cryoturbation), and biota (dominant species coverage). All water aspects represented independent hydrological gradients of the landscape, with evidently differing responses from individual species to community level. The predictive performance of SDMs mostly improved, after considering the water-related predictors in the models: the average AUC values for vascular plants increased from 0.807 to 0.836 (P ≤ 0.01), for mosses from 0.687 to 0.727 (not significant), and lichens from 0.667 to 0.668 (ns). The predictive performance of all species richness models also improved significantly (P ≤ 0.001): mean R2 for vascular plants improved from 0.427 to 0.568, for mosses from 0.097 to 0.193, and for lichens from 0.271 to 0.335. In the NMDS plots, water disturbance and resources were within the most strongly correlating variables (R² = 0.232 and 0.191). While controlling all other key environmental gradients, these aspects proved to be crucial environmental drivers from species to community level. We showed that the three water aspects represent independent variables in tundra. Each water aspect has different impacts on the distribution of individual species, spatial variation in species richness, and community composition of three taxonomical groups. Our results highlight the role of water as a multifaceted driver of vegetation, and acknowledging uncertainties in the anticipated hydrological changes in tundra, there are possibly ecological surprises ahead for these vulnerable environments. (1) McLaughlin et al. (2017). Glob Chang Biol (2) Nabe‐Nielsen et al. (2017). Ecol Evol (3) le Roux et al. (2013). Glob Chang Biol