Species interactions such as facilitation and competition play a crucial role in driving species range shifts. However, density-dependence as a key feature of these processes has received little attention in both empirical and modelling studies. Herein, we used a novel, individual-based treeline model informed by rich in-situ observations to quantify the contribution of density-dependent species interactions to alpine treeline dynamics, an iconic biome boundary recognized as an indicator of global warming. We found that competition and facilitation dominate in dense versus sparse vegetation scenarios, respectively. The optimal balance between these two effects was identified at an intermediate vegetation thickness where the treeline elevation was the highest. Further, treeline shift rates decreased sharply with vegetation thickness and the associated transition from positive to negative species interactions. We thus postulate that vegetation density must be considered when modeling species range dynamics to avoid inadequate predictions of its responses to climate warming.

This dataset includes the code of the Sygera Treeline Model and the raw data for model results for validation and model experiments. 

The Sygera Treeline Model is a new individual-based treeline model, which shares some key features and processes with the polar treeline simulator LAVESI (v. 1.01) (Kruse et al. 2016), and two classical forest gap models, FORCLIM (Bugmann 1994) and JABOWA (Botkin 1993). It is designed to explore how alpine treeline formation and dynamics under different vegetation scenarios with the background of climate change. In this model, trees were placed on a hypothetical mountain slope where the temperature gradually decreased with elevation, and had been impacted by alpine vegetation. All trees experience three physiological processes of establishment, growth, and mortality, and these processes were influenced by abiotic (temperature, drought) and biotic (intra- and interspecific interaction) environmental factors. 

The another file includes the the raw data files that were analyzed to produce the statistics and figures reported in the paper. It includes model validation and model experiments.

We validated only the last 50 years (1963–2012) of treeline shifts by comparing them with observed shift in test sets (plots LZ1, LZ3, RW2, RW3, BM1, and BM2). Average relative treeline elevation (in 2012) and elevational treeline shifts were obtained after 100 repeated simulations. Besides, a spatial analysis was performed by computing Ripley’s K(t) (Ripley 1977) from the x and y coordinates of adult trees in both simulated and measured treeline plots. The complete spatial randomness was also tested based on chi-squared tests, which indicate a cluster pattern when P < 0.05. The R package “spatstat” version 3.0-7 (Gabriel 2017) was used for all spatial analyses. 

In order to exclude direct effects of climate warming, we quantified the independent contributions of interspecific facilitation and competition, as well as their combined effects on treeline dynamics. Manipulative experiments were carried out with different model scenarios and vegetation scenarios in LZ2, as a typical treeline site. Firstly, we ran three model scenarios by gradually adding facilitation and competition processes into the model. In scenario A, we assumed that treeline dynamics were only determined by climate, thus we excluded the processes of vegetation facilitation and competition (i.e., the formula 13 in Appendix). The facilitation of alpine vegetation on tree growth and establishment was simulated in scenario B. Finally, both facilitation and competition were included in scenario C, with the formula 13 in Appendix set. Next, vegetation scenarios with different vegetation cover (0–1, with an interval of 0.1) and heights (0–300 cm, with an interval of 20 cm) were integrated into scenarios B and C to quantify the density-dependent species interactions. Various model scenarios and vegetation scenarios were simulated 50 times to reproduce the average treeline position in 2012 and treeline shift elevations over the last 50 years. Finally, the difference between scenarios A-C determines the total effect of interspecific interactions by eliminating the direct effects of climate. Similarly, the differences in treeline elevations and shifts between scenarios B and A represent the contribution of facilitation, whereas C-B represents the contribution of competition.