1. Ecological processes such as seed dispersal or plant–plant interactions and environmental constraints such as climate or soil heterogeneity are known to influence establishment, and thus the spatial patterns of plant communities and populations. In this study, we hypothesized that key functional traits such as the specific leaf area (SLA), reproductive ratio (reproductive/vegetative biomass), seed mass, and maximum plant height would influence the spatial patterns of individual species in annual, gypsophilous plant communities, and that these effects would be modulated by both the soil surface structure (biocrust) and climate (precipitation) conditions.

2. We mapped the spatial patterns of all plants found in six 1 x 1 m plots (more than 1000 individuals per plot) in both the seedling (autumn) and adult stages (spring) under two biocrust experimental conditions (intact vs disturbed biocrust) during two consecutive years which were contrasted in term of precipitation (dry year and wet year). To assess the spatial patterns of seedlings and adults, we fitted four different spatial point pattern models (i.e., Poisson, inhomogeneous Poisson, Poisson cluster, and inhomogeneous Poisson cluster processes) to each of the 242 populations of the 27 most abundant species that had more than 15 individuals per plot.

3. Most seedling populations exhibited clustered spatial patterns that persisted in the adult stage, which suggests that short-distance dispersal is an adaptive trait for soil specialists such as gypsophilous plants. One-third of the populations fitted an inhomogeneous model best, but the physical structure of the biocrust was not related to them. More importantly, we found a connection between the functional strategies of species and the spatial distribution of plants. In particular, during the dry year, irrespective of the biocrust conditions, species with a high SLA and high Rep/Veg mainly exhibited clustered spatial patterns, whereas low SLA and low Rep/Veg were associated with random distributions. Species with heavy and light seed masses had random and clustered patterns, respectively. In both the dry and wet years, species with lower maximum heights had clustered patterns, whereas taller species exhibited random patterns. In addition, species with heavier seeds and greater maximum heights had the largest cluster sizes.

4. Our results confirm that the spatial patterns of seedlings and adult plants are significantly determined by the functional strategy of each species.

This data set was collected in a flat zone on the top of a gypsum hill in El Espartal (Valdemoro, Madrid, Central Spain, 40°11'11.5" N 3°37'47.0" W). Six plots (plot, "P1" to "P6"), either 1x 1 m or 1 x 1.2 m  were randomly distributed in a flat area of around 0.4 ha avoiding small holes or depressions, where the main microtopographic differences were created by the Biological Soil Crust (BSC). In three of the plots, the BSC was kept intact ("intact"), and in the other three plots, the BSC was thoroughly destroyed (i.e., "perturbed"). We monitored every plant species (sp) that emerged in each plot in two consecutive years (year), which we designated as"year1" and "year2".  Each year, we sampled each plot twice to register the spatial patterns of plants in two different stages (stage), i.e. "seedling" or "adult". In total, we performed 4 censuses (time 1 to 4) in each plot. During each census, we mapped the x and y coordinates of every plant species in the six plots. . Four functional traits, i.e., specific leaf area (SLA), reproductive to vegetative ratio (reproductive.ratio), seed mass (seed.mass) and maximum plant height (max.height)  were estimated based on at least 10 healthy undamaged individuals per species, which were randomly collected during the spring outside sampling plots surrounding our study area (see Peralta et al., 2019 for more details). All traits were measured according to the protocols described by Cornelissen et al. (2003) and Pérez-Harguindeguy et al. (2013).