Plant communities are not merely snapshots of species located in the same area. Instead, they are dynamic entities connected through complex biotic interactions. Understanding the interplay of processes that bind species within a community is vital for comprehending species maintenance and may assist us in predicting how communities respond to the perturbations of a changing world. This Ph.D. aims to unveil how abiotic conditions, plant strategies, and biotic interactions shape plant communities growing in stressful environments. Plant species can thrive in stressful environments through contrasted strategies, either showing species-specific adaptive strategies or reliance on facilitative interactions with neighboring species. While some species rely on adaptive traits for overcoming specific stressful conditions, although it may imply unaffordable costs in other environments, others stress-sensitive species may show generalist strategies to deal with stressful environments and take advantage of facilitative interactions with neighboring species. Facilitation occurs when the presence of one species enhances the establishment of another, and as with any interaction, its outcomes vary over time due to changes in the environment, the ontogenetic state of the involved plants, or the intervention of third species through indirect interactions. Therefore, the maintenance of stress-sensitive species, or the co-occurrence of adult plants in general, should provide long-term benefits for both of them or a more equitable performance when they grow associated. For plant species that rely on facilitation, their resilience to perturbations may be affected by their dependence on specific nurses (i.e., species from which they benefit). Here, the more alternative nurses they have, the more likely they will switch nurses and thus survive (i.e., facilitation rewiring). Here, we assess different processes shaping plant communities in stressful environments by exploring (a) contrasted species-specific strategies to thrive in these environments and the potential trade-offs among them, (b) potential mechanisms of long-term co-occurrence among adult plant species in these communities, and (c) the limitations to switch among nurse species in environments with contrasted stresses. To do so, we explore plant communities in two semi-arid environments in Southeast Spain varying in stress levels. Specifically, we studied plant communities growing in stressful gypsum outcrops and their immediate surrounding of limestones, which constitute a milder environment. To address our aims, we approached the exploration of these communities from a broad perspective. We quantify species' affinity to the stressful substrate as their percentage of cover in the stressful environment relative to the milder environment. Then, we relate this affinity to several species traits and test whether species dependence on facilitation is related to this affinity. To do so, in both communities, we quantify the tendency of each species to recruit under nurse plants (i.e., species reliance on facilitation) and whether it depends on the species' affinity for the substrate. To assess weather facilitative effects remain throughout the lifespan of the interacting species, we focused on adult individuals. We explored whether plants of two different species show a more equitable performance (i.e., size asymmetry) when they are co-occurring in multispecific patches compared to their conspecifics living in isolation. We propose this as a mechanism of long-term co-occurrence, as it may reduce the likelihood of species competitively outperforming each other. Finally, we assessed whether facilitation rewiring is constrained and which nurse traits produce such constraints in contrasted environments. Our findings show that species living in stressful soils employ various coping strategies to face such conditions. On the one hand, species with high gypsum affinity showed traits that allowed them to overcome stressful gypsum constraints. However, we did not find collateral costs to their success in other environments as a result of these adaptations. On the other hand, rare and stress-sensitive species rely on facilitation for establishing, resulting in multispecific patches where species balance their performance, enhancing their co-occurrence in the long term. Finally, we explore general patterns in facilitation rewiring. Nurses’ traits can shape facilitative interactions, limiting the rewiring capacity to nurse species harboring similar traits. Nurses’ species selection varied between environments. Specifically, we found that in gypsum environments, facilitation rewiring was constrained to nurse species with gypsum-specialized traits, while in milder environments, rewiring occurs with nurse species that are close relatives of the preferred nurses. Altogether, these results show that functional or phylogenetic species redundancy is key to ensuring the survival of species requiring facilitation, thus contributing to the resilience of the entire community.