Exploring patterns of riparian vegetation assemblages to reveal human-induced landscape change along southern semi-arid Mediterranean streams

Abstract

This study investigates the composition and ecological structure of riparian vegetation in southern Mediterranean semi-arid streams (north-eastern Algeria) to test how environmental filtering and anthropogenic pressures shape community assemblages. Vegetation relevé plots and a combination of landscape metrics have been used to quantify the spatial configuration and express the ecological condition of these riparian areas. Multivariate statistical process including Jaccard’s distances matrix, hierarchical clustering procedures, Distance-based Linear Models (DistLM), and Generalized Linear Models (GLMs) were used to unveil the ecological status of these areas. We have identified a total of 25 woody plants (trees and shrubs) distributed in two dissimilar assemblages (SIMPROF test) promoted mainly by topographical variables, proximity to human activities, damming and aridity intensity, which explained 44.4% of total variance in woody composition according to DistLM ordination. Group 1, undisturbed landscape dominated (IndVal > 0.7, p < 0.001) by natural riparian plants such as Populus alba (stat = 0.874), Cytisus purgans (0.788), Juniperus oxycedrus (0.788), and Rubus ulmifolius (0.766), mainly align with sites located at mid to high altitudes, further from human settlements, and in undammed streams. Group 2, a human induced landscape dominated by Rhus tripartita, and associated with lower altitudes, close proximity to human development, and the presence of dams. Our best‐supported GLM showed that woody species richness increased at sites with permanent stream flow (β = 0.33, p = 0.01) and higher elevation (β = 0.25, p = 0.009), but declined in riparian zones affected by damming structures. The significant effect of proximity to human habitation on species composition, but not on species richness (not retained in the best model), indicating that the complementary use of diversity metrics (species richness and composition) is crucial to properly capture the effects of human disturbance.