Spatial and Environmental Drivers of Macroinvertebrate Diversity in Groundwater‐Dependent Pan Wetlands Across a Transboundary Aquifer System
Busisiwe P. Gumede, Linton F. Munyai, Naicheng Wu, Collins Oduro, Tatenda DaluABSTRACT
Groundwater‐dependent pan wetlands are keystone habitats for aquatic biodiversity in semi‐arid landscapes, yet the processes that structure their invertebrate communities remain poorly resolved. We surveyed 221 pan wetlands across the Khakhea–Bray Transboundary Aquifer (KBTA), South Africa, a landscape of low relief (1027–1231 m a.s.l.), and asked whether macroinvertebrate assemblages track local physicochemical conditions, landscape position, or both. Pans were assigned to four elevation classes, used here as a surrogate for hydrogeomorphic setting and its hydroperiod consequences rather than as a climatic elevation gradient, because the total relief is an order of magnitude smaller than in studies where elevation acts as a direct driver. Waters were predominantly neutral to alkaline (mean pH 7.6–8.6), consistent with a strong groundwater signature. Dissolved oxygen, water temperature, water depth, ammonium, nitrate and phosphate differed significantly among classes, whereas pH, oxidation–reduction potential, conductivity, total dissolved solids and surface area did not. Assemblage composition differed among classes, but the effect was weak: global ANOSIM detected no separation, and similarity percentage analysis attributed between‐class dissimilarity (67.9%–73.4%) to shifts in the relative abundance of a few dominant taxa, chiefly Anostraca, Spinicaudata and Notostraca, rather than to species turnover. Taxa richness differed significantly among classes, whereas abundance, Simpson, Shannon–Wiener and evenness did not, and within‐class variation exceeded among‐class differences. Shallow, short‐hydroperiod pans were dominated by large branchiopods, whereas deeper, longer‐hydroperiod pans supported proportionally more predatory aquatic insects and molluscs, consistent with hydroperiod and predation, rather than elevation, acting as the operative filters. Boosted regression trees ranked geographic position, particularly longitude, above local physicochemical variables as a correlate of taxa richness; because the surveyed pans are strongly clustered in space, this result is interpreted as evidence of unmeasured, spatially structured habitat variation rather than as a predictive relationship. Conserving these systems will require maintaining groundwater regimes and the full range of hydroperiods across the pan network under coordinated transboundary management, and monitoring assemblage composition rather than richness alone.