Unsafe but Ecologically Diverse: Water Quality, Planktonic Communities, and Cyanotoxin Risks in Multi-Use Waterpans of Semi-Arid Narok County, Kenya

Waterpans are shallow, man-made rainwater harvesting structures that are the primary water source for domestic use, livestock, and wildlife across African drylands. Despite their proliferation, they remain among the least-studied freshwater ecosystems globally. This study provides an integrated assessment of physicochemical water quality, phytoplankton and zooplankton communities, and cyanobacterial contamination risk across twelve waterpans representing agricultural-livestock (ALV), livestock (LIV), and livestock-wildlife (LWL) land uses in Narok County, Kenya, sampled during dry and wet seasons. Multivariate analyses including PERMANOVA, Elements of Metacommunity Structure (EMS), beta diversity partitioning, generalised dissimilarity modelling (GDM), and temporal beta diversity indices were used to identify the relative roles of season, land use, waterpan morphometry, and environmental filtering in structuring plankton metacommunities.

Waterpan identity, more than land-use, explained the greatest share of water quality variation (PERMANOVA R² = 0.307, p < 0.001), followed by season (R² = 0.108) and land use (R² = 0.104). Land-use effects on water quality were more pronounced in the dry season, when LWL waterpans showed elevated oxygenation and electrical conductivity, while ALV and LIV waterpans had higher dissolved inorganic nitrogen concentrations. Phytoplankton and zooplankton richness both increased significantly during the wet season, though density and evenness were broadly stable. Cyanophyta and Dinophyta co-dominated phytoplankton across all land-use types, with Cyanophyta and Bacillariophyta most prominent during the dry season; the prevalence of toxin-producing Microcystis aeruginosa and Anabaena spp. indicates a persistent risk of microcystin contamination. Season was the dominant structuring factor for both plankton guilds (phytoplankton: R² = 0.064, p = 0.001; zooplankton: R² = 0.075, p = 0.001), with land use contributing a smaller but significant effect in the full model; however, land-use centroid models were not significant, indicating that land-use differences reflected community dispersion rather than compositional centroids. Dry-season assemblages were markedly more heterogeneous among waterpans (phytoplankton total β = 0.799; zooplankton total β = 0.754) than wet-season assemblages, with turnover the dominant component in both seasons. The wet-to-dry seasonal transition generated the highest temporal community turnover in both guilds (p ≤ 0.017), underscoring the critical ecological role of desiccation. GDM identified ammonium, temperature, waterpan area, SRP, and dissolved oxygen as the primary nonlinear drivers of compositional change, with deterministic environmental filtering outweighing spatial isolation in both guilds. These findings reveal that waterpans in ASALs are simultaneously microbiologically unsafe, seasonally dynamic, and ecologically diverse, and that site-specific rather than land-use-based management frameworks are needed. Integration of waterpans into national water quality monitoring, One Health policy, and nature-based restoration programmes, including livestock exclusion fencing and riparian buffer planting are some of the management options urgently required.