Isoscapes reveal ENSO-driven variability in evaporation and source water contributions in a semi-arid urban tap water supply

This study tests the hypothesis that ENSO drives spatiotemporal changes in isotope patterns and water source contributions within Gaborone's (Botswana) tap water system. Spatially distributed tap water samples were collected between 2022 and 2024, along with other hydrological endmembers. El Niño (warmer/drier) conditions led to higher (lower) δ 18 O ( d -excess), while La Niña (cooler/wetter) conditions led to lower (higher) δ 18 O ( d -excess), reflecting changes in atmospheric moisture sources and evaporation dynamics. Cluster analysis identified four isotopically distinct groups. Cluster 1, with the lowest d -excess (−19.01 to −8.28 ‰), was dominant in the western (2023) and eastern (2023/24) regions. Cluster 2 (−8.13 to −1.41 ‰) was prominent in the south and southeast (2022), and later expanded to the central, northern, and western regions for all other seasons. Cluster 3 (−1.73 to +4.85 ‰) prevailed in the north and south-central regions (2022/23). Cluster 4 (+5.59 to +14.08 ‰) was minimally represented and linked to groundwater use. MixSIAR showed Dikgatlhong Dam as the primary source (46.2 % ± 18.8 %), followed by Gaborone and Bokaa Dams, groundwater, and Letsibogo Dam. Isotopic changes during El Niño highlight intensified evaporative losses and the use of diverse source-water strategies to meet urban demand, offering a valuable framework for examining ENSO-driven variability in tap water systems worldwide. Isotope mixing analysis of water source contributions within Gaborone's (Botswana) tap water distribution system. • ENSO modulates δ 18 O and d -excess in tap water, indicating shifts in atmospheric moisture source and evapoconcentration. • Clustering of δ 18 O and d -excess reveals spatiotemporal variability in urban water source blending. • MixSIAR modeling quantifies Dikgatlhong Dam (∼46 %) as primary input to Gaborone's tap water isotopic signal.