Future Quantitative Water Security Under Climate and Demographic Change in Selected Peruvian Tributaries of the Lake Titicaca Basin: A Spatially Explicit SWAT–CMIP6–SSP Assessment
Jhonatan Hinojosa Mamani, Julio Cesar Huanca Marin, Yalmar Temistocles Ponce Atencio, Jose Adrian Ramos Choque, Juan Mauricio Pilco Churata, Eva Clara Blanco Yupanqui, Adderly Mamani Flores, Abraham Coila Torres, Gerardo Hugo Flores Mestas, Jorge Apaza TiconaClimate change and demographic dynamics may jointly alter future water availability and demand in high-Andean watersheds, yet their combined spatial effects remain insufficiently quantified in the Lake Titicaca Basin. This study integrated the Soil and Water Assessment Tool (SWAT), bias-adjusted CMIP6 climate projections, Shared Socioeconomic Pathway (SSP) population projections, and a network-consistent demand-to-availability Water Stress Index (WSI) to assess quantitative water security in three selected Peruvian tributary basins of Lake Titicaca—the Ramis, Coata, and Huancané—during 2025–2100. SWAT simulations were forced with ACCESS-CM2 and MPI-ESM1-2-HR under SSP1-2.6 and SSP3-7.0, while spatially explicit FuturePop projections were used to estimate domestic water demand across 177 sub-basins. Projected water availability showed substantial interannual and inter-model variability, with Huancané exhibiting the highest supply, Ramis intermediate conditions, and Coata the lowest availability. Under SSP1-2.6, population declined by 2100 in all three basins, whereas SSP3-7.0 increased population by 64.5% in Ramis, 17.2% in Coata, and 66.9% in Huancané. Coata exhibited the narrowest basin-scale supply–demand margin and reached a slight deficit of −0.29 hm3 year−1 by 2100 under SSP3-7.0. However, the network-consistent basin-outlet WSI remained below the critical threshold of 1 in all three basins. Spatially, 98.3–98.9% of the 177 sub-basins were classified as very low stress across the representative years and scenarios under the principal 10% non-allocable streamflow reserve assumption, while WSI ≥ 1 hotspots remained limited to 0–2 sub-basins and were concentrated in Coata. Increasing the non-allocable reserve from 10% to 30% changed hotspot occurrence by no more than one sub-basin at any evaluated horizon. These results show that quantitative water stress is highly localized and sensitive to the spatial correspondence between accumulated demand and routed-streamflow availability, emphasizing the importance of network-consistent supply–demand assessments for long-term water-resource planning in high-Andean systems.