Exploratory modeling and analysis to inform adaptive water management under deep uncertainty in the Peruvian Andes

Mountains are an important source of freshwater for ecosystems and the livelihoods of around two billion people worldwide. However, this abundance of water resources is jeopardized by climatic and socioeconomic changes, particularly in the tropical Andes which are considered one of the most vulnerable regions to these impacts worldwide. Those changes not only impact water availability for local populations but also contribute to the declining health of the aquatic ecosystems reducing their provision of services to society. Adaptive water management has emerged as a concept to address such social and environmental challenges. However, there is limited information on how such an adaptive approach can be systematically implemented. Other important challenges in mountain regions include data scarcity (e.g. hydroclimatic or socioeconomic data), knowledge gaps (e.g. groundwater contribution to total runoff), and uncertain future climatic and socioeconomic changes (obtained from e.g. climate models or socioeconomic projections). All these knowledge gaps and limitations lead to deep uncertainties in water policies, the condition where experts do not know or cannot agree on which initial conditions and corresponding results are most relevant.

To deal with such deep uncertainties we tested the Exploratory Modeling and Analysis (EMA) framework to support adaptive water management. Therefore, we set a case study in the high-Andean Pitumarca catchment in the glaciated headwaters of the Vilcanota basin, Southern Peru. Three policy options were assessed along a large set of uncertainties to achieve water security for human and environmental needs by 2050. A total of 12,000 simulations were run, driven by three climate scenarios (SSP1-1.9, SSP1-2.6, and SSP5-8.5) and 15 climate models, and a wide range of irrigation and domestic water use scenarios.

Results from the applied EMA framework show that in 43% of simulations (5,182) the water system failed to supply water for human and environmental needs mostly driven by the way of implementation of water policies than by climatic or socioeconomic changes. The implemented framework also contributed to identify that a two combination of improvements of irrigation efficiencies and reservoir schemes can avoid system failures under a wide range of changes and uncertainties. These results highlight the importance of focused policy actions to deal with climatic and socioeconomic changes. Such a framework facilitates moving from traditionally broad problems that center on the impact of climate change to more specific and locally tailored questions, e.g. which reservoir scheme should be implemented to avoid system failure. In order to reduce uncertainties and optimize local water use, EMA should be combined with other methods such as citizen science, sensitivity analysis, joint knowledge production. Furthermore, EMA should be implemented through semi-distributed glacio-hydrological models to fully combine the advantages of different approaches and assess the spatio-temporal occurrence of water demand. We also encourage the use of socio-hydrological models where socioeconomic and environmental factors can actively interact.