Solutions to long-term water supply–demand balance challenges: Decision making under uncertainty in the Marquis catchment

The Caribbean accounts for seven of the world’s top 36 water-stressed countries and FAO defines many Caribbean countries as water scarce, with less than 1000 m3 freshwater resources per capita. Climate change will bring rising temperatures, increases in evaporation and less rainfall. This means the Caribbean is likely to experience more intense and frequent droughts. At the same time, the lack of capacity to accurately assess basin water budgets and supply-demand balances, the absence of a structured methodology and the paucity of data in most of the Caribbean countries has hindered evidence-based decision making to address drought challenges to water supply systems. Based on a pilot study in the Marquis catchment in Saint Lucia, the aim of this paper is to present a Caribbean focused methodology that can be used to assess the resilience of water supply systems to drought, including the testing of adaptation options under different socio-economic and climate scenarios. Resilience metrics are established in relation to water service performance and water resources management. Hydrological and water resources models of the study area are developed, validated and coupled in order to simulate the behaviour of the Marquis water supply system against current climate and hydrologic variables. The system is then tested against a range of climate change (e.g. droughts) and socio-economic development scenarios (population growth, development planning, new infrastructure etc.). Their impacts on the resilience metric as well as the current and future gap in the supply-demand balance are assessed. A suite of adaptation options are defined and tested in order to provide an evidence-base for the prioritisation of options that provide the greatest resilience and robustness. Finally, the strengths and weaknesses of the methodology are discussed with particular reference to technical capacity and overcoming the absence of long-term data records