This working paper explains the methodology used to generate the hydrological metrics and indicators in the Aqueduct Water Risk Atlas. For a detailed list of data sources, see the Aqueduct Global Maps 2.1 Metadata Document.
The availability of freshwater resources to meet human demands has emerged as a top-tier global issue for both environment and development. However, many decision-makers lack the technical expertise to fully understand hydrological information. In response to growing concerns from private sector actors around water availability, water quality, climate change, and increasing demand, we employed the composite index approach as a robust communication tool to translate hydrological data into intuitive indicators of water- related risks.
We grouped 12 indicators into a framework identifying spatial variation in water risks. For 6 of the 12 indicators, we used an ensemble of time series estimators, spatial regression, and a sparse hydrological model to generate novel data sets of water supply and use. We adapted the remaining six indicators from existing publications. We chose aggregation methods to maximize transparency and communicability, and to allow for dynamic weighting to reflect different users’ sensitivities to water-related risks. We are currently unable to validate overall index scores because no datasets of water-related losses exist in the public domain. Data availability, specifically for major infrastructure (e.g., inter-basin transfers) and in-situ water quality and river gauge measurements, is the primary constraining factor in model accuracy.
The resulting Aqueduct Water Risk Atlas (Aqueduct) is a publicly available, global database and interactive tool that maps indicators of water-related risks. Aqueduct enables comparison across large geographies to identify regions or assets deserving of closer attention. This paper documents the methodology used to generate the hydrological metrics and indicators in the Aqueduct Water Risk Atlas (see Reig et al. 2013 for explanation of the Aqueduct framework and indicator selection). The data and maps are publicly available so that others may build off this effort.