Prior to the creation of the global Aqueduct Water Risk Atlas, water risk indicators (Table 1) were developed and tested in a number of river basins worldwide. The results of these basin studies helped inform and shape the global Aqueduct Water Risk Framework. Complete guidelines and processes for indicator selection, data collection, calculations, and mapping techniques are described in the Aqueduct Water Risk Framework available online.1 This study focuses on the specific characteristics of the indicator data and calculation in the Colorado River Basin (CRB).
The data selection and validation process for the Colorado River Basin Study involved three steps: (1) a literature review, (2) identification of data sources in the public domain, and (3) the compilation and expert review of selected data sources. Calculation of 5 of the 12 indicators requires the creation of original datasets to estimate water availability and use at a sub-basin scale. The hydrological catchments used in the exercise are merged HUC-8 sub-basins from the National Watershed Boundary Dataset2 by the U.S. Department of Agriculture, Natural Resources Conservation Service (USDA–NRCS) and the Binational Watersheds Database by the U.S. Geological Survey (USGS) Border Environmental Health Initiative (BEHI).3 Computation of the original datasets was completed by ISciences, L.L.C.
Two measures of water use were used in this study: total withdrawal, the total amount of water abstracted from freshwater sources for human use, and consumptive use, the portion of withdrawn water that evaporates or is incorporated into a product thus is no longer available for further use. Withdrawals for the Colorado River’s Upper Basin are estimated from recorded Bureau of Reclamation (BOR) consumptive use for the year 2005 by multiplying by ratios of USGS estimates of withdrawals by sector divided by consumptive use by sector, most recently generated for the year 1995. For the Lower Basin, USGS county-level withdrawals for 2005 were spatially disaggregated by sector based on regressions with spatial datasets selected to maximize the correlation with the reported withdrawals and re-aggregated to hydrological catchments using sector specific spatial correlates (irrigated areas for agricultural, nighttime lights for industrial, and population for domestic withdrawals). BOR data was recorded at the hydrological catchment scale and did not require spatial disaggregation. Both total withdrawal and consumptive use were coded at the hydrological catchment scale.
Two metrics of water supply were computed: total blue water and available blue water. Total blue water approximates natural river discharge and does not account for withdrawals or consumptive use. Available blue water is an estimate of surface water availability minus upstream consumptive use. Modeled estimates of water supply were calculated using a catchment-to-catchment flow accumulation approach developed by ISciences, L.L.C., which aggregates water by catchment and transports it to the next downstream catchment. Water supply is computed from runoff (R), the water available to flow across the landscape from a particular location, and is calculated as the remainder of precipitation (P) after evapotranspiration (ET) and change in soil moisture storage (ΔS) are accounted for (i.e., R = P – ET – ΔS). The runoff data is courtesy of Livneh et al.4 Rainfall and the calibrated parameters are used to generate runoff values for 1950 to 2010.
The remainder of this document contains definitions, formulas, and data sources for the Colorado River Basin Study.