METHOD FOR QUANTIFYING VOLUME OF RECHARGED WATER IN AN AQUIFER STORAGE AND RECOVERY (ASR) SITE USING MULTIPLE-REGRESSION ANALYSIS OF MIXING SCENARIOS AND CONSTRAINT OF TOTAL PORE VOLUME ANALYSIS

Cities and other entities store available water in unconfined and confined aquifers to hold for periods of high demand. Managing storage and recovery requires a way to measure and map the concentration and distribution of injected or recharged water. The fraction of recharged water that is recovered during production cycles can be determined where the chemical compositions of recharged and ‘native’ groundwaters are distinct and residence time is short relative to geochemical reaction rates. In a case study, available groundwater produced from a karst limestone aquifer during low-demand seasons is chemically distinct from native groundwater in the host confined sandstone aquifer. Data for identifying and monitoring the recovered volume fraction (F = C/C_o) mostly were limited, however, to a few routinely measured water-quality constituents or parameters (pH, alkalinity, hardness, total iron). PHREEQC was used to simulate the mixing of end-member solutions, defined by complete charge-balanced chemical analyses, which reflect the range of sample concentrations. A four-parameter multiple-regression equation has volume mixing fraction (e.g., 10-, 20-, …, 50-, etc., percent) as the dependent variable and pH, alkalinity, hardness, and total iron as independent variables with the combined results of several end-member mixing scenarios calculated by PHREEQC. Volume fraction (F) for 183 water samples was determined using the 4-parameter regression equation; F also was determined using various 3-parameter regression equations for 104 additional samples with one of the constituents unreported.

A contoured ‘synoptic’ map of volume-fraction distribution (samples collected over 28-day period) was constrained by porosity, injection-zone thickness, and gaged record of stored (recharged – recovered) water. Porosity had been mapped from a set of TCMR, density, and sonic porosity logs for a ~60-m-thick injection interval across a ~12.23-km² site. Volume of recharged water in place is given by V_i = A_i × bn_i × F_i, for i intersected areas (A_i) of shapefile polygons of porosity-thickness product (bn_i) and recharged-water volume fraction (F_i). Trial-and-error adjustment of volume fraction contours yielded a non-unique but constrained match (<3 % error) of the gaged 0.112 km³ volume of recharged water in place.