An equation is derived by which advective groundwater velocity in a confined aquifer may be estimated by a single‐well tracer test in which a single tracer pulse is allowed to drift from the well and then pumped back to the well and sampled to obtain a breakthrough curve. Although similar in methodology to preexisting methods, this method differs in that it takes into account ambient groundwater movement during the pumpback phase. Using sodium chloride as a tracer solution, a series of small‐scale tests were run in a laboratory sand tank model to test the theory. Results of linear flow tracer tests through the model, simulating unperturbed regional advective flow at known velocities, were compared to results of single‐well drift‐and‐pumpback tests conducted during linear flow through the model. Advective velocities computed by both types of tests were identical, thus proving the validity of the equation.

No abstract

The U.S. Geological Survey's Amargosa Tracer Calibration Site in southern Nevada has been used for three different recirculating tracer tests using as tracers: (1) tritium, (2) sulfur‐35, and (3) tritium and bromide together. Although the physical setup, well spacings, and thus apparent scale were the same in all tests, the recirculating rates and pore pressures were different. Apparent dispersivities found in the three tests differed considerably, revealing an inverse relationship between computed dispersivity and recirculation rate. These differences are believed to be caused primarily by changes in fissureaperture widths and thus, hydraulic conductivities, with resulting changes in flow rates and directions caused by changes in pore pressure between different tests in the presence of high ambient pore pressure. The results of this study indicate that forced‐gradient tracer tests for determination of dispersivity should not be performed when ambient pore pressure and/or testing pore pressure is a significant percentage of overburden pressure. In addition, apparent dispersivity calculated by forced‐gradient tracer tests can differ considerably from that of unstressed natural situations, and consequent solute‐transport estimates are likely to be in error.

A tritium profile obtained from a core located on the Tipton Till Plain near West Lafayette, Indiana, showed a “tritium reversal” (corresponding to a low‐high‐low activity of tritium relative to depth) at 7 m, indicating the present position of recharge water derived from 1963/64 precipitation. The average recharge rate calculated using the tritium data is 3.5 cm/yr for the “mass‐balance” method and 4.7 cm/yr for the “transit‐time” method. The average of 4.1 cm/yr is in reasonable agreement with Arihood's (1982) estimate of 5.1 cm/yr for tills, based on numerical modeling of data from a water‐budget study in the White River basin of Indiana. Two other cores, located on a nearby topographically lower slope, showed no definite reversal in tritium activity, probably because of lateral flow components at these locations. To be applied, the tritium technique should assess the contribution of summer and winter precipitation to recharge because tritium activities tend to peak in spring and summer precipitation. Contrary to the usual assumptions that recharge to aquifers in temperate zones of the northern hemisphere occurs between October and April, stable isotopic data of well water in the study area indicate that an appreciable amount of annual recharge (about 34%) actually occurs between April and October.

In order to determine the suitability of bromide as a surrogate for tritium for ground‐water tracing in fractured and fissured dolomite, apparent relative retardation coefficients of bromide with respect to tritium were computed by the cumulative relative mass‐time method from breakthrough curves of both substances, when used simultaneously as tracers in a two‐well recirculating tracer test. Throughout the 33.78‐day test, relative retardation differences of only a few hundredths were detected for most of the testing time. Tritium was less retarded in flow through highly transmissive openings than in more diffuse flow. Bromide was generally retarded relative to tritium except at the pumping well where the concentrations of both were diluted an order of magnitude below those in the monitoring wells, and tritium was retarded slightly more than bromide. Apparent relative retardation was found to change with transmissivity, concentration, flow‐path length and time, but not enough to rule out bromide as a suitable surrogate for, or even better than, tritium in dolomite if the limitations of each tracer are known.