Regional ground-water flow in upper and middle Paleozoic rocks in southeastern Utah and adjacent parts of Arizona, Colorado, and New Mexico

Abstract

A somewhat hydrologically isolated regional ground-water flow system in Paleozoic rocks that includes part of the Upper and Lower Colorado River Basins was simulated. The area of the ground-water system is about 60,000 square miles. Aquifers and confining units in rocks of Paleozoic age were defined: the sandstone and red-bed aquifer consists of all rocks of Permian age and rocks of Early Pennsylvanian age; the limestone and dolomite aquifer consists of all rocks of Mississippian and Devonian age. Between these two units is a confining unit consisting of rocks of Late Pennsylvanian age. Permeability measurements and equivalent freshwater heads derived from drill-stem tests were examined for both aquifers. The permeability measurements of each aquifer were lognormally distributed. The geometric mean of permeability measurements for the sandstone and red-bed aquifer was 4 millidarcies per centipoise, and the geometric mean of permeability measurements for the limestone and dolomite aquifer was 424 millidarcies per centipoise. A large probability exists that geometric mean permeability of the sandstone and red-bed aquifer is one-hundredth of the geometric mean permeability of the limestone and dolomite aquifer. Equivalent freshwater-head differences between the aquifers (the difference between the means was 521 feet) and density differences between the aquifers indicated predominantly downward flow. The only substantial, estimable osmotic-pressure gradients were chemical and were caused by differences in concentrations of dissolved solids between the aquifers in the brine area; those chemical osmotic pressure gradients tended to move water downward. An adequate simulation of regional flow in the sandstone and red-bed aquifer was achieved by using the boundary integral-equation method. Based on the data available, the following approximations that were tested by simulation seemed to be reasonable: (1) The aquifer has the same transmissivity everywhere. (2) No interaquifer flow occurs for the aquifer. (3) Ground water was the same density everywhere. (4) A no-flow boundary occurs along the southern edge of the Uncompaghre uplift. (5) Discharge occurs primarily from the Paleozoic outcrops in stream channels, near the center of the modeled area. (6) Discharge occurs secondarily from the Paleozoic outcrops at Marble Canyon. In the flow model of the sandstone and red-bed aquifer, discharge from the Paleozoic outcrops in stream channels near the center of the modeled area was controlled by the altitude of the stream. Discharge or recharge at Dark Canyon was not part of the flow model because Dark Canyon probably is a discharge area for local flow from the Abajo Mountains. Discharge from the sandstone and red-bed aquifer in Marble Canyon was controlled by the altitudes of the Paleozoic outcrops. A brine that has concentrations of dissolved solids as large as 374 grams per liter is present in the limestone and dolomite aquifer in the Paradox basin area. Substantial flow out of the brine area is unlikely. Attempts to use the boundary integral-equation method to simulate reasonable flow in the limestone and dolomite aquifer failed. The principal problem in simulating accurate flow in the aquifer was simulating equivalent freshwater heads that were small enough to match measured equivalent freshwater heads. To do this, additional discharge was needed. Simulating variabledensity flow will not fulfill this need and therefore should not be a primary concern. A geostatistical interpolation of measured equivalent freshwater head indicates that, except for the boundary along the southern edge of the Uncomphgre Uplift, water generally flows from the boundaries to the center of the aquifer.