Descriptions of anisotropy and heterogeneity and their effect on ground-water flow and areas of contribution to public supply wells in a karst carbonate aquifer system

Abstract

Delineation of areas of contribution to wells tapping a karst carbonate aquifer system can be extremely difficult using conventional approaches designed for isotropic and homogeneous aquifers, because ground-water flow tends to be through solution-enhanced conduits. Nonradial flow along preferential zones can result in inaccurate estimates of flow paths and traveltimes. Because of the large variability in factors affecting contributing areas and an imperfect understanding of how these factors can vary, the estimation of contributing areas is an approximation at best. To better understand the effects of aquifer anisotropy and heterogeneity on areas of contribution, an exploratory modeling approach was used. MODFLOW, a numerical flow model, and MODPATH, a particle tracking program, were used to generate time-related areas of contribution for six hypothetical carbonate aquifer system types. The six types were conceptualized to approximate different types of aquifer anisotropy and heterogeneity. These include: (1) an isotropic and homogeneous single-layer system; (2) an anisotropic in a horizontal plane single layer system; (3) a discrete vertically fractured single-layer system; (4) a multilayered system; (5) a doubly porous single-layer system; and (6) a vertically and horizontally interconnected heterogeneous system. The simulated aquifer anisotropy was 5:1 (Kxx/Kyy) determined from TENSOR2D results. The simulated discrete vertical fracture network represents locations inferred from mapped photolineaments. The simulated enhanced flow zones were determined from borehole video and geophysical logs. Areas of contribution were simulated for two prototype regions. The two prototypes were selected to be representative of the hydrologic diversity within the study area and were designated the Central Swamp and Lake Terrace regions. Localized conditions in pumping, production well distribution, and aquifer transmissivity affect the size, shape, and orientation of areas of contribution to public supply wells. The simulated areas of contribution are 60 percent larger in the Central Swamp region where pumpage is more than double and transmissivity is about half that of the Lake Terrace region. Although these factors are important, this study focused on the effects from hydrogeologic factors common to karst carbonate aquifer systems. This study indicates that the distribution and type of aquifer anisotropy and heterogeneity will affect the size, shape, and orientation of areas of contribution in a karst carbonate aquifer system. The size of the 50-year time-related areas of contribution ranged from 8.2 to 39.1 square miles in the Central Swamp region and from 4.0 to 18.3 square miles in the Lake Terrace region. Simulations showed that the size of areas of contribution is primarily affected by simulated withdrawal rates, effective porosity of the carbonate rock, and transmissivity. The shape and orientation of the simulated areas of contribution primarily result from aquifer anisotropy, well distribution, flow along solution-enhanced zones, and short-circuiting of flow through fracture networks. Comparisons also were made between protection zones delineated using analytical models and areas of contribution delineated using numerical models. The size of the 5-year time-related protection zone in the Central Swamp region using an analytical model was almost twice as large as the numerically simulated area of contribution, and more than eight times larger than the numerically simulated area of contribution in the Lake Terrace region. The differences in size are primarily the result of how the flow field is approximated. The analytical method assumes only lateral flow to wells but numerical methods allow particles to move laterally and vertically. Additionally, multiple-well-interference effects resulting from the close proximity of several pumping wells cause individual capture zones to converge or diverge, depending on the difference in pumping rates and orientation among the wells. Such an interpretation is not available from analytical methods. The simulated distributions of aquifer anisotropy and heterogeneity, in this study, were highly conceptualized, but were based on plausible occurrences of anisotropy and heterogeneity inherent in carbonate aquifer systems.