Characterizing the vertical variations of hydraulic conductivity of an unconsolidated aquifer based on the direct-push point-source injection tests
摘要
Hydraulic conductivity (K) and its spatial variability are critical factors governing groundwater flow and solute transport processes. Traditional methods, such as pumping tests and slug tests, often treat the aquifers as homogeneous to derive equivalent K values, making it difficult to accurately characterize the aquifer medium's intrinsic spatial heterogeneity. Recently, the direct-push point-source injection test method, relying on casing-free direct advancement technology, enables rapid and high-resolution point-source injection through one miniature screen on the side of the probe. However, the interpretation of data from this type of tests currently relies primarily on one empirical model, and the accuracy and reliability of its estimates lack systematic evaluation. This study selected the North Campus Research Site (NCRS), a typical unconsolidated aquifer, to comparatively evaluate the differences between the site-specific model and the traditional empirical model in quantitatively estimating hydraulic conductivity. The estimation effectiveness of the point-source injection test, soil permeameter tests, and slug tests across layers with different hydraulic conductivities was analyzed and discussed. Furthermore, a numerical model was used to quantitatively analyze the impact of hydraulic conductivity estimates from different methods on simulated injection water level dynamics. This study indicates that the high-resolution response data from the direct-push point-source injection tests can qualitatively indicate vertical variations in aquifer media permeability. The traditional empirical model can lead to significant overestimation or underestimation of hydraulic conductivity, whereas the site-specific model can effectively improve estimation accuracy. Compared to permeameter tests and slug tests, the point-source injection test can provide hydraulic conductivity estimates more conveniently and efficiently. However, for low-permeability layers, its estimates st