Predicting Groundwater Flow and Solute Transport in the Heterogeneous Aquifer Sandbox Using Different Parameter Estimation Methods

Protection and management of groundwater resources demand high-resolution distributions of hydraulic parameters (e.g., hydraulic conductivity (K) and specific storage (Ss)) of aquifers. In the past, these parameters were obtained by traditional analytical solutions (e.g., Theis (1935) or Cooper and Jacob (1946)). However, traditional methods assume the aquifer to be homogeneous and yield the equivalent parameter, which are averages over a large volume and are insufficient for predicting groundwater flow and solute transport process (Butler & Liu, 1993). For obtaining the aquifer heterogeneity, some scholars have used kriging (e.g., Illman et al., 2010) and hydraulic tomography (HT) (e.g., Yeh & Liu, 2000; Zhu & Yeh, 2005) to describe the K distribution.

In this study, the laboratory heterogeneous aquifer sandbox is used to investigate the effect of different hydraulic parameter estimation methods on predicting groundwater flow and solute transport process. Conventional equivalent homogeneous model, kriging and HT are used to characterize the heterogeneity of sandbox aquifer. A number of the steady-state head data are collected from a series of single-hole pumping tests in the lab sandbox, and are then used to estimate the K fields of the sandbox aquifer by the steady-state inverse modeling in HT survey which was conducted using the SimSLE algorithm (Simultaneous SLE, Xiang et al., 2009), a built-in function of the software package of VSAFT2. The 40 K core samples from the sandbox aquifer are collected by the Darcy experiments, and are then used to obtain K fields through kriging which was conducted using the software package of Surfer 13. The role of prior information on improving HT survey is then discussed. The K estimates by different methods are used to predict the process of steady-state groundwater flow and solute transport, and evaluate the merits and demerits of different methods, investigate the effect of aquifer heterogeneity on groundwater flow and solute transport.

According to lab sandbox experiments results, we concluded that compared with kriging, HT can get higher precision to characterize the aquifer heterogeneity and predict the process of groundwater flow and solute transport. The 40 K fields from the K core samples are used as priori information of HT survey can promote the accuracy of K estimates. The conventional equivalent homogeneous model cannot accurately predict the process of groundwater flow and solute transport in heterogeneous aquifer. The enhancement of aquifer heterogeneity will lead to the enhancement of the spatial variability of tracer distribution and migration path, and the dominant channel directly determines the migration path and tracer distribution.