The Hsueh-Shan tunnel is the fifth longest road tunnel in the world. During the excavation, the tunnel encountered several events of groundwater inrush, causing serious delay of the construction. Data on groundwater discharge to the tunnel were gathered from the monitoring system and their spatial and temporal variations were analyzed. The results of the integrated analysis of groundwater discharge and local precipitation indicated that the discharge increased rapidly when the cumulative rainfall exceeded 85 mm. The groundwater level recession rate after a rainfall event was found to be independent of rainfall intensity. A hydrogeological conceptual model was developed to simulate the long-term groundwater discharge to the tunnel. Sensitivity analysis was first conducted to identify sensitive parameters, and then the calibration process was accomplished by the automated parameter estimation method. The calibrated model was then used to evaluate the potential impact of tunnel excavation on the Feitsui reservoir; the average percentage loss of inflow to the Feitsui reservoir from 2006 to 2010 is estimated to be 1.74%. The developed model can provide a tool for evaluating the regional hydrogeologic setting and the influence of tunnel construction on water resources.

A study in eastern Taiwan evaluated the importance of montane water contribution (MC) to adjacent valley-plain groundwater (VPG) in a tectonic suture zone. The evaluation used a ternary natural-tracer-based end-member mixing analysis (EMMA). With this purpose, VPG and three end-member water samples of plain precipitation (PP), mountain-front recharge (MFR), and mountain-block recharge (MBR) were collected and analyzed for stable isotopic compositions (δ²H and δ¹⁸O) and chemical concentrations (electrical conductivity (EC) and Cl⁻). After evaluation, Cl⁻ is deemed unsuitable for EMMA in this study, and the contribution fractions of respective end members derived by the δ¹⁸O–EC pair are similar to those derived by the δ²H–EC pair. EMMA results indicate that the MC, including MFR and MBR, contributes at least 70% (679 × 10⁶ m³ water volume) of the VPG, significantly greater than the approximately 30% of PP contribution, and greater than the 20–50% in equivalent humid regions worldwide. The large MC is attributable to highly fractured strata and the steep topography of studied catchments caused by active tectonism. Furthermore, the contribution fractions derived by EMMA reflect the unique hydrogeological conditions in the respective study sub-regions. A region with a large MBR fraction is indicative of active lateral groundwater flow as a result of highly fractured strata in montane catchments. On the other hand, a region characterized by a large MFR fraction may possess high-permeability stream beds or high stream gradients. Those hydrogeological implications are helpful for water resource management and protection authorities of the studied regions.

A three-dimensional variable-density finite element model was developed to study the combined effects of overabstraction and seawater intrusion in the Pingtung Plain coastal aquifer system in Taiwan. The model was generated in different layers to represent the three aquifers and two aquitards. Twenty-five multilayer pumping wells were assigned to abstract the groundwater, in addition to 95 observation wells to monitor the groundwater level. The analysis was carried out for a period of 8 years (2008–2015 inclusive). Hydraulic head, soil permeability, and precipitation were assigned as input data together with the pumping records in different layers of the aquifer. The developed numerical model was calibrated against the observed head archives and the calibrated model was used to predict the inland encroachment of seawater in different layers of the aquifer. The effects of pumping rate, sea-level rise, and relocation of wells on seawater intrusion were examined. The results show that all layers of the aquifer system are affected by seawater intrusion; however, the lengths of inland encroachment in the top and bottom aquifers are greater compared with the middle layer. This is the first large-scale finite-element model of the Pingtung Plain, which can be used by decision-makers for sustainable management of groundwater resources and cognizance of seawater intrusion in coastal aquifers.

Changes in groundwater level during earthquakes have been reported worldwide. In this study, field observations of co-seismic groundwater-level changes in wells under different aquifer conditions and sampling intervals due to near-field earthquake events in Taiwan are presented. Sustained changes, usually observed immediately after earthquakes, are found in the confined aquifer. Oscillatory changes due to the dynamic strain triggered by passing earthquake waves can only be recorded by a high-frequency data logger. While co-seismic changes recover rapidly in an unconfined aquifer, they can sustain for months or longer in a confined aquifer. Three monitoring wells with long-term groundwater-level data were examined to understand the association of co-seismic changes with local hydrogeological conditions. The finite element software ABAQUS is used to simulate the pore-pressure changes induced by the displacements due to fault rupture. The calculated co-seismic change in pore pressure is related to the compressibility of the formation. The recovery rate of the change is rapid in the unconfined aquifer due to the hydrostatic condition at the water table, but slow in the confined aquifer due to the less permeable confining layer. Fracturing of the confining layer during earthquakes may enhance the dissipation of pore pressure and induce the discharge of the confined aquifer. The study results indicated that aquifer characteristics play an important role in determining groundwater-level changes during and after earthquakes.

The analysis of pre-existing landslides and landslide-prone hillslopes requires an estimation of maximum groundwater levels. Rapid increase in groundwater levels may be a dominant factor for evaluating the occurrence of landslides. System identification—use of mathematical tools and algorithms for building dynamic models from measured data—is adopted in this study. The fluid mass-balance equation is used to model groundwater-level fluctuations, and the model is analytically solved using the finite-difference method. Entropy-based classification (EBC) is used as a data-mining technique to identify the appropriate ranges of influencing variables. The landslide area at Wushe Reservoir, Nantou County, Taiwan, is chosen as a field test site for verification. The study generated 65,535 sets of numbers for the groundwater-level variables of the governing equation, which is judged by root mean square errors. By applying cross-validation methods and EBC, limited numbers of validation samples are used to find the range of each parameter. For these ranges, a heuristic method is employed to find the best results of each parameter for the prediction model of groundwater level. The ranges for governing factors are evaluated and the resulting performance is examined.

Parameter structure identification is formulated in terms of solving an inverse problem, which allows for a determination of an appropriate level of parameter structure complexity, and the identification of its pattern and the associated parameter values. With the increasing complexity of parameter structure identification in groundwater modeling, demand for robust, fast, and accurate optimizers is on the rise among researchers from groundwater hydrology fields. A novel global optimizer, differential evolution (DE), has been proposed to solve the parameter-structure-identification problem. The Voronoi tessellation is adopted for the automatic parameterization. The stepwise regression method and the error covariance matrix are used to determine the optimal structure complexity. Numerical experiments with a continuous hydraulic conductivity distribution are conducted to demonstrate the proposed methodology. The results indicate that the DE can identify the global optimum effectively and efficiently. A sensitivity analysis of the control parameters and mutation schemes implemented in the DE is employed to examine their influence on the objective function. The comparison between DE and genetic algorithm shows the advantage of DE in terms of robustness and efficiency. The proposed methodology is also applied to a real groundwater system, Pingtung Plain in Taiwan, and the properties of aquifers are successfully identified.

Recent advances in borehole geophysical techniques have improved characterization of cross-hole fracture flow. The direct detection of preferential flow paths in fractured rock, however, remains to be resolved. In this study, a novel approach using nanoscale zero-valent iron (nZVI or ‘nano-iron’) as a tracer was developed for detecting fracture flow paths directly. Generally, only a few rock fractures are permeable while most are much less permeable. A heat-pulse flowmeter can be used to detect changes in flow velocity for delineating permeable fracture zones in the borehole and providing the design basis for the tracer test. When nano-iron particles are released in an injection well, they can migrate through the connecting permeable fracture and be attracted to a magnet array when arriving in an observation well. Such an attraction of incoming iron nanoparticles by the magnet can provide quantitative information for locating the position of the tracer inlet. A series of field experiments were conducted in two wells in fractured rock at a hydrogeological research station in Taiwan, to test the cross-hole migration of the nano-iron tracer through permeable connected fractures. The fluid conductivity recorded in the observation well confirmed the arrival of the injected nano-iron slurry. All of the iron nanoparticles attracted to the magnet array in the observation well were found at the depth of a permeable fracture zone delineated by the flowmeter. This study has demonstrated that integrating the nano-iron tracer test with flowmeter measurement has the potential to characterize preferential flow paths in fractured rock.

The quantification of vertical groundwater fluxes across semi-confining layers is fundamental to evaluate groundwater recharge and discharge rates to and from aquifer systems. Methods to estimate vertical groundwater fluxes from temperature–depth profiles have been available since the 1960s. While some methodologies assume steady-state conditions, changes in land-surface temperatures as well as hydrogeological conditions can lead to transient heat flow conditions. Indeed, many studies have indicated that transient temperatures in deeper confined aquifers are widespread. A study is presented that uses transient-temperature–depth curves obtained from groundwater observation wells in the Chianan coastal plain in southern Taiwan. In this area, sedimentary aquifer systems consist of a stack of alternating sand and mud layers, over several hundred meters in thickness. Groundwater has been abstracted from these aquifers for decades, resulting in large hydraulic gradients between the shallow and deeper aquifers. Hence, vertical groundwater flow is likely enhanced across finer-grained, semi-confining units. A set of temperature–depth profiles is available from this area. Constrained by these profiles, numerical models of one-dimensional transient heat transfer were used to infer vertical fluxes of 3.3 × 10⁻⁸ to 3.9 × 10⁻⁸ m/s using thermal data from 2013 to 2016. An analytical solution was also employed that assumes steady-state conditions. Calculated fluxes using the latter approach were lower, at approximately 1.1 × 10⁻⁸ to 1.6 × 10⁻⁸ m/s. The study suggests that vertical fluxes derived from using Bredehoeft and Papadopulos’s analytical solutions result in underestimates of actual vertical seepage rates across aquitards.