Coastal aquifers are important sources of freshwater for domestic, agricultural and industrial purposes in many countries. Currently, coastal aquifers worldwide are under increasing threat of seawater intrusion (SWI). SWI is a phenomenon where seawater intrudes into a coastal aquifer and contaminates fresh groundwater. The occurrence of SWI is predominantly a human-induced process due to groundwater overexploitation, although natural factors such as sea-level rise (SLR) may also contribute to SWI problems. Other transient factors, such as tides, waves and storm surges, and land use change may also affect SWI.

Previous studies identify two different types of SWI: passive and active. Where passive SWI arises from a reduction in the watertable elevation, there will be an accompanying decrease in the depth of freshwater discharging at the shoreline and SWI occurs despite persistent freshwater discharge to the sea, and the seawater wedge intrudes rather slowly. At certain condition where freshwater discharge to the sea ceases, a transition occurs from passive SWI to active SWI at the shoreline. Active SWI is driven by the combined effects of the density difference between freshwater and seawater, and the inland-sloping hydraulic gradient, whereas in passive SWI, these forces are opposed. The processes associated with active SWI have received considerably less attention compared to situations of passive SWI despite several investigations demonstrating that active SWI is occurring in many areas. It is hypothesized that there will be an extensive salinization under active SWI not only at the lower domain of coastal aquifer but also at the upper domain of coastal aquifer.

In this thesis, numerical modelling and physical experiments are combined to examine quantitatively and systematically the transient SWI processes, including the geometry of the freshwater-saltwater interface, the width of the mixing zone, the flow processes and the related effects (e.g., watertable salinization (WTS), advection and density) that may be occurred, in response to freshwater head decline (FHD), both under passive and active SWI situations. The effects of freshwater head declines (rather than SLR) on WTS are considered, because these are expected to develop faster and be larger in magnitude than SLR, and are therefore more likely to generate more aggressive SWI situations. This study is motivated by the fact that to date, no experimental and numerical studies exist that provides an assessment of transient active SWI processes and explores the interface characteristics in terms of freshwater-saltwater interface geometry, particularly under active SWI conditions. The results of this study is very important in giving insight into transient SWI processes (especially for active SWI) in response to FHD and can be used as a basis and a reference in understanding the transient behaviour of active SWI, given the lack of transient active SWI literatures exist and some evidences which showed that active SWI is occurring and threatens the coastal groundwater resources around the world.

Firstly, physical experiments and numerical modelling are combined to examine the occurrence of WTS associated with SWI in response to an inland FHD. Comparing the laboratory and numerical modelling results offers insight into both the veracity of the laboratory set up and the assumptions of the numerical code. WTS is examined in the absence of tidal effects, and the analysis focuses on the influence of watertable decline on WTS. An important outcome of this work is SWI can cause WTS in unconfined coastal aquifer setting and may induce unsaturated zone soil salinization in the vicinity of the sea boundary. Physical and numerical model of the laboratory experiment show that significant WTS might happen under active SWI caused by the associated cessation of seaward freshwater discharge and under passive SWI, minor WTS might occur under conditions of high dispersivity, hydraulic conductivity and low freshwater discharge to the sea.

Secondly, the first work is extended by investigating characteristics of active SWI. In this work, numerical modelling is used to characterise active SWI in various idealised unconfined coastal aquifer settings, following an inland FHD. Relationships between key features of active SWI (e.g., interface characteristics and SWI response time-scales) and the parameters of the problem (e.g., inland FHD, freshwater-seawater density contrast, dispersivity, hydraulic conductivity, porosity, aquifer thickness) are investigated, thereby extending the previous work in Chapter 2. The outcome of the second work is the SWI response time-scales are affected by both the initial and final boundary head differences between the inland and the sea boundary. The freshwater-saltwater interface is found to be steeper under stronger advection (i.e., caused by the inland FHD), higher dispersivity and hydraulic conductivity, and lower aquifer thickness, seawater density and porosity. The interface movement is faster and the mixing zone is wider with larger hydraulic conductivity, seawater-freshwater density difference and aquifer thickness, and with lower porosity. Advection effects become more dominant on the interface movement relative to density effects as SWI becomes more active. It is found that larger differences of density between freshwater and saltwater may increase the width of the mixing zone, particularly under active SWI conditions. The results also show that the dimensionless forms of mixed convection ratio and Peclet number used in this study cannot generally reflect the interface characteristics under active SWI.

Thirdly, the research is expanded by conducting a regional 3D numerical simulation of the real world SWI case. In this work, SEAWAT is used to model the condition of SWI that occurs in Uley South Basin (USB). This study purposes to quantify the interrelationship between climate variability, anthropogenic stresses and the extent of seawater in USB. A regional three-dimension SWI model of USB is conducted that adopts the spatial heterogeneity in parameters obtained from during a prior calibration effort using a single-density model. This study also aims to investigate the seawater effects on the groundwater head behaviour near the coast. There is a chance that SWI may affect the groundwater head response around the coast and makes it different to the groundwater behaviour over the basin. The results of this work show that the effect of pumping on the extent of SWI in the Quaternary (QL) and Tertiary Sand (TS) layers of USB are shown to be larger, relative to SWI arising from climate variability. The effect of pumping on SWI is 3.5% higher on average than climate impacts in terms of total mass of salt in the aquifer. The numerical model results also demonstrate that including seawater in the numerical model slightly modifies the groundwater behaviour near the coast where the relative effects of pumping on groundwater head behaviour are larger compared to that without seawater in the model.

Fourthly, an SWI work is presented where vertical leakage in layered coastal aquifers is investigated using a sharp and dispersive interface numerical model and combined with a physical model. We consider SWI in layered aquifers where upward freshwater leakage occurs and three alternatives (i.e., Case 1 involves freshwater bypassing any overlying saltwater, Case 2 assumes no upward freshwater leakage where there is overlying saltwater and Case 3 converts upward freshwater leakage into saltwater) are used in the sharp interface numerical model to treat the upward freshwater leakage through aquitards. Steady-state and transient predictions using both sharp-interface and dispersive models are assessed, and compared to the results of sand-tank experiments. The outcome of this work is Case 1 (i.e., freshwater bypassing any overlying saltwater) produces optimal matches to both SEAWAT results and sand-tank observations in terms of saltwater wedge locations relative to the two other cases, even though all sharp-interface models over-predict the extent of saltwater both under steady-state and transient conditions. Streamlines from SEAWAT model show that upward freshwater leakage tends to flow around and bypass overlying saltwater.