Modelling seawater intrusion (SWI) has evolved from a tool for understanding to a water management need. Yet, it remains a challenge. Difficulties arise from the assessment of dispersion coefficients and the complexity of natural systems that results in complicated aquifer geometries and heterogeneity in the hydraulic parameters. Addressing such difficulties is the objective of this thesis. Specifically, factors that may affect the flow and transport in coastal aquifers and produce heterogeneous salinity distributions are studied.
First, a new paradigm for seawater intrusion is proposed since the current paradigm (the Henry problem) fails to properly reproduce observed SWI wedges. Mixing is represented by means of a velocity dependent dispersion tensor in the new proposed problem. Thereby, we denote it as "dispersive Henry problem". SWI is characterized in terms of the wedge penetration, width of the mixing zone and influx of seawater. We find that the width of the mixing zone depends basically on dispersion, with longitudinal and transverse dispersion controlling different parts of the mixing zone but displaying similar overall effects. The wedge penetration is mainly controlled by the horizontal permeability and by the geometric mean of the dispersivities. Transverse dispersivity and the geometric mean of the hydraulic conductivity are the leading parameters controlling the amount of salt that enters the aquifer.
Second, the effect of heterogeneity was studied by incorporating heterogeneity in the hydraulic permeability into the modified Henry problem. Results show that heterogeneity causes the toe to recede while increases both the width and slope of the mixing zone. The shape of the interface and the saltwater flux depends on the distribution of the permeability in each realization. However, the toe penetration and the width of the mixing zone do not show large fluctuations. Both variables are satisfactorily reproduced, in cases of moderate heterogeneity, by homogeneous media with equivalent permeability and either local or effective dispersivities.
Third, the effect of aquifer geometry in horizontally large confined aquifers was analyzed. Lateral slope turned out to be a critical factor. Lateral slopes in the seaside boundary of more than 3% cause the development of horizontal convection cells. The deepest zones act as preferential zones for seawater to enter the aquifer and preferential discharging zones are developed in the upwards lateral margins. A dimensionless number, Nby, has been defined to estimate the relative importance of this effect.
All these factors can be determinant to explain the evolution of salinity in aquifers such as the Main aquifer of the Llobregat delta. Finally, a management model of this aquifer is developed to optimally design corrective measures to restore the water quality of the aquifer. The application of two different optimization methodologies, a linear and a non-linear optimization method, allowed (1) to quantify the hydraulic efficiency of two potential corrective measures: two recharge ponds and a seawater intrusion barrier; (2) to determine the water necessary to be injected in each of these measures to restore the water quality of the aquifer while minimizing changes in the pumping regime and (3) to assess the sustainable pumping regime (with and without the implementation of additional measures) once the water quality has been restored. Shadow prices obtained from linear programming become a valuable tool to quantify the hydraulic efficiency of potential corrective measures to restore water quality in the aquifer