English abstract
Seawater intrusion (SWI) is one of the real threats to groundwater resources around the
globe, specifically in highly populated coastal areas as groundwater is heavily exploited.
Gaining a better understanding of SWI in coastal aquifers is very important for the
efficient and sustainable management of fresh groundwater resources. Saline water
dynamic is affected by many hydrological and hydrogeological parameters that include:
hydraulic conductivity of the aquifer, recharge rate, abstraction rate, aquifer geometry,
geological settings, and density of seawater among others. Many studies have been
conducted to address different aspects of SWI using both physical experiments and
mathematical (analytical and numerical) modeling. This includes various initial and
boundary conditions, and other hydrological settings under different scales. There is no
study investigated the dynamics of saline water under managed aquifer recharge (MAR)
practice using a sand tank experiment similar to the one presented in this thesis. The
main objective of this thesis is to investigate the dynamics of saline water in a coastal
unconfined aquifer under MAR practices. The effectiveness of MAR in mitigating SWI
was explored under different hydrological conditions that include various values of
hydraulic conductivities, hydraulic gradients, injection rates, location of injection wells
from the coastline, and depth of injection (well screen). This was achieved using both
sand tank experiment and numerical modeling. For the numerical modeling part, a
density-dependent flow and transport code, SEAWAT, was selected.
Investigating the effect of hydraulic conductivity and salinity concentration using sand
tank experiments is challenging and requires tremendous efforts. Hence, numerical
modeling is an excellent tool to study seawater dynamics under various conditions
because it is cheaper and less time demanding compared with the sand tank physical
experiment. A conceptual model that mimics the designed sand tank was developed and
calibrated. Sensitivity and grid resolution analyses were performed to ensure the
reliability of the developed model when simulating observed sand tank runs and other
different scenarios. Results of the sand tank experiments showed that injecting water
close to the toe of the saline water interface was more effective to reduce the SWI by
32% to 61.4% compared with the situation when MAR was practiced away from it
depending on the injected volume (0.5 L to 1.4 L). Injection of water within the
dispersion zone (behind the interface) retarded the re-intrusion of saline water as it built
a hydrological barrier against intrusion. Injection of water near the sea boundary resulted
in large immediate loss of the injected water through the coastal boundary. The results
revealed that injection at shallow depth was more effective in reducing the salinity in the
intruded zone of the aquifer compared with the scenario when the injection was
performed deeper (near the bottom of the aquifer). For example, injecting 1.06 L of
freshwater near the toe of saline water interface at shallow depth (22.5 cm) reduced
more salinized volume by 25.5% compared to when the same injection was at deeper
depth (42 cm). With injecting more volume, the saline water interface receded further
seaward direction (increasing the injected volume by 32%, the retreated distance
increased by 43%). However, it is very important to monitor the height of the water table
mound under injection as a high water mound may result in a steeper hydraulic gradient.
As a consequence, the rapid discharge of the injected water occurred through the
constant head boundaries (in this case acted as a discharge boundary) and hence less
vii
impact in controlling SWI. In a real situation, abstraction wells can be the discharge (or
drain) facility and so injection and abstraction rates must be balanced to increase the
effectiveness of injection in mitigation of SWI problem. As that, optimization of MAR
(in terms of the location of injection, rate and volume of injection, depth of injection,
duration of injection, along with maintaining water balance for the aquifer) is necessary
to avoid ineffectiveness or failure of MAR to mitigate SWI problem.
The calibrated numerical simulation (Root-Mean-Square Error (RMSE) is 0.3) showed
that the efficiency of MAR is affected by aquifer's conductivity. Our numerical
simulation showed that applying MAR to highly conductive porous media drain out
quickly the injected water resulting in a smaller water table mound and hence small
impact in controlling the SWI. Moreover, the rate of SWI was high under higher
hydraulic conductivity. Therefore, the permeability and the injection rate and volume
need to be well studied for any site. The results also revealed that the denser is the
intruding saline water, the more challenging is the effect of MAR in controlling the
SWI. When the density of the intruding saline water increases by 30%, more of the
injected volume required (increased by 100%, from 1.06 L to 2.12 L) in order for the
saline water interface to recede seaward direction by more than 20%. In general, the
SWI intensity and its response to any mitigation solution is a site-specific and depends
on a combination of factors. The MAR system to control SWI must be well planned and
assessed before implementation to ensure its effectiveness and success. The study
illustrated the importance of conducting laboratory scale experiments that indeed helps a
lot to gain a better understanding and enriching our knowledge about real-life challenges
like the SWI. Sand tank experiments allow change, mimic, and manipulate the different
possible conditions/variables that you may face when dealing with reality.
Misunderstanding or oversimplification of reality would result in the failure of efforts to
solve real problems in fields. Definitely, the field condition is more complicated than the
one simulated in our sand tank experiments, but obtaining insights about the effect of
each factor and parameter is important for comprehensive understanding and hence more
successful implementations in reality. The work presented in this thesis to be advanced
further to study the seawater dynamics in different layered heterogeneities and
hydrological conditions (e.g., abstraction rate, and tidal fluctuation at the sea boundary)
under MAR practices.
