Impacts of Sea-Level Rise and Recharge Fluctuations on Cutoff Wall Effectiveness for Freshwater Lens Development and Seawater Intrusion Mitigation in Unconfined Island Aquifers

Abstract

Sea-level rise (SLR) and regional precipitation pattern change cause island subsurface freshwater, typically shaped like a thin lens, to be at higher risk of contamination from seawater intrusion (SWI). Installing a cutoff wall is considered a feasible strategy for protecting coastal fresh groundwater from SWI. However, the performance of the cutoff wall in managing freshwater lens (FWL) development and mitigating SWI into island aquifers under SLR and aquifer recharge (RCH) fluctuations remains inadequately quantified. This study investigates how water table elevation (WTE), FWL depth, thickness, and SWI extent, measured by aquifer salt mass and freshwater volume, in an island aquifer equipped with cutoff walls, respond to SLR and RCH fluctuations. It focuses on a two-dimensional, variable-density island groundwater simulation model based on hydrogeological conditions of San Salvador Island, Bahamas. The results demonstrate that RCH critically influences cutoff wall effectiveness for FWL development and SWI mitigation, with higher RCH amplifying gains in WTE, FWL metrics, freshwater storage, and aquifer salt removal, but this influence diminishes with wall depth increasing. SLR elevates WTE in a stable manner associated with its magnitude but negligibly affects the cutoff wall performance in FWL enhancement and SWI mitigation. Under simultaneous SLR and RCH fluctuations, SLR can offset the WTE reduction caused by reduced RCH, but the joint effects of SLR and RCH on FWL metrics, freshwater storage and aquifer salt removal align with their individual impacts. Moreover, cutoff walls are more efficient in low-RCH settings, yielding greater relative improvements in FWL development and SWI mitigation per unit wall depth increase.

1. Introduction

Seawater intrusion (SWI), a global environmental issue, is a phenomenon of subsurface freshwater–seawater interface migrating landward and the subsequent salinization of coastal aquifers. It is widely acknowledged that SWI is driven by excessive groundwater extraction and climate change, which result in the recession in the hydraulic gradient between seaward-discharging freshwater and landward-moving seawater [1]. The serious effects of SWI include endangering coastal aquifer ecosystems and restricting subsurface freshwater access for coastal communities [2,3].

Island aquifers, a specific type of coastal aquifer, are usually sustained solely by precipitation recharge, forming lens-shaped freshwater bodies atop denser seawater [4] (Figure 1a). These freshwater lenses are typically thin due to the limited land area, low recharge rates, or/and high hydraulic conductivity [5,6], which restrict water accumulation. Due to their shape and size, island freshwater lens (FWL) is therefore highly vulnerable to SWI. However, it is important to note that, for many island communities, freshwater aquifers are the sole accessible and economically viable source of freshwater [4,7,8]. Protecting island subsurface freshwater resources from SWI contamination is therefore of critical importance.

To date, various countermeasures have been developed to protect subsurface freshwater from SWI, mainly including: (1) hydraulic barriers, which enhance the seaward hydraulic gradient by recharging freshwater into aquifers, extracting saltwater near the freshwater–saltwater transition zone, or a combination of both approaches [9,10,11]; (2) subsurface physical barriers, such as cutoff walls and subsurface dams, which are impermeable or low-permeability structures built near the shoreline and perpendicular to groundwater flow paths, effectively retaining groundwater and limiting SWI [12]; (3) land reclamation, which increases the area available for precipitation recharge and extends the pathway for SWI to reach freshwater aquifers [13]; (4) groundwater pumping optimization, striking an acceptable balance between local water demand and SWI mitigation by adjusting pumping patterns [14,15,16,17]; (5) rational land-use planning, promoting precipitation infiltration and reducing stormwater runoff by measures like installing permeable pavements and restoring mangrove ecosystems [18].

Most SWI mitigation measures may be less effective or inappropriate for islands due to the islands’ unique hydrogeological and socio-environmental factors. For example, scarcity of freshwater resources in islands limits the application of hydraulic barriers that depend on artificial recharge. Subsurface dams, which require an impervious bottom layer to prevent seawater from bypassing the barrier through underlying strata, are usually ineffective, as island aquifers often lack such layers and instead consist of highly permeable materials like coral sands or porous limestones [19,20,21]. Land reclamation demands advanced techniques and substantial investments in equipment, filling materials and continuous maintenance, making it an impractical option for islands with limited budgets. Moreover, islands are typically small, therefore leaving little room for implementing groundwater pumping optimization strategies to alleviate SWI through relocating pumping wells. Limited land availability also restricts large-scale surface interventions, such as implementing green infrastructure or other recharge-enhancing measures, rendering the rational land-use planning less effective.

Beyond those aforementioned SWI mitigation measures, cutoff walls are widely adopted to limit SWI and protect groundwater in vulnerable coastal aquifers. They are vertically oriented, low-permeable structures placed in the upper part of the coastal aquifer, close and parallel to the shoreline, with a gap between the wall bottom and the underlying aquitard to allow for groundwater discharge [9] (Figure 1b). Their effectiveness has been consistently demonstrated through numerous laboratory and numerical studies [9,22,23,24]. The protective mechanism of a cutoff wall is to force freshwater to flow beneath it, increasing the velocity and repulsive force against the intruding saltwater wedge. The efficacy of this mechanism, however, is highly dependent on the aquifer’s hydrogeological conditions. For instance, aquifer stratification can disrupt flow dynamics and reduce the freshwater velocity at the wall’s base, thereby lessening its repulsion ability compared to a homogeneous aquifer [25]. Moreover, based on numerical simulations and a systematic sensitivity analysis, the study of [22] demonstrated that the protective effect of cutoff walls on coastal groundwater extractions is stronger for cases when the extractions are located at relatively small distances from the coast, relatively large depths, and in aquifers with small velocity ratio, weak mixing and high anisotropy. The work of [26] investigated the impact of cutoff walls on both fresh and saline submarine groundwater discharge behavior, finding that the cutoff wall construction generally reduces all components of submarine groundwater discharge fluxes.

Unlike other SWI mitigation measures, cutoff walls remain a feasible and effective option for preventing SWI into island aquifers. Installing cutoff walls prevents SWI into island aquifers without requiring extensive land areas, as needed for rational land-use planning or for relocating pumping wells in groundwater pumping optimization. It further relaxes specific geological conditions, such as the impervious bottom layer that is required for subsurface dams. The construction of cutoff walls also involves manageable costs and technical demands compared to more complex alternatives like hydraulic barriers and land reclamation.

The effectiveness of cutoff walls in repelling SWI and enhancing subsurface freshwater storage in island aquifers has been well-demonstrated in some studies recently [27,28]. The investigation of [27] compared the effects of cutoff walls and full-section physical barriers on freshwater lenses in both circular and strip-shaped islands using numerical simulations, laboratory-scale experiments, and analytical solutions. The findings showed that cutoff walls with lower permeability and greater thickness significantly enhance freshwater storage as anticipated, and when constructed to a sufficient depth, the cutoff wall can achieve comparable FWL to those created by full-section physical barriers, demonstrating its cost-effectiveness for mitigating SWI in oceanic islands. In [28], based on a two-dimensional (2D) cross-sectional model, the responses of island aquifers equipped with the cutoff wall (IAECW) to ocean surge inundation were simulated. The simulation results revealed that the presence of cutoff walls can increase freshwater storage in island aquifers and reduce seawater infiltration by up to 40% during inundation.

Global climate change, a well-documented and ongoing phenomenon, is driving sea-level rise (SLR) worldwide while altering regional precipitation and evapotranspiration patterns. The Intergovernmental Panel on Climate Change (IPCC) reported that relative to the 1995–2014 baseline, the global mean sea level would rise 0.15–0.23 m under the low-emission scenario (SSP1–1.9) and 0.20–0.29 m under the high-emission scenario (SSP5–8.5) by 2050, with the SLR rate expected to accelerate in the future [29]. SLR reduces the hydraulic gradient that drives subsurface freshwater discharging seaward while simultaneously increasing the hydrostatic pressure from the ocean on the aquifer’s submarine boundary, thereby destabilizing the aquifer’s equilibrium and exacerbating SWI [30,31,32]. Climate-induced shifts in precipitation and evapotranspiration patterns significantly influence aquifer recharge (RCH) on islands, which may either increase or decrease depending on localized climatic conditions, placing significant pressure on FWL [33,34,35].

A key limitation of cutoff walls is their statically permanent nature once installed. Given the inevitability of SLR and RCH fluctuations, it is critical to investigate how these factors influence a cutoff wall’s performance in mitigating SWI and shaping freshwater zones in coastal aquifers. However, to the best of the authors’ knowledge, no study has yet investigated how SLR and RCH influence the performance of cutoff walls in mitigating SWI and preserving groundwater resources on islands.

Motivated by this, the main purpose of this study is to assess the impact of cutoff walls on island groundwater and evaluate their effectiveness in alleviating SWI under different levels of SLR and RCH using numerical simulations. This study focuses on a 2D, variable-density, dispersive, island aquifer simulation model based on hydrogeological conditions observed in the island aquifer of San Salvador Island, Bahamas. Specifically, a predefined set of SWI simulation scenarios is designed to investigate how RCH, SLR and the cutoff wall depth (D) affect water table elevations and FWL in the IAECW and the effectiveness of the cutoff wall in mitigating SWI extent, measured by subsurface freshwater volume and aquifer salt mass. The MODFLOW family SEAWAT model is employed to simulate the resilience of IAECW to SWI in different cases.

2. Materials and Methods

2.1. Study Area

The goal of this study is to investigate the impacts of SLR and RCH fluctuations on the cutoff wall performance in FWL development and SWI mitigation in unconfined island aquifers, using the San Salvador Island aquifer as a case study. San Salvador Island is located within the Bahamian Archipelago (Figure 2), about 600 km east-southeast of Miami, and sits on a small, isolated carbonate platform [36,37]. This is a strip island, and it is about 20 km long north-to-south and has an average width west-to-east of approximately 8 km [38]. The topography is dominated by consolidated carbonate dune ridges, with elevations a few meters above sea level [39]. Characterized by a subtropical climate, San Salvador Island has an annual temperature ranging between 22 and 28 °C [37] and annual precipitation and potential evaporation of 1000–1250 mm/year and 1250–1375 mm/year, respectively [40].

2.2. Numerical Simulation Model Development for SWI

This study applies the SEAWAT model to simulate the SWI process in the island aquifer. SEAWAT couples the groundwater flow model (MODFLOW) and the solute transport model (MT3DMS) to solve the variable-density flow equations using a finite-difference numerical approach [41]. Since the SEAWAT groundwater model can account for water density variations that depend on salt concentration, it is well-suited for simulating flow in aquifers characterized by freshwater–seawater interactions. Governing equations for groundwater flow and solute transport are presented in Appendix A.

To efficiently evaluate how SLR and RCH affect the cutoff wall performance in FWL and SWI extent in San Salvador Island aquifer, a simplified 2D, variable-density, dispersive, vertical ‘cross-section’ model is adopted. This 2D cross-section model represents a homogeneous, isotropic, unconfined island aquifer as a rectangular domain measuring 1000 m in length, 63 m in height and 1 m in width. The aquifer domain is discretized into a finite-difference regular grid with cells sized 2 m × 4 m at the top layer and 2 m × 1 m in the remaining layers, in the lateral and vertical directions, respectively. Two additional grid columns are used to represent the boundary conditions at the leftmost and rightmost ends of the domain so that the finite-difference grid is made up of 502 columns and 60 layers, for a total of 30,120 cells.

Figure 3 shows a conceptualisation of the aquifer domain along with the numerical model grid and its boundary conditions. A no-flow boundary is prescribed at the model bottom. The model top is a specified flux boundary, reflecting the RCH from precipitation. RCH is uniform over the island width and constant over time for each SWI simulation case. At the left and right boundaries, a constant head of $h_{\mathrm{s}}$ is prescribed over the water column, which represents the sea level relative to the datum. At the same boundaries, a constant salt concentration C of 35.0 g/L is imposed, which represents the salt content in seawater. Two cutoff walls are installed symmetrically at both sides of the aquifer in the columns immediately bordering the sea boundaries (Figure 3, brown cells). The walls extend from the ground surface to a certain depth D below the initial sea level, and each wall has a simulated thickness of 2 m, corresponding to the width of a single column in the discretized model grid. The wall cells are assigned a very low hydraulic conductivity (HK) (1.0 × 10⁻⁵ m/day) to represent their effectively impermeable properties in the numerical simulation model [23,28].

To model the effects of SLR and RCH fluctuations on FWL dynamics and SWI extent in IAECW at the steady state, the transient simulation of flow and solute transport are run for a sufficiently long period of time to reach the steady state. The groundwater flow equation is solved using the Preconditioned Conjugate Gradient solver, with both the head-change convergence criterion and the residual convergence criterion set to 1 × 10⁻⁵, and Solute transport was solved using the Generalized Conjugate Gradient package within MT3DMS, with a relative concentration convergence criterion of 1 × 10⁻⁶. These solver tolerances were selected to ensure numerical stability and solution accuracy for density-dependent flow conditions. A baseline scenario is first developed to simulate the island FWL under steady-state conditions of natural groundwater recharge from precipitation, no SLR and no cutoff wall. This serves as the initial condition to model the aquifer freshwater distribution under various scenarios of cutoff wall installation, SLR and RCH fluctuations. For the simulations involving cutoff wall, SLR and RCH fluctuations, SEAWAT is run until a steady state is reached, which is typically between 45 and 150 years depending on the simulated scenarios. Correspondingly, the required CPU time for each simulation varies from a minimum of about 20 min to a maximum of over 1 h.

Table 1. Model parameters used for SWI simulation in the 2D island aquifer model.

Model Component	Parameters	Units	Values
Groundwater Flow	RCH	m/year	0.1~0.3
	Effective porosity	\	0.3
	Specific yield	\	0.15
	Horizontal HK	m/day	50.0
	HK transversal anisotropy ratio	\	1.0
	HK vertical anisotropy ratio	\	1.0
Solute Transport	Longitudinal dispersivity	m	1.0
	Transversal dispersivity	m	0.1
	Vertical dispersivity	m	0.01
	Molecular diffusion coefficient	m2/s	1.0 × 10−9
	Aquifer recharge concentration	g/L	0
Density Dependence	Freshwater density	kg/m3	1000
	Seawater density	kg/m3	1025
	Density/concentration slope 1	\	0.7143
Cutoff Wall	HK	m/day	1.0 × 10−5

¹ The water density ${\rho }_w$ [kg/m³] varies linearly with the salt concentration $C$ [kg/m³] through the equation ${\rho }_w=1000+0.7143·C$ [41].

provides a list of the relevant parameters adopted in the simulation model introduced above. A horizontal HK of 50 m/day is assigned based on field studies of the Lucayan Limestone in the Bahamas, which report values typically ranging from 10 to 100 m/day, and an isotropy ratio of 1.0 is assumed, due to young carbonate platforms often exhibiting high primary porosity and poorly developed vertical stratification. These parameters are set based on published works [35,42] that have used islands in the Bahamian archipelago as case studies. Details on how the values of $h_{\mathrm{s}}$, RCH, and D are selected for conducting SWI simulations in the island aquifer model are provided in the following section.

2.3. SWI Simulation Scenarios

This work is devoted to identifying how FWL and SWI extent in IAECW respond to climate change, with a focus on the influence of $h_{\mathrm{s}}$, RCH fluctuations and D on this response, based on SWI simulations using SEAWAT. Inherently, numerous SWI simulations need to be conducted, with the three independent variables (IDVs) of D, $h_{\mathrm{s}}$ and RCH being varied continuously. To draw conclusions while limiting computational costs, our investigation is based on the analysis of a predefined set of SWI simulation scenarios (i.e., SEAWAT model runs), expressed as a prescribed ensemble of variable sets. The comparison of differences in water table elevation, FWL geometry, freshwater storage and aquifer salinization will be conducted as the basis to draw general conclusions.

Table 2. IDV values considered for the SWI simulations.

Variables	Discrete Values
D (m)	0	9	15	21	27	33
$h_{\mathrm{s}}$ (m)	0		0.1		0.2
RCH (m/year)	0.1		0.2		0.3

provides a description of the ensemble of IDV sets (D, $h_{\mathrm{s}}$, RCH) used in this study. The cutoff wall depth D varies over 6 discrete values (0, 9, 15, 21, 27 and 33 m), representing scenarios ranging from no cutoff wall installed (D = 0 m) to a sufficiently deep cutoff wall (D = 33 m), where D reaches up to nearly half of the aquifer’s thickness. For $h_{\mathrm{s}}$, there are 3 potential values, 0 m, 0.1 m, and 0.2 m, representing the sea level at the datum and two different magnitudes of SLR relative to the datum, respectively [29]. RCH varies across three discrete values, 0.1 m/year, 0.2 m/year, and 0.3 m/year, in which 0.2 m/year represents the current RCH in the Bahamian archipelago, while 0.1 m/year and 0.3 m/year respectively represent cases where RCH decreases or increases by 50% due to climate change.

A full factorial combination of these variable values leads to an ensemble of 6 × 3 × 3 = 54 SWI simulation scenarios, and thus as many SEAWAT model runs. The SWI simulation scenario with no cutoff wall, no SLR and RCH equal to 0.2 m/year is considered the baseline scenario.

2.4. Indicators for Quantifying FWL Dynamics and SWI Extent

To quantify the impact of D, $h_{\mathrm{s}}$ and RCH on the FWL and SWI extent in island aquifers, this study develops eight indicators: (1) absolute change in average water table elevation $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$, measuring the overall shift in the groundwater level; (2) relative change in average water table elevation $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{r}}$, providing a normalized measure of the overall water table shift for comparing changes across different scenarios; (3) absolute change in maximum water table elevation $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$, tracking changes in the highest groundwater level; (4) relative change in maximum water table elevation $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{r}}$, offering a normalized perspective on the peak water level change for comparisons across different scenarios; (5) relative change in FWL depth $∆d$, quantifying how much the freshwater–saltwater interface deepens or shallows; (6) relative change in FWL thickness $∆T$, measuring the change in the maximum vertical extent of the subsurface freshwater resource; (7) relative change in fresh groundwater volume $∆FV$, measuring the change in freshwater storage in the aquifer; (8) relative change in total salt mass in the aquifer $∆SM$, quantifying the extent of aquifer salinization under SWI.

$∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$ is calculated as follows:

$$∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}={WTE}_{\mathrm{a}\mathrm{v}\mathrm{g}}-{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},0}$$

(1)

where ${WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},0}$ represents the average water table elevation at the steady state in the baseline scenario, while ${WTE}_{\mathrm{a}\mathrm{v}\mathrm{g}}$ is the average water table elevation at the steady state for other SWI simulation scenarios, which depends on the IDV set. The average water table elevation is calculated by integrating the water table elevations across all top-layer island aquifer cells and then dividing by the number of these grid cells.

$∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{r}}$ is calculated as follows:

$$∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{r}}={\displaystyle \frac{{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g}}-{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},0}}{{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},0}}}·100 [\%]$$

(2)

$∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$ is given by the following equation:

$$∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}={WTE}_{\mathrm{m}\mathrm{a}\mathrm{x}}-{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},0}$$

(3)

where ${WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},0}$ represents the maximum water table elevation at the steady state in the baseline scenario, while ${WTE}_{\mathrm{m}\mathrm{a}\mathrm{x}}$ is the maximum water table elevation at the steady state for other SWI simulation scenarios, which depends on the IDV set.

$∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{r}}$ is given by the following equation:

$$∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{r}}={\displaystyle \frac{{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x}}-{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},0}}{{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},0}}}·100 [\%]$$

(4)

$∆d$ is calculated by

$$∆d={\displaystyle \frac{d_{\mathrm{c}}-d_0}{d_0}}·100 [\%]$$

(5)

where $d_0$ represents the FWL depth at the steady state in the baseline scenario, and $d_{\mathrm{c}}$ is the FWL depth at the steady state for other SWI simulation scenarios. The FWL depth, in this study, is defined as the distance from the freshwater–saltwater interface at the island center to the ground surface (see Figure 1a). This study defines groundwater with a salt concentration of 1.0 g/L or less as freshwater. The 1.0 g/L threshold is adopted here as an operational and analytical benchmark, and it is a common convention in island hydrogeology and numerical modeling of the freshwater–saltwater interface, which can effectively represent the transition between potable fresh groundwater and the brackish mixing zone [17,43].

$∆T$ is given by the following equation:

$$∆T={\displaystyle \frac{T_{\mathrm{c}}-T_0}{T_0}}·100 [\%]$$

(6)

where $T_0$ represents the FWL thickness at the steady state in the baseline scenario, while $T_{\mathrm{c}}$ is the FWL thickness at the steady state for other SWI simulation scenarios. The FWL thickness, in this study, refers to the distance from the water table to the freshwater–saltwater interface at the island center (see Figure 1a).

$∆FV$ is given by

$$∆FV={\displaystyle \frac{FV-{FV}_0}{{FV}_0}}·100 [\%]$$

(7)

where ${FV}_0$ represents the steady-state freshwater volume in the baseline scenario, while $FV$ denotes the steady-state freshwater volume for other SWI simulation scenarios. Subsurface freshwater volume is calculated by spatial integration of the pore volume in those grid cells where the simulated salt concentration is no more than 1.0 g/L.

$∆SM$ is calculated by the following equation:

$$∆SM={\displaystyle \frac{SM-{SM}_0}{{SM}_0}}·100 [\%]$$

(8)

where ${SM}_0$ represents the steady-state salt mass in the aquifer baseline scenario, while $SM$ denotes the steady-state salt mass in the aquifer for other SWI simulation scenarios. The salt mass in the aquifer is calculated by integrating the salt concentration multiplied by the pore volume across all model grid cells.

The IAECW scenarios exhibiting higher values for $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$, $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{r}}$, $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$, $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{r}}$, $∆d$, $∆T$, and $∆FV$, coupled with a lower $∆SM$, are interpreted as having a greater seaward hydraulic gradient, a deeper and thicker FWL, greater freshwater storage, and less aquifer salinization, showing greater resilience to SWI and climate changes.

3. Results and Discussion

3.1. Impact of RCH and SLR on FWL and SWI Extent in Island Aquifers

The distribution of groundwater concentrations under different RCH and sea level conditions, prior to the construction of the cutoff wall, is shown in Figure 4. It can be observed that the subsurface freshwater resource forms a lens-shaped body in all profiles, with its thickness ranging from a maximum at the island’s center to zero at the shoreline, and the width of the freshwater–saltwater transition zone gradually increases from the island center toward the shoreline, which is consistent with published theoretical solutions [44]. Moreover, Figure 4 shows that in the absence of the cutoff wall, the FWL is more sensitive to RCH than to SLR. A decrease in RCH from 0.2 to 0.1 m/year causes a significant shrinkage of the lens, while an increase to 0.3 m/year causes a noticeable expansion. In contrast, a rise in $h_{\mathrm{s}}$ from 0 to 0.2 m results in only a minor reduction in lens size.

Table 3. Summary of steady-state indicator values for the island aquifer at the baseline case.

Indicators	${\mathit{W}\mathit{T}\mathit{E}}_{\mathbf{a}\mathbf{v}\mathbf{g},0}$ (m)	${\mathit{W}\mathit{T}\mathit{E}}_{\mathbf{m}\mathbf{a}\mathbf{x},0}$ (m)	${\mathit{d}}_0$ (m)	${\mathit{T}}_0$ (m)	${\mathit{F}\mathit{V}}_0$ (m2)	${\mathit{S}\mathit{M}}_0$ (kg/m)
Value	0.29	0.37	15.06	12.43	8328	1,692,912

shows a summary of steady-state indicator values for the island aquifer at the baseline case. These reference values are provided to facilitate interpretation of the relative change plots by establishing a consistent benchmark against which scenario-based deviations can be quantitatively evaluated.

Figure 5 presents the indicator values that quantify FWL thickness and depth under different SLR and RCH scenarios in the absence of the cutoff wall, and Figure 6 shows the corresponding WTE and SWI indicator values. All indicator values in Figure 5 and Figure 6 are calculated relative to the corresponding indicator values in the baseline scenario ($h_{\mathrm{s}}$ = 0 m, RCH = 0.2 m/year) for comparison.

As shown in Figure 5, in the absence of the cutoff wall, progressive growth is observed in both $d_{\mathrm{c}}$ and $T_{\mathrm{c}}$ as RCH increases from 0.1 to 0.3 m/year when $h_{\mathrm{s}}$ is given. For instance, at $h_{\mathrm{s}}$ = 0 m, $d_{\mathrm{c}}$ rises from 10.37 m to 15.26 m to 18.25 m, while $T_{\mathrm{c}}$ expands from 7.28 m to 12.43 m to 15.22 m, with similar trends occurring at higher sea levels ($h_{\mathrm{s}}$ = 0.1 m and 0.2 m). However, the FWL system exhibits significantly higher sensitivity to reductions in RCH than to increases. Specifically, when comparing against baseline conditions ($h_{\mathrm{s}}$ = 0 m, RCH = 0.2 m/year), a decrease in RCH to 0.1 m/year causes substantial declines of 32.02% in FWL depth and 41.40% in FWL thickness, but increasing RCH to 0.3 m/year results in comparatively smaller increases of just 19.61% for FWL depth and 22.42% for FWL thickness under the same sea level conditions ($h_{\mathrm{s}}$ = 0 m). In contrast, SLR consistently reduces FWL metrics; for instance, at RCH = 0.2 m/year, $d_{\mathrm{c}}$ declines from 15.26 m ($h_{\mathrm{s}}$ = 0 m) to 15.09 m ($h_{\mathrm{s}}$ = 0.1 m) to 14.66 m ($h_{\mathrm{s}}$ = 0.2 m), while $T_{\mathrm{c}}$ drops from 12.43 m ($h_{\mathrm{s}}$ = 0 m) to 12.26 m ($h_{\mathrm{s}}$ = 0.1 m) to 11.86 m ($h_{\mathrm{s}}$ = 0.2 m).

Figure 5 also intuitively indicates that the FWL metrics, $d_{\mathrm{c}}$, $∆d$, $T_{\mathrm{c}}$, and $∆T$, are predominantly controlled by RCH, where for a given RCH, these indicator values cluster tightly, showing minimal variation across different SLR scenarios. This demonstrates that RCH fluctuations have a stronger influence on the FWL than SLR. For instance, when RCH is held constant at 0.2 m/year, a $\pm$0.1 m/year variation in RCH at $h_{\mathrm{s}}$ = 0 m leads to significant changes in FWL dynamics: a 0.1 m/year decrease in RCH reduces FWL depth by 32.02% and thickness by 41.40%, while a 0.1 m/year increase in RCH raises FWL depth by 19.61% and thickness by 22.42%. In contrast, SLR changing from 0 m to 0.2 m at RCH = 0.2 m/year results in minor FWL depth declines by 1.12% (at $h_{\mathrm{s}}$ = 0.1 m) and 3.94% (at $h_{\mathrm{s}}$ = 0.2 m), while FWL thickness decreases by 1.38% (at $h_{\mathrm{s}}$ = 0.1 m) and 4.55% (at $h_{\mathrm{s}}$ = 0.2 m). These results highlight the greater sensitivity of FWL to RCH fluctuations than to SLR.

The results for $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$, $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{r}}$, $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$ and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{r}}$ under different combinations of SLR and RCH (Figure 6a–d) reveal that: (1) Both SLR and intensified RCH increase average and maximum WTE in island aquifers, but SLR plays a more significant role in affecting WTE than RCH. For instance, compared to the case of RCH = 0.2 m/year and $h_{\mathrm{s}}$ = 0 m, a $\pm$0.1 m/year RCH variation and the sea level keeping unchanged cause modest WTE changes (30.48% decrease in average WTE and 29.90% decrease in maximum WTE for RCH = 0.1 m/year; 23.08% and 22.58% increases respectively for RCH = 0.3 m/year), whereas SLR changing from 0 m to 0.2 m at RCH = 0.2 m/year drives fairly larger shifts, with average WTE rising by 34.69% ($h_{\mathrm{s}}$ = 0.1 m) and 69.39% ($h_{\mathrm{s}}$ = 0.2 m), while maximum WTE decreases by 27.00% ($h_{\mathrm{s}}$ = 0.1 m) and 53.95% ($h_{\mathrm{s}}$ = 0.2 m). (2) WTE rise precisely matches SLR magnitude if RCH remains constant. For instance, at RCH = 0.2 m/year, both average and maximum WTE increase by 0.10 m at $h_{\mathrm{s}}$ = 0.1 m and by 0.20 m at $h_{\mathrm{s}}$ = 0.2 m relative to the scenario (RCH = 0.2 m/year, $h_{\mathrm{s}}$ = 0 m). (3) Average WTE exhibits greater sensitivity to SLR and RCH variations than maximum WTE, which is evident from the comparative analysis of $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{r}}$ and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{r}}$, highlighting the differential response of these parameters to hydrological forcing.

In Figure 6e,f, it is interesting to observe that although both RCH recession and SLR contribute to declines in fresh groundwater volume and increases in aquifer salt mass, RCH fluctuations impose a more significant impact than SLR. For instance, compared to the case of RCH = 0.2 m/year and $h_{\mathrm{s}}$ = 0 m, a $\pm$0.1 m/year RCH variation (with sea level held constant) causes larger changes in fresh groundwater volume and aquifer salt mass: $∆FV$ drops by 41.86% and $∆SM$ rises by 7.32% at RCH = 0.1 m/year, while $∆FV$ increases by 29.15% and $∆SM$ declines by 5.57% at RCH = 0.3 m/year. In contrast, SLR changing from 0 m to 0.2 m at RCH = 0.2 m/year drives smaller shifts: $∆FV$ decreases by 0.72% at $h_{\mathrm{s}}$ = 0.1 m and 3.36% at $h_{\mathrm{s}}$ = 0.2 m, while $∆SM$ increases by just 0.26% at $h_{\mathrm{s}}$ = 0.1 m and 0.50% at $h_{\mathrm{s}}$ = 0.2 m. That also indicates that fresh groundwater volume and aquifer salt mass respond more strongly to reductions in RCH than to increases.

The simulation results of specified hydrogeological scenarios in Figure 5 and Figure 6, overall, show that in the absence of the cutoff wall, both SLR and RCH fluctuations significantly influence groundwater dynamics and SWI extent in island aquifers. In particular, RCH fluctuations have a more pronounced impact than SLR on FWL depth, thickness, fresh groundwater volume and aquifer salt mass, and FWL depth, thickness, fresh groundwater volume and aquifer salt mass exhibit greater sensitivity to RCH reductions than increases, which is consistent with established theoretical solutions [5,6,45]. Compared to RCH, SLR plays a dominant role in driving WTE changes, with WTE rising linearly in proportion to SLR magnitude under constant RCH. This differential response explains why FWL metrics (depth, thickness), fresh groundwater volume, and aquifer salt mass show larger variations under RCH fluctuations but smaller shifts under SLR, where SLR-induced WTE rise enhances the hydraulic gradient and partially counteracts SWI progression.

3.2. Impact of RCH on Cutoff Wall Performance in FWL Development and SWI Mitigation

Figure 7 presents how RCH influences the performance of the cutoff wall in FWL development and SWI mitigation in IAECW, under a constant sea level of 0 m. Figure 7a–Figure 7d, respectively, depicts the profiles of $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$, $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{r}}$, $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$, and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{r}}$ plotted against D for varying RCH conditions. Figure 7a and Figure 7c, respectively, shows that in all RCH cases, the $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$ and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$ profiles experience a gradual increase up to D of 9 m before rising more sharply with further deepening of the cutoff wall, with larger RCH conditions consistently producing higher values of both $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$ and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$ than smaller RCH. It is also noted that both $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$ and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$ curves initially exhibit wide separation across RCH conditions, but this separation progressively narrows with increasing D (Figure 7a,c).

The $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{r}}$ (Figure 7b) and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{r}}$ profiles (Figure 7d) mimic the trends observed in their absolute counterparts ($∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$ and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$, respectively). When D extends from 0 m to 9 m, and to 33 m, $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{r}}$ grows from −30.49% to −6.26%, and to 179.92% at RCH = 0.1 m/year; from 0% to 20.92%, and to 195.42% at RCH = 0.2 m/year; and from 23.08% to 42.17%, and to 208.98% at RCH = 0.3 m/year. Similarly, D extends from 0 m to 9 m, and to 33 m, $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{r}}$ climbs from −29.90% to −15.52%, and to 121.31% at RCH = 0.1 m/year; from 0% to 11.93%, and to 136.79% at RCH = 0.2 m/year; and from 22.58% to 33.19%, and to 150.47% at RCH = 0.3 m/year.

The observations in Figure 7a–d demonstrate that: (1) in all cases of RCH, the deeper cutoff wall leads to greater lift of WTE, with particularly enhanced efficiency of increasing the seaward hydraulic gradient when its depth exceeds the natural FWL depth (see Figure 5a). (2) For any given D, higher RCH results in more pronounced increases in both average and maximum WTE in IAECW systems than lower RCH, because greater volumetric water recharge enhances hydraulic head build-up upstream of the cutoff wall, and in the meantime, the cutoff wall traps excessive recharge. (3) RCH is the primary control on WTE increases for shallow cutoff walls (e.g., D < 15 m), but its influence diminishes with cutoff wall depth increasing. The progressive narrowing of intervals among curves demonstrates that D gradually supersedes RCH as the dominant factor governing WTE lift in deeper wall configurations. (4) Relative to pre-construction conditions, cutoff walls induce greater relative WTE increases in low-RCH IAECW systems than in high-RCH cases, underscoring their greater efficacy in lifting WTE in low-RCH island aquifers, and thus greater resilience enhanced against SWI. (5) Under all RCH conditions, average WTE in IAECW exhibits stronger relative growth than maximum WTE, suggesting that the cutoff wall promotes more uniform hydraulic head increase by effectively redistributing water across the enclosed area.

Figure 7e and Figure 7f, respectively, illustrates that, across all RCH conditions, both $∆d$ and $∆T$ increase with D, where higher RCH values produce greater $∆d$ and $∆T$ values. The growth in $∆d$ and $∆T$ is gradual for D < 9 m, but becomes sharp once the cutoff wall extends deeper beyond that depth. The separation between the $∆d$ (or $∆T$) curves under different RCH conditions is initially wide, but they narrow as D increases. This narrowing occurs because the $∆d$ (or $∆T$) curves for lower RCH conditions rise more rapidly than those for higher RCH, and the curves almost approach parallelism at greater depths (D > 27 m). Specifically, with D extending from 0 m to 33 m, $∆d$ progresses from −32.02% to 111.30% (RCH = 0.1 m/year), from 0% to 128.16% (RCH = 0.2 m/year), and from 19.61% to 144.60% (RCH = 0.3 m/year). Similarly, with D increasing from 0 m to 33 m, $∆T$ increases from −41.40% to 141.86% (RCH = 0.1 m/year), from 0% to 163.02% (RCH = 0.2 m/year), and from 22.42% to 183.61% (RCH = 0.3 m/year).

Figure 8 displays the responses of groundwater salt concentration and FWL to RCH fluctuations (RCH = 0.2 m/year for Figure 8a–c; RCH = 0.1 m/year for Figure 8d–f; RCH = 0.3 m/year for Figure 8g–i) for IAECW cases with three different cutoff wall depths (D = 9, 15, and 33 m). The results indicate that: (1) in all cases of RCH, deeper walls (particularly beyond natural FWL depth) robustly enhance FWL depth and thickness in IAECW; (2) the influence of RCH on the cutoff wall’s capacity to promote FWL depth and thickness diminishes with D increasing, eventually leveling off; (3) low-RCH aquifers respond more sensitively to cutoff walls’ depth, experiencing larger $∆d$ or $∆T$ changes per unit D. This underscores the higher efficiency of cutoff walls in enhancing FWL depth and thickness under low-RCH environments compared to high-RCH ones.

In Figure 7g, $∆FV$ profiles in all RCH cases increase with D, with higher RCH leading to larger $∆FV$ values. $∆FV$ profiles rise slowly up to D = 9 m before steepening as the cutoff wall extends deeper. The initial wide separation among $∆FV$ curves progressively narrows and the curves beyond D = 25 m become almost parallel to each other. With D increasing from 0 m to 9 m, and to 33 m, $∆FV$ increases from −41.86% to −10.38%, and to 243.08% at RCH = 0.1 m/year; from 0 to 27.14%, and to 267.60% at RCH = 0.2 m/year; and from 29.15% to 56.72%, and to 286.53% at RCH = 0.3 m/year. These reveal that: (1) in all cases of RCH, deeper cutoff wall enhances freshwater storage in island aquifers, with particularly high efficiency occurring when D exceeds the natural FWL depth (see Figure 5a), beyond which freshwater volume shows near-linear growth with increasing D; (2) for a given D, island aquifers with higher RCH can store more freshwater than those with lower RCH; (3) the influence of RCH on the cutoff wall’s performance in freshwater storage depends on D, showing pronounced effects at shallow depths (e.g., D < 9 m), but progressively diminishes and levels off as D increases to sufficient depth; (4) for a given increase in D, lower RCH cases exhibit greater changes in $∆FV$, demonstrating that cutoff walls provide more substantial freshwater storage efficiency in low-RCH IAECW compared to high-RCH ones.

Figure 7h shows that the $∆SM$ profiles negatively correlate with increasing D across all RCH conditions, with higher RCH resulting in smaller $∆SM$ values. $∆SM$ profiles decline gradually at D = 0~9 m before undergoing a sharp decline. The initial wide separation among $∆SM$ curves progressively narrows and the curves become almost parallel to each other at D > 25 m. To be specific, as D increases from 0 m to 9 m, and to 33 m, $∆SM$ declines from 7.32% to 1.70%, and to −42.40% at RCH = 0.1 m/year; from 0 to −4.83%, and to −46.12% at RCH = 0.2 m/year; and from −5.57% to −9.90%, and to −49.35% at RCH = 0.3 m/year. These indicate that: (1) in all cases of RCH, deeper walls more effectively reduce aquifer salt mass and mitigate aquifer salinization, with particularly high efficiency occurring when D exceeds the natural FWL depth, beyond which salt mass decreases nearly linearly with increasing D; (2) for any given D, higher RCH conditions enable greater salt removal compared to lower RCH scenarios; (3) the influence of RCH on salt removal efficacy varies with D: its impact is most pronounced at shallow depths (D < 9 m) but progressively diminishes and levels off beyond a critical depth; (4) low-RCH aquifers show greater responsiveness to cutoff wall installation (larger changes in $∆SM$ per unit D increase), highlighting that walls provide greater salt removal in IAECW with lower RCH compared to those with higher RCH.

The observations in Figure 7 and Figure 8, in general, demonstrate that RCH imposes a substantial influence on the cutoff wall performance in FWL development and SWI mitigation in island aquifers. Under the same conditions of cutoff wall depth, compared to lower RCH scenarios, higher RCH can lead to more pronounced increases in average and maximum WTE, FWL depth and thickness, freshwater storage, and aquifer salt removal. The impact of RCH on cutoff wall performance is dependent on D. RCH makes a significant difference to WTE increases, FWL development, freshwater storage and aquifer salt removal under shallow cutoff walls, but its influence diminishes with the cutoff wall extending deeper, eventually leveling off. Island aquifers characterized by lower RCH show greater responsiveness to cutoff wall installation, experiencing greater relative increases in WTE, FWL depth and thickness, freshwater storage and aquifer salt removal per unit D increase. This indicates that cutoff walls provide higher efficiency in promoting FWL development and mitigating SWI extent in IAECW systems with lower RCH.

3.3. Impact of SLR on Cutoff Wall Performance in FWL Development and SWI Mitigation

Figure 9 presents how SLR affects the performance of the cutoff wall in FWL development and SWI mitigation in island aquifers, under a constant RCH (i.e., 0.2 m/year). Figure 9a–d depicts the profiles of $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$, $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{r}}$, $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$, and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{r}}$, respectively, plotted against D for different SLR conditions. Figure 9a and Figure 9c, respectively, shows that in all SLR cases, the $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$ and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$ profiles experience a gradual increase up to D of 9 m before rising more sharply with further deepening of the cutoff wall, and larger SLR magnitudes consistently produce higher values of both $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$ and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$ than smaller SLR. Another notable observation is that both $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$ and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$ profiles remain nearly parallel across all SLR conditions throughout the range of D, with their separation approximately equal to the SLR magnitude. Specifically, as D increases from 0 m to 9 m, and to 33 m, $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{r}}$ grows from 0 m to 0.06 m, and to 0.56 m at $h_{\mathrm{s}}$ = 0 m; from 0.10 m to 0.16 m, and to 0.66 m at $h_{\mathrm{s}}$ = 0.1 m; and from 0.20 m to 0.26 m, and to 0.76 m at $h_{\mathrm{s}}$ = 0.2 m. With D increasing from 0 m to 9 m, and to 33 m, $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{r}}$ climbs from 0 m to 0.04 m, and to 0.50 m at $h_{\mathrm{s}}$ = 0 m; from 0.10 m to 0.15 m, and to 0.61 m at $h_{\mathrm{s}}$ = 0.1 m; and from 0.20 m to 0.25 m, and to 0.71 m at $h_{\mathrm{s}}$ = 0.2 m.

In Figure 9b,d, the $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{r}}$ and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{r}}$ profiles mimic the trends observed in their absolute counterparts $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$ (Figure 9a) and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$ (Figure 9c), respectively. When D extends from 0 m to 9 m, and to 33 m, $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{r}}$ grows from 0% to 20.92%, and to 195.42% at $h_{\mathrm{s}}$ = 0 m; from 34.69% to 56.30%, and to 231.09% at $h_{\mathrm{s}}$ = 0.1 m; and from 69.39% to 91.67%, and to 266.82% at $h_{\mathrm{s}}$ = 0.2 m. Similarly, with D increasing from 0 m to 9 m, and to 33 m, $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{r}}$ climbs from 0% to 11.93%, and to 136.79% at $h_{\mathrm{s}}$ = 0 m; from 27.00% to 39.36%, and to 164.47% at $h_{\mathrm{s}}$ = 0.1 m; and from 53.95% to 66.75%, and to 192.22% at $h_{\mathrm{s}}$ = 0.2 m.

The observations in Figure 9a–d demonstrate that: (1) a deeper cutoff wall leads to greater lift of WTE, with particularly enhanced efficiency when its depth exceeds the natural FWL depth; (2) larger SLR produces greater increases in WTE in the IAECW, and the impact of SLR on WTE in the IAECW is linearly dependent on SLR magnitude; (3) under all SLR conditions, average WTE in IAECW exhibits stronger relative growth than maximum WTE, indicating that the cutoff wall promotes more uniform hydraulic conditions by effectively redistributing water across the enclosed area.

Figure 9e and Figure 9f, respectively, illustrates that, across all SLR conditions, both $∆d$ and $∆T$ increase with D, and the $∆d$ and $∆T$ curves nearly overlap while rising linearly with D beyond 9 m. Only minor differences are observed between various $h_{\mathrm{s}}$ conditions at D < 9 m, where higher $h_{\mathrm{s}}$ yields slightly lower $∆d$ and $∆T$ profiles. Specifically, with D extending from 0 m to 33 m, $∆d$ progresses from 0% to 128.16% ($h_{\mathrm{s}}$ = 0 m), from −1.12% to 128.54% ($h_{\mathrm{s}}$ = 0.1 m), and from −3.94% to 128.69% ($h_{\mathrm{s}}$ = 0.2 m). Similarly, $∆T$ increases from 0% to 163.02% ($h_{\mathrm{s}}$ = 0 m), from −1.38% to 163.09% ($h_{\mathrm{s}}$ = 0.1 m), and from −4.54% to 165.29% ($h_{\mathrm{s}}$ = 0.2 m).

Figure 10 presents the responses of groundwater salt concentration and FWL to SLR ($h_{\mathrm{s}}$ = 0 m for Figure 10a–c; $h_{\mathrm{s}}$ = 0.1 m for Figure 10d–f; $h_{\mathrm{s}}$ = 0.2 m for Figure 10g–i) for IAECW cases with three different cutoff wall depths (D = 9, 15, and 33 m). These simulation results indicate that: (1) in all cases of SLR, deeper walls (particularly beyond natural FWL depth) robustly enhance FWL depth and thickness in IAECW; (2) the impact of SLR on the cutoff wall’s effectiveness in enhancing FWL depth and thickness is limited and contingent on wall depth. For walls shallower than the natural FWL depth, greater SLR leads to slightly smaller FWL depth and thickness compared to IAECW with lower SLR. Once the wall depth surpasses the natural FWL depth, the influence of SLR on the wall’s performance becomes negligible.

Likewise, in Figure 9g, $∆FV$ profiles across all SLR cases increase with D, showing only minor differences among various $h_{\mathrm{s}}$ conditions at D = 0~9 m (higher $h_{\mathrm{s}}$ yields slightly lower $∆FV$ profiles), and the curves nearly overlap with each other, rising linearly with D beyond 9 m. Figure 9h reveals an opposite trend for $∆SM$ profiles, which decline with D, exhibiting subtle variations among different $h_{\mathrm{s}}$ values at D = 0~9 m (higher $h_{\mathrm{s}}$ resulting in slightly higher $∆SM$ profiles) until the curves become almost on top of each other and drop linearly along with increasing D once D exceeds 9 m. The $∆FV$ and $∆SM$ changes are described as follows.

As D increases from 0 m to 9 m, and to 33 m, at $h_{\mathrm{s}}$ = 0 m, $∆FV$ rises from 0 to 27.14%, and to 267.60% while $∆SM$ declines from 0 to −4.83% to −46.12%; at $h_{\mathrm{s}}$ = 0.1 m, $∆FV$ changes from −0.72% to 25.82% to 267.41% while $∆SM$ decreases from 0.26% to −4.74%, and to −46.09%; and at $h_{\mathrm{s}}$ = 0.2 m, $∆FV$ increases from −3.36% to 24.64%, and to 267.33% while $∆SM$ drops from 0.50% to −4.65%, and to −46.06%. These indicate that: (1) across all cases of SLR, deeper cutoff walls more effectively enhance freshwater volume and reduce salt mass, especially when their depth exceeds the natural FWL depth, in which both freshwater storage and aquifer salt removal vary linearly with D; (2) SLR only affects the cutoff wall performance in freshwater storage and salt removal for D shallower than natural FWL depth, where larger SLR results in less freshwater storage and salt removal, but SLR’s impact becomes negligible once D surpasses natural FWL depth.

Overall, the observations in Figure 9 and Figure 10 demonstrate two key effects of SLR on cutoff wall performance in island aquifer settings. Larger SLR produces greater WTE increases in the IAECW, and this impact is stable and linearly dependent on SLR magnitude. However, SLR imposes a nearly negligible influence on the wall’s effectiveness in enhancing FWL depth, FWL thickness, freshwater storage, and aquifer salt removal. This is because SLR raises both the inland WTE and sea level, thereby preserving the relative head difference created by the cutoff wall, the primary driver for pushing the saltwater interface seaward. The key mechanisms for aquifer storing freshwater and flushing salt, namely recharge rate and the flow paths diverted by the wall, are independent of the absolute sea level. As a result, with SLR shifting the WTE upward, it does not diminish the cutoff wall’s capacity to actively enhance freshwater lens thickness, increase freshwater storage, or remove aquifer salt. This underscores that the wall’s utility is resilient to SLR.

3.4. Joint Impact of RCH and SLR on Cutoff Wall Performance in FWL Development and SWI Mitigation

Figure 11 presents how RCH and SLR jointly affect the performance of the cutoff wall in FWL development and SWI mitigation in island aquifers. Figure 11a–d depicts the profiles of $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$, $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{r}}$, $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$, and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{r}}$ as functions of D for different combinations of RCH and SLR, respectively. In Figure 11a,c, the $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$ and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$ profiles rise with increasing D, experiencing the gradual increase at shallow depths (D < 9 m) and sharp increase once the cutoff wall extends deeper than 9 m. Among all cases examined herein, the scenario with RCH = 0.3 m/year and $h_{\mathrm{s}}$ = 0.2 m yields the highest ${∆WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$ and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$ values across D, and the baseline case (RCH = 0.2 m/year, $h_{\mathrm{s}}$ = 0 m) consistently produces the lowest values. Compared to the baseline case, the case with RCH = 0.1 m/year and $h_{\mathrm{s}}$ = 0.1 m shows slightly higher ${∆WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$ values and nearly identical $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$ values at shallow depths (D < 9 m), beyond which both metrics increase more rapidly with D. The remaining two cases, the scenario with RCH = 0.1 m/year, $h_{\mathrm{s}}$ = 0.2 m and the scenario with RCH = 0.3 m/year, $h_{\mathrm{s}}$ = 0.1 m, exhibit intermediate behaviors, with their ${∆WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$ and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$ curves positioned between those previously mentioned cases. It is observed that the ${∆WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$ and $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$ curves for the case with RCH = 0.1 m/year and $h_{\mathrm{s}}$ = 0.2 m initially lie below those for RCH = 0.3 m/year and $h_{\mathrm{s}}$ = 0.1 m, but with D increasing, the curves of RCH = 0.1 m/year, $h_{\mathrm{s}}$ = 0.2 m exhibit a much more remarkable rise, eventually converging to and even surpassing the corresponding curves for RCH = 0.3 m/year and $h_{\mathrm{s}}$ = 0.1 m under deeper cutoff wall conditions.

Figure 11b,d displays similar trends observed in Figure 11a and Figure 11c, respectively, but under the same conditions of RCH, SLR, and D (D > 0), the values of ${∆WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{r}}$ are consistently greater than those of $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{r}}$.

The observations in Figure 11a–d demonstrate that: (1) a combination of SLR and enhanced RCH promotes cutoff wall effectiveness in lifting both average and maximum WTE, a trend amplified by the magnitudes of both SLR and RCH increase; (2) in the combination of RCH fluctuations and SLR, RCH fluctuations play a dominant role in governing WTE in IAECW with small D, but with D increasing, the influence of SLR becomes more pronounced, eventually surpassing RCH as the dominant control on WTE under larger D conditions; (3) SLR acts as a counteracting mechanism against the adverse impacts of reduced RCH on WTE in IAECW, which depends on D and SLR magnitude. In other words, SLR can either partially or fully offset the negative impact caused by lower RCH; specifically, a full offset, where the WTE is maintained or elevated despite low RCH, is more likely to occur in IAECW with larger D and/or higher SLR.

Figure 11e and Figure 11f, respectively, demonstrates that, across all RCH and SLR combinations, the $∆d$ and $∆T$ profiles under the same RCH conditions follow nearly identical trends. These $∆d$ and $∆T$ profiles rise with increasing D, climbing slowly at D < 9 m but rising sharply once the cutoff wall extends deeper beyond 9 m, with higher RCH values producing greater $∆d$ and $∆T$ values. Initially, $∆d$ (or $∆T$) curves under different RCH conditions diverge from each other, but with D increasing, the curves for lower RCH conditions exhibit steeper growth rates compared to those under higher RCH conditions. This response progressively reduces the inter-curve separation, with the profiles becoming almost parallel when the cutoff wall reaches sufficient depth. Figure 12 presents how groundwater salt concentration distribution and FWL respond to combined SLR and RCH fluctuations for three IAECW cases with different cutoff wall depths (D = 9, 15, and 33 m).

These indicate that: (1) under simultaneous SLR and RCH fluctuations, SLR exerts a negligible influence on FWL depth and thickness in IAECW, with RCH emerging as the primary driver of lens geometry; (2) the sensitivity of cutoff wall performance to RCH variations diminishes as D increases, eventually leveling off; (3) lower-RCH aquifers respond more strongly to cutoff walls (larger $∆d$ or $∆T$ changes per unit D), again underscoring the higher efficiency of cutoff walls in low-RCH environments compared to high-RCH ones.

Figure 11g and Figure 11h, respectively, illustrates that the $∆FV$ and $∆SM$ profiles under the same RCH conditions are almost identical, with higher RCH results in larger $∆FV$ values but smaller $∆SM$ values. In Figure 11g, $∆FV$ profiles for all cases increase with D, rising gradually up to D = 9 m before steepening as the cutoff wall extends deeper. Initially, the curves are widely spaced but converge progressively, and become almost parallel to each other beyond D = 25 m. In Figure 11h, the $∆SM$ profiles decline with increasing D across all cases, dropping slowly at D = 0~9 m before undergoing a sharp decline at D = 9~33 m. Similar to the $∆FV$ profiles, the distinct curves at shallow depth gradually cluster towards each other and approach near-parallel alignment beyond D = 25 m.

The observations in Figure 11g,h align with those seen in Figure 7g,h and Figure 9g,h, revealing that under simultaneous SLR and RCH fluctuations, RCH dominates freshwater storage and salt removal in IAECW, while SLR has a negligible impact. The influence of RCH on cutoff wall’s performance in freshwater storage and salt removal and how D affects this influence are the same as findings in Section 3.2. Furthermore, for any given SLR scenario, IAECW in lower-RCH environments exhibit greater changes in $∆FV$ and $∆SM$ per unit of wall depth, indicating a higher efficiency of cutoff walls in freshwater storage and salt removal in such settings compared to high-RCH IAECW.

Observations in Figure 11 and Figure 12, overall, reveal that both SLR and RCH significantly impact WTE in IAECW, but their relative influence depends on D. In IAECW with shallow D, RCH fluctuations dominate WTE variations, whereas SLR exerts a greater influence than RCH in systems with deep D. Notably, SLR can partially or fully offset the WTE-lowering effect caused by reduced RCH in shallow IAECW, depending on SLR magnitude. Compared to RCH, SLR has negligible effects on FWL depth, FWL thickness, freshwater storage, and aquifer salt removal in IAECW. In contrast, RCH’s influence on these FWL and SWI metrics in IAECW is significant and depth-dependent: strongest at shallow depths, diminishing and leveling off as D increases. In the presence of SLR, IAECW with lower RCH exhibit greater responsiveness to cutoff wall installation, showing larger relative increases in FWL depth, thickness, freshwater storage, and salt removal per unit increase in D, indicating higher efficiency of cutoff walls in such settings.

4. Conclusions

This study explored the influence of RCH and SLR, both separately and in combination, on the cutoff wall performance in managing FWL development and mitigating SWI in island aquifers. The variable-density 2D island groundwater models based on hydrogeological conditions representative of Bahamian island aquifers were developed using SEAWAT to simulate aquifer response to different configurations of SLR, RCH and cutoff wall depth. To quantify the difference made by SLR, RCH and D to the groundwater system and SWI extent in the island aquifer, eight indicators were developed, including $∆T$, $∆d$, $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{a}}$, $∆{WTE}_{\mathrm{a}\mathrm{v}\mathrm{g},\mathrm{r}}$, $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{a}}$, $∆{WTE}_{\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{r}}$, $∆FV$, and $∆SM$. The investigation has been based on the analysis of nine distinct island hydrogeological scenarios representing unique combinations of SLR and RCH, each simulated across six cutoff wall depths, which has required as many as 54 steady-state SEAWAT model runs to identify the indicator values under different conditions.

The simulation results demonstrated that RCH significantly influences cutoff wall effectiveness in developing FWL and mitigating SWI in island aquifers. Higher RCH amplifies increases in average/maximum WTE, FWL depth/thickness, freshwater storage, and aquifer salt removal compared to lower RCH, for a given wall depth. This influence is inherently wall-depth-dependent. Specifically, the impact of RCH on cutoff wall effectiveness, for both FWL development and SWI mitigation, is more pronounced with D shallower than the natural FWL depth, and gradually diminishes and eventually stabilizes with D exceeding the natural FWL depth. Therefore, from a cost–benefit perspective, the optimal design depth for a cutoff wall should be slightly greater than the natural FWL depth to ensure robust performance considering RCH fluctuations. Accurately identifying the natural FWL depth remains a primary challenge for the practical implementation of cutoff wall strategies. SLR primarily elevates WTE within the IAECW in a stable manner associated with the SLR magnitude, but has a negligible impact on FWL development, freshwater storage, or salt removal enhancement, indicating that IAECW is resilient to SLR-induced SWI.

When SLR and RCH fluctuations are imposed concurrently, RCH dominates WTE variations in IAECW systems with shallow D, while SLR dominates in systems with deep D. Under simultaneous SLR and RCH fluctuations, SLR can partially or fully offset the WTE reduction caused by reduced RCH, depending on SLR magnitude and D; specifically, a full offset, where the WTE was maintained or elevated despite low RCH, was more likely to occur in IAECW with larger D and/or higher SLR. The joint effects of SLR and RCH on FWL depth, thickness, freshwater storage and aquifer salt removal align with their individual impacts.

Regardless of SLR, island aquifers characterized by lower RCH show greater responsiveness to cutoff wall installation, experiencing greater relative increases in WTE, FWL depth and thickness, freshwater storage and aquifer salt removal per unit D increase. This indicates a higher cutoff wall efficiency for FWL development and SWI mitigation in low-RCH settings.

While this study provides valuable insights into the performance of cutoff walls in FWL development and SWI mitigation under various SLR and RCH conditions, several limitations involved in this study should be acknowledged: (a) the investigation depended on an idealized 2D cross-section island groundwater model that assumes homogeneous and isotropic aquifer conditions. The omission of three-dimensional (3D) spatial complexities, aquifer heterogeneities (such as karst conduits), and transient dynamics on shoreline boundaries (like tidal forcing and storm events) may lead to an underestimation of the mixing zone width and salt transport velocity. Although the 2D model cannot fully capture the spatial complexities of real-world 3D systems, this approach was intentionally selected to ensure computational viability and maintain feasibility within the scope of this study. (b) The cutoff walls were idealized as nearly impermeable, and potential real-world leakage, construction challenges, and long-term material degradation were not incorporated. Adopting the idealized parameterization was to quantify the upper-bound performance of the cutoff wall system. (c) All SWI simulations were conducted under steady-state conditions to identify long-term equilibrium responses to SLR and RCH fluctuations. However, real-world climate impacts are inherently transient. Future research should utilize transient simulations to explore the time-lagged response of the freshwater–saltwater interface, particularly under seasonal recharge variability and episodic extreme weather events. (d) This study lacks field-scale validation. The model parameters were based on the established literature for San Salvador Island, but local-scale heterogeneities in hydraulic conductivity and porosity could significantly affect the predictive accuracy for a specific site.

Overall, the findings in this study are intended to be generic to island aquifers, even though they are based on the general characteristics of a specific site. As such, this analysis serves as a crucial initial step towards developing more sophisticated models that can effectively evaluate the cutoff wall performance in real-world island aquifers for FWL development and SWI mitigation. Future work should focus on developing field-calibrated models and advanced 3D simulations to provide site-specific management strategies. Additionally, exploring the interaction between cutoff walls and other climatic stressors, such as storm-induced surges and land surface inundation, and evaluating how aquifer characteristics affect the cutoff wall performance, such as island size, HK, dispersivity, porosity and anisotropy, would further enhance our understanding of island aquifer resilience.