Assessing recharge, discharge, and sea-level rise impacts in a coastal aquifer using a new approach

Abstract

This study examines a coastal aquifer located along the eastern sandy coast of Buenos Aires Province, Argentina, which serves as the primary water source for both local and seasonal tourist populations. The inhabitants in these resorts swell fivefold during the summer months, placing additional pressure on the aquifer’s resources. The study area, a 60 km-long coastal barrier, features a shallow freshwater aquifer that has been systematically monitored over several years, yielding a comprehensive conceptual understanding of its behavior. Water table data from wells positioned 100, 300, and 600 m from the shoreline, alongside rainfall and tidal records, were used to explore short-term variations, particularly the anthropogenic effects during the summer tourist influx. Due to data limitations attempts with conventional models produced results inconsistent with prior knowledge of the aquifer. To assess the dynamic response of this aquifer under these conditions, a simple, intuitive, and computationally efficient method was developed using a modified electrical–hydraulic analogy integrated with an electrical circuit analysis program (ECAP).The proposed methodology successfully quantified recharge and discharge dynamics, including flow reversals in the intertidal zone and delayed responses to rainfall in urbanized areas. Also, it was simulated sea-level rise effects on the discharge to the sea and inland. Additionally, it enables estimation of aquifer transmissivity or storativity, given one of these parameters is known. Importantly, this approach requires no software licensing and offers a practical tool for both short-and long-term dynamic analysis in data-limited contexts. It is well-suited for preliminary assessments, educational purposes and when conventional numerical models are impractical or unreliable. This study contributes to the sustainable management of the San Clemente del Tuyú aquifer, providing a transferable framework for other coastal aquifers facing similar data constraints. It highlights the value of the modified electrical–hydraulic analogy combined with an ECAP in complementing more complex modeling techniques.

1 Introduction

In recent years, technological innovations have significantly expanded the capacity to monitor and manage coastal groundwater resources. High-frequency sensor networks, satellite-based remote sensing, and autonomous monitoring platforms now provide unprecedented spatiotemporal coverage of key variables, including groundwater levels, salinity dynamics, and land subsidence (Famiglietti et al., 2011; Han and Currell, 2022; Wang et al., 2023). The integration of Internet of Things (IoT) devices, wireless telemetry, and cloud-based databases enables real-time monitoring of aquifer responses, reducing data gaps and enhancing detection of short-term events such as storm surges and tidal pulses (Nhemachena et al., 2020). Furthermore, machine learning and surrogate modeling approaches are increasingly applied to optimize groundwater management strategies under climate stressors, offering greater computational efficiency than full numerical simulations (Jiang et al., 2024). Together, these innovations create new opportunities to better understand and sustainably manage coastal groundwater systems under the dual pressures of climate change and human exploitation. However, these technological advancements require significant investment, limiting their applicability in regions where resources are scarce.

Beyond technological innovation, combining experimental and modeling approaches remains fundamental for predicting and monitoring environmental and engineering processes. Experiments provide detailed insights into hydrodynamic behavior, while models allow extrapolation and scenario testing across spatial and temporal scales. Their integration enhances understanding of flow resistance mechanisms and supports the development of simplified yet robust predictive tools (Errico et al., 2019; Lama et al., 2020; Box et al., 2021; Västilä et al., 2013; Mewis, 2021). This complementary perspective demonstrates that methodological innovation often depends more on integration than on model complexity.

In this context, sensitivity analysis provides a complementary and cost-effective means to improve the understanding and management of coastal groundwater systems. By identifying the parameters that most influence model performance and system dynamics, it helps refine monitoring strategies and prioritize data collection, especially where technological or financial resources are limited. Its versatility has been demonstrated across diverse water-related studies: Sánchez-Canales et al. (2012) emphasized the need for precise inputs in evaluating water provisioning services, while Beheshti et al. (2024) optimized sampling strategies for long-term monitoring. Other examples include assessing Nature-Based Solutions (Lama et al., 2021), developing simplified flood-mitigation models (Pirone et al., 2024), and analyzing vegetation management effects on flood dynamics (Giovannini et al., 2025). Together, these studies highlight the value of sensitivity analysis for strengthening model-based decision-making under climate-driven coastal change.

Long, uninterrupted time series are essential for applying complex models capable of capturing the detailed behavior of aquifers. Yet, in many cases, resources for data collection are insufficient: automatic monitoring stations are limited, records are short, and they often contain gaps and interruptions (Candela et al., 2013; Van Camp et al., 2013; Asmael et al., 2015; Rama et al., 2018; Ivković et al., 2018). Under these conditions, complex models forced with sparse or low-quality data may produce results that fail to reflect the physical reality of the system, rendering them meaningless (Voss, 2011a, 2011b).

In certain cases, only a limited number of parameters are truly relevant to the research question. Frequently, parameters are incorporated solely to allow the model to run; however, in practice, such detailed data are rarely available, particularly when they involve the temporal evolution of the variable. Highly parameterized models are not necessarily superior when only specific aspects of the aquifer are relevant. Any model is a simplified representation of hydrogeological reality; the initial objective should be to reproduce the system as simply as possible while capturing its essential behavior. Overly complex models can become difficult to interpret and of limited practical value (ibid.).

When groundwater dynamics must be assessed under conditions of limited data availability, simplified models can serve as a practical alternative, despite yielding less detailed results. This was exemplified by the case of the San Clemente del Tuyú coast (Argentina), where data scarcity led to inconsistent outcomes from complex modeling approaches such as MODFLOW, which failed to align with known aquifer characteristics. In a subsequent modeling effort, a finite element approach was employed using a combination of software tools—Argus ONE, SATC3D, and VisIt—integrated with custom Fortran code. However, this attempt also proved unsuccessful, as the model was unable to replicate the observed response of the water table to annual recharge events. It was concluded that the failure was largely due to the particular morphological features of the system, which rendered this modeling approach unsuitable. Therefore, a more tailored method was required to characterize the aquifer’s dynamic response to external forcings—including tidal fluctuations, precipitation, and groundwater abstraction—as well as to estimate recharge and discharge volumes.

One promising alternative is the electrical–hydraulic analogy, which has long been used in hydrogeology. Early groundwater modeling employed physical-electrical models composed of resistors, capacitors, and variable voltages and currents to simulate flow (Verruijt, 1970; Custodio and Llamas, 1976). This analogy has proven effective in teaching groundwater concepts, with over 95% of surveyed students reporting improved understanding of flow dynamics (Nave, 2008; Kasi et al., 2013; Guaraglia and Pousa, 2014). This approach is straightforward, building on basic physics principles familiar to most students and researchers, thus reducing the time required to achieve proficiency. In this study, this analogy was adopted to simulate groundwater dynamics without constructing a physical model or developing specialized software, as is necessary with classical numerical methods such as finite difference or finite element approaches.

The methodology involves translating aquifer properties into an equivalent electrical circuit and simulating its behavior using an electric circuit analysis program (ECAP). Time series of measured groundwater level, rainfall, and sea level can be directly input to drive the models, and outputs are easily exported to spreadsheets. Several electrical circuit solvers are freely available (LTspice 24.1.10 [Software], 2025; TINA-TI [Software], 2025), and simulations with fewer than 10,000 elements—as in this study—can be run in under a minute on a standard personal computer. The traditional analogy-based aquifer model was further modified to incorporate an independent current source representing rainfall intensity. This innovative configuration allowed simultaneous assessment of precipitation and groundwater extraction impacts on water table dynamics.

The primary objective of this study was to evaluate the dynamic behavior and sustainable management of the San Clemente del Tuyú coastal aquifer (Buenos Aires Province, Argentina). Specifically, the study aims to assess recharge, discharge, and anthropogenic influences on the aquifer system, including tidal effects, rainfall variability, seasonal pumping, and the potential impacts of sea-level rise. Also, a secondary goal was to provide a practical and transferable framework for estimating hydraulic parameters and supporting groundwater management in coastal aquifers facing similar data constraints.

2 Study area

In this area, most localities lack centralized drinking water supply systems, and residents depend on individual domestic wells without water treatment. A small portion of the population, such as that of San Clemente del Tuyú (Figure 1A), is connected to a water supply network. The resource is extracted from a pumping field situated outside the urban area.

Figure 1

According to Thornthwaite (1948) climatic classification, the area is humid and mesothermal, with little or no water deficiency and a summer concentration of thermal efficiency below 48%. The climate features a dry season during the cold months of the Southern Hemisphere (April–September) and a wet season in the warm months (October–March). Maximum temperatures occur in the rainy season, with an average of 22.2 °C in January, while the dry season is the coldest, with an average minimum temperature of 7.6 °C in July. Mean annual precipitation ranges from 900 to 1,000 mm, with 60% falling in months of highest potential evapotranspiration; however, the greatest groundwater recharge occurs during the dry season (Carretero and Kruse, 2012). Tides are mixed and predominantly semidiurnal, with ranges below 2 m (Servicio de Hidrografía Naval, 2008).

The study area comprises two geomorphological environments: the sand-dune barrier and the continental plain. The continental plain, located west of the coastal barrier, lies below 5 m above sea level (m a.s.l.), whereas the sand-dune barrier remains exposed to the sea (Figure 1A).

The hydrogeology consists of deep and shallow systems (Figure 1B). Data from the deep system are limited due to the presence of low-permeability units intercalated with sandy, high-salinity water layers (Consejo Federal de Inversiones, 1990). The shallow system consists of a freshwater phreatic aquifer within the sand-dune barrier, extending from Punta Rasa to Punta Médanos. This aquifer is 2–4 km wide, 60 km long, and 4–10 m thick. Its unique morphology has historically complicated modeling efforts. Geologically, it comprises Holocene sand and shelly sand (Violante et al., 2001) overlying a clayey aquitard/aquiclude with intercalated high-salinity sand lenses. Average transmissivity is approximately 100 m²/d, with a storage coefficient of 10% (Carretero, 2011).

The sand-dune barrier is the primary groundwater recharge zone. Flow is conducted over a short distance, discharging in two directions: eastward to the sea and westward to the continental plain. This hydrogeological configuration is bounded by two interfaces: brackish water–freshwater landward and freshwater–saltwater seaward (Figure 1B). The water table is recharged naturally through direct infiltration of rainfall surplus. The response of water levels to recharge is on the order of 2–4 h under natural conditions (Rodrigues Capítulo, 2015). Regional water budget estimates indicate that water surplus ranges from 268 to 444 mm, representing the only source of recharge for the groundwater system (Carretero and Kruse, 2014; Carretero et al., 2014). On average, infiltration accounts for approximately 37% of precipitation. Seasonal fluctuations of the water table are driven by the distribution of surpluses, with a general rise during the coldest months and a decline during the warmest months (Rodrigues Capítulo et al., 2018).

Groundwater in the sand-dune barrier is primarily low-salinity Ca–HCO₃ type, while in the continental plain it is Na–Cl type with high salinity. Elevated concentrations of total Fe (up to 6.21 mg/L) and Mn²⁺ (up to 1.25 mg/L), due to the mineralogical composition of the sand forming the aquifer, pose significant water quality concerns (Carretero et al., 2013,2022).

The pumping field, located south of the urban area, consists of seven horizontal Ranney-type wells (yield: 7 m³/h each) and 21 well-point systems, each comprising 10 wells (yield: 13 m³/h each). Extraction depths range from 4 to 6 m. These isolated, shallow, small-diameter wells are interconnected via a common aspiration pipe. Although the influence radius between wells is estimated at 50 m, the actual spacing is only 25 m. To minimize interference, wells are ideally operated alternately, but this is often not followed during the summer peak demand period (Carretero et al., 2019a). The wells occupy an area of approximately 0.23 km². Annual extraction is around 250,000 m³, with increasing trends. Maximum pumping occurs in summer, particularly in January and February (1,500 m³/d), coinciding with peak tourist influx, while minimum values occur in May–June (250–300 m³/d). Average extraction during the remaining months ranges from 400 to 500 m³/d (Carretero and Kruse, 2010; Carretero et al., 2024). Extracted water is appropriately treated to ensure potability before distribution.

3 Materials and methods

3.1 Equations to convert the hydraulic problem into an electrical one

3.1.1 Continuity equation for a confined aquifer

For a confined aquifer, which always remains saturated, the continuity equation can be established through the concept of storativity (S), defined as the volume of water released by the aquifer per unit area and per unit decline in piezometric level (Custodio and Llamas, 1976). Thus,

where h is the hydraulic head, x the horizontal distance, t the time and T the transmissivity of the aquifer.

3.1.2 Electric lines

In 1854, Sir William Thomson (Lord Kelvin) derived the following equation to account for voltage (V) changes in a transmission line, considering neither inductive effects nor dielectric losses (Olyslager, 1999)

where R and C are, respectively, the resistance and capacitance per unit length of the transmission line, x is a distance measured along the line and t the time. As resistance and capacitance are distributed along the line, they are considered magnitudes per unit length.

Equations 1, 2 are similar parabolic, diffusion-type differential equations occurring in entirely different areas of physics. Equation 1 refers to flow and levels in hydrodynamics, whereas Equation 2 relates to currents and voltages in an electric transmission line. The similitude of both equations, however, allows a hydraulic-electric analogy to be set up. In Feynman’s words, “The same equations have the same solutions” (Feynman et al., 1966).

3.1.3 Hydraulic-electric analogy

Consider first a discretized x-t plane in which the integer subscripts i and n refer to the increments in the coordinates x and t, respectively. Equation (1) can then be approximated by Equation 3 (LeVeque, 2007).

which can be rearranged as

A similar approach may be applied to discretize Thomson’s equation. It is perhaps more physically intuitive to consider the transmission line as divided into successive sections, and to use the subscripts i and n to refer to the increments in the length of the line and the time, respectively. In doing so, Equation 2 can be approximated by

where V_{i − 1}, V_i and V_i + 1 represent the voltages at different points along the line, and V_n and V_n + 1 represent the voltages at different times.

In order to have a more physical point of view of the problem, a similar result can be obtained by using Kirchhoff’s second rule, namely that the sum of the currents into any node must be zero. Figure 2A shows a node with two equal resistances R and a grounded capacitance C, through which currents I₁, I₂ and I₃ pass due to the voltage drops V_i − 1− V_i, V_i + 1 – V_i and V_i.

Figure 2

Application of Kirchhoff’s second rule to the node at a voltage V_i gives Equation 6.

where, as before, the integers i and n stand for the space and time increments, and t is the time. Kirchhoff’s second rule is applied to concentrated electric parameters of a defined size. They are not distributed parameters as in the case of the line and, for this reason, the term ∆x is absent from this equation.

The circuit in Figure 2B represents a piece of aquifer of length ∆x. In the direction of the x axis exist three different hydraulic levels h_{i + 1}, h_i and h_i − 1. The differences in levels produce flows Q₁ and Q₃ entering the piece of aquifer. Flow Q₂ which is the sum of Q₁ and Q₃ is stored in the soil of this piece of aquifer. Depending on the relative differences of levels h_{i + 1}, h_i and h_i − 1, flows will enter or exit the piece of aquifer. In this piece of aquifer, 1/T represents the resistance of the soil to the groundwater flow; S, the capacity of the soil to store water.

When comparing Equations 4, 5, the following hydraulic-electric analogs result in Equation 7,

The time t is the same for both the hydraulic and the electric cases.

If in Figure 2B an independent flow could be added (rain intensity) to the piece of aquifer at the point h_i. In the electrical counterpart this independent flow is represented by a current source I_s, as in Figure 2C. So far, it has been described how to model in one dimension (x), a portion of an aquifer with a length ∆x whose transmissivity is T and storativity S, when it is subject to changes in the water table and rainfall. By extending the concepts developed previously, it is possible to model a portion of an aquifer in two or three dimensions. In order to extend the one-dimensional model to two and three-dimensional models, is only required to change the cell shown in Figure 2C to the cells in Figures 2D,E respectively.

The electric-hydraulic analogy and the equations to convert the hydraulic parameters into electrical ones are summarized in Table 1.

3.2 The electric model of the complete aquifer

The next step was the construction of the electric circuit to represent the complete aquifer. From Table 1, once the storativity (S), the transmissivity (T) and the size of the step were fixed, the values of R and C could be calculated and the model built.

The aquifer was divided into a series of adjacent cells containing elemental electric circuits as shown in Figure 2C. In the particular case of a coastal aquifer in which the flow parallel to the coast is low and the vertical flow was disregarded, a very simple one-dimensional model seemed enough to describe the aquifer dynamics in a direction perpendicular to the shore. One way of building a discrete model is with cells of constant thickness along a modeling line. By choosing an adequate increment, it was easy to locate the observation wells (where the data was recorded) at the nodes of the model. The aquifer could thus be modeled by cascading a series of such elementary circuits, as illustrated in Figure 2F by one cell and a dashed box. In the case of a homogeneous aquifer, the box contains a certain number of cells identical to the first one, but cells with different values of R and C could be used in cases of inhomogeneities in T or S along the x dimension.

Points denoted by n_A and n_B in Figure 2F were the outermost nodes where the boundary conditions for the model had to be established. Boundaries were represented here by two voltage sources V_A and V_B. For example, to represent a tide exciting the aquifer, a voltage source analog to the tide level was connected to node A. Another voltage source representing the water table measured inland is connected to node B. The dashed box had to have as many cells as needed to represent the aquifer adequately (see below Ferris equation, Equation 8).

The electric model of the aquifer could be solved by most ECAPs. In each cell, R represented the opposition of the soil to the groundwater flow; C, the soil capacity to store water; and I_s, the rain intensity by unit area, connected at each node of the model. The values of the water levels measured in meters were introduced into the model as Volts. In order to express the transmissivity in usual units (m²/d), day (d) was adopted as the time unit. The unit of the current sources, which represent the rain intensity by square meter, was Ampere (A) and the value of the current was numerically equal to the values of the measured rain intensity expressed in m³/m²d. In hydrology, the units used for precipitation intensity (mm), levels (m), and transmissivity (m²/d) were standardized across the model, with meters (m) used for all lengths and days (d) used for time.

As in any discrete model, an error appears when attempting to represent an aquifer with a finite number of cells. As the number of cells increases this error becomes smaller, but computational time increases. It was thus convenient to find a limited number of cells that made the model reliably represent the aquifer. The truncation effects due to the finite number of cells were evaluated by exciting the model with 1 m height sinusoidal water level oscillations of different periods, similar to those of the tide; they were represented in the model by voltage sinusoids of 1 V in amplitude. The water table at different distances from the shore (voltages along the circuit) was computed with the model, and the lag in time (t₁) and amplitude attenuation (h/h₀) predicted by Ferris’ equations (Ferris, 1952) were calculated (Equation 8).

where h₀ was the amplitude of the tide, h the amplitude of the water table at different distances from the shore, x the distance from the shore (node n_A) to the observation well and t₀ the period of the tide.

When the number of circuit’s cells was not enough to correctly represent the aquifer, the lag and attenuation as a function of distance calculated by the model, deviated from what the Ferris’ equations predicted, then for the model to be reliable, the number of cells has to be increased. For the coastal aquifer section of interest, which was less than 2 km wide, it was verified that with 2,000 cells (1 m in length) the truncation effects were unnoticeable. Therefore, this number of cells was adopted.

3.2.1 Application of the model to an unconfined aquifer

The electrical model is valid for horizontal flows that are uniform throughout a vertical section. While it is directly applicable to confined aquifers, it can also be extended to unconfined aquifers, similarly to the transient equations derived by Theis (1935), provided that the water table declines by less than approximately 10% of the aquifer thickness—as is the case for the coastal aquifer described below. When flow occurs parallel to the shoreline, which is not accounted for by the one-dimensional model, discrepancies between measured and modeled values may arise. However, in the study area, the dominant flow is horizontal, perpendicular to the coastline and characterized by low hydraulic gradients, which is typical of plains. For this reason, it was considered that applying the one-dimensional model would not have a significant impact on the discharge estimates. In aquifers that do not meet these conditions, the model can be extended to two or three dimensions, as shown in Figure 2 of Section 3.1.3, Hydraulic–electric analogy.

3.3 The software

An ECAP, known as PSPICE (Keown, 1991), was employed in this study. Other software, such as the free LTspice and TINA, could also be used (TINA was successfully tested giving similar results than PSPICE). These programs are highly reliable and have been standard tools in Electrical Engineering for decades. By constructing an electrical model of an aquifer, as described above, it is possible to use these programs in which time-dependent voltages (water level) and currents (rain intensity) could be introduced as inputs at any place along the electric circuit (aquifer). The resulting time series of voltages and currents can be analyzed within the program and exported to spreadsheets or other software for further processing.

3.4 Data available

Water-table records were obtained from three data loggers equipped with absolute pressure sensors. Two of these loggers were installed at monitoring wells P300 and P600, located along a transect 300 m and 600 m from the shoreline, respectively. The transect is approximately perpendicular to the shoreline. Measurements were recorded hourly, and the monitoring periods for P300 and P600 extended from December 2014 to March 2015 and from December 2016 to April 2017. At well P100, situated 100 m from the shoreline along the same line as P300 and P600, water levels were recorded hourly for 30 days in August 2015, and manually during the installation and removal of the data loggers at P300 and P600. A data logger at monitoring well ST13, located in a highly urbanized coastal area, was installed in December 2013 and operated until February 2014.

Rainfall data were collected using a rain gauge located near the monitoring wells. Additionally, barometric pressure records were used to correct the absolute pressure sensor measurements. Hourly tide data were obtained from the Servicio de Hidrografía Naval for the Santa Teresita and San Clemente del Tuyú tide stations.

4 Results

4.1 Calibration of the model

In the first test, the electrical model of the study area was driven by three field-recorded time series: sea level, inland water table, and rainfall. Initially, the model was applied to simulate the water table at P100, as this well is the closest to the sea and the influence of the tide could be clearly observed. The tide was applied at node 1, defined as the seaward node, while the inland end of the model (node 2000) was fixed at a reference voltage (water level), simulating the inland boundary of the aquifer at mean sea level. Rain intensity (Is) was applied to all model cells, representing homogeneous rainfall infiltration across the aquifer. Previous studies reported approximate values of transmissivity (100 m²/d) and storativity (0.10), which, together with the parameters in Table 1, corresponded to model settings of R = 0.01 Ω and C = 0.1 F (Ω = ohm, F = farad).

A homogeneous initial water table equal to the first recorded observation at P100 (1.63 m a.s.l.) was imposed across the aquifer. This water level was introduced into the circuit as an initial voltage, charging each capacitor. Spatial and temporal variations in water table and flow were evaluated by running the ECAP. The voltage at node 101, corresponding to P100, was recorded over time and compared with the measured water table.

In the first simulation, only the influence of the tide on the water table at P100 was considered; rainfall was deliberately disregarded by setting all Is = 0. The hourly recorded tide level (in meters) was applied to node 1 as a voltage source, providing the sole excitation to the aquifer model.

The resulting voltage at node 101, representing the water table at P100, is shown in Figure 3 (“Tide only”). The actual observed water table is also plotted (“Measured”) for comparison. The modeled water table in this scenario, generated by the ECAP, was considerably lower than the measured values.

Figure 3

In the second simulation, rainfall intensity (Is) was incorporated using a fixed infiltration ratio (IR) of 37% (see Study Area), resulting in a much closer agreement between measured and modeled water levels (“Tide and rain” in Figure 3). Given the high permeability and low topographic slope of the terrain, surface runoff is negligible when considering the rainfall-infiltration ratio.

A strong correlation between observed and modeled water tables at P100 was achieved during the 30-day monitoring period. The range of measured water table fluctuations was 88 cm. The maximum absolute difference between measured and modeled water table was 8 cm, and the Root Mean Square Error (RMSE) was 6.1 cm. Remaining discrepancies may be attributed to variations in the infiltration ratio. Overall, the electrical model was deemed a reliable tool for predicting the water table in this aquifer, and the chosen values of R, C, and IR were found to be appropriate. These results validate the model and encourage its application to other data sets, independent of those used in this validation.

4.2 Sensitivity analysis

A sensitivity analysis (Table 2) was conducted to assess how model parameters and boundary conditions influence model outputs. The sensitivity coefficient for the One-Factor-at-a-Time (OAT) analysis (Saltelli, 1999; Chow et al., 2018; Van Griensven et al., 2002) was computed as shown in Equation 9. This approach enables an unambiguous attribution of changes in the model output to perturbations in individual input parameters. However, it should be noted that the resulting sensitivity metrics are only valid in the vicinity of the baseline case, as the influence of a given parameter may depend on the specific values assigned to the remaining parameters.

Where y is the model output, xᵢ is the i-th parameter, and Sᵢ represents the sensitivity of the output to the i-th parameter.

For the San Clemente area, the baseline configuration was defined using the following parameters: transmissivity T = 105, storativity S = 0.09, total rainfall = 260 mm, infiltration constant = 0.33, inland water-table elevation = 1 m, sea level = 1 m, initial water-table elevation = 1 m, simulation period = 91 days, and time step = 1 h. These values were selected to reflect the characteristic conditions observed in the study area.

The sensitivity analysis provides valuable guidance for prioritizing future data-collection efforts and refining subsequent model development. In particular, rainfall and storativity exhibit strong influences on simulated discharge, while sea level exerts a substantial impact on the water-table elevation. In contrast, storativity, transmissivity, and rainfall show relatively low influence on the water-table response. Transmissivity and sea level both demonstrate moderate effects on discharge. Overall, the parameters that demand the most attention during model calibration are rainfall, sea level, and storativity.

4.3 Model applications

4.3.1 Long-term tide influence on the water table

The extent to which the tide influences the water table at the pumping field, located between 500 and 1,000 m inland, was previously unknown. To assess the long-term effects of tides, the water table was simulated at various distances from the shoreline, with sea level as the sole excitation. The model was driven by a 7.5-year tide record from January 1987 to June 1994, as more recent uninterrupted datasets were unavailable. Rainfall was excluded in this simulation. The results, shown in Figure 4, indicate that maximum water table fluctuations due to tidal influence are approximately ±0.025 m for a well located 600 m (P600) from the shoreline.

Figure 4

4.3.2 Impact of the pumping field on the water table at 300 m from the shore

In the San Clemente del Tuyú aquifer, the risk of saline intrusion is particularly relevant during the summer, when population density increases. For two summer periods (December 2014–March 2015 and December 2016–April 2017), water table levels at an observation well 300 m from the shore (P300) were recorded (Figure 5A), this adds up to a total of 225 days of independent data, in addition to the 30 days used to calibrate the model. To study aquifer recharge, the water table at P300 was simulated using rainfall as the only excitation, with an infiltration ratio (IR) between 34 and 56%. The boundary conditions at the seaward and inland nodes (nodes 1 and 2,000, respectively) were fixed at constant voltages (water levels). When testing different constant voltages as boundary conditions, the model’s best fit still showed discrepancies with the measured water table. Results for summer 2014–2015 are presented in Figure 5B. During the initial phase of the simulation, the model output (“Rain only”) was below the measured water table (“Measured”), while in the later period it exceeded the measured values. Differences are most pronounced in slow water table variations, whereas rapid changes (high-frequency components) were well captured. After day 70, “Rain only” exceeds the observed water table “Measured” by approximately 0.15 m. The model was driven with tide records to assess whether the differences could be attributed to tidal oscillations, but this input did not modify the model results for the P300 water table.

Figure 5

Based on prior studies, the pumping field, located at least 500 m from the shore, was expected to influence the water table at P600 but not at P300. To evaluate these assumptions, the water table time series at P600 was incorporated as an excitation at node 601. The simulation including both rainfall and the P600 water table produced a significantly improved fit to the measured water table at P300 (“Rain and W.T.”). The observed water table fluctuated by 60 cm over the summer, while the maximum absolute difference between measured and modeled values was 5 cm, and the Root Mean Square Error (RMSE) was 2.37 cm. These results suggest that summer pumping lowers the water table at P600, which in turn affects P300. This interaction had not been identified in previous studies.

4.3.3 Assessment of the hydraulic parameters

The methodology can also be applied to estimate hydraulic parameters, such as transmissivity (T) or storativity (S). The temporal response of the water table to infiltration depends on the hydraulic diffusivity (T/S) of the aquifer. Therefore, there is a unique T/S value that, when applied in the model, reproduces the observed water table changes.

Assuming that either T or S is known, the model can be calibrated by fixing the known parameter (e.g., T) and adjusting the other (S) until the simulated water table closely reproduces the observed time series. The model shows high sensitivity to both parameters, as illustrated in Figure 6 for the case of S, where transmissivity was fixed at T = 105 m²/d and storativity was varied between 0.06 and 0.12 (for clarity, only the model outputs corresponding to three values of S are shown). Values of S = 0.09 and 0.10 yield the best fits, in agreement with previous studies reporting storativity values of approximately 0.10 for the aquifer.

Figure 6

As shown in Figure 6, the modeled and measured curves (S ~ 0.09, 0.10) coincide only partially. In such cases, the S value that best matches the water table response during rainfall events should be considered the most probable, as rainfall drives the primary changes in the water table. Over time, other factors, such as pumping at nearby wells, may alter water levels, as observed after day 41.

4.4 Evaluation of recharge and discharge

4.4.1 Assessment of the delayed recharge

In areas with delayed recharge, the lag between rainfall and the rise in the water table can be estimated using the model. This was tested at well ST13, located in the same coastal region but within a highly urbanized area with similar hydrogeological characteristics. Water levels in this area, where surface runoff is significant and vertical infiltration is minimal, were recorded concurrently with rainfall. The lag between rainfall, modeled water table, and observed water table is shown in Figure 7.

Figure 7

The observed delay between modeled and measured peaks (~4 days) may be due to horizontal groundwater flow, rather than direct vertical recharge from rainfall at the well. It is assumed that rainfall infiltrates a permeable zone located more than 150 m west of ST13, and groundwater naturally flows toward the well and the sea. The model output represents the aquifer’s behavior under natural conditions, without urbanization impacts.

Impermeabilization in urban areas reduces aquifer recharge due to increased surface runoff. Recharge still occurs but in a delayed manner. The model captures the natural water table response, while measurements reflect conditions altered by anthropogenic factors. For rainfall events smaller than 4 mm, the model shows small rises in water table, whereas measurements do not, because such rainfall does not produce significant infiltration in urbanized areas. Similarly, modeled water table declines are slower, as individual rainfall events sustain the level, while measured levels drop abruptly due to lack of continued water input.

4.4.2 Discharge evaluation toward the sea and the pumping field

In addition to calculating voltages (water levels), the electrical model also computes currents, which correspond to water flows. This capability allows estimation of discharge toward the sea and the pumping field. For the summer of 2014–2015, Figure 8 shows water tables at P600 and P300 and modeled flows at 50 m and 550 m from the coast. The purpose of this figure is to qualitatively assess the correlation between flows and water table levels. In this case, positive flow indicates movement toward the sea, while negative flow indicates movement toward the pumping field. The flow direction assignments made here are deliberately at odds with those discussed in Section 4.4.3 and are implemented primarily to enhance the visualization and interpretation of the relationship between flow patterns and water table elevations.

Figure 8

Despite initial transient differences, the flow toward the sea and the water table at P300 display similar patterns. Similarly, the flow toward the pumping field and the water table at P600 show comparable behavior. The model qualitatively confirms that the flow near the pumping field correlates well with the water table at P600, which is influenced by aquifer exploitation.

Quantitative analysis indicates that during a 91-day period in summer 2014–2015, total rainfall over the modeled section of aquifer (1 m wide, 600 m long, 10 m deep, perpendicular to the shore) amounted to 157.8 m³, with total infiltration of 58.4 m³. Average daily flows computed by the model were 0.422 m³/d at 550 m, 0.279 m³/d at 50 m, and 0.311 m³/d at 5 m. In the first case, the flow was toward P600, and in the others toward the sea. From these flows, the total water discharged toward the sea was 28.26 m³, and toward the pumping field, 38.44 m³. The total estimated discharge was 66.7 m³, approximately 8.3% higher than the estimated infiltration. During the summer of 2016–2017 (126 days), discharge exceeded infiltration by 25.5%.

These discharge values indicate that water reserves are depleted during summer and replenished during winter (Table 3). This pattern aligns with previous regional studies. Notably, the population can increase up to fivefold during the summer due to seasonal tourism, leading to higher water demand. Such seasonal fluctuations highlight the importance of considering transient demographic pressures when assessing groundwater availability and designing sustainable management strategies. Furthermore, average discharge estimates derived from groundwater flow maps (Carretero, 2011) are consistent with the model results, confirming the reliability of the modeled flows.

4.4.3 Effect of tides on discharge into the sea

The dynamic horizontal flow in the aquifer was evaluated over a four-day period at 5 m (Q5) and 50 m (Q50) from the coast (Figure 9). The water table at 600 m from the coast (pumping area) was held constant at the summer 2014–2015 average value of 1.33 m a.s.l. The model was driven solely by the tidal level over these 4 days, representing an average tide for the region. Rainfall was set to zero to avoid interference.

Figure 9

Figures 9A,B show flow oscillations. At Q(5), flows ranged from approximately +3 to −4 m³/d. Positive values indicate reverse flow, i.e., from the sea toward the continent, a phenomenon possible in the intertidal zone of gently sloping beaches such as the study site. At Q(50), oscillations were smaller, between +0.2 and −0.8 m³/d.

Figure 9A illustrates that discharge decreases at high tide and increases at low tide. Figure 9B shows a delay of roughly 7 h between tidal excitation and the flow response. The average discharge calculated at Q(5) was 0.24 m³/d, while at Q(50) it was 0.31 m³/d. The expectation that discharge near the sea (Q5) would be larger than at 50 m from the coast, as in Section 4.3.2, is contradicted here. This apparent discrepancy can be attributed to flow reversal and the short averaging period, which is insufficient to filter out the effect of truncation of the time series.

4.4.4 Discharge changes due to climate change

Along the Buenos Aires Province coast, a relative sea level rise of approximately 1.5 mm/year has been observed (Dennis et al., 1995), with accelerations below 0.15 mm/year. Using the same summer 2014–2015 data employed to estimate aquifer discharge toward the sea and pumping field, simulations were performed by adjusting only the average sea level in the model to evaluate changes in aquifer discharge.

If sea level continues to rise at the observed rate, Figure 10 illustrates projected changes over roughly 130 years. Discharge to the sea is expected to decrease by 27%, while flow toward the pumping field would increase slightly by 4%. This trend indicates that reduced discharge to the sea would elevate the water table, increasing continental water contribution. In the pumping area, this effect may be beneficial; however, in urbanized areas, elevated water tables could present significant problems.

Figure 10

5 Discussion

Climate change is increasingly altering groundwater–seawater interactions in coastal zones through sea-level rise, changing precipitation patterns, and modifications of the hydrological cycle, which in turn enhance seawater intrusion and submarine groundwater discharge (Richardson et al., 2024; Xu et al., 2024). These processes threaten freshwater availability and coastal ecosystems, particularly in low-lying areas with shallow, vulnerable aquifers. Recent advances highlight the need for dynamic, coupled models that overcome the limitations of traditional approaches with static ocean–aquifer boundary conditions (Jin et al., 2025). However, in data-scarce regions, complex models may be impractical or unreliable. In this context, electrical–hydraulic analogies provide a simple, physically consistent, and computationally efficient alternative for analyzing aquifer dynamics.

This study presents a novel approach for investigating aquifer behavior. The electrical–hydraulic analogy, long recognized in hydrogeology, was applied here in conjunction with an electrical circuit analysis program (ECAP), and the classical electric–hydraulic cell analogy was modified to incorporate rainfall effects. This enhancement extends the analogy’s applicability to a wider range of hydrogeological phenomena.

Electrical analogies have been employed to address heat transfer problems (Guaraglia and Pousa, 1999; Guaraglia et al., 2001). Additionally, prior studies have utilized electrical–hydraulic analogies; however, none have applied them in the manner presented in this work. Stallman (1963) established electric-analog design criteria for cylindrical geometries to analyze pumping tests in Grand Island, Nebraska, USA. Goudarzi et al. (2018) evaluated hydraulic–electric flow analogies to predict porous media properties, and Ji et al. (2021) proposed a method for estimating network fracture conductivity. More recently, Gao et al. (2025) implemented a three-dimensional hydraulic–electric analogy in a block-diagram environment (Simulink) integrated with MATLAB. These examples demonstrate the continued utility of this traditional analogy for modeling diverse phenomena.

The results of this study demonstrate that this methodology provides a clear understanding of the dynamic behavior of a coastal aquifer. A simple one-dimensional model was sufficient to simulate tides, water tables, rainfall recharge, pumping effects, and discharge. For those familiar with the physical concepts underlying Equations 1, 2, the combination of analogy and ECAPs directly reveals aquifer dynamics.

For the study area, the method confirmed the significant influence of pumping on the water table during summer seasons. Results are consistent with prior knowledge of the San Clemente aquifer and regional studies. Although constant transmissivity (T) and storage (S) values were assumed, the method can be extended to aquifers with variable T and S. This study represents the first simulation in the region to quantify discharge variations to the sea due to tidal effects and sea level rise from climate change. Modeled discharge values (0.24 and 0.31 m³/d/m) align with previous estimates (0.21–0.33 m³/d per meter of coastline; Carretero et al., 2019b). Importantly, the model reveals discharge variability over the tidal cycle, offering insights into aquifer dynamics.

A limitation of this approach is that hydrogeologists accustomed to commercial software may initially find the electrical–hydraulic analogy unfamiliar. Interpreting results requires basic knowledge of electric circuits; without it, misinterpretation or erroneous conceptual models may occur. Nevertheless, this method can be a valuable tool when applied in multidisciplinary teams including researchers experienced in electronic circuit modeling, suitable for both preliminary assessments and advanced studies of aquifer dynamics. However, to make the method easier to use for those with some knowledge of electrical circuits, an example of what the file running in ECAP would look like was added as Supplementary material.

ECAPs are inherently designed for dynamic phenomena and can be applied to time-varying excitations, incorporating tools for frequency analysis and correlation. These capabilities allow easy identification and quantification of the influence of individual factors (tides, rainfall, pumping) at specific aquifer locations without requiring additional software.

These capabilities reflect the principles of sensitivity analysis, which evaluates the relative importance of hydrological and management variables. Identifying the most influential parameters—such as precipitation, discharge, or groundwater extraction—supports more efficient monitoring, prioritization of data collection, and adaptive management under uncertain conditions, including climate variability. Its broad application in water resource studies (Sánchez-Canales et al., 2012; Lama et al., 2021; Beheshti et al., 2024; Pirone et al., 2024; Giovannini et al., 2025) highlights its value as a practical decision-support tool for coastal aquifer management.

This work demonstrates the method’s potential as a simple, free alternative to existing simulation programs. To date, no other models have successfully reproduced the behavior of this coastal aquifer.

6 Conclusion

A new method combining a classical electric–hydraulic analogy with electrical circuit analysis software proved effective in improving understanding of groundwater dynamics in a coastal aquifer with limited data availability.

This approach allowed estimation of aquifer response over both long-term (years) and short-term (hours) periods, including the effects of tides, rainfall, and pumping. It also enabled computation of discharges toward the pumping field and the sea, and simulation of the impact of tides and sea level rise on aquifer discharge. Notably, the method does not require licensing fees for software.

In addition, integrating sensitivity analysis with this approach highlights which factors exert the greatest influence on aquifer behavior, supporting more efficient monitoring and informed management decisions. The combined methodology provides valuable information for water resource authorities managing the pumping field and underscores the practical significance of the approach for coastal aquifers, which serve as the sole water source for towns along the 60 km study area.

References

• 1

  Asmael N. M. Dupuy A. Huneau F. Hamid S. Le Coustumer P. (2015). Groundwater modeling as an alternative approach to limited data in the northeastern part of Mt. Hermon (Syria), to develop a preliminary water budget. Water7, 3978–3996. doi: 10.3390/w7073978

  • CrossRef
  • Google Scholar

• 2

  Beheshti K. M. Smith R. S. Raimondi P. Reed D. (2024). Sensitivity analysis to reduce sampling effort of a long-term monitoring program. Ecol. Indic.158:111585. doi: 10.1016/j.ecolind.2024.111585

  • CrossRef
  • Google Scholar

• 3

  Box W. Järvelä J. Västilä K. (2021). Flow resistance of floodplain vegetation mixtures for modelling river flows. J. Hydrol.601:126593. doi: 10.1016/j.jhydrol.2021.126593

  • CrossRef
  • Google Scholar

• 4

  Candela L. Elorza F. J. Tamoh K. Jiménez-Martínez J. Aureli A. (2013). Groundwater modelling with limited data sets: the Chari–Logone area (Lake Chad Basin, Chad). Hydrol. Process.28, 3714–3727. doi: 10.1002/hyp.9901

  • CrossRef
  • Google Scholar

• 5

  Carretero S. (2011) Comportamiento hidrológico de las dunas costeras en el sector nororiental de la provincia de Buenos Aires (doctoral thesis). Facultad de Ciencias Naturales y Museo, La Plata, Argentina

  • Google Scholar

• 6

  Carretero S. Dapeña C. Kruse E. (2013). Hydrogeochemical and isotopic characterisation of groundwater in a sand-dune phreatic aquifer on the northeastern coast of the province of Buenos Aires. Isot. Environ. Health Stud.49, 399–419.

  • Google Scholar

• 7

  Carretero S. Braga F. Kruse E. Tosi L. (2014). Temporal analysis of the changes in the sand-dune barrier in the Buenos Aires Province, Argentina, and their relationship with the water resources. Appl. Geogr.54, 169–181. doi: 10.1016/j.apgeog.2014.07.008

  • CrossRef
  • Google Scholar

• 8

  Carretero S. Kruse E. (2010). “Areal exploitation of groundwater in coastal dunes, Buenos Aires, Argentina” in Global groundwater resources and management. ed. PaliwalB. S. (Jodhpur: Scientific Publishers), 385–398.

  • Google Scholar

• 9

  Carretero S. Kruse E. (2012). Relationship between precipitation and watertable fluctuation in a coastal dune aquifer: northeastern coast of the Buenos Aires province, Argentina. Hydrogeol. J.20, 1613–1621. doi: 10.1007/s10040-012-0906-2

  • CrossRef
  • Google Scholar

• 10

  Carretero S. Kruse E. (2014). Impacto de la urbanización en la infiltración en un área costera, Argentina [Impact of urbanization on infiltration in a coastal area, Argentina]. Tecnol. Cienc. Agua6, 5–24. Available online at: https://revistatyca.org.mx/index.php/tyca/article/view/1217

  • Google Scholar

• 11

  Carretero S. Perdomo S. Capitulo L. R. Kruse E. (2024). Effect of pumping in a coastal aquifer of limited thickness, Buenos Aires, Argentina. Groundw. Sustain. Dev.26:101302. doi: 10.1016/j.gsd.2024.101302

  • CrossRef
  • Google Scholar

• 12

  Carretero S. Rapaglia J. Perdomo S. Albino Martínez C. Rodrigues Capítulo L. Gómez L. et al . (2019b). A multi-parameter study of groundwater–seawater interactions along Partido de La Costa, Buenos Aires Province, Argentina. Environ. Earth Sci.78:513. doi: 10.1007/s12665-019-8532-5

  • CrossRef
  • Google Scholar

• 13

  Carretero S. Rodrigues Capítulo L. Dapeña C. Fabiano M. Kruse E. (2022). A chemical and isotopic approach to investigate groundwater dynamics in a coastal aquifer. Catena213:106229. doi: 10.1016/j.catena.2022.106229

  • CrossRef
  • Google Scholar

• 14

  Carretero S. Rodrigues Capítulo L. Kruse E. (2019a). Decision tree as a tool for the management of coastal aquifers of limited saturated thickness. Q. J. Eng. Geol. Hydrogeol.53, 189–200. doi: 10.1144/qjegh2018-136

  • CrossRef
  • Google Scholar

• 15

  Chow C. Ramirez J. Keiler M. (2018). Application of sensitivity analysis for process model calibration of natural hazards. Geosciences8:218. doi: 10.3390/geosciences8060218

  • CrossRef
  • Google Scholar

• 16

  Consejo Federal de Inversiones (1990). Evaluación del Recurso Hídrico Subterráneo de la Región Costera Atlántica de la Provincia de Buenos Aires Región I Punta Rasa-Punta Médanos. Informe Final Tomo IV Caracterización climática y Balance Hidrológico, CFI, CABA. 177.

  • Google Scholar

• 17

  Custodio E. Llamas M. R. (1976). Hidrología subterránea: Omega.

  • Google Scholar

• 18

  Dennis K.C. Schnack E.J. Mouzo F.H. Orona C.R. (1995). Sea-level rise and Argentina: potential impacts and consequences, J. Coast. Res., pp. 205–223. Available online at: http://www.jstor.org/stable/25735709 (Accessed March 11, 2025).

  • Google Scholar

• 19

  Errico A. Lama G. F. C. Francalanci S. Chirico G. B. Solari L. Preti F. (2019) Validation of global flow resistance models in two experimental drainage channels covered by Phragmites australis (common reed). In Proceedings of the 38th IAHR World Congress – Water Connecting the World, pp. 1313–1321

  • Google Scholar

• 20

  Famiglietti J. S. Lo M. Ho S. L. Bethune J. Anderson K. J. Syed T. H. et al . (2011). Satellites measure recent rates of groundwater depletion in California's Central Valley. Geophys. Res. Lett.38. doi: 10.1029/2010GL046442

  • CrossRef
  • Google Scholar

• 21

  Ferris J. G. (1952). Cyclic fluctuations of water level as a basis for determining aquifer transmissibility (No. Note 1). US Geological Survey.

  • Google Scholar

• 22

  Feynman R. P. Leighton R. B. Sands M. (1966). The Feynman Lectures on Physics.Vol. II Reading, Massachusetts: Addison-Wesley.

  • Google Scholar

• 23

  Gao Y. Ismail M. A. M. Li T. Li B. Zhang J. (2025). Hydroelectric simulation of the phreatic water response of mining cracked soil based on microbial solidification. Open Geosciences, 17:20220728. doi: 10.1515/geo-2022-0728

  • CrossRef
  • Google Scholar

• 24

  Giovannini M. R. M. Lama G. F. C. Scopetani L. Francalanci S. Signorile A. Saracino R. et al . (2025). Evaluating the impacts of riparian plants on flood hazard within vegetated rivers. J. Flood Risk Manag.18:e70063. doi: 10.1111/jfr3.70063

  • CrossRef
  • Google Scholar

• 25

  Goudarzi B. Mohammadmoradi P. Kantzas A. (2018). Direct pore-level examination of hydraulic-electric analogy in unconsolidated porous media. J. Pet. Sci. Eng.165, 811–820. doi: 10.1016/j.petrol.2018.01.045

  • CrossRef
  • Google Scholar

• 26

  Guaraglia D. O. Pousa J. L. (1999). An electrical model of heat flow in soil. Soil Sci. Soc. Am. J.63, 457–463. doi: 10.2136/sssaj1999.03615995006300030006x

  • CrossRef
  • Google Scholar

• 27

  Guaraglia D. O. Pousa J. L. (2014). Introduction to modern instrumentation for hydraulics and environmental sciences, vol. 433. Warsaw/Berlin: De Gruyter Open.

  • Google Scholar

• 28

  Guaraglia D. O. Pousa J. L. Pilan L. (2001). Predicting temperature and heat flow in a sandy soil by electrical modeling. Soil Sci. Soc. Am. J.65, 1074–1080. doi: 10.2136/sssaj2001.6541074x

  • CrossRef
  • Google Scholar

• 29

  Han D. Currell M. J. (2022). Review of drivers and threats to coastal groundwater quality in China. Sci. Total Environ.806:150913. doi: 10.1016/j.scitotenv.2021.150913,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 30

  Ivković M. Todorović A. Plavšić J. (2018). Improved input to distributed hydrologic model in areas with sparse subdaily rainfall data using multivariate daily rainfall disaggregation. J. Hydroinf.20, 784–797. doi: 10.2166/hydro.2018.053

  • CrossRef
  • Google Scholar

• 31

  Ji G. Chen S. Zhou T. Huang Z. Li F. (2021). A new method of predicting network fracture conductivity based on the similitude principle of water and electricity. Phys. Fluids33. doi: 10.1063/5.0073291

  • CrossRef
  • Google Scholar

• 32

  Jiang S. Liu C. Dwivedi D. (2024). GeoFUSE: A high-efficiency surrogate model for seawater intrusion prediction and uncertainty reduction. arXiv [Preprint]. arXiv:2410.20118. doi: 10.48550/arXiv.2410.20118

  • CrossRef
  • Google Scholar

• 33

  Jin J. Espino M. Fernández-García D. Folch A. (2025). Coupling of numerical groundwater–ocean models to improve understanding of the coastal zone. Ocean Sci.21, 1407–1424. doi: 10.5194/os-21-1407-2025

  • CrossRef
  • Google Scholar

• 34

  Kasi M. Chi Y. Padmanabhan G. (2013). Integrating electrical analogy and computer modeling of groundwater flow for teaching flownet concepts. Int. J. Eng. Pedagog.3, 39–46. doi: 10.3991/ijep.v3i4.3052

  • CrossRef
  • Google Scholar

• 35

  Keown J. L. (1991). PSpice and circuit analysis. Indianapolis: Merrill.

  • Google Scholar

• 36

  Lama G. F. C. Errico A. Francalanci S. Solari L. Chirico G. B. Preti F. (2020). “Hydraulic modeling of field experiments in a drainage channel under different riparian vegetation scenarios” in Innovative Biosystems engineering for sustainable agriculture, forestry and food production (Cham: Springer), 69–77.

  • Google Scholar

• 37

  Lama G.F.C. Giovannini R.M. Errico A. Mirzaei S. Chirico G.B. Preti F. (2021). The impacts of nature based solutions (NBS) on vegetated flows’ dynamics in urban areas, 2021 IEEE international workshop on metrology for agriculture and forestry (MetroAgriFor), pp. 58–63. doi:10.1109/MetroAgriFor52389.2021.9628438

  • CrossRef
  • Google Scholar

• 38

  LeVeque R. J. (2007). Finite difference methods for ordinary and partial differential equations: Steady-state and time-dependent problems. Washington: Society for Industrial and Applied Mathematics.

  • Google Scholar

• 39

  LTspice 24.1.10 [Software] (2025). Analog devices. Available online at: https://www.analog.com/en/resources/design-tools-and-calculators/ltspice-simulator.html (Accessed March 11, 2025).

  • Google Scholar

• 40

  Mewis P. (2021). Estimation of vegetation-induced flow resistance for hydraulic computations using airborne laser scanning data. Water13:1864. doi: 10.3390/w13131864

  • CrossRef
  • Google Scholar

• 41

  Nave C.R. (2008) Hydrogeology teaching by means of an electrical analogy. Georgia State University HyperPhysics project. Available online at: http://hyperphysics.phy-astr.gsu.edu/hbase/electric/watcir.html (Accessed March 11, 2025).

  • Google Scholar

• 42

  Nhemachena C. Nhamo L. Matchaya G. Nhemachena C. R. Muchara B. Karuaihe S. T. et al . (2020). Climate change impacts on water and agriculture sectors in southern Africa: threats and opportunities for sustainable development. Water12:2673. doi: 10.3390/w12102673

  • CrossRef
  • Google Scholar

• 43

  Olyslager F. (1999). Electromagnetic waveguides and transmission lines, vol. 51. Oxford: OUP.

  • Google Scholar

• 44

  Pirone D. Cimorelli L. Pianese D. (2024). The effect of flood-mitigation reservoir configuration on peak-discharge reduction during preliminary design. J. Hydrol. Reg. Stud.52:101676. doi: 10.1016/j.ejrh.2024.101676,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 45

  Rama F. Miotlinski K. Franco D. Corseuil H. (2018). Recharge estimation from discrete water table datasets in a coastal shallow aquifer in a humid-subtropical climate. Hydrogeol. J.26, 1887–1902. doi: 10.1007/s10040-018-1742-1

  • CrossRef
  • Google Scholar

• 46

  Richardson C. M. Davis K. L. Ruiz-González C. Guimond J. A. Michael H. A. Paldor A. et al . (2024). The impacts of climate change on coastal groundwater. Nat. Rev. Earth Environ.5, 100–119. doi: 10.1038/s43017-023-00500-2

  • CrossRef
  • Google Scholar

• 47

  Rodrigues Capítulo L. (2015). Evaluación geohidrológica en la región costera oriental de la provincia de Buenos Aires. Caso de Estudio: Pinamar [Geohydrological evaluation in the eastern coast of the province of Buenos Aires. Case of Study Pinamar] (PhD thesis), Facultad de Ciencias Naturales y Museo, La Plata, Argentina

  • Google Scholar

• 48

  Rodrigues Capítulo L. Carretero S. C. Kruse E. E. (2018). Impact of afforestation on coastal aquifer recharge. Case study: eastern coast of the province of Buenos Aires, Argentina. Environ. Earth Sci.77:74. doi: 10.1007/s12665-018-7258-0

  • CrossRef
  • Google Scholar

• 49

  Saltelli A. (1999). Sensitivity analysis: could better methods be used?J. Geophys. Res. Atmos.104, 3789–3793. doi: 10.1029/1998JD100042

  • CrossRef
  • Google Scholar

• 50

  Sánchez-Canales M. Benito A. L. Passuello A. Terrado M. Ziv G. Acuña V. et al . (2012). Sensitivity analysis of ecosystem service valuation in a Mediterranean watershed. Sci. Total Environ.440, 140–153. doi: 10.1016/j.scitotenv.2012.08.032

  • CrossRef
  • Google Scholar

• 51

  Servicio de Hidrografía Naval (2008). Tabla de Mareas [Tide Table]. Pub. H-610, Armada Argentina, Buenos Aires.

  • Google Scholar

• 52

  Stallman R. W. (1963). Electric analog of three-dimensional flow to wells and its application to unconfined aquifers. Water Supply Paper 1536-H. Washington: US Government Printing Office.

  • Google Scholar

• 53

  Theis C. V. (1935). The relation between the lowering of the piezometric surface and the rate and duration of discharge of a well using ground-water storage. Trans. Am. Geophys. Union16, 519–524.

  • Google Scholar

• 54

  Thornthwaite C. W. (1948). An approach toward a rational classification of climate. Geogr. Rev.38, 55–94. doi: 10.2307/210739

  • CrossRef
  • Google Scholar

• 55

  TINA-TI [Software] (2025). A SPICE-based analog simulation program, developed by Texas Instruments in cooperation with DesignSoft, Inc. Available online at: https://www.ti.com/tool/TINA-TI (Accessed February 1, 2025).

  • Google Scholar

• 56

  Van Camp M. Mjemah I. C. Al Farrah N. Walraevens K. (2013). Modeling approaches and strategies for data-scarce aquifers: example of the Dar Es Salaam aquifer in Tanzania. Hydrogeol. J.21, 341–356. doi: 10.1007/s10040-012-0908-5

  • CrossRef
  • Google Scholar

• 57

  Van Griensven A. Francos A. Bauwens W. (2002). Sensitivity analysis and auto-calibration of an integral dynamic model for river water quality. Water Sci. Technol.45, 325–332. doi: 10.2166/wst.2002.0271,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 58

  Västilä K. Järvelä J. Aberle J. (2013). Characteristic reference areas for estimating flow resistance of natural foliated vegetation. J. Hydrol.492, 49–60. doi: 10.1016/j.jhydrol.2013.04.015

  • CrossRef
  • Google Scholar

• 59

  Verruijt A. (1970). Theory of groundwater flow. 1era Edn. London: Macmillan International Higher Education.

  • Google Scholar

• 60

  Violante R.A. Parker G. Cavallotto J.L. 2001 Evolución de las llanuras costeras del este bonaerense entre bahía Samborombón y la laguna Mar Chiquita durante el Holoceno Rev. Asoc. Geol. Argent. 56, 51–66. Available online at: https://revista.geologica.org.ar/raga/issue/view/18/1 (Accessed March 11, 2025).

  • Google Scholar

• 61

  Voss C. I. (2011a). Editor’s message: groundwater modeling fantasies—part 1, adrift in the details. Hydrogeol. J.19, 1281–1284. doi: 10.1007/s10040-011-0789-z

  • CrossRef
  • Google Scholar

• 62

  Voss C. I. (2011b). Editor’s message: groundwater modeling fantasies—part 2, down to earth. Hydrogeol. J.19, 1455–1458. doi: 10.1007/s10040-011-0790-6

  • CrossRef
  • Google Scholar

• 63

  Wang X. He X. Li H. Luo W. (2023). Complex effects of tides on coastal groundwater revealed by high-resolution integrated flow modeling. Water Resour. Res.59:e2022WR033942. doi: 10.1029/2022WR033942

  • CrossRef
  • Google Scholar

• 64

  Xu X. Liang X. Song J. Gao H. Ye X. (2024). Impacts of sea-level rise on the coastal groundwater table simulated by an earth system model with a land–ocean coupling scheme. Earths Future12:e2024EF004479. doi: 10.1029/2024EF004479

  • CrossRef
  • Google Scholar