Integrated modeling of surface water and groundwater interactions in the Eastern Mitidja Plain, North Algeria: investigating the impact of human activities and climate change

Groundwater (GW) holds major importance in many countries, often serving as the sole or primary source of drinking water. In these regions, GW is not just a resource but a cornerstone of national water security, which is essential for preserving ecosystems, providing water for human needs, and promoting agriculture. The rising challenges induced by climate change and precipitation highlight the importance of GW, making its sustainable management critical. The significance of GW in ensuring water quality and supply has been further emphasized by studies, highlighting its role in public health and environmental stability (Wang et al. 2021; Isinkaralar et al. 2024). This issue is addressed by implementing GW models to gain a comprehensive understanding of GW dynamics and forecast the responses of aquifers to diverse stresses, both globally (Wang et al. 2021; Al-Amri et al. 2022; Li et al. 2023) and regionally, such as within the Mediterranean (Diaz et al. 2020; Daoud et al. 2022). However, surface water (SW) and GW were traditionally managed as separate components of this cycle; by the time, their interaction was explored (Winter et al. 1998). Over time, however, their interaction was increasingly explored, demonstrating that the exchange of water between SW and GW influences both resources' quantity and quality worldwide (Winter et al. 1998; Brunner et al. 2017; Banerjee & Ganguly 2023). The interactions between SW and GW are dynamic and complicated. They are influenced by a variety of factors, including climate change (Wang et al. 2021) and human activities (Jiang et al. 2022).

The harmful impacts of extraction and pollution, particularly through contaminants like polycyclic aromatic hydrocarbons, significantly complicate water management due to their ecological and human health risks (Ambade et al. 2022). It underscores the importance of understanding their interdependencies, which enable the assessment of water availability, the creation of sustainable water allocation policies, and the protection of water quality. This has become increasingly important as climate change has more noticeable effects. Adaptive water management solutions are necessary because changes in precipitation patterns, altered evaporation rates, and shifts in hydrological regimes can drastically alter the dynamics of these interactions. However, despite the abundance of studies exploring the interactions between SW and GW, there is still a major gap in our understanding of effectively modeling these interactions and evaluating the overall effects of climate change and human activities on it. Even though some field measurements can be taken to investigate GW–SW interactions (Brunner et al. 2017; Banerjee & Ganguly 2023). Integration of computer models is required (Foglia et al. 2018). This limited understanding hampers our ability to make informed decisions regarding water resource management.

This study focuses on the El Hamiz watershed in the Eastern Mitidja aquifer (North Algeria) as part of the GW contract project. This study is the first of its kind in Algeria to model interactions between SW and GW aiming to develop sustainable GW management strategies in accordance with the United Nations Sustainable Development Goals (SDGs). This study focuses on the Eastern Mitidja alluvial aquifer, which has recently experienced excessive depletion, evaluating the SW–GW interactions influenced by human activities and contemporary climate conditions. Therefore, the primary objective of this study is to address this research gap by using the latest version of the three-dimensional finite difference model (MODFLOW 6) and its specialized packages for modeling these interactions (Langevin et al. 2017). This model represents rivers, streams, and other SW features within the domain and simulates water exchange between the GW system and SW, accounting for factors such as infiltration and evaporation. Spanning from 1981 to 2022, the study involved rigorous calibration and validation to ensure model accuracy before simulating future outcomes under climate change. We generated five scenarios using Coupled Model Intercomparison Project version 5 (CMIP5) and Coupled Model Intercomparison Project phase 6 (CMIP6) forecasts. The initial scenario simulated the continuity of current pumping rates, the second scenario assumed a 25% increase in pumping, and the third scenario simulated a 20% decrease. The fourth scenario amalgamated the preceding two, concluding with an optimistic assessment of the Algerian government's sustainable GW development strategies. This approach incorporates desalination technology, projected to augment the water supply by 60% of the water demand. Finally, this study intends to improve our understanding of SW and GW dynamics, providing insights to guide water resource management decisions and support the sustainable use and preservation of the region's water resources.

This study addresses a crucial gap in the integration of SW–GW interactions within semi-arid regions, specifically in the El Hamiz watershed, Algeria. Despite prior research, limited focus has been given to modeling these interactions under climate change and human impact. By using MODEFLOW 6, this study fills that gap, offering a framework adaptation to similar regions worldwide. The findings support more accurate GW management and inform evidence-based policies, contributing to the United Nations SDGs in water security, climate resilience, and environmental sustainability.

The study area experiences a semi-arid Mediterranean climate, characterized by wet winters (September–May) and dry summers (June–August). Data collected at the El Hamiz climatic station show that the average annual rainfall varies between 300 and 1,000 mm, the mean annual average precipitation is about 700 mm, and the average annual temperature is 25 °C.

For the geological context, several studies were done by Glangeaud (1932), Bonneton (1977), Partners (1983), Benziada (2003), and Toubal (1998) (Yahiaoui et al. 2023). The synthesis of these studies reveals that the El Hamiz sub-basin is characterized by the presence of two essential morpho-structural units: (i) Blidean Atlas is situated in the upstream part of the El Hamiz sub-basin. It consists mostly of fossil-free Lower Cretaceous shales and marly limestone outcrops. These shales extend to the south and southeast under the clay formation and create the base of the deposits such as marly limestones, sandstones, sandy-clay, and conglomerates. (ii) The Mitidja plain is located in the downstream part of the El Hamiz sub-basin and is formed by subsidence and sedimentation. It is covered by Pliocene to Quaternary alluvial deposits. Partners (1983) divided these formations into five stratigraphic divisions from oldest to youngest: Pliocene-Plaisancian, Pliocene-Astian, El Harrach, Mitidja, and recent alluvial deposits.

For the hydrogeological context, the quaternary alluvial aquifer is the main reservoir for GW storage in the study region. It is composed of sand, gravel, and rollers alternating with silt and clay. The substratum of the aquifer is trained by the blue marls of the Plaisancian. The thickness of the aquifer and GW depth vary from 100 to 200 m and from 4 to 30 m, respectively (Khous et al. 2019).

Based on bibliographic data, piezometric measurements, and pumping tests, the transmissivity varies from 1.5 × 10⁻² to 2.1 × 10⁻² m²/s and the storage coefficient ranges from 2 to 10% (Boufekane et al. 2020). Moreover, the wells of the study area have revealed a pumping rate from 5 to 40 l/s (AGIRE 2023). Rainfall is an important source of the aquifer recharge.

In this study, the working database was created from a piezometric companion conducted in June 2022 and data collected from the National Agency for Water Resources (ANRH) and The National Agency for Integrated Water Resources Management (AGIRE) as follows: historical piezometric measurements, results from pumping tests in 160 wells, results of geoelectrical section interpretation, cartographic documents (geological map, topographic map, land use/land cover map, slope map, and elevation model), El Hamiz river's data (cross-section produced by ANRH, water level, and flow rate), meteorological data, and CMIP5 and CMIP6 projections (precipitation and temperature).

The flow model is simulated and calibrated using the Model Muse Interface. It is a free and open-source graphic user interface that works with the different versions of USGS GW models, while the temporal units chosen are seconds and the spatial ones are meters (Winston 2019).

The ground elevation, which is derived from 30 × 30 m Digital Elevation Model (DEM) data, was used to represent the top of the aquifer and interpolated within the Model Muse to match the model grid. The assessment of accurate boundary conditions is crucial in GW flow modeling as it significantly influences GW movement. The model boundaries include hydraulic features like GW divides and physical features such as SW impermeable rocks (Anderson et al. 1992).

We used the Constant Head Designation (CHD) package to simulate dynamic head boundaries between the study area and the Mediterranean Sea. Most of the study area's outer boundary was treated as a no-flow boundary, with regional GW inflow from the southeastern limits based in McDonald's (1982) study and piezometric data, indicating a flow rate of 12,400 m³/day.

The well package was implemented using data from 164 wells provided by the National Integrated Resources Management Agency of Algiers (ANRH), with pumping rates ranging from 3 to 3,360 m³/day. The river package simulated the El Hamiz River using data from the El Hamiz hydrometric station and ANRH cross-sections.

Hydraulic conductivity was extrapolated from pumping test data and geological profiles, while the recharge package simulated flow across the model's top layer. Recharge was estimated as 5–20% of average rainfall using a chloride balance approach, as reported in the Eastern Mitidja GW contract (AGIRE 2023). Measured rainfall data collected from ANRH and projections from CMIP5/6 were key inputs in the model.

MODFLOW 6 solvers are managed by the iterative model solution application using linear and nonlinear methods to solve flow and transport simulations (Langevin et al. 2017).

Calibration is the process by which the simulated head value is modified to align with the measured head value. The head values from 1981 to 1982 were taken for calibrating the model for this study.

In steady-state simulation, the goal is to balance total inflow and outflow, minimizing mass balance error. The model is calibrated by iteratively adjusting hydrogeologic parameters to reduce the difference between computed and observed head values under steady-state conditions.

The model is run for 4 years, from 2019 to 2022, under transient-state conditions. There were eight stress periods, with two periods occurring each year: wet and dry. Piezometric data and pumping data were collected from 164 distinct wells in ANRH and AGIRE.

After calibration and validation, a future forecast was conducted using the calibrated GW model to examine the impact of climate change on the GW system and its connection to SW bodies. This analysis applied multiple future projections, namely the Representative Concentration Pathways RCP4.5 and RCP8.5, derived from the CMIP5. RCP4.5 represents a scenario with moderate greenhouse gas emission stabilization, while RCP8.5 represents a high greenhouse gas emission trajectory. Additionally, SSP2 is the Shared Socioeconomic Pathway, a future with moderate challenges to mitigation and adaptation, and SSP5 is a scenario with high challenges and uncertainties in addressing climate change impacts, both from the CMIP6, an updated and improved iteration of climate models built upon the legacy of CMIP5 for comprehensive climate assessments. These pathways were selected for predicting the effects of climate change (Almazroui et al. 2020; Mesgari et al. 2022; Ha et al. 2023). By comparing the different results, five scenarios were proposed for a 40-year period until 2060. The different characteristics of each scenario are presented in Table 1.

The steady-state simulation adjusted hydraulic conductivity values, derived from literature and geological data, to achieve a minimal error relative to the 1982 piezometric head of September. The result is a hydraulic conductivity value of K_x = K_y = 1 × 10⁻³ m/s and K_z = 1 × 10⁻⁴ m/s and 1.33 × 10⁻⁵ m/s, with recharge at 20% of the rainfall. The model showed a south-north GW flow with a coefficient of determination (R²) of 0.95, RMSE of 0.6 m, and NSE of 0.98, indicating an accurate representation of GW flow. The transient model was validated using 100 piezometric head observations, as detailed in Table 2.

The SSP5 and SSP2 models significantly impact piezometric levels, particularly in the northeastern region, due to reduced rainfall and recharge. Conversely, RCP4.5 and RCP8.5 models predict a potential 5-m rise in the northwestern and southern regions. However, these models also suggest a possible decline in piezometric levels, ranging from 10 m in 2040 to 20 m in 2060, if the region experiences prolonged drought.

Comparing the simulation scenarios from 2019 to 2020 for recharge and discharge, we conclude that scenarios with high pumping (scenario 2), low recharge (scenario 3), and their combination (scenario 4) enhance the GW recharge through an important hydraulic gradient. This promotes water flow from the El Hamiz River to the aquifer. In contrast, the lowest recharge is observed in scenario 5, which involves a reduced pumping rate combined with seawater desalination and new irrigation techniques to balance water supply demand.

The reduction in the recharge rate has a varied impact, with less drastic variations near the El Hamiz River. Cumulative effects further decrease aquifer levels. Otherwise, scenario 5 offers a different approach that mitigates the negative effects of increased pumping wells and low recharge rates, particularly noticeable in regions with many water supply wells.

• (ii) For the period 2040–2060, the negative effects on water persist due to the continued increase in pumping and decrease in recharge. Scenario 2 shows a more pronounced decrease in piezometric levels, with variations reaching −4.96 m. However, scenario 3 is less affected, with some areas decreasing by about −0.40 m. In scenario 5, the situation improves as piezometric levels increase by up to +13.17 m.

The results of this study indicate that the El Hamiz sub-basin faces significant challenges primarily due to excessive GW extraction and decreasing replenishment rates. The observed declines in GW levels across several scenarios suggest potential consequences, including water scarcity, higher extraction costs, and gradual aquifer degradation. These issues underscore the urgent need for a comprehensive management strategy. Recommended actions include optimizing water use, enforcing GW management regulations, and integrating alternative water sources such as desalination and optimized irrigation techniques.

This study employed the MODFLOW6 model to investigate the dynamics of SW and GW interactions within the El Hamiz River and their impact on the surrounding areas. The analysis based on simulations covering the period from 2019 to 2060 yielded several important conclusions.

The interaction between SW and GW in the study area exhibits a significant variation in the annual scale. The recharge from the aquifer to the El Hamiz River, which averaged 2.05 m³/s in 1982, reversed by 2019, with a flow of 1.03 m³/s from the river to the aquifer. This reversal is attributed to overexploitation and decreased rainfall, resulting in a lower recharge rate.

Scenario analyses revealed that both climate change and human activities profoundly affect the aquifer at the first impression. The results of the two projections, CMIP5 and CMIP6, are barely different. We also noticed that a substantial decline in the water table in the northeastern region is primarily driven by heavy pumping activities. In contrast, the northwestern and southern regions demonstrated increased piezometric levels attributed to reduced pumping and increased recharge from the El Hamiz River and El Hamiz Dam. The second scenario, on the other hand, predicted a more severe drop in the northeast, owing largely to increased water extraction. Interestingly, despite the projected decreases in rainfall in the SSP2 and SSP5 forecasts, the northwestern and southern regions exhibited resistance, with water levels either stabilizing or increasing. The fourth scenario highlighted the greater impact of pumping activities on GW levels compared to climate change in the second and third scenarios, while the fifth scenario shows the effectiveness of saltwater desalination and improved irrigation techniques in sustaining GW levels. This underscores the critical need to reduce GW extraction as a key strategy, aligning with the SDGs particularly SDG 6 (Clean Water and Sanitation) and SDG 13 (Climate Action), emphasizing the importance of sustainable water management practices and climate-resilient strategies to ensure the long-term viability of GW resources. In terms of the interaction between SW and GW, the findings show a gradual decrease in the recharge rate from the river to the aquifer. This decline is likely due to reduced precipitation and overexploitation. As urbanization progresses, there is a notable expansion of urban land and an accompanying increase in irrigated areas. Consequently, there is a rise in GW extraction. This intensifies SW–GW interactions, primarily driving flow from rivers to GW, and may lead to environmental problems, such as the seawater intrusion in the northern part of the aquifer. The study emphasizes the importance of integrating sustainable water management practices and adaptive strategies, in line with the SDGs, to ensure the long-term viability and resilience of the El Hamiz River basin.

The study was conducted by the Water and Environment Engineering Laboratory (GEE) at the Higher National School of Hydraulics (ENSH). It is carried out within the framework of the SWATCH project (Prima project) funded by the Directorate General for Scientific Research and Technological Development (DGRSDT), Algeria.

This research is a part of the project ‘Mitidja Groundwater Contract’ conducted by AGIRE.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.