The site investigation was conducted following an initial reference to the soil map of the municipality of Pobè in Republic of Benin, as depicted in Figure 1 . Subsequently, on-site visits were carried out to identify sampling locations, prioritizing areas where the phenomenon is most pronounced within the study area. The materials were gathered within the Pobè municipality, situated in the southeast of Benin, specifically within the districts of Issaba and Ahoyèyè. Table 1 provides the geographical coordinates of each designated sampling site.

Considering the required tests (identification and mechanical tests), a combination of disturbed and undisturbed samples was collected. The undisturbed samples were cored using cubic boxes designed for the collection of undisturbed samples, a compaction hammer, and wooden containers for transportation. In each study area, four (04) undisturbed samples were cored: one from 0 to 0.50 m depth, a second from 0.5 to 1.0 m, a third from 1.0 to 1.50 m, and the final one from 1.50 to 2.0 m depth.

Soil characterization was conducted using geotechnical tests for both physical and mechanical identification.

The disturbed soil samples were subjected to the following physical tests:

The oedometer test as per NF EN ISO 17892-5 [24] and the direct shear test in as per NF EN ISO 17892-10 [25] were conducted on undisturbed soil samples subjected to drying and wetting cycles.

This study employed both oedometer and direct shear tests to investigate the mechanical behavior of undisturbed expansive clay under cyclic wetting and drying conditions. These tests were performed using specialized equipment, following established standards, to simulate natural moisture variations and their impact on shear strength, compressibility, and consolidation characteristics.

The oedometer tests were conducted in accordance with the NF EN ISO 17892-5 (2017) standard to assess the compressibility and consolidation behavior of expansive clay. The testing apparatus consisted of a standard oedometer device, composed of a rigid steel frame, a vertical loading system (using weights or pneumatic pressure), and consolidation rings to hold the soil specimens ( Figure 2 ). These rings, with a diameter of 75 mm and a height of 20 mm, allowed for the application of vertical loads, simulating field stress conditions, and measuring the soil’s response to consolidation.

In addition, a precision balance with an accuracy of 0.01 g was used to monitor the mass of the soil samples throughout the wetting and drying cycles, which was essential for determining moisture content changes. Vertical deformations were continuously recorded using dial gauges or electronic displacement sensors, providing real-time data on the consolidation behavior of the soil samples during each loading phase. The soil specimens underwent a controlled drying process in an oven, while a water bath was employed during the wetting phase to achieve full saturation, replicating the effects of natural moisture infiltration.

The testing procedure began with the careful preparation of undisturbed soil samples. These samples were trimmed to fit within the oedometer consolidation rings, ensuring minimal disturbance to the natural structure. Once prepared, the samples were subjected to a series of incremental vertical loads, ranging from 12.5 kPa to 800 kPa, simulating real-life stress conditions. At each load increment, vertical deformations were measured over a period of 24 hours, allowing for the calculation of the compressibility index (Cc) and the swelling index (Cs) based on the stress-strain curves obtained during the loading and unloading cycles.

To simulate natural environmental moisture variations, the soil samples were subjected to multiple cycles of wetting and drying. During the wetting phase, the samples were gradually submerged in water until full saturation was reached, while during the drying phase, the samples were placed in an oven to simulate the effects of evapotranspiration. Moisture content and vertical deformations were measured at each cycle, providing a detailed understanding of how the soil’s compressibility and swelling behavior evolved over time.

The direct shear tests were performed following the NF EN ISO 17892-10 (2018) standard to determine the shear strength parameters of the soil samples, specifically the internal friction angle ( $\phi$) and cohesion (Cu). The direct shear apparatus used in this study consisted of a horizontally split shear box, allowing the upper half of the box to move relative to the stationary lower half, thereby imposing a horizontal shear force on the soil sample while being confined under a vertical load ( Figure 3 ).

The load application system allowed for the precise application of normal stresses to the soil samples using dead weights or a hydraulic piston. Horizontal displacements were measured using high-precision displacement transducers, and shear forces were recorded through a load cell, enabling accurate calculation of the shear strength parameters. As in the oedometer tests, a precision balance and drying oven were used to regulate the moisture content of the samples before and after testing, ensuring accurate control of the sample conditions.

The experimental procedure began with the preparation of undisturbed soil samples, which were carefully trimmed to fit the dimensions of the shear box (60 mm in diameter and 20 mm in height). These samples were saturated to replicate field conditions of maximum moisture content. After saturation, normal stresses were applied to the soil samples at levels of 50, 100, and 200 kPa, simulating in-situ loading conditions. Horizontal shear stress was incrementally applied, and the resulting horizontal displacements were recorded.

The shear stress versus horizontal displacement curves were analyzed to determine the peak shear strength at each level of normal stress. From these results, the internal friction angle and cohesion were calculated using the Mohr-Coulomb failure criterion. This allowed for the evaluation of the degradation in shear strength parameters after several wetting-drying cycles.

Wetting Procedure

The wetting process was designed to simulate the moisture infiltration that occurs during the rainy season. The following steps were followed:

1) Sample Preparation: Undisturbed soil samples were carefully extracted and placed within shear rings and oedometer rings. These rings were pre-weighed both empty and after coring. For each sample type, a control core was extracted to determine the wetting water content.

2) Placement in Jars: The cored samples were placed inside jars, each labeled with material details, the weight of the empty ring, and the combined weight of the ring and soil. Control samples for moisture content and saturation control were placed in a separate jar, while the samples intended for testing were placed in another. A perforated support plate was installed inside the jar to allow water to circulate freely and moisten the samples from beneath.

3) Water Addition: Water was poured into the jar until the samples were fully submerged, and the jar was then sealed. This ensured uniform wetting of the samples.

4) Monitoring of Saturation: Control samples were weighed every hour for 48 hours. A mass-versus-time curve was plotted for each control sample to monitor changes. Once three consecutive weight measurements showed no significant variation (less than 0.2 g), the degree of saturation was considered 100%. After wetting, one control core was placed in a drying oven to determine the moisture content, while the remaining cores were prepared for the drying phase.

Drying Procedure

Following the wetting process, the drying phase was initiated to simulate the natural drying conditions that occur during the dry season. The soil samples were subjected to sunlight exposure at temperatures ranging between +26˚C to +36˚C. The drying procedure followed these steps:

1) Exposure to Sunlight: After wetting, the samples were exposed to direct sunlight for 9 hours each day. The temperature was monitored, with morning temperatures around +26˚C and afternoon temperatures reaching up to +36˚C, based on a 3-day period of temperature measurements.

2) Target Moisture Content: The drying process continued until the soil samples reached a moisture content of 14%. This target moisture content was chosen based on literature data for expansive soils in the region, specifically from observations in the depression of La Lama during the dry season.

3) Control Core Measurements: A control core was used to monitor the moisture content during the drying process. The mass and moisture content of the control cores were measured to ensure accurate replication of the drying conditions.

Figure 4 illustrates the increase in soil sample mass over time during the wetting procedure.

Figure 5 showcases the drying kinetics, depicting moisture content evolution over time.

The results of the oedometer and direct shear tests provided critical insights into the effects of cyclic wetting and drying on the mechanical properties of expansive clay. These tests allowed for the calculation of key parameters such as the compressibility index (Cc), the swelling index (Cs), and the shear strength parameters (internal friction angle φ and cohesion Cu). The data were further analyzed using numerical simulations to model the soil’s behavior under cyclic moisture conditions, offering essential data for improving pavement design in regions with expansive clay subgrades.

To assess the influence of various drying-wetting cycles on the studied soil, oedometer tests and direct shear tests were conducted on the undisturbed samples. Each sample underwent the following cycles:

The results of the grain-size distribution tests for the clayey soils collected in the Issaba area are presented in Figure 6 . The classification of these soils is based on the American soil classification (HRB) and the GTR classification (NF P11-300). The soil characteristics obtained from various identification tests, including particle size distribution, plasticity index, and liquid limit, are shown in Table 2 . According to the HRB classification, the soil samples are classified as A7-5, indicating a clayey soil, and according to the GTR classification, they are primarily categorized as A4, except for the soil at a depth of 1.5 to 2.0 m, which belongs to class A3.

The progressive degradation of soil samples over the course of multiple cycles of wetting and drying is evident in Figures 7-10 . By Cycle 3, the samples exhibit significant swelling post-wetting and severe cracking post-drying ( Figure 10 ), indicating progressive soil damage with increasing cycle numbers.

The results of the direct shear tests are summarized in Table 3 . A clear trend of decreasing shear strength with increasing drying-wetting cycles is observed. Cohesion ( $C_u$) generally increases between Cycle 0 and Cycle 2 but decreases significantly by Cycle 3, as shown in Figure 11 . The friction angle ( ${\phi }_u$) also follows a decreasing trend ( Figure 12 ), confirming that repeated cycles of shrink-swell progressively weaken the soil’s resistance to shear forces.

These experimental findings illustrate the impact of cyclic moisture variations on the mechanical properties of the expansive clay. The shear strength of the soil, measured in terms of cohesion and internal friction angle, decreases notably after three cycles, suggesting a degradation of the soil’s structural integrity.

Table 4 presents the results of oedometer tests, showing variations in compressibility and swelling index across different cycles. The evolution of the compressibility index over cycles is depicted in Figure 13 . There is an initial decrease in compressibility from Cycle 0 to Cycle 1, followed by an increase from Cycle 1 to Cycle 2. This could be attributed to the breakdown of soil structure and the formation of new cracks. By Cycle 3, the compressibility stabilizes but remains higher than in Cycle 0, indicating residual structural degradation.

The swelling index, shown in Figure 14 , follows a similar trend, with a sharp increase after Cycle 2, suggesting a higher swelling potential after prolonged exposure to shrink-swell conditions. This behavior is further confirmed by the evolution of the oedometer modulus ( Figure 15 ), which shows a decline after Cycle 1, indicating a reduction in soil stiffness.

In conjunction with the experimental findings, a two-dimensional finite element (FE) model was developed using PLAXIS 2D Foundation software to simulate the pavement system’s behavior on expansive clay subjected to cyclic shrink-swell. The model configuration is shown in Figure 16 , and material properties for the pavement layers are provided in Tables 5-7 .

The simulation results in Figures 17-19 , is directly responsible for the observed increases in pavement displacement in the numerical model.

The experimental and numerical results together offer a comprehensive view of the mechanical behavior of expansive clays under cyclic wetting and drying. The experimental tests show that repeated cycles lead to significant changes in soil properties, including reduced cohesion, increased compressibility, and a higher swelling index. These findings are supported by the numerical model, which predicts an increase in pavement displacement as the soil beneath it degrades.

In practical terms, this study highlights the need for robust design strategies when dealing with pavements constructed on expansive clay subgrades. The numerical simulations align well with the experimental observations, showing that cyclic environmental conditions, particularly in regions with significant seasonal moisture variations, can lead to progressive pavement failure if not properly mitigated.

The findings from both the laboratory tests and numerical modeling demonstrate that cyclic shrink-swell can cause significant degradation in the mechanical properties of expansive clays, ultimately affecting pavement performance. To mitigate these effects, soil stabilization techniques or design modifications may be necessary to improve the long-term performance of pavements in areas prone to expansive clay behavior. Additionally, the integration of experimental data with numerical models provides a powerful tool for predicting pavement performance and planning for sustainable infrastructure development in such challenging environments.

One of the critical aspects to consider when analyzing the behavior of expansive clay soils is the impact of environmental conditions, particularly temperature and drying rates. Seasonal variations, ambient temperatures, and the rate at which moisture is lost or gained in soils play a significant role in altering the mechanical response of soils subjected to drying-wetting cycles.

Previous studies have highlighted that expansive clay soils respond sensitively to temperature fluctuations, particularly concerning swelling and shrinkage behavior. Basma et al. (1996) observed that higher temperatures accelerate the drying rate of soils, leading to faster contraction, which can cause increased surface cracking and the formation of microcracks within the soil matrix. These cracks affect the permeability and mechanical strength of the soil, contributing to the progressive degradation of its physico-mechanical properties.

Regarding the drying rate, studies such as Soltani et al. (2020) show that rapid drying cycles exacerbate internal stresses in clay soils, increasing the potential for differential shrinkage. Faster drying leads to a rapid reduction in water content, accelerating pore desaturation and encouraging uneven contraction in different parts of the soil. This phenomenon affects the shear strength and consolidation properties of soils, as demonstrated by our oedometer and direct shear test results.

In our study, temperatures during the drying phases ranged between 26˚C and 36˚C, simulating the diurnal variations observed under natural climate conditions in the study area. These temperature ranges are comparable to those described by Louati et al. (2021), who found that similar daytime temperatures intensify soil shrinkage, leading to an increase in surface cracking and, consequently, changes in shear strength properties.

Drying rates influenced by temperature also impact the swelling behavior of clay soils. When soils are re-saturated after rapid drying cycles, the samples tend to exhibit differential swelling, often more pronounced than in earlier cycles. This observation aligns with the findings of Basma et al. (1996), who noted a temporary increase in swelling capacity in soils that underwent rapid desiccation cycles. As the number of cycles increases, the intensity of swelling gradually decreases, as observed in our experimental results, suggesting a possible stabilization of the soil’s behavior over time.

The drying rate plays a central role in the distribution of internal stresses within the soil. Slow drying, as simulated in some field studies, allows for more uniform contraction of clay layers, reducing microstructural damage within the soil. Conversely, rapid drying, observed during the phases of our study, tends to exacerbate internal stresses, promoting the formation of cracks. This behavior is particularly critical when predicting the durability of infrastructure built on expansive clay soils in arid or semi-arid regions, where prolonged drought periods are common.

The influence of temperature and drying rates on the shrink-swell behavior of expansive clay soils is evident. Our experimental observations confirm the conclusions of previous studies, suggesting that rapid drying cycles combined with high temperatures lead to a faster degradation of the soil’s mechanical properties, particularly in terms of shear strength and compressibility. This underscores the importance of considering environmental conditions when assessing the behavior of expansive soils to ensure the durability and stability of infrastructure in at-risk areas.

The reliability of the numerical simulations was assessed by comparing the model’s results with both field data and documented case studies. This validation process is crucial for ensuring that the finite element model accurately simulates the behavior of expansive clay soils under cyclic shrink-swell conditions, providing credible insights for real-world applications and infrastructure planning.

Field data from a monitored pavement section in Pobè, Benin were employed to validate the outcomes of the numerical simulations. This site was selected due to the presence of expansive clay soils with similar characteristics to those modeled in this study. The site has been subjected to long-term cyclic wetting and drying, providing an excellent benchmark for comparison.

In addition to field data, the model was further validated using case studies from the literature. A comparative analysis was conducted with the results of studies by Basma et al. (1996) and Sayem et al. (2016), which examined pavement performance on expansive soils under cyclic environmental loads. The numerical model’s predictions demonstrated strong alignment with the documented field results from these studies, providing additional confidence in the model’s accuracy.

Despite demonstrating robust accuracy through comparisons with both field data and literature-based case studies, the numerical model has some limitations:

Conducting a sensitivity analysis is a crucial step in evaluating the robustness and reliability of numerical models, particularly in the context of complex soil-pavement interactions. This analysis identifies which parameters exert the greatest influence on the performance of pavement systems constructed over expansive clay soils subjected to cyclic shrink-swell conditions. By quantifying the impact of each parameter, the model can be optimized for enhanced predictive accuracy, allowing for more informed decision-making in geotechnical engineering.

The sensitivity analysis focused on a set of critical input parameters that govern the mechanical behavior of expansive soils and their interaction with overlying pavement structures. The parameters selected for this analysis include:

To assess the sensitivity of the model, each of these key parameters was systematically varied, while all other parameters were held constant. This approach isolates the influence of each variable on the model outputs, specifically vertical displacement, shear strength, and stress distribution within the pavement system. Multiple simulations were conducted for each parameter, with incremental adjustments, to capture non-linear effects and interdependencies.

The sensitivity of the model was evaluated using two primary performance indicators:

The sensitivity analysis yielded the following insights:

Young’s Modulus (E):

The numerical model exhibited significant sensitivity to variations in Young’s Modulus across both the pavement layers and the expansive clay subgrade. Higher values of E resulted in reduced vertical displacements, indicating that stiffer materials lead to enhanced structural performance and resistance to deformation. Conversely, lower values of E led to greater deformations, particularly in the pavement layers, reflecting a reduction in system rigidity. Figures 17-19 illustrate the strong correlation between Young’s Modulus and vertical displacement across different stages of the cyclic testing.

Cohesion (C_u):

Cohesion was identified as a critical factor influencing the shear strength of the soil. Reductions in cohesion significantly increased vertical displacement and reduced the soil’s ability to resist shear forces, particularly in the deeper soil layers. Figure 11 shows the marked decline in system stability as cohesion diminishes over successive shrink-swell cycles.

Internal Friction Angle ( ${\phi }_u$):

The internal friction angle exhibited a moderate impact on the system’s overall behavior. Although its effect on vertical displacement was less pronounced than cohesion, its influence became more significant after multiple shrink-swell cycles. This suggests that the internal friction angle plays a key role in the long-term shear resistance of the soil. The correlation between internal friction angle and shear strength is depicted in Figure 12 .

Poisson’s Ratio (v):

While variations in Poisson’s Ratio produced relatively minor effects on vertical displacement, they did influence lateral deformation and stress distribution within the pavement system. Higher values of $\nu$ led to greater lateral expansion under stress, which could potentially contribute to surface cracking in the pavement layers.

Saturated Permeability ( $K_s$):

Saturated permeability emerged as one of the most influential parameters, as it controls the rate of moisture infiltration into the expansive clay. Increased permeability led to faster wetting, which exacerbated the cyclic shrink-swell behavior, resulting in higher vertical displacements. Figure 14 demonstrates the relationship between permeability and moisture infiltration rate, highlighting its significant impact on swelling and settlement behavior.

Oedometer Modulus ( $E_{oed}$):

The oedometer modulus had a profound effect on the compressibility of the soil. Lower values of $E_{oed}$ resulted in greater settlement, particularly in the early stages of the wetting-drying cycles. As shown in Figure 15 , the reduction in soil stiffness over successive cycles indicates progressive degradation of the expansive clay’s mechanical properties.

Based on the sensitivity analysis, the following parameters were identified as the most critical for accurately modeling the performance of pavement systems on expansive clay soils:

The results of the sensitivity analysis underscore the importance of precise calibration of key parameters in the numerical model to ensure reliable predictions of pavement performance over expansive clay subgrades. Parameters such as Young’s Modulus, Cohesion, Saturated Permeability, and Oedometer Modulus emerged as the most critical, demonstrating the strongest influence on vertical displacement, shear strength, and overall pavement stability.

This analysis highlights the necessity for field calibration of these parameters to account for site-specific conditions and environmental factors. Further research should focus on integrating real-time environmental data, such as temperature and moisture variations, to refine the model and improve its predictive capabilities for long-term infrastructure performance.

Soil stabilization has been widely recognized as an effective method to reduce the shrink-swell potential of expansive soils. The application of stabilizing agents, such as lime, cement, fly ash, and chemical polymers, can significantly improve the mechanical properties of the soil, making it more resistant to moisture-induced volume changes. Key stabilization techniques include:

Lime stabilization is one of the most effective and commonly employed techniques for treating expansive clays. The chemical reaction between lime and clay minerals results in pozzolanic activity, which increases the soil’s modulus of elasticity and cohesion while reducing its plasticity and swell potential. Lime treatment decreases soil permeability, thereby limiting moisture infiltration and subsequent swelling. The long-term effectiveness of lime stabilization in mitigating the effects of cyclic shrink-swell behavior has been well-documented in both laboratory and field studies [26] .

Cement and fly ash are widely used to stabilize expansive soils, particularly in high-traffic areas. Cement improves the load-bearing capacity of the soil by forming a rigid matrix, while fly ash acts as a pozzolanic material that enhances the soil’s resistance to moisture-related volume changes. These stabilizing agents significantly improve shear strength and reduce swelling potential, making them suitable for areas prone to significant traffic-induced stresses [27] . This combined approach also enhances the durability of pavements on expansive soils, contributing to long-term structural integrity.

Geosynthetics, including geogrids and geotextiles, are increasingly used to reinforce expansive soils, providing additional tensile strength and stabilizing the soil mass. When integrated into pavement systems, geosynthetics help to distribute loads more evenly across the subgrade, reducing differential settlement and mitigating the impact of soil swelling. Their effectiveness under cyclic loading conditions has been confirmed in numerous studies, making them an effective solution for enhancing pavement performance on expansive soils [28] .

Recent advancements in soil stabilization have introduced chemical polymers as environmentally friendly alternatives for improving expansive soils. Polymers form a flexible yet durable matrix within the soil, reducing plasticity and improving moisture resistance. These polymer-based stabilization techniques exhibit reduced swell-shrink behavior and increased resilience to cyclic moisture variations [29] . Given their non-toxic and environmentally sustainable nature, polymers are particularly attractive for projects subject to stringent environmental regulations.

Beyond soil stabilization, certain pavement design modifications can play a crucial role in managing the challenges posed by expansive soils. These design adjustments focus on optimizing the pavement structure to accommodate soil movements and protect the pavement system from damage.

Increasing the thickness of pavement layers, particularly the base and subbase, can help distribute the applied loads more evenly across the expansive subgrade, thereby minimizing cracking and differential settlement caused by shrink-swell behavior. Thicker pavement layers act as a buffer, reducing the stress transferred to the subgrade and enhancing long-term pavement performance [30] . This approach is particularly effective in regions with high traffic volumes, where additional load-bearing capacity is necessary to withstand cyclic soil movements.

Flexible pavements are better suited to accommodate the movements of expansive soils compared to rigid pavements. Their inherent flexibility allows for minor deformations without significant cracking, thus reducing the risk of damage due to soil swelling and shrinkage. Additionally, the materials used in flexible pavements can be designed to resist moisture infiltration, further protecting the subgrade from cyclic moisture variations [31] . This design strategy provides enhanced durability and reduces maintenance costs over time.

Effective drainage systems are essential for managing the moisture content of expansive soils. Proper drainage prevents excessive moisture from infiltrating the subgrade, thereby reducing the risk of swelling and shrinking. The installation of sub-surface drains, permeable layers, and moisture barriers (e.g., geomembranes) can significantly mitigate the impact of moisture fluctuations on pavement performance [32] . These systems are particularly crucial in areas with pronounced seasonal wet and dry cycles.

The incorporation of transition zones or buffer layers into pavement design is an effective method for absorbing differential movements caused by expansive soils. These zones typically consist of less expansive materials, such as granular soils or treated subgrade layers, that act as buffers between the expansive soil and the pavement structure. Transition zones help to minimize cracking and other forms of pavement distress by absorbing and redistributing soil movements [33] . This approach is especially beneficial in areas with severe shrink-swell cycles.

The selection of appropriate soil improvement strategies and pavement design modifications depends on several factors, including the degree of soil expansivity, traffic loads, and regional environmental conditions. In highly expansive regions, a combination of soil stabilization (e.g., lime or cement) and design modifications (e.g., thicker pavement layers or flexible pavement systems) may be required to ensure long-term pavement performance.

Field trials and numerical simulations provide valuable tools for evaluating the effectiveness of these strategies. Finite element models can simulate the behavior of stabilized soils and modified pavement systems under cyclic loading, helping to identify critical parameters that influence performance and guide the selection of the most appropriate improvement techniques. This integrative approach, combining empirical data and advanced modeling, allows for a more tailored and efficient response to the challenges posed by expansive soils.