Statistical modeling of meiganga soils mechanical parameters: regression and correlation approach

In Meiganga, as in many tropical areas, lateritic soils are both an abundant resource and a source of uncertainty for engineers. Their behavior, inherited from a long history of weathering in a humid climate, often defies conventional normative frameworks, while heavy investigative equipment remains difficult to access. Based on this observation, this study aims to characterize the physical and mechanical properties of Meiganga soils and, on this basis, to develop empirical models for predicting their shear strength parameters from simple physical measurements. Seventy-two samples, taken from nine sites during the dry and rainy seasons, were subjected to physical tests and direct shear tests under unconsolidated, undrained conditions. The results reveal a strong dominance of the fine fraction (< 0.08 mm: 28.8–82.9%), high plasticity (mean Ip ≈ 28%) and marked water variability. Apparent cohesion ranged from 18.9 to 32.1 kPa and internal friction angle from 19.5° to 31.2°. Statistical analyses (Pearson correlations, polynomial regressions) show that the friction angle is mainly controlled by the water content and fine grain size (ω, d0.08, Sr), while cohesion depends mainly on the plasticity index and the fine fraction (Ip, d0.08). The models developed offer good predictive performance (R² up to 0.97; RMSE < 4 kPa for cohesion and < 3° for φ). These tools provide a reliable solution for the preliminary and rapid estimation of shear parameters for lateritic soils in tropical contexts where data is limited.