Regression modeling to predict flexible pavement rutting on Uganda’s road from Ntungamo to Kabale

Rutting remains a predominant distress in flexible pavements, compromising ride quality, accelerating structural fatigue, and escalating maintenance costs. The Ntungamo–Kabale highway in southwestern Uganda has experienced premature rutting within seven years of reconstruction, raising concerns about the durability of tropical pavement systems. This study developed and validated a multiple linear regression (MLR) model to predict rut depth as a function of key structural, traffic, and environmental variables. Field and laboratory data were collected from 50 pavement sections, encompassing asphalt thickness, binder content, aggregate gradation, subgrade moisture content, California Bearing Ratio (CBR), and Equivalent Single Axle Loads (ESALs). Stepwise regression identified subgrade moisture, CBR, and ESALs as significant predictors (p < 0.05). The final model achieved a coefficient of determination R ²  = 0.82 and a Root Mean Square Error (RMSE) of 0.21 mm, with validation producing comparable accuracy (R ²  = 0.78; RMSE = 0.18 mm). Moisture and axle-load intensity were dominant contributors to rutting, whereas binder content, gradation, and asphalt thickness were statistically insignificant. The model provides a robust and data-efficient tool for rut depth prediction, enabling improved drainage design, subgrade stabilization, and axle-load regulation in tropical pavement environments.