Abstract The base course is the principal structural layer of flexible pavements, and its accurate characterization is vital for predicting pavement response and long-term performance. In tropical regions, where temperature, moisture, and material variability differ significantly from temperate design conditions, existing empirical and mechanistic–empirical models often fail to represent in-situ behaviour. This study compares the performance of the stress-dependent AUSTROADS model and the Smith–Witczak (SW) equivalent-modulus model using field data from twelve flexible pavement sections in Nigeria. The influence of mechanistic parameters on pavement strain and deformation behaviour was also evaluated through global sensitivity and uncertainty analyses. Measured tensile strain at the asphalt bottom (εₜ), compressive strain at the subgrade top (εᵥ), and Falling Weight Deflectometer (FWD) back-calculated moduli (E2et, E2ev) were used to assess model fidelity. Results show that the SW model reproduced field strains with higher accuracy and lower computational demand than AUSTROADS. The latter consistently underestimated εₜ and overestimated εᵥ, whereas SW predictions closely matched measured responses, showing no significant difference at p<0.05. Discrepancies between predicted and observed moduli were primarily governed by the regression slope constant K₂, reflecting stress sensitivity in granular bases. The importance of incorporating K₂ as a variable was evidenced in the pronounced effects of K2 on E2et and E2ev with correlation coefficients of 0.95 and 0.94, respectively. The refined SW model provides a rational, field-calibrated, and environmentally responsive framework for characterizing tropical pavements.
Owolabi et al. (Sun,) studied this question.