A study on the correlation between soil properties and subgrade stiffness using the long-term pavement performance data

Abstract

This paper presents the statistical analysis results for correlating the stiffness characteristics of subgrade with subgrade soil properties using the long-term pavement performance database in the USA. The stiffness of subgrade soils was characterised by the falling weight deflectometer deflection basin parameters and subgrade modulus calculated using the backcalculation technique. For the statistical approach, both linear fixed- and mixed-effects models were applied to estimate the correlation between subgrade stiffness and soil properties. It is found from this study that the base curvature index and subgrade modulus are highly correlated with subgrade soil properties and variation of water content. The results also indicate that the mixed-effect model can accurately estimate the stiffness of subgrade soils using the soil properties and water content. Region-specific factors considered as the sources of random effects in the mixed-effects model can help to improve the degree of correlation.