The exponential relationships reported in the plots may be used to convert the dielectric values to air void values as prescribed in previous studies [1–3]. The AC pavement composite permittivity reduces, along with the reflection coefficient, as the volumetric proportion of air increases as compared to the remaining components. However, the method relies on an empirical fit, determined on a case-by-case basis, since the permittivity of the remaining components depends on the mix design (aggregate type, binder content, etc.). Long term studies in Finland concluded that this empirical fit is an exponential relationship [1]. The exponential fits, using a sufficient amount of cores, can be used to map the air void content variation in a similar manner to the dielectric maps shown in Fig. 4. Only 4 cores were feasible due to various factors involved with testing the final lift of an in-service pavement. More cores are needed for more stable exponential coefficients, although the limited cores show that the predicted relationships are similar for the measured dielectric range in this case-study. Since both regressions predict air void content at a maximum difference of 0.56%, which is within the uncertainty of the core measurement precision of 0.7%, use of either the initial or repeat run regression predictions are appropriate.
