Effectiveness evaluation and influencing factor analysis of pavement seal coat treatments using random forests.

• Study on effectiveness of seal coat pavement treatments and the influencing factors. • Performance based study aimed at evaluating effectiveness of seal coat treatments. • Random forests method employed to investigate influencing factors. • Influencing factors in construction, environment, and existing pavement conditions. Seal coat or chip seal treatment is a cost-effective preventive pavement maintenance technique frequently used by highway agencies to extend asphalt pavement service life. Many studies were conducted to understand the best practice for a seal coat treatment. However, a lot of the studies focused on the material side and in the laboratory test stage, while there is a great need for field performance based studies to evaluate the effectiveness of seal coat treatments. The performance of seal coats depends on construction techniques, adhesive characteristics of asphalt emulsion, and properties of aggregates, and hence, many factors during the seal coat construction can impact the effectiveness of a seal coat treatment. The main purpose of this study is to evaluate the effectiveness of seal coat treatments and analyze the effects of influencing factors on the seal coat construction. The data employed are 28 pavement sections in Texas with relatively complete records of pavement performance from 2000 to 2016, which are retrieved from the Long-Term Pavement Performance (LTPP) database. Statistical analyses and visualization methods are applied to compare the performance before and after the treatment of a seal coat. Random forests are conducted to identify the effects of selected variables using an importance score indicator. The results show that seal coat treatments have an effect on reducing the cracking of pavement surfaces. The average air temperature and the shortwave surface value are important to the cracking performance. The pavement condition and seal coat thickness are vital to the rutting and IRI performance. [ABSTRACT FROM AUTHOR]