A comparison between three approaches for modelling deterioration of five pavement surfaces

Reported in this paper are the process and outcomes of deterioration modelling for five types of pavement bituminous surfacing for use in identifying and prioritising resurfacing needs. Surface condition data is collected using subjective assessment surveys and the condition ratings of distresses are combined into a single measure referred to as surface inspection rating (SIR). Deterioration models of SIR as a function of surface age are developed for the five networks using three modelling approaches. They include regression analysis, logistic regression analysis and Markov chains. A comparison between the output models of the three approaches is presented and assessed in terms of predictions reasonableness and accuracy considering their closeness to actual data and practical local experience and engineering judgement. Overall the findings indicate that the three modelling approaches provide similar predictions and deterioration rates for most surfacing types. A comparison of performance between the five networks is also presented.