A high proportion of the rural arterial network in Victoria /Australia are low volume roads built of sealed granular pavements, which are important routes for freight movement between rural centres. Investigation into rehabilitation of these arterials is triggered when roughness reaches a certain threshold level. To assist road agencies in their long term planning, a project has been initiated to develop absolute deterministic deterioration models for these roads. A representative sample network of low volume arterial roads has been selected and all relevant data including pavement condition are collected. The network covers a representative range of traffic loading, subgrade reactivity level and environmental factors. For each highway section, raw longitudinal profile data from at least four years was used to determine roughness progression over time. All profile data was aligned and then cleaned and filtered to ensure that the same length of road profile was compared over time. To remove the influence of maintenance activities, only sections with positive progression was included. Multiple linear regression analysis was used to develop models for these sections to predict pavement roughness over time as a function of a number of contributing variables. The output of the analyses was used to evaluate the significance and contribution of the different factors including traffic and environmental conditions. This paper provides a description of data preparation and analysis. It was observed that higher traffic loading and soil reactivity, poor drainage and climates with high seasonal variation increase roughness progression rate.