This report details findings from the New Zealand Long-Term Pavement Performance Programme that aims at the development of deterioration models. Earlier work, completed between 2005 and 2007, resulted in prediction models mostly for thin, flexible chipsealed pavements. This report documents the development of models for dense-graded and porous asphalt surfaces (OGPA). The research was successful in developing pavement deterioration models for crack initiation of densegraded asphalt surfaces, and ravelling initiation for OGPA surfaces, and has confirmed the validity of the rutting model that was developed earlier. For both the crack and ravelling initiation models, continuous probabilistic models were developed that predict the probability of the defect to occur. These models use data that is readily available on network level databases, and can therefore be applied on asset management applications such as the New Zealand dTIMS system. The models were tested on the network data and had a significant success rate (up to 75 percent) in predicting the behaviour of the surfaces. Based on this finding, it is recommended that the models are adopted within the New Zealand dTIMS system. Further work required includes refining the models on the basis of individual sections’ LTPP data. In addition, this research has highlighted a number of practical aspects that require further investigation, and the need for the development of maintenance best-practice guidelines.