This paper studies the long term structural behaviour of a Victorian railway viaduct under train loading and temperature variation. A multi-sensing, self-sustaining and remotely controlled data acquisition system combines fibre Bragg grating strain sensors with acoustic emission sensors for the study of both global dynamic deformation and local masonry deterioration. A statistical analysis of fibre Bragg grating signals reveals regions with permanent change in the dynamic deformation of the bridge over the last two years, whereas in other locations the deformation follows a seasonal cyclic pattern. In order to decouple changes in structural behaviour due to real mechanical damage from normal seasonal effect, the paper studies the ambient temperature effect on the dynamic deformation of the bridge, showing a clear linear dependence. In particular, when temperature increases, the dynamic strain due to train loading decreases uniformly in the longitudinal direction. In the transverse direction, where the thermal expansion is not constrained, the decrease is smaller. Decoupling damage from normal seasonal effect is of critical importance for the development of reliable early warning structural alert systems for infrastructure networks. The paper further studies local masonry deterioration at four critical location by combining data from the two sensing technologies: fibre optic and acoustic emission sensors.