Building up a historical picture of the deterioration of the total population of defects on a bridge can help reduce the risk of failure and downtime, effectively optimise the use of maintenance resources and extend the lifecycle of components. Extensive long range ultrasonic (LRU) data was gathered for different temperatures on sound steel samples and for sound and defective aluminium samples across broad frequency sweeps to determine the effect of temperature on the interpretation of LRU data when included in a structural health monitoring (SHM) database. Granular parameter sets were statistically analysed to determine the influence of temperature on the wave form and the amplitude of LRU inspections. Sensitivity to temperature was modelled and validated to ameliorate or remove the impact of temperature on waveforms. This allowed the development of comparable SHM wave form data sets. However, amplitude variations were noted for ultrasonic testing (UT) data collected at different temperatures which challenges the use of amplitude-linked parameters as defect or deterioration recognition criteria. A neural network (NN) was trained to determine accurately at which temperature UT data was acquired. This classification NN was used to validate the inclusion of temperature adjusted datasets within the SHM database. The findings of how to reduce and correct the effect of fluctuating bridge temperature on the interpretation of LRU data were incorporated into a structural health monitoring system being developed for bridges by the Wi- Health Project (283283), a collaborative project, funded by the European Union under the FP7 framework. These findings will improve the effectiveness of defect and deterioration detection and will optimise the resources needed to maintain bridges; both from the perspective of the level of staffing needed and extending the lifecycle of bridge components by the prioritisation of the maintenance schedule.