Conference Agenda

Monitoring the long-term performance of blue-green infrastructure (BGI) is underexplored in the published literature. This oversight risks overestimating its benefits in the long run and ignoring its maintenance requirements. To address this gap, a workflow is proposed to investigate temporal dynamics in BGI hydrological performance. The workflow utilises Time Series Data Mining (TSDM), combining clustering, segmentation, and dynamic time warping techniques with statistical tests. This workflow first identifies pairs of statistically similar rainfall events from different periods. Then, the corresponding BGI underdrain flows for these similar rainfall event pairs are analysed using the volume-based underdrain flow-to-rainfall ratio (UF-P ratio) and time-based metrics, which include centroid delay, cross-correlation, and DTW. When using the workflow in a green roof in Oslo, only six rainfall pairs were found to be statistically similar. The UF-P ratio for these pairs increased on average from 0% to around 6.6% over time, albeit with high variability (SD > 16%); the changes over time were not statistically significant. Although the change in the UF-P ratio was not statistically significant due to the small sample size, the results suggest performance drifts over time. This result also highlights the need to monitor BGI to track its hydrological performance.