The Relationship Between Initiation of Landslides and Rainfall Intensity-Duration Thresholds in South-East Queensland, Australia

Rainfall attributes to slope instability by increasing soil moisture, reducing soil's matric suction and elevating pore water pressure. As such, rainfall thresholds are often used to predict the likelihood of slope failures by establishing the minimum rainfall conditions or parameters necessary to initiate a landslide. However, the reliability and accuracy of thresholds need to be rigorously validated prior to adopting them in operational warning systems. This study aims to develop empirical rainfall thresholds for the initiation of shallow landslides in South-East Queensland (SEQ), Australia, where rainfall-induced sediment-related disasters occur annually. The current study examines 104 rainfall-induced shallow landslides that occurred during 1974-2018. The corresponding rainfall conditions were analysed objectively from rainfall data to derive the thresholds using the quantile regression method, separating rainfall events by the absence of rainfall for 24 h. The thresholds were determined for the different quantiles (i.e. 2nd, 10th, 50th and 90th), and the 2nd percentile quantile was considered the rainfall threshold for SEQ. In order to render comparable rainfall thresholds for various regions with normalised rainfall intensity, the study also proposed IMAP-D thresholds for SEQ, in terms of mean annual precipitation (MAP). Further, the developed I-D threshold is validated using the physical-based real-time monitoring system at Maleny, Queensland, Australia. The validation provides an optimal balance between the maximisation of accurate predictions and the minimisation of inaccurate predictions, fostering the operational use of validated rainfall thresholds in the operational early warning system for regional shallow landslide forecasting.