A prototype landslide early warning system in Rize (Turkey): analyzing recent impacts to design a safer future

Abstract

In July 2021, Rize province (Turkey) was struck by two intense rainstorms that caused two widespread landslide disasters with a short turnaround between them. Dozens of landslides were triggered, resulting in casualties, damages, and interruption of services. The objective of this technical note is to investigate if the knowledge and technical level available before 2021 was enough to calibrate an effective landslide early warning system (LEWS) and to propose a prototypal version for future landslide risk management in the province. A landslide dataset for the period 1990-2020 was compiled, coupled with the available rainfall records, and used to define a LEWS composed of a forecasting model based on I-D rainfall thresholds and a newly developed warning model calibrated on the severity of the past events. Afterwards, the year 2021 events were used to test the model, highlighting points of strength and weaknesses of the proposed prototype: the prototype would have been an effective tool to assist the management of landslide risk during the recent disastrous events, and it is also expected to be useful in the future, as climate change projections highlight that similar meteorological events are likely to happen again, with a slightly higher frequency. The prototype could be implemented in a short time and a roadmap is proposed to keep the system running while gradually updating it with more complex features. For instance, the rainfall forecasts segment of the system is identified as one of the priorities to be improved. The lesson learnt in this case study demonstrates the effectiveness of purposely calibrated warning systems and their utility in risk management procedures.