Work

Potential and constraints of ERA5-driven dynamical downscaling to estimate landslide-triggering rainfall conditions, evaluated for the Marche region in Italy

Year: 2026
Type: article
Abstract: Study region Marche region, central Italy: a mountainous and hilly area between the Apennines and the Adriatic Sea, highly prone to rainfall-induced shallow landslides. Study focus This study investigates whether three convection-permitting dynamical downscaling products available for Italy of the ERA5 global reanalysis, namely MERIDA-HRES (grid spacing ~4 km), MOLOCH-MORE (~1.7 km), and VHR-REA (~2.2 km), can be used to reconstruct landslide-triggering rainfall conditions and derive rainfall thresholds. A catalogue of 360 rainfall-triggered shallow landslides recorded in the region between 2002 and 2024 is analysed, and triggering conditions are reconstructed from both a regional rain gauge network (145 stations) and the three products. New hydrological insights for the region Compared with rain gauges, all products underestimate cumulated rainfall and overestimate rainfall duration, leading to biased triggering conditions and lower thresholds, especially for short durations. The evaluation of two recent severe events reveals a marked contrast in performance: the intense, localized, September 2022 event is severely underestimated by all products, whereas the severe, widespread, May 2023 event is better reproduced. A further analysis confirms that this limitation is inherited from the driving ERA5 reanalysis. Overall, downscaled reanalyses show potential for regional landslide early warning, but they should be used as a complement to, rather than a replacement for, ground-based observations. Moreover, the rainfall thresholds cannot be transferred from gauge-based to model-based rainfall without careful recalibration.
Language: English
Open Access status: gold
Estimated APC paid: $2,600
OpenAlex ID: W7226036809
FWCI:
Cites: 52
Cited by:
Related to: 10