Mimic expert judgement through automated procedure for selecting rainfall events responsible for shallow landslide: A statistical approach to validation. (January 2016)

• Record Type: Journal Article
• Authors: Vessia, Giovanna; Pisano, Luca; Vennari, Carmela; Rossi, Mauro; Parise, Mario
• Is Part Of: Computers & geosciences. Volume 86(2016)
• Page Start: 146
• Page End: 153
• Publication Date: 2016-01
• Subjects: Rainfall thresholds -- Shallow landslide -- Automated procedure -- Statistical tests -- LANDTRAIN code; Environmental policy -- Periodicals; 550.5
• DOI: 10.1016/j.cageo.2015.10.015
• Languages: English
• ISSNs: 0098-3004

Abstract

Abstract: This paper proposes an automated method for the selection of rainfall data (duration, D, and cumulated, E ), responsible for shallow landslide initiation. The method mimics an expert person identifying D and E from rainfall records through a manual procedure whose rules are applied according to her/his judgement. The comparison between the two methods is based on 300 D – E pairs drawn from temporal rainfall data series recorded in a 30 days time-lag before the landslide occurrence. Statistical tests, employed on D and E samples considered both paired and independent values to verify whether they belong to the same population, show that the automated procedure is able to replicate the expert pairs drawn by the expert judgment. Furthermore, a criterion based on cumulated distribution functions (CDFs) is proposed to select the most related D – E pairs to the expert one among the 6 drawn from the coded procedure for tracing the empirical rainfall threshold line. Highlights: A code to calculate duration and cumulated rainfall related to occurred landslides. Six duration and cumulated values are drawn by setting W and Δ T input variables. An automated method for non-expert users to undertake expert tasks. Automated and expert procedures are compared through statistical tests.