Towards landslide space-time forecasting through machine learning - the influence of rainfall parameters and model setting

Aug 3, 2026

Landslide susceptibility maps provide only spatial predictions because they are derived from static predisposing factors, leaving temporal assessment insufficient. This research addresses the gap by combining a static susceptibility index with dynamic rainfall variables—seasonality and cumulative rainfall over multiple reference periods—inside a Random Forest framework. Model behavior is interpreted using Out-of-Bag Error and Partial Dependence Plots to evaluate variable importance and consistency with shallow landslide triggering mechanisms, especially short intense rainfall.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages20

Reading Time50 min

AccessFree to download

Introduction

Motivation and problem background

Forecasting for early warning and rainfall thresholds

Study contribution and research goals

TYPE Original Research PUBLISHED 13 April 2023 DOI 10.3389/feart.2023.1152130 OPEN ACCESS EDITED BY Hyuck-Jin Park, Sejong University, Republic of Korea REVIEWED BY Yi Wang, China University of Geosciences Wuhan, China Davide Tiranti, Agenzia Regionale per la Protezione Ambientale del Piemonte (Arpa Piemonte), Italy *CORRESPONDENCE Nicola Nocentini, nicola. nocentini@uniﬁ . it RECEIVED 27 January 2023 ACCEPTED 04 April 2023 PUBLISHED 13 April 2023 CITATION Nocentini N, Rosi A, Segoni S and Fanti R (2023), Towards landslide space-time forecasting through machine learning: the inﬂuence of rainfall parameters and model setting. Front. Earth Sci. 11:1152130 . doi: 10.3389/feart.2023.1152130 COPYRIGHT © 2023 Nocentini, Rosi, Segoni and Fanti. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Frontiers in Earth Science Towards landslide space-time forecasting through machine learning: the inﬂuence of rainfall parameters and model setting Nicola Nocentini 1*, Ascanio Rosi 2, Samuele Segoni 1 and Riccardo Fanti 1 1Department of Earth Sciences, University of Florence, Florence, Italy, 2Department of Geosciences, University of Padua, Padua, Italy Landslide susceptibility assessment using machine learning models is a popular and consolidated approach worldwide. The main constraint of susceptibility maps is that they are not adequate for temporal assessments: they are generated from static predisposing factors, allowing only a spatial prediction of landslides. Recently, some methodologies have been proposed to provide spatiotemporal landslides prediction starting from machine learning algorithms (e.g., combining susceptibility maps with rainfall thresholds), but the attempt to obtain a dynamic landslide probability map directly by applying machine learning models is still in the preliminary phase. This work provides a contribution to ﬁx this gap, combining in a Random Forest (RF) algorithm a static indicator of the spatial probability of landslide occurrence (i.e., a classical susceptibility index) and a number of dynamic variables (i.e., seasonality and the rainfall amount cumulated over different reference periods) . The RF implementation used in this work allows the calculation of the Out-of-Bag Error and depicts Partial Dependence Plots, two indices that were used to quantify the variables’ importance and to comprehend if the model outcomes are consistent with the triggering mechanism observed in the case of study (Metropolitan City of Florence, Italy) . The goal of this research isnot to set up a landslide probability map, but to 1) understand how to populate training and test datasets with observations sampled over space and time, 2) assess which rainfall variables are statistically more relevant for the identiﬁcation of the time and location of landslides, and 3) test the dynamic application of RF in a forecasting model for the spatiotemporal prediction of landslides. The proposed dynamic methodology shows encouraging results, consistent with the actual knowledge of the physical mechanism of the triggering of shallow landslides (mainly inﬂuenced by short and intense rainfalls) and identiﬁes some benchmark conﬁgurations that represents a promising starting point for future regional-scale applications of machine learning models to dynamic landslide probability assessment and early warning. KEYWORDS machine learning, landslides, random forest, susceptibility, variables’ importance, landslide probability map, cumulative rainfall, dynamic analysis 01 [frontiersin.org](frontiersin.or

FAQs

Why are conventional landslide susceptibility maps limited for temporal forecasting?

They are built from static predisposing factors, which supports only spatial prediction and not temporal evolution of landslide probability.

What modeling approach does the research propose for space-time landslide forecasting?

It uses a Random Forest model that integrates a static susceptibility index with dynamic variables such as seasonality and cumulative rainfall over different reference periods.

How is the influence of rainfall-related inputs assessed in the study?

The work applies Out-of-Bag Error and Partial Dependence Plots to quantify variables’ importance and interpret whether model outputs match the triggering mechanism observed in the case study.

Aug 3, 2026

Upload by Olivia Brown