Estimation of Landslide Volume by Machine Learning and Remote Sensing Techniques in Himalayan Regions

Aug 2, 2026

Upload by Skyler

Topographical and geological conditions are widely regarded as key drivers of landslides, yet estimating landslide volume on rock slopes via empirical equations remains difficult. A data-science workflow is developed using machine learning to improve volume estimation accuracy and reliability. XGBoost is applied to predict potential landslide volume in Gyirong, China, based on combined geomorphic (slope unit area and mean elevation) and geological (fault density and geological index) factors. Model performance is benchmarked against GBDT, AdaBoost, and random forest using MAPE and R-squared. Results show XGBoost yields the highest accuracy (R-squared 0.986) and lowest error (MAPE 8.19%), outperforming empirical models. Accurate remote-area volume estimation supports disaster management and reduces human and infrastructure losses.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages16

Reading Time40 min

AccessFree to download

Introduction

Challenges in landslide volume estimation

Role of susceptibility maps and empirical models

Machine learning approaches for landslide analysis

Landslides DOI 10. 1007/s10346-025-02532-9 Received: 12 September 2024 Accepted: 24 April 2025 © The Author(s) 2025 Chongcai Xu · Congchao Bian · Teng Yu · Chenchen Qiu Estimation of landslide volume by machine learning and remote sensing techniques in Himalayan regions Abstract Topographical and geological conditions are typically regarded as the primary causes of landslides. However, accurately estimating landslide volumes on rock slopes using empirical equations remains challenging. In contrast, data science approaches, such as machine learning, leverage advanced data integration and processing capabilities, significantly enhancing the accuracy and reliability of landslide volume estimations. As such, an resemble method, XGBoost, was chosen in our study to estimate the potential landslide volume in Gyirong, China. A factor combination was proposed in this study. They are related to geomorphic (area of slope units (S) and mean elevation of slope units (El)) and geological (faults density (Fd) and geological index (GI)) conditions. The performance of the developed model was compared with three other machine learning models, including gradient boosting (GBDT), adaptive boosting (AdaBoost) and random forest (RF) based on the mean absolute percentage error (MAPE) and determination of coefficient (R-squared) . The results demonstrate that XGBoost achieves the highest prediction accuracy, with an R-squared value of 0.986, and a mean absolute percentage error (MAPE) reduces to 8.19%. Additionally, the prediction outcomes ofthe machine learning model, using the proposed factor combination, were compared with several empirical models. Once again, XGBoost model exhibits the lowest error relative to measured values, highlighting the superiority of machine learning in landslide volume prediction and validating the effectiveness of the selected factors. Overall, the accurate estimation of landslide volume in remote areas can benefit the disaster management and decrease losses of human lives and properties. Keywords Landslide volume estimation · Rock slopes · Machine learning model · Satellite images · Lithology Introduction Landslides can cause severe damage to infrastructure and loss of human lives due to the movement of vast quantities of rock down slopes under the influence of gravity (Lacasse et al. 2009; Zhao et al. 2024) . Furthermore, the accumulation of crushed rock on slopes serves as a potential source of material for future debris flows (Blahut et al. 2010; Qiu et al. 2024), leading to secondary damage to infrastructure and further degradation of the ecological environment. (Petrakov et al. 2007; Jomelli et al. 2015; Perov et al. 2017) . It was reported from the Center for Research on Epidemiology of Disasters (CRED) that, all over the world, almost 17% of fatalities caused by natural disasters was caused by landslide. The mountainous areas are taking 36% of the earth land and over 10% of population that are still living in these areas (Gerrard 1990) . It is still a challenge to provide an early warning for landslides to protect properties and lives and mitigate the loss (Singhroy 2009) . This is because landslide warning normally requires a better understanding of the rock types and orientation of bed rocks (Guzzetti et al. 2009; Gaziev 2013). Furthermore, climate change in the past several decades (IPCC, 2013) leads to an increasing numbers of extreme weather event (O’Gorman 2015; Qiu et al. 2022), which also increased the occurrence frequency of landslide events (Kirschbaum and Adler 2013) . Accurate susceptibility maps of landslide can be helpful to provide timely warning for local people and therefore reduce casualties (Stanley and Kirschbaum 2017; Chen and Li 2020). The prediction of landslide volume is a crucial component and an essential supplement to landslide risk assessment (Lacasse et al. 2009) . However, accurately determining landslide volume is challenging, as it is closely related to the i

FAQs

Why is estimating landslide volume on rock slopes challenging?

Accurate volume estimation is difficult because landslide volume is tightly linked to the internal structure and geomorphic conditions of rock slopes, and empirical equations have limitations when failure mechanisms differ.

What machine learning method is used to estimate landslide volume in Gyirong, China?

The study uses XGBoost to estimate potential landslide volume in Gyirong, China.

How does XGBoost perform compared with other machine learning and empirical models?

XGBoost achieves the highest prediction accuracy with R-squared of 0.986 and the lowest error, with MAPE reduced to 8.19%, outperforming other tested machine learning models and showing lower relative error than empirical models.

Aug 2, 2026

Upload by Skyler