Landslide susceptibility mapping using GIS-based statistical and machine learning modeling in the city of Sidi Abdellah, Northern Algeria

Landslides are one of the most common and damaging geological hazards that constrain the urban planning and development of many cities in northern Algeria. Therefore, landslide susceptibility maps (LSMs) constitute an essential tool for effective hazard management and long-term development planning...

Full description

Saved in:
Bibliographic Details
Published in:Modeling earth systems and environment Vol. 9; no. 2; pp. 2477 - 2500
Main Authors: Hamid, Bourenane, Massinissa, Braham, Nabila, Guessoum
Format: Journal Article
Language:English
Published: Cham Springer International Publishing 01.06.2023
Springer Nature B.V
Subjects:
Accuracy
Chemistry and Earth Sciences
Computer Science
Density
Distance
Earth and Environmental Science
Earth Sciences
Earth System Sciences
Ecosystems
Environment
Field investigations
Field tests
Geographic information systems
Geographical information systems
Geological hazards
Geology
Land use
Land use management
Land use planning
Landslides
Landslides & mudslides
Learning algorithms
Lithology
Machine learning
Mapping
Math. Appl. in Environmental Science
Mathematical Applications in the Physical Sciences
Mathematical models
Original Article
Physics
Regression analysis
Regression models
Risk management
Risk reduction
Satellite imagery
Seismicity
Statistical analysis
Statistical models
Statistics for Engineering
Susceptibility
Thematic mapping
Urban areas
Urban planning
Algeria
Validation
GIS
Susceptibility map
Machine learning model
Statistical model
Landslide
Algeria
ISSN:2363-6203, 2363-6211
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Landslides are one of the most common and damaging geological hazards that constrain the urban planning and development of many cities in northern Algeria. Therefore, landslide susceptibility maps (LSMs) constitute an essential tool for effective hazard management and long-term development planning in landslide-prone areas. The aim of this work is to prepare and compare the LSMs by applying GIS - based statistical and machine learning models for the new city of Sidi Abdellah and surrounding areas (Northern Algeria). We implemented the statistical models of the frequency ratio (FR), statistical index (SI), and weights of evidence (WoE) models, and the machine learning models represented by a logistic regression (LR) model for landslide susceptibility prediction. An historical landslide inventory map was produced using the interpretation of Google Earth satellite images, available historical records, and geological field investigations. The obtained landslides were randomly divided into the training (70%) and validation (30%) processes. Furthermore, 12 influencing factors for landslide occurrence (including precipitation, slope, elevation, distance to drainage, aspect, land use, density of streams, distance to road, lithology, distance to fault, seismicity, and density of roads) were selected to prepare thematic maps and were considered for susceptibility analysis. Subsequently, landslide susceptibility assessment and mapping are performed by considering the inventoried landslide events and their related predisposing factors using LR, SI, WoE, and FR models in GIS. The accuracy of the four models was verified, validated, and compared using the area under curve (AUC) value of the Receiver Operating Characteristics Curves (ROC) method. The validation results showed that all the used statistical models provided a good accuracy in predicting landslide susceptibility than the machine learning models, with the SI model having a higher AUC value of 80.1% than the WoE (78.2%), FR (78.2%), and LR (64.2%) models. Based on these results, we conclude that the established maps can be used as useful tools for land use planning and risk reduction in the urban area of Sidi Abdellah.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:2363-6203
2363-6211
DOI:10.1007/s40808-022-01633-x