<abstract xmlns="http://www.w3.org/1999/xhtml"><p>The purpose of the present study was to compare the prediction performances of three statistical methods, namely, information value (IV), weight of evidence (WoE) and frequency ratio (FR), for landslide susceptibility mapping (LSM) at the east of Constantine region. A detailed landslide inventory of the study area with a total of 81 landslide locations was compiled from aerial photographs, satellite images and field surveys. This landslide inventory was randomly split into a testing dataset (70%) for training the models, and the remaining (30%) was used for validation purpose. Nine landslide-related factors such as slope gradient, slope aspect, elevation, distance to streams, lithology, distance to lineaments, precipitation, Normalized Difference Vegetation Index (NDVI) and stream density were used in the landslide susceptibility analyses. The inventory was adopted to analyse the spatial relationship between these landslide factors and landslide occurrences. Based on IV, WoE and FR approaches, three landslide susceptibility zonation maps were categorized, namely, “very high, high, moderate, low, and very low”. The results were compared and validated by computing area under Road the receiver operating characteristic (ROC) curve (AUC). From the statistics, it is noted that prediction scores of the FR, IV and WoE models are relatively similar with 73.32%, 73.95% and 79.07%, respectively. However, the map, obtained using the WoE technique, was experienced to be more suitable for the study area. Based on the results, the produced LSM can serve as a reference for planning and decision-making regarding the general use of the land.</p></abstract>

The purpose of the present study was to compare the prediction performances of three statistical methods, namely, information value (IV), weight of evidence (WoE) and frequency ratio (FR), for landslide susceptibility mapping (LSM) at the east of Constantine region. A detailed landslide inventory of the study area with a total of 81 landslide locations was compiled from aerial photographs, satellite images and field surveys. This landslide inventory was randomly split into a testing dataset (70%) for training the models, and the remaining (30%) was used for validation purpose. Nine landslide-related factors such as slope gradient, slope aspect, elevation, distance to streams, lithology, distance to lineaments, precipitation, Normalized Difference Vegetation Index (NDVI) and stream density were used in the landslide susceptibility analyses. The inventory was adopted to analyse the spatial relationship between these landslide factors and landslide occurrences. Based on IV, WoE and FR approaches, three landslide susceptibility zonation maps were categorized, namely, “very high, high, moderate, low, and very low”. The results were compared and validated by computing area under Road the receiver operating characteristic (ROC) curve (AUC). From the statistics, it is noted that prediction scores of the FR, IV and WoE models are relatively similar with 73.32%, 73.95% and 79.07%, respectively. However, the map, obtained using the WoE technique, was experienced to be more suitable for the study area. Based on the results, the produced LSM can serve as a reference for planning and decision-making regarding the general use of the land.

Landslide Susceptibility Assessment in Constantine Region (NE Algeria) By Means of Statistical Models

Studia Geotechnica et Mechanica

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

The purpose of the present study was to compare the prediction performances of three statistical methods, namely, information value (IV), weight of evidence...

Landslide Susceptibility Assessment in Constantine Region (NE Algeria) By Means of Statistical Models

Article Category: Research Article

Online veröffentlicht: 21. Dez. 2018

Seitenbereich: 208 - 219

Eingereicht: 01. Juni 2018

Akzeptiert: 08. Okt. 2018

DOI: https://doi.org/10.2478/sgem-2018-0024

Schlüsselwörter
geographic information system, probabilistic methods, information value, weight of evidence, frequency ratio

© 2018 Nabil Manchar, Chaouki Benabbas, Riheb Hadji, Foued Bouaicha, Florina Grecu published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Landslide Susceptibility Assessment in Constantine Region (NE Algeria) By Means of Statistical Models

Article Category: Research Article

Online veröffentlicht: 21. Dez. 2018

Seitenbereich: 208 - 219

Eingereicht: 01. Juni 2018

Akzeptiert: 08. Okt. 2018

DOI: https://doi.org/10.2478/sgem-2018-0024

Schlüsselwörtergeographic information system, probabilistic methods, information value, weight of evidence, frequency ratio

© 2018 Nabil Manchar, Chaouki Benabbas, Riheb Hadji, Foued Bouaicha, Florina Grecu published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Schlüsselwörter
geographic information system, probabilistic methods, information value, weight of evidence, frequency ratio