The susceptibility measures the probability of occurrence of a phenomenon under study. It estimates how much the predisposing factors affect past phenomena in order to predict future scenarios. Numerous tools and software are available for susceptibility analysis. Some of them are focused on spatial susceptibility or susceptibility zoning and are developed on both proprietary and open-source Geographical Information Systems, such as: GRASS r.landslide (Bragagnolo et al., 2020), ArcGIS ArcSDM (Kemp et al. 2002) and Toolboxes (Jebur et al., 2015), R LAND-SE (Rossi & Reichenbach, 2016), R RSAGA (Brenning et al., 2018). In this work, a new open-source tool for susceptibility zoning has been implemented for the QGIS software, named Susceptibility Zoning plugin (SZ-plugin). It allows to map the susceptibility of a specific object of study in the selected area. In particular, the plugin has been developed for landslide susceptibility mapping in the context of the project Silk Road Disaster Risk Reduction of the Belt and Road Initiative (Lei et al., 2018) to prevent and mitigate the risk induced by landslides on the infrastructure. The code has been written in Python3 using third-party libraries and plugins such us: scikit-learn (Pedregosa et al., 2011), NumPy, Matplotlib (Caswell et al., 2020 ; Hunter, 2007), GDAL, QGIS Plugin Builder, Plotly and Pandas. The plugin version 1.0 maps the susceptibility through severalstatistical methods: Weight of Evidence, Frequency Ratio, Logistic Regression, Decision Tree, Support Vector Machine, Random Forest. The analysis provides a measure of the probability of occurrence expressed by the Susceptibility Index (SI). Moreover, the SZ-plugin allows to evaluate the prediction capacity of the model by the representation of the Receiver Operating Characteristic (ROC) curves based on the training and validating datasets respectively and to cross-validate the result by simple random selection or k-fold method. The result is an efficient tool which supports the susceptibility zoning from the spatial analysis to the validation of the method and results. The code repository, further versions and upgrades are available on GitHub: https://github.com/CNR-IRPI-Padova/SZ .
CNR-IRPI-Padova/SZ: SZ plugin / Giacomo Titti; Alessandro Sarretta; Luigi Lombardo. - (2021).
CNR-IRPI-Padova/SZ: SZ plugin
Giacomo Titti;
2021
Abstract
The susceptibility measures the probability of occurrence of a phenomenon under study. It estimates how much the predisposing factors affect past phenomena in order to predict future scenarios. Numerous tools and software are available for susceptibility analysis. Some of them are focused on spatial susceptibility or susceptibility zoning and are developed on both proprietary and open-source Geographical Information Systems, such as: GRASS r.landslide (Bragagnolo et al., 2020), ArcGIS ArcSDM (Kemp et al. 2002) and Toolboxes (Jebur et al., 2015), R LAND-SE (Rossi & Reichenbach, 2016), R RSAGA (Brenning et al., 2018). In this work, a new open-source tool for susceptibility zoning has been implemented for the QGIS software, named Susceptibility Zoning plugin (SZ-plugin). It allows to map the susceptibility of a specific object of study in the selected area. In particular, the plugin has been developed for landslide susceptibility mapping in the context of the project Silk Road Disaster Risk Reduction of the Belt and Road Initiative (Lei et al., 2018) to prevent and mitigate the risk induced by landslides on the infrastructure. The code has been written in Python3 using third-party libraries and plugins such us: scikit-learn (Pedregosa et al., 2011), NumPy, Matplotlib (Caswell et al., 2020 ; Hunter, 2007), GDAL, QGIS Plugin Builder, Plotly and Pandas. The plugin version 1.0 maps the susceptibility through severalstatistical methods: Weight of Evidence, Frequency Ratio, Logistic Regression, Decision Tree, Support Vector Machine, Random Forest. The analysis provides a measure of the probability of occurrence expressed by the Susceptibility Index (SI). Moreover, the SZ-plugin allows to evaluate the prediction capacity of the model by the representation of the Receiver Operating Characteristic (ROC) curves based on the training and validating datasets respectively and to cross-validate the result by simple random selection or k-fold method. The result is an efficient tool which supports the susceptibility zoning from the spatial analysis to the validation of the method and results. The code repository, further versions and upgrades are available on GitHub: https://github.com/CNR-IRPI-Padova/SZ .I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
