Interactive Web-Based Approach for Spatial Modeling of Landslide Susceptibility with R-Shiny
Abstract
The research explores the application of an interactive web-based prototype for modeling landslide susceptibility using GIS Free Open Source Software (FOSS). The combined approach links the shallow slope stability model (SHALSTAB) from the System for Automated Geoscientific Analyses Geographical Information System (SAGA GIS) with R-Shiny using R programming environment. The client-server architecture of the framework integrates multiple components where Nginx with R-Studio Server provide computing-intensive services which are delivered to the client (i.e., Chrome or Firefox). Specifically, the Shiny web app prototype consists of consecutive and easy-to-follow steps organized into modeling and validation pages. The modeling page is used for model development and represents a graphical interface for interactive parameter manipulation and running SHALSTAB model using a single upload of a georeferenced digital elevation model (DEM). In contrast, the validation page is used for model tuning and assessment that is accomplished by different measures of accuracy (i.e., F1-score, accuracy and kappa) derived from a confusion matrix associated with the binary classification of the landslides (i.e., presence, absence). The mapping results from the modeled susceptibility are generated by customizable widgets and Shiny components through interactive visualization, modeling, calibration, printed reports and file downloads for external software such as Google Earth Pro or SAGA GIS. The value of this approach is that it allows a real-time implementation using site-specific datasets that are applied to different areas of interest. In addition, the prototype expands its usability to a wider community (i.e., non-R, non-GIS users) such as a classroom teaching activity that can enable research-informed learning. The prototype with an example of datasets from the Clearwater National Forest, Idaho can be accessed through the website https://geogis.bgsu.edu/apps/shalstabcv.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2022
- Bibcode:
- 2022AGUFMNH25D0483G