Geoprocessing functions available in several GIS-based software, like ArcGIS, also allow the application of several statistical quantitative techniques described above for landslide susceptibility evaluation however, for the application of physically based models or the use of AI, most previous research shows the realisation of hybrid platforms, which are often very complex. GIS is a milestone in environmental management for natural hazards, disasters, and global climate change it allows management of geodata, even big volumes, and displays them spatially. Ĭonditioning factors are derived from geographical data (geodata) their dissemination in digital format has been favoured thanks to Geographical Information System (GIS) technology. Recently, methods based on Artificial Intelligence (AI), and specifically Machine Learning, are wide-spread examples of Machine Learning techniques are Artificial Neural Networks, the Kernel-based Support Vector Machine (SVM), Tree-based decision trees, or Fuzzy Clustering. These factors are analysed in light of several techniques, grouped into statistical, such as bivariate (Frequency Ratio or Weight of Evidence) and multivariate (Logistic Regression or Discriminant method), or deterministic, such as empirical or physically based. Quantitative techniques rationalise the process, carrying out a numerical evaluation through different methodologies, all united by the analysis of factors responsible for triggering the phenomenon, named “conditioning factors” they usually include lithology, aspect, and land use. Qualitative techniques rely on the experience and judgement of experts and are based on the analysis of quasi-static variables they are considered subjective. Methodologies for the assessment of landslide susceptibility can be classified into qualitative and quantitative, both of which aim to define the susceptibility as low, moderate, or high. The paper sets the future development of MATLAB as a fully implemented platform for landslide susceptibility, based on any available methods. Moreover, it is discussed how raster resolution affects the processing time. An application of these preliminary operations to a study area affected by shallow landslides in the past is shown results show how geodata can be managed as easily as in GIS, as well as being displayed in a fashionable way too. Specifically, it is discussed how to build matrices of parameters, needed to assess susceptibility, by using grid cell mapping units, and mapping them bypassing GIS. This paper describes how MATLAB can be used to derive the most common landslide conditioning factors, by managing the geographic data in their typical formats: raster, vector or point data. GIS-based risk management platforms are thus sometimes hybrid, requiring relatively complex adaptive procedures before exchanging data among different environments. According to the adopted methodology, after an initial phase conducted on the GIS platform, data need to be transferred to specific software, e.g., MATLAB, for analysis and elaboration. These are usually derived from geographic data commonly handled through Geographical Information System (GIS) technology. Most of the methods for landslide susceptibility assessment are based on mathematical relationships established between factors responsible for the triggering of the phenomenon, named the conditioning factors.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |