A data-mining approach towards damage modelling for El Niño events in Peru

Abstract

Compound natural hazards like El Niño events cause high damage to society, which to manage requires reliable risk assessments. Damage modelling is a prerequisite for quantitative risk estimations, yet many procedures still rely on expert knowledge, and empirical studies investigating damage from compound natural hazards hardly exist. A nationwide building survey in Peru after the El Niño event 2017which caused intense rainfall, ponding water, flash floods and landslidesenables us to apply data-mining methods for statistical groundwork, using explanatory features generated from remote sensing products and open data. We separate regions of different dominant characteristics through unsupervised clustering, and investigate feature importance rankings for classifying damage via supervised machine learning. Besides the expected effect of precipitation, the classification algorithms select the topographic wetness index as most important feature, especially in low elevation areas. The slope length and steepness factor ranks high for mountains and canyons. Partial dependence plots further hint at amplified vulnerability in rural areas. An example of an empirical damage probability map, developed with a random forest model, is provided to demonstrate the technical feasibility.

Community Perception and Communication of Volcanic Risk from the Cotopaxi Volcano in Latacunga, Ecuador

“…The iteratively constructed RIESGOS demonstrator for a multi-risk information system is based on a modular and scalable concept in which the different hazards, the related exposure models, and vulnerability schemas are each represented by one individual web service. These independent and distributed web-services (managed and maintained by individual research institutions) are based on the quantitative methodologies developed within the RIESGOS framework for multi-risk analysis (i.e., [69,[89][90][91][92][93][94]). Therefore, their integration into the RIESGOS demonstrator simulates the multi-risk environment of Latacunga.…”

Uncertainty Reduction of Unlabeled Features in Landslide Inventory Using Machine Learning t-SNE Clustering and Data Mining Apriori Association Rule Algorithms

“…To achieve such spatial pattern visualization, dimensionality reduction machine learning, with a t-distributed stochastic neighbor embedding (t-SNE) algorithm, offers a promising and practical capacity to analyze the structure of unlabeled slopes, leading to pattern identification and matching to known locations with similar formations [5]. t-SNE has been used to visualize high-dimensional datasets, such as remote sensing products and open data, by the clustered embedding of high-dimensional data into lower-dimensional space, such as a two-or three-dimensional map [6][7][8]. In addition, t-SNE is regarded as a very efficient technique for detecting potential errors in a reference dataset through visual analysis of a t-SNE plot [9].…”

“…Moreover, t-SNE returns highly compressed data, making it suitable for identification of large margins within a dataset. However, t-SNE may be unsuitable for datasets with recurring step-like temporal profiles [6].…”