4.5 Article

Machine learning for sinkhole risk mapping in Guidonia-Bagni di Tivoli plain (Rome), Italy

期刊

GEOCARTO INTERNATIONAL
卷 37, 期 27, 页码 16687-16715

出版社

TAYLOR & FRANCIS LTD
DOI: 10.1080/10106049.2022.2113455

关键词

Sinkhole; susceptibility risk; machine learning; maximum entropy algorithm; Guidonia-Tivoli plain

资金

  1. Ministry of the Ecological Transition

向作者/读者索取更多资源

This study presents a sinkhole susceptibility and risk assessment in Guidonia-Bagni di Tivoli plain, Italy, using a machine learning model. The results show that lithology, travertine thickness, groundwater, and land use are the main factors affecting sinkhole formation. The risk map indicates that 2% of the study area is at higher risk, especially in the main urban fabric. The study demonstrates the potential of machine learning models in predicting sinkhole areas and providing useful information for urban planning and geohazard risk management.
This work presents a sinkhole susceptibility and risk assessment mapping in Guidonia-Bagni di Tivoli plain (Italy), a travertine sinkhole-prone area where sudden occurrences of sinkholes have happened in past and recent times. We collected a point-like sinkhole inventory and we considered a series of different sinkhole-controlling and precursory factors over the study area, related to its geo-litho-hydrological setting and to its terrain deformational scenario, i.e. ground motion rates derived from InSAR COSMO-SkyMed imagery. A sinkhole susceptibility map was produced through a machine learning model, namely Maximum Entropy algorithm (MaxEnt). Results highlight that the most determining factors for sinkhole formation are the lithology, the travertine thickness, groundwater and the land use. The sinkhole susceptibility map was then combined with data on vulnerability and elements-at-risk economic exposure in order to provide a sinkhole risk map of the area. The outcomes show that areas at higher risk covers about 2% of the total study area and primarily relies on the zoning of the main urban fabric. In particular, it is worth to highlight that 5% of the whole road-network pavement and 27% of all the residential buildings fall into High and Very High risk classes. Overall, results of this work demonstrate capabilities of machine learning models to assess sinkhole susceptibility for predicting potential sinkhole areas, and provide a sinkhole risk map, along with information on urban environment, as a useful tool for urban planning and geohazard risk management.

作者

我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。

评论

主要评分

4.5
评分不足

次要评分

新颖性
-
重要性
-
科学严谨性
-
评价这篇论文

推荐

暂无数据
暂无数据