Sinkholes are common phenomena in the world that occur as a result of collapse processes due to natural and/or anthropogenic causes. Sinkholes consist of three-dimensional funnel-shaped depressions, predominantly circular on the surface, deep from centimeters to several meters. Sinkholes in urban areas, also called “anthropogenic” sinkholes, can be very dangerous from an engineering point of view, causing instability or damaging buildings and infrastructures or even leading to the death of people. In Naples (Italy), the presence of a dense underground cavity network, generated as a result of ancient and historical quarrying of bedrock volcanic tuff (used as building material), promotes the generation of sinkholes occurrence. In this work, sinkhole susceptibility analysis was conducted for the production and the comparison of two different sinkhole susceptibility maps by means of statistical-based algorithms (Random Forest and Maximum Entropy). Twelve environmental variables have been used for the susceptibility assessment, such as groundwater depth, bedrock depth and maps of density and distance from different predisposing factors (aqueducts, roads, sewers, anthropic cavities and underground railroad networks). Both produced maps present good predictive performance and indicate a very high sinkhole susceptibility in the city center of Naples, in agreement with the high density of underground cavities, supporting the importance of the latter as predisposing factor. The road network, considered in this work as representative of secondary aqueduct and sewer systems generally located under such infrastructures, also appears to be an important variable. This study aims to represent a useful tool to improve the development of sinkhole susceptibility maps and to support the local government to protect its cultural heritage.

Comparison of two machine learning algorithms for anthropogenic sinkhole susceptibility assessment in the city of Naples (Italy) / Bausilio, G.; Annibali Corona, M.; Di Martire, D.; Di Napoli, M.; Francioni, M.; Guerriero, L.; Tufano, R.; Calcaterra, D.. - III:(2022), pp. 1112-1123. [10.1201/9781003308867-88]

Comparison of two machine learning algorithms for anthropogenic sinkhole susceptibility assessment in the city of Naples (Italy)

Bausilio G.
;
Annibali Corona M.;Di Martire D.;Guerriero L.;Tufano R.;Calcaterra D.
2022

Abstract

Sinkholes are common phenomena in the world that occur as a result of collapse processes due to natural and/or anthropogenic causes. Sinkholes consist of three-dimensional funnel-shaped depressions, predominantly circular on the surface, deep from centimeters to several meters. Sinkholes in urban areas, also called “anthropogenic” sinkholes, can be very dangerous from an engineering point of view, causing instability or damaging buildings and infrastructures or even leading to the death of people. In Naples (Italy), the presence of a dense underground cavity network, generated as a result of ancient and historical quarrying of bedrock volcanic tuff (used as building material), promotes the generation of sinkholes occurrence. In this work, sinkhole susceptibility analysis was conducted for the production and the comparison of two different sinkhole susceptibility maps by means of statistical-based algorithms (Random Forest and Maximum Entropy). Twelve environmental variables have been used for the susceptibility assessment, such as groundwater depth, bedrock depth and maps of density and distance from different predisposing factors (aqueducts, roads, sewers, anthropic cavities and underground railroad networks). Both produced maps present good predictive performance and indicate a very high sinkhole susceptibility in the city center of Naples, in agreement with the high density of underground cavities, supporting the importance of the latter as predisposing factor. The road network, considered in this work as representative of secondary aqueduct and sewer systems generally located under such infrastructures, also appears to be an important variable. This study aims to represent a useful tool to improve the development of sinkhole susceptibility maps and to support the local government to protect its cultural heritage.
2022
9781003308867
Comparison of two machine learning algorithms for anthropogenic sinkhole susceptibility assessment in the city of Naples (Italy) / Bausilio, G.; Annibali Corona, M.; Di Martire, D.; Di Napoli, M.; Francioni, M.; Guerriero, L.; Tufano, R.; Calcaterra, D.. - III:(2022), pp. 1112-1123. [10.1201/9781003308867-88]
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11588/889481
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact