Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Research note: Integrating big data to predict tree root blockages across sewer networks
Highlights We modeled pipe-tree-environment systems to predict root blockages in pipes. Pipe and environmental conditions modulate tree root blockages. A few tree genera and species pose greater blockage risk under certain growth conditions. Our approach advances urban arboriculture and sewerage management.
Abstract Tree root blockages can occur when tree roots enter sewer pipes, resulting potentially in high economic and environmental costs. However, comprehensive understanding of which tree species and environmental factors (i.e., soil, hydrological and landscape properties) may most impact sewer blockages remains unclear. We integrated sewer pipe, tree and environmental data from 103 suburbs across seven local government areas to identify the most significant factors affecting root blockages along 902 km of sewer infrastructure in Sydney, Australia. Distributed Random Forest (DRF) models were used to predict 2,942 root blockages that occurred between 2010 and 2019 along 22,192 sewer pipe segments, and to relate these to 90,858 tree stems of 651 tree species, while accounting for other pipe and environmental variables. We found that tree root blockages were clustered across the urban landscape, creating “blockage hotspots”. Tree stem abundance was positively correlated with number of blockages. We found that urban morphology and pipe characteristics were more important than tree characteristics (for example stem size, useful life expectancy) in explaining root blockages. We also found that the number of tree stems and the abundance of the Chinese banyan (Ficus macrocarpa) were the most important variables predicting frequency of root blockages per unit pipe length. The most important landscape variables were slope, aspect, soil salinity and available water content, while most pipe-related variables also explained significant deviance in the DRF model. This study is among the first to mine large datasets representing urban infrastructure, tree and environmental datasets to predict their role on sewer blockages and toward the proactive management of sewerage infrastructure. Furthermore, our approach may be applied to improve urban greening efforts, thus minimising future conflicts with, and costly disruptions to, underground infrastructure.
Research note: Integrating big data to predict tree root blockages across sewer networks
Highlights We modeled pipe-tree-environment systems to predict root blockages in pipes. Pipe and environmental conditions modulate tree root blockages. A few tree genera and species pose greater blockage risk under certain growth conditions. Our approach advances urban arboriculture and sewerage management.
Abstract Tree root blockages can occur when tree roots enter sewer pipes, resulting potentially in high economic and environmental costs. However, comprehensive understanding of which tree species and environmental factors (i.e., soil, hydrological and landscape properties) may most impact sewer blockages remains unclear. We integrated sewer pipe, tree and environmental data from 103 suburbs across seven local government areas to identify the most significant factors affecting root blockages along 902 km of sewer infrastructure in Sydney, Australia. Distributed Random Forest (DRF) models were used to predict 2,942 root blockages that occurred between 2010 and 2019 along 22,192 sewer pipe segments, and to relate these to 90,858 tree stems of 651 tree species, while accounting for other pipe and environmental variables. We found that tree root blockages were clustered across the urban landscape, creating “blockage hotspots”. Tree stem abundance was positively correlated with number of blockages. We found that urban morphology and pipe characteristics were more important than tree characteristics (for example stem size, useful life expectancy) in explaining root blockages. We also found that the number of tree stems and the abundance of the Chinese banyan (Ficus macrocarpa) were the most important variables predicting frequency of root blockages per unit pipe length. The most important landscape variables were slope, aspect, soil salinity and available water content, while most pipe-related variables also explained significant deviance in the DRF model. This study is among the first to mine large datasets representing urban infrastructure, tree and environmental datasets to predict their role on sewer blockages and toward the proactive management of sewerage infrastructure. Furthermore, our approach may be applied to improve urban greening efforts, thus minimising future conflicts with, and costly disruptions to, underground infrastructure.
Research note: Integrating big data to predict tree root blockages across sewer networks
Ossola, Alessandro (Autor:in) / Yu, Mengran (Autor:in) / Le Roux, Jaco (Autor:in) / Bustamante, Heriberto (Autor:in) / Uthayakumaran, Luther (Autor:in) / Leishman, Michelle (Autor:in)
10.09.2023
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Sewer Performance Reporting: Factors That Influence Blockages
| Online Contents | 2011
Proactive sewer maintenance to reduce blockages and collapses
| British Library Conference Proceedings | 2002
Spatiotemporal Modelling of Sediment-Related Blockages in the Sewer System of Bogotá (Colombia)
| Springer Verlag | 2018
| Europäisches Patentamt | 2020
Integrating Sewer Condition with Sewer Management
| British Library Conference Proceedings | 1999