A data-driven analysis, and its limitations, of the spatial flood archive of Flanders, Belgium to assess the impact of soil sealing on flood volume and extent.
Soil sealing increases surface runoff in a watershed and decreases infiltration into the soil. Consequently, urbanization poses a significant challenge for watershed management to mitigate faster runoff accumulation downstream and associated floods. Hydrological models are often employed to assess the impact of land-use dynamics on flood events. Alternatively, data-driven approaches combining time series of land use geodatasets and georeferenced flooded zones also allow to assess the relationship between soil sealing and flood severity. This study presents such data-driven analysis using a spa... Mehr ...
Verfasser: | |
---|---|
Dokumenttyp: | Artikel |
Erscheinungsdatum: | 2020 |
Reihe/Periodikum: | PLoS ONE, Vol 15, Iss 10, p e0239583 (2020) |
Verlag/Hrsg.: |
Public Library of Science (PLoS)
|
Schlagwörter: | Medicine / R / Science / Q |
Sprache: | Englisch |
Permalink: | https://search.fid-benelux.de/Record/base-27471398 |
Datenquelle: | BASE; Originalkatalog |
Powered By: | BASE |
Link(s) : | https://doi.org/10.1371/journal.pone.0239583 |
Soil sealing increases surface runoff in a watershed and decreases infiltration into the soil. Consequently, urbanization poses a significant challenge for watershed management to mitigate faster runoff accumulation downstream and associated floods. Hydrological models are often employed to assess the impact of land-use dynamics on flood events. Alternatively, data-driven approaches combining time series of land use geodatasets and georeferenced flooded zones also allow to assess the relationship between soil sealing and flood severity. This study presents such data-driven analysis using a spatially explicit archive of flooded areas dating back to 1988 in the Flanders region of Belgium, which is characterized by urban sprawl. This archived data, along with time series of rainfall and land use, were analyzed for three middle-sized river subbasins using two machine learning methods: boosted regression trees and support vector regression. The machine learning methods were found suitable for this type of analysis, since their flexibility allows for spatially explicit models with larger sample sizes. However, the relationship between soil sealing and flood volume and extent could not be conclusively confirmed by our models. This may be due to data limitations, such as the limited number of recorded historical floods, inaccuracies in recorded historical flood polygons and inconsistencies in the land use classifications. It is therefore stressed that continued consistent monitoring of floods and land use changes is required.