Towards urban flood susceptibility mapping using data-driven models in Berlin, Germany
Identifying urban pluvial flood-prone areas is necessary but the application of two-dimensional hydrodynamic models is limited to small areas. Data-driven models have been showing their ability to map flood susceptibility but their application in urban pluvial flooding is still rare. A flood invento...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Taylor & Francis Group
2022-12-01
|
| Series: | Geomatics, Natural Hazards & Risk |
| Subjects: | |
| Online Access: | https://www.tandfonline.com/doi/10.1080/19475705.2022.2097131 |
| _version_ | 1811296975920824320 |
|---|---|
| author | Omar Seleem Georgy Ayzel Arthur Costa Tomaz de Souza Axel Bronstert Maik Heistermann |
| author_facet | Omar Seleem Georgy Ayzel Arthur Costa Tomaz de Souza Axel Bronstert Maik Heistermann |
| author_sort | Omar Seleem |
| collection | DOAJ |
| description | Identifying urban pluvial flood-prone areas is necessary but the application of two-dimensional hydrodynamic models is limited to small areas. Data-driven models have been showing their ability to map flood susceptibility but their application in urban pluvial flooding is still rare. A flood inventory (4333 flooded locations) and 11 factors which potentially indicate an increased hazard for pluvial flooding were used to implement convolutional neural network (CNN), artificial neural network (ANN), random forest (RF) and support vector machine (SVM) to: (1) Map flood susceptibility in Berlin at 30, 10, 5, and 2 m spatial resolutions. (2) Evaluate the trained models' transferability in space. (3) Estimate the most useful factors for flood susceptibility mapping. The models' performance was validated using the Kappa, and the area under the receiver operating characteristic curve (AUC). The results indicated that all models perform very well (minimum AUC = 0.87 for the testing dataset). The RF models outperformed all other models at all spatial resolutions and the RF model at 2 m spatial resolution was superior for the present flood inventory and predictor variables. The majority of the models had a moderate performance for predictions outside the training area based on Kappa evaluation (minimum AUC = 0.8). Aspect and altitude were the most influencing factors on the image-based and point-based models respectively. Data-driven models can be a reliable tool for urban pluvial flood susceptibility mapping wherever a reliable flood inventory is available. |
| first_indexed | 2024-04-13T05:57:16Z |
| format | Article |
| id | doaj.art-e41e7c6d900a45a490a3a921e245be2f |
| institution | Directory Open Access Journal |
| issn | 1947-5705 1947-5713 |
| language | English |
| last_indexed | 2024-04-13T05:57:16Z |
| publishDate | 2022-12-01 |
| publisher | Taylor & Francis Group |
| record_format | Article |
| series | Geomatics, Natural Hazards & Risk |
| spelling | doaj.art-e41e7c6d900a45a490a3a921e245be2f2022-12-22T02:59:36ZengTaylor & Francis GroupGeomatics, Natural Hazards & Risk1947-57051947-57132022-12-011311640166210.1080/19475705.2022.2097131Towards urban flood susceptibility mapping using data-driven models in Berlin, GermanyOmar Seleem0Georgy Ayzel1Arthur Costa Tomaz de Souza2Axel Bronstert3Maik Heistermann4Institute of Environmental Science and Geography, Chair for Hydrology and Climatology, University of Potsdam, Potsdam, GermanyInstitute of Environmental Science and Geography, Chair for Hydrology and Climatology, University of Potsdam, Potsdam, GermanyInstitute of Environmental Science and Geography, Chair for Hydrology and Climatology, University of Potsdam, Potsdam, GermanyInstitute of Environmental Science and Geography, Chair for Hydrology and Climatology, University of Potsdam, Potsdam, GermanyInstitute of Environmental Science and Geography, Chair for Hydrology and Climatology, University of Potsdam, Potsdam, GermanyIdentifying urban pluvial flood-prone areas is necessary but the application of two-dimensional hydrodynamic models is limited to small areas. Data-driven models have been showing their ability to map flood susceptibility but their application in urban pluvial flooding is still rare. A flood inventory (4333 flooded locations) and 11 factors which potentially indicate an increased hazard for pluvial flooding were used to implement convolutional neural network (CNN), artificial neural network (ANN), random forest (RF) and support vector machine (SVM) to: (1) Map flood susceptibility in Berlin at 30, 10, 5, and 2 m spatial resolutions. (2) Evaluate the trained models' transferability in space. (3) Estimate the most useful factors for flood susceptibility mapping. The models' performance was validated using the Kappa, and the area under the receiver operating characteristic curve (AUC). The results indicated that all models perform very well (minimum AUC = 0.87 for the testing dataset). The RF models outperformed all other models at all spatial resolutions and the RF model at 2 m spatial resolution was superior for the present flood inventory and predictor variables. The majority of the models had a moderate performance for predictions outside the training area based on Kappa evaluation (minimum AUC = 0.8). Aspect and altitude were the most influencing factors on the image-based and point-based models respectively. Data-driven models can be a reliable tool for urban pluvial flood susceptibility mapping wherever a reliable flood inventory is available.https://www.tandfonline.com/doi/10.1080/19475705.2022.2097131Urban pluvial flood susceptibilityconvolutional neural networkdeep learningrandom forestsupport vector machinespatial resolution |
| spellingShingle | Omar Seleem Georgy Ayzel Arthur Costa Tomaz de Souza Axel Bronstert Maik Heistermann Towards urban flood susceptibility mapping using data-driven models in Berlin, Germany Geomatics, Natural Hazards & Risk Urban pluvial flood susceptibility convolutional neural network deep learning random forest support vector machine spatial resolution |
| title | Towards urban flood susceptibility mapping using data-driven models in Berlin, Germany |
| title_full | Towards urban flood susceptibility mapping using data-driven models in Berlin, Germany |
| title_fullStr | Towards urban flood susceptibility mapping using data-driven models in Berlin, Germany |
| title_full_unstemmed | Towards urban flood susceptibility mapping using data-driven models in Berlin, Germany |
| title_short | Towards urban flood susceptibility mapping using data-driven models in Berlin, Germany |
| title_sort | towards urban flood susceptibility mapping using data driven models in berlin germany |
| topic | Urban pluvial flood susceptibility convolutional neural network deep learning random forest support vector machine spatial resolution |
| url | https://www.tandfonline.com/doi/10.1080/19475705.2022.2097131 |
| work_keys_str_mv | AT omarseleem towardsurbanfloodsusceptibilitymappingusingdatadrivenmodelsinberlingermany AT georgyayzel towardsurbanfloodsusceptibilitymappingusingdatadrivenmodelsinberlingermany AT arthurcostatomazdesouza towardsurbanfloodsusceptibilitymappingusingdatadrivenmodelsinberlingermany AT axelbronstert towardsurbanfloodsusceptibilitymappingusingdatadrivenmodelsinberlingermany AT maikheistermann towardsurbanfloodsusceptibilitymappingusingdatadrivenmodelsinberlingermany |