Quantifying Flood Depth: Remote Estimation of Flood Depths with Fast Response Tools

Flood depth modelling has seen significant improvements in recent years. Historically, it relied heavily on large datasets, which were difficult to gather and handle. Simple, fast response methods, such as RICorDE and FwDET, are potentially suitable for assessments of flood scenarios where detailed hydraulic data might not be available or necessary. In this study, the performance of flood depth estimation algorithms RICorDE v1.0.1 and FwDET v2.1 was assessed for the 2020 Fort McMurray ice jam flood event. To assess their performance, the models were calibrated using a hand drawn case study event flood extent, and the resulting depth outputs from the fast response models were compared against a calibrated HEC-RAS model from Alberta Environment and Protected Areas (EPA) as well as high water mark (HWM) depths collected for the case study event, also from EPA. In the comparative assessment, the fast response RICorDE model achieved an R² value of 0.69 when compared to the calibrated HEC-RAS model, while FwDET achieved a value of 0.86. In relation to the surveyed HWM-derived depths, the fast response tools established point connections with fewer data points than the calibrated HEC-RAS model. In the surveyed depths assessment, RICorDE exhibited a Root Mean Square Error (RMSE) value of 0.58 meters, and FwDET had an RMSE of 0.20 meters. The calibrated HEC-RAS model, to which both fast response models were compared, presented a high point connection with an RMSE difference of 0.41 meters. Additionally, we assessed the Height Above Nearest Drainage (HAND) maps integrated into the fast response model RICorDE as a basis for flood depth estimations. HAND maps represent the elevation of a point on the landscape relative to the nearest stream or drainage network. In their assessment, the discrepancies in elevation points when creating HAND maps using either a stream or drainage network as reference points was evaluated. Specifically, the study investigated the impact of generating HAND maps based on these features on the accuracy of estimating flood extent. The findings indicate that defining drainage as a continuous streamline using flow accumulation algorithms yielded a more precise depiction of flood extent, in contrast to maps where drainage is delineated along the boundary between water and land. In conclusion, the study demonstrated that fast response flood models can generate flood depth estimations with high correlation to observations and physical models, especially in areas with flat topography and high-quality DEM data. These estimations are highly dependent on accurate delineation of flood extent, which can also be obtained from minimal data using HAND models.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.