TY - JOUR

T1 - An ensemble neural network model for real-time prediction of urban floods

AU - Berkhahn, Simon

AU - Fuchs, Lothar

AU - Neuweiler, Insa

N1 - Funding information: We thank Aaron Peche, Robert Sämann and Tim Berthold for discussion and technical support. Furthermore, we thank the anonymous reviewers and Slobodan Djordjevic for challenging comments which helped to improve the present study. The research is being conducted with financial support of the state of Lower Saxony and the BMBF funded research project EVUS (EVUS – Real-Time Prediction of Pluvial Floods and Induced Water Contamination in Urban Areas) [BMBF, 03G0846A].

PY - 2019/8

Y1 - 2019/8

N2 - The real-time forecasting of urban flooding is a challenging task for the following two reasons: (1) urban flooding is often characterized by short lead times, (2) the uncertainty in precipitation forecasting is usually high. Standard physically based numerical models are often too slow for the use in real-time forecasting systems. Data driven models have small computational costs and fast computation times and may be useful to overcome this problem. The present study presents an artificial neural network based model for the prediction of maximum water levels during a flash flood event. The challenge of finding a suitable structure for the neural network was solved with a new growing algorithm. The model is successfully tested for spatially uniformly distributed synthetic rain events in two real but slightly modified urban catchments with different surface slopes. The computation time of the model in the order of seconds and the accuracy of the results are convincing, which suggest that the method may be useful for real-time forecasts.

AB - The real-time forecasting of urban flooding is a challenging task for the following two reasons: (1) urban flooding is often characterized by short lead times, (2) the uncertainty in precipitation forecasting is usually high. Standard physically based numerical models are often too slow for the use in real-time forecasting systems. Data driven models have small computational costs and fast computation times and may be useful to overcome this problem. The present study presents an artificial neural network based model for the prediction of maximum water levels during a flash flood event. The challenge of finding a suitable structure for the neural network was solved with a new growing algorithm. The model is successfully tested for spatially uniformly distributed synthetic rain events in two real but slightly modified urban catchments with different surface slopes. The computation time of the model in the order of seconds and the accuracy of the results are convincing, which suggest that the method may be useful for real-time forecasts.

KW - Artificial neural network

KW - Ensemble neural network

KW - Real-time forecast

KW - Urban flooding

UR - http://www.scopus.com/inward/record.url?scp=85067078458&partnerID=8YFLogxK

U2 - 10.15488/12043

DO - 10.15488/12043

M3 - Article

AN - SCOPUS:85067078458

VL - 575

SP - 743

EP - 754

JO - Journal of hydrology

JF - Journal of hydrology

SN - 0022-1694

ER -