Flood Hazard Rating Prediction for Urban Areas Using Random Forest and LSTM
- PDF / 4,671,482 Bytes
- 13 Pages / 595 x 786 pts Page_size
- 9 Downloads / 282 Views
pISSN 1226-7988, eISSN 1976-3808 www.springer.com/12205
DOI 10.1007/s12205-020-0951-z
Water Resources and Hydrologic Engineering
Flood Hazard Rating Prediction for Urban Areas Using Random Forest and LSTM Hyun Il Kim
a
and Byung Hyun Kimb
a
Member, Nakdong River Flood Control Office, Busan 49300, Korea Member, Dept. of Civil Engineering, Kyungpook National University, Daegu 41566, Korea
b
ARTICLE HISTORY
ABSTRACT
Received 28 May 2020 Revised 15 July 2020 Accepted 27 July 2020 Published Online 9 October 2020
A flood hazard rating prediction model was developed that is based on a long short-term memory (LSTM) neural network and random forest. The target area was Samseong District in Seoul, which has a history of severe flooding. The Storm Water Management Model was used to generate training data for the LSTM model to predict the total overflow as the rainfall input data. Two-dimensional numerical analysis was performed to calculate inundation and flow velocity maps for training the random forest, which was used to generate a map of the predicted flood hazard rating of grid units given the total accumulative overflow of the target area. To confirm the goodness of fit, the proposed model was used to predict a flood hazard rating map for a rainfall event observed on July 27, 2011. The prediction accuracy for the flood hazard rating of each grid was 99.86% when the debris factor was considered and 99.99% when the debris factor was not considered.
KEYWORDS Urban flooding Observed rainfall Hazard rating Machine learning Flood prediction Random forest
1. Introduction Climate change and urbanization are increasing the hazard of urban flooding. For areas with complex drainage systems and close to or below sea level, flooding damage is likely to increase even more. Flood hazard prediction is very important for understanding flooding damage that can occur in urban basins in advance. Approaches include one- and two-dimensional numerical analysis models. However, such models have several disadvantages, such as the time needed to adjust parameters, collect data, and preprocess and post-process the input and output data. Recent studies have focused on data-based flood analysis models that use machine learning, which can provide the flood analysis results more quickly (Mosavi et al., 2018). Shen (2018) has noted the increasing application of machine learning and deep learning to water resources and hydrologic analysis; as in most scientific fields, they are expected to produce excellent prediction and analysis results. Using a one-dimensional numerical simulation to calculate the urban runoff of a given rainfall can take a long time to input data, adjust parameters, and calculate the results. Thus, this approach is difficult to apply to predicting runoff quickly and CORRESPONDENCE Byung Hyun Kim ⓒ 2020 Korean Society of Civil Engineers
[email protected]
securing sufficient time to prepare countermeasures against urban flooding. To solve this problem, various studies have applied machine learning to the analysis of urb
Data Loading...
