Objective This study aims to explore the application of machine learning as a novel method in flash flood simulation and improve the accuracy of flash flood simulation and early warning in the Pingjiang River Basin. Methods A total of 25 typical flood events and their corresponding 5-minute high-resolution rainfall data from 1997 to 2018 in the flash flood occurrence area were selected. Three models—based on the HEC-HMS hydrological model, the support vector machine (SVM) model, and the HEC-HMS-SVM coupled model were constructed to compare and evaluate their accuracy and stability in flood hydrograph simulation. Results (1) The HEC-HMS model demonstrated excellent performance in simulating “single-peak type” floods, with good extrapolation capability, and the overall qualification rate reached 92% (Class A accuracy). (2) The SVM model had an overall qualification rate of 84% (Class B accuracy), but it showed high sensitivity to the time of peak occurrence, with a significant difference between the calibration period and the validation period, and its stability was relatively weak. (3) The coupled model exhibited the best overall performance. The qualification rate during the validation period increased to 100% (25% higher than SVM), the overall qualification rate increased by 8% and 16% compared with HEC-HMS and SVM, respectively, and the fit of the flood process was significantly improved. Conclusion The HEC-HMS-SVM coupled model can effectively improve the accuracy of flash flood simulation and provide more reliable technical support for flash flood disaster prevention and control.
ShuaiW, ZhanX H, LvW L. Study on the application of torrential rain analysis and evaluation system for mountain torrential disaster prevention[J]. Science and Technology & Innovation, 2024(1):187-189.
[3]
熊朕,田宏岭.我国山洪灾害监测现状与发展趋势[J].灾害学,2019,34(3):140-145.
[4]
XiongZ, TianH L. A review and trend: flash flood disaster monitoring in China[J]. Journal of Catastrophology, 2019,34(3):140-145.
GeK Y, WanX Y, XuH J, et al. Study on flood simulation of data deficient mountainous areas based on the HEC-HMS model [J]. Water Power, 2024,50(3):19-24.
ZhaiX Y, GuoL, LiuR H, et al. Impact assessment of antecedent soil moisture conditions and rain-fall variability on flash flood warning index at catchment scale[J]. Geographical Research, 2019,38(12):2957-2965.
[9]
HeM X, HogueT S. Integrating hydrologic modeling and land use projections for evaluation of hydrologic response and regional water supply impacts in semi-arid environments[J]. Environmental Earth Sciences, 2012,65(6):1671-1685.
[10]
KhalidK, AliM F, Abd RahmanN F, et al. Application on one-at-a-time sensitivity analysis of semi-distributed hydrological model in tropical watershed[J]. International Journal of Engineering and Technology, 2016,8(2):132-136.
LiuX Y, ZhangH L, LuoZ Y, et al. Response of runoff and sediment yields to land use change in Fu River watershed based on SWAT model[J]. Research of Soil and Water Conservation, 2024,31(3):79-89,100.
ZhuangQ Y, CuiY X, ZhangY H, et al. Effects of climate and land use change on runoff and total phosphorus in the Ulungu River basin based on SWAT model[J]. Research of Soil and Water Conservation, 2024,31(6):1-10.
FengS W, DengF F, LiJ. Application of HEC-HMS model in flood forecasting for upper reaches of Lijiang River basin[J]. Guangxi Water Resources & Hydropower Engineering, 2023(3):19-23.
WangY X, LiuB, WangW P, et al. Simulation of flood in Three Gorges Region based on HEC-HMS model[J]. Journal of Changjiang River Scientific Research Institute, 2024,41(6):76-83.
ZhongF Q, HuoA D, ZhaoZ X, et al. Impact of urbanization on flood regime in gullied Loess Plateaus based on HEC-HMS[J]. Yellow River, 2024,46(2):67-72,79.
ChenE Y, ZhangS Y, ZhangB. Flood simulation using HEC-HMS hydrological model at watershed in Taihang Mountain[J]. Water Sciences and Engineering Technology, 2023,(2):14-16.
ZhaoZ, FengM Q, HouZ L, et al. Comparative study on different rainfall loss methods of the HEC-HMS hydrological model[J]. Journal of China Hydrology, 2024,44(2):83-88.
DuanX H, LiL J, LiJ Z. Study on the influence of topographic data accuracy on watershed flood simulation[J]. Journal of China Hydrology, 2024,44(1):33-36,49.
LiJ Z, LiL J, ZhangT, et al. Effect of DEM data sources and resolutions on watershed flood simulations[J]. Journal of Hydroelectric Engineering, 2023,42(3):26-40.
ZhangR, ChaiZ Y, ZhangT, et al. Research progress of flood forecasting based on machine learning models[J]. Water Resources and Hydropower Engineering, 2023,54(11):89-101.
JiZ S, ZhangG W, HuangW. Flood process forecasting method of east Tiaoxi based on LS-SVM[J]. Zhejiang Hydrotechnics, 2021,49(2):5-8.
[37]
SahooA, SamantarayS, GhoseD K. Prediction of flood in Barak River using hybrid machine learning approaches: a case study[J]. Journal of the Geological Society of India, 2021,97(02):186-198.
LiY X, LiuH Y. Spatiotemporal variation of rainfall erosivity at Pingjiang basin in upstream of Ganjiang River[J]. Bulletin of Soil and Water Conservation, 2020,40(1):1-8,23.
TianL Y, LiuH Y, WangY W. Research on runoff characteristics of Zhangshui basin of Ganjiang River based on quantile regression[J]. Journal of Nanchang Institute of Technology, 2023,42(1):19-27.
LiuS H, LiuH Y, LuoP, et al. Spatial and temporal distribution characteristics of rainfall of different magnitudes in Pingjiang basin of upper Ganjiang River in recent 30 years[J]. Research of Soil and Water Conservation, 2022,29(3):106-114.
[46]
杨欢.基于HEC-HMS模型的半干旱黄土沟壑区小流域设计洪水研究[D].兰州:兰州大学,2019.
[47]
YangH. Research on design flood of small watershed in semi-arid Loess Gully Region based on HEC-HMS model[D]. Lanzhou: Lanzhou University, 2019.
CuiD W, JinB. Improved support vector machine regression model and its application to annual runoff forecasting[J]. Journal of Hydroelectric Engineering, 2015,34(2):7-14.
ChenF, ChenX W, XieJ B. Sensitivity analysis of paramerters of the rainflood similation by using HEC-HMS hydrological model and its application[J]. Journal of Water Resources and Water Engineering, 2012,23(5):119-122.
ChengX, MaX X, WangW S, et al. Applicability research of HEC-HMS model parameter regionalization in small basin of Henan Province[J]. Journal of China Hydrology, 2022,42(1):40-46,102.