Please wait a minute...
浙江大学学报(工学版)
Journal of ZheJiang University (Engineering Science)  2021, Vol. 55 Issue (5): 1010-1018    DOI: 10.3785/j.issn.1008-973X.2021.05.022
    
Flood prediction by multi-hydrological models with forecasting ability analysis
Deng-song LIU1(),Lu GE1,Yue-ping XU1,Shan-liang ZHANG2,Yen-ming CHIANG1,*()
1. Institute of Hydrology and Water Resources, College of Civil Engineering and Architecture, Zhejiang University, Hangzhou 310058, China
2. Power China HuaDong Engineering Co. Ltd, Hangzhou 310058, China
Download: HTML     PDF(1142KB) HTML
Export: BibTeX | EndNote (RIS)      

Abstract  

The forecast ability of the conceptual model was evaluated by the lag and route method to solve the selection problem of appropriate forecast lead time for both black box and conceptual models with hourly time scale. With the same forecast lead time, the flood forecasting was carried out by using the Xin'an jiang model, soil moisture accounting and routing (SMAR) model, back propagation (BP) neural network and Elman neural network and the performance was analyzed according to the set pair analysis method. The forecast ability of the conceptual models was relatively stable in specific watersheds and was corresponding to the size of the watershed. According to the lag and route method, the forecast lead time values for Jinhua and Longquan watersheds were 7 hours and 3 hours, respectively. Results indicated that the BP model and the Elman model produced the best and the worst performance, respectively, for flood forecasting in Longquan basin under the same forecast lead time. Moreover, the Xin'an jiang model and the SMAR model produced the best and the worst performance, respectively, for flood forecasting in Jinhua basin. The set pair analysis method is able to carry out comprehensive evaluation from single criterion to multiple criteria, indicating that the method has better applicability and higher efficiency.

Key words： black-box model      conceptual model      lag and route method      set pair analysis      flood forecasting     
Received: 22 July 2020      Published: 10 June 2021
CLC:  TV 213  
Fund:  浙江省重点研发计划择优委托资助项目（2021C03017）；国家重点研发计划政府间/港澳台重点专项资助项目（2016YFE0122100）；浙江省自然基金重点资助项目（LZ20E090001）
Corresponding Authors: Yen-ming CHIANG     E-mail: liuds@zju.edu.cn;chiangym@zju.edu.cn
Service
E-mail this article
Add to my bookshelf
Add to citation manager
E-mail Alert
RSS
Articles by authors
Deng-song LIU
Lu GE
Yue-ping XU
Shan-liang ZHANG
Yen-ming CHIANG
Cite this article:

Deng-song LIU,Lu GE,Yue-ping XU,Shan-liang ZHANG,Yen-ming CHIANG. Flood prediction by multi-hydrological models with forecasting ability analysis. Journal of ZheJiang University (Engineering Science), 2021, 55(5): 1010-1018.

URL:

http://www.zjujournals.com/eng/10.3785/j.issn.1008-973X.2021.05.022     OR     http://www.zjujournals.com/eng/Y2021/V55/I5/1010


基于多水文模型预报能力分析的洪水预报

为了解决概念性和黑箱模型在小时尺度模拟效果对比中预见期选择的问题，提出利用滞后演算法求解概念性模型预报能力. 在两模型共同预报能力条件下，引入集对分析理论对比新安江模型、土壤水核算与演算（SMAR）模型、倒传递（BP）和Elman神经网络模型模拟效果. 概念性模型预报能力与流域面积成正相关且特定流域的预报能力相对固定，其中金华江和龙泉溪流域的预见期可以分别取第7、3 h；在相同预见期条件下，龙泉溪流域表现最好的是BP模型而最差的是Elman模型，金华江流域表现最好的是新安江模型而最差的是SMAR模型；引入集对分析方法可以从单指标转向更多指标进行综合评价，具有较好的适用性和高效性.


关键词： 黑箱模型,  概念性模型,  滞后演算法,  集对分析,  洪水预报 
Fig.1 Principle of lag and route method
Fig.2 Training results of Xin'an jiang model with different population numbers and iterations in Longquan River
评价指标 Ivf NSE Qer Ter
1级 ≤5 ≥0.9 ≥80 ≥80
2级 5~10 0.8~0.9 50~80 50~80
3级 10~50 0~0.8 0~50 0~50
Tab.1 Standard of evaluation index grade
Fig.3 Stations and boundary of Longquan and Jinhua study area
Fig.4 NSE Results of different lag times for river networks in Longquanxi basin
Fig.5 NSE Results of different lag times for river networks in Jinhuajiang basin
Fig.6 Comparison of fitting effects of runoff for BP, Elman, SMAR and XAJ models in Longquanxi basin
Fig.7 Comparison of fitting effects of runoff for BP, Elman, SMAR and XAJ models in Jinhuajiang basin
流域 模型 Ivf NSE Qer Ter
龙泉溪 BP 1.46 0.96 66.67 100.00
Elman 21.46 0.88 44.44 66.67
SMAR 1.86 0.85 33.33 55.56
XAJ 4.58 0.93 55.56 66.67
金华江 BP 3.85 0.88 50.00 50.00
Elman 0.67 0.84 75.00 25.00
SMAR 14.88 0.67 50.00 12.50
XAJ 3.26 0.85 75.00 62.50
Tab.2 Evaluation indexes of varies models in different basins
流域 模型 等级 联系数
龙泉溪 BP 1 0.72
Elman 3 0.55
SMAR 2 0.78
XAJ 2 0.86
金华江 BP 2 0.76
Elman 2 0.81
SMAR 3 0.40
XAJ 2 0.87
Tab.3 Results of set pair analysis method in different areas
[1]   MULVANEY T J On the use of self registering rain and flood gauges[J]. Proceedings of the Institution of Civil Engineers of Ireland, 1850, 4: 1- 8
[2]   周琦, 朱跃龙, 陆佳民, 等 组合式水文模型建模方法综述[J]. 国外电子测量技术, 2020, 39 (2): 11- 18
ZHOU Qi, ZHU Yue-long, LU Jia-min, et al Overview of combined hydrological modeling methods[J]. Foreign Electronic Measurement Technology, 2020, 39 (2): 11- 18
[3]   RUMELHART D E, HINTON G E, WILLIAMS R J Learning representations by back-propagating errors[J]. Nature, 1986, 323 (9): 533- 536
[4]   CHEN P A, CHANG L C, CHANG F J Reinforced recurrent neural networks for multi-step-ahead flood forecasts[J]. Journal of Hydrology, 2013, 497: 71- 79
doi: 10.1016/j.jhydrol.2013.05.038
[5]   赵人俊. 流域水文模拟: 新安江模型与陕北模型[M]. 北京: 水利电力出版社, 1984: 106-130.
[6]   O"CONNELL P E, NASH J E, FARRELL J P River flow forecasting through conceptual models part II-the Brosna catchment at Ferbane[J]. Journal of Hydrology, 1970, 10 (4): 317- 329
doi: 10.1016/0022-1694(70)90221-0
[7]   BADRZADEH H, SARUKKALIGE R, JAYAWARDENA A W Hourly runoff forecasting for flood risk management: application of various computational intelligence models[J]. Journal of Hydrology, 2015, 3 (529): 1633- 1643
[8]   江衍铭, 张建全, 明焱 集合神经网络的洪水预报[J]. 浙江大学学报: 工学版, 2016, 50 (8): 1471- 1481
CHIANG Yen-ming, ZHANG Jian-quan, MING Yan Flood forecasting by ensemble neural networks[J]. Journal of Zhejiang University: Engineering Science, 2016, 50 (8): 1471- 1481
[9]   崔巍, 顾冉浩, 陈奔月, 等 BP与LSTM神经网络在福建小流域水文预报中的应用对比[J]. 人民珠江, 2020, 41 (2): 74- 84
CUI Wei, GU Ran-hao, CHEN Ben-yue, et al Comparison of BP and LSTM neural network for hydrologic forecasting of a small watershed in Fujian[J]. Pearl River, 2020, 41 (2): 74- 84
[10]   HARRIS A, HOSSAIN F Investigating the optimal configuration of conceptual hydrologic models for satellite-rainfall-based flood prediction[J]. IEEE Geoscience and Remote Sensing Letters, 2008, 5 (3): 532- 536
doi: 10.1109/LGRS.2008.922551
[11]   BRAM V A F, BRUNO S, VALENTIJN R N P Internal validation of conceptual rainfall-runoff models using baseflow separation[J]. Journal of Hydrology, 2010, 381 (1/2): 158- 173
[12]   张漫莉 改进的HBV模型与新安江模型在武江流域洪水预报中的应用比较[J]. 人民珠江, 2014, 35 (1): 34- 37
ZHANG Man-li Comparison of improved HBV model and Xin'anjiang model in flood forecasting of Wujiang river basin[J]. Pearl River, 2014, 35 (1): 34- 37
doi: 10.3969/j.issn.1001-9235.2014.01.009
[13]   TAYYAB M, ZHOU J Z, ZENG X F, et al. Optimal application of conceptual rainfall-runoff hydrological models in the Jinshajiang river basin, China [C]// Proceedings of RSHS and ICGRHWE 2015. Portugal: [s. n], 2015, 368: 227-232.
[14]   DE VOS N J, RIENTJES T H M Multi-objective performance comparison of an artificial neural network and a conceptual rainfall-runoff model[J]. Hydrological Sciences Journal, 2007, 52 (3): 397- 413
doi: 10.1623/hysj.52.3.397
[15]   BANIHABIB M E, REVELLI R Performance of conceptual and black-box models in flood warning systems[J]. Cogent Engineering, 2016, 3: 1- 13
doi: 10.1080/2331186X.2016.1139438
[16]   虞慧. 新安江模型在中型水库洪水预报中的应用研究[D]. 南昌: 江西师范大学, 2010.
YU Hui. Study on the application of Xin’anjiang model for forecasting in medium-sized reservoir[D]. Nanchang: Jiangxi Normal University, 2010.
[17]   KENNEDY J, EBERHART R. Particle swarm optimization [C]// 1995 IEEE International Conference on Neural Networks Proceedings. Piscataway: IEEE, 1995: 1942-1948.
[18]   ZHANG L, YANG X Y Applying a multi-model ensemble method for long-term runoff prediction under climate change scenarios for the Yellow River Basin, China[J]. Water, 2018, 301: 1- 15
[19]   盖永岗, 熊立华, 万民, 等. 基于模糊数学的水文预报综合评价方法 [C]// 环境变化与水安全——第五届中国水论坛论文集. 南京: 中国水利水电出版社, 2007: 141-144.
GAI Yong-gang, XIONG Li-hua, WAN Min, et al. Comprehensive evaluation method of hydrological forecast based on fuzzy mathematics [C]// Environmental Change and Water Security: the 5th Chinese Water Forum. Nanjing: China Water and Power Press, 2007: 141-144.
[20]   WEIJS S V, SCHOUP G, GIESEN N Why hydrological predictions should be evaluated using information theory[J]. Hydrology and Earth System Sciences, 2010, 14: 2545- 2558
doi: 10.5194/hess-14-2545-2010
[21]   肖洁, 罗军刚, 张晓 基于DS证据理论的多模型洪水预报方案优选[J]. 西北农林科技大学学报: 自然科学版, 2013, 41 (9): 193- 199
XIAO Jie, LUO Jun-gang, ZHANG Xiao DS evidence theory based scheme optimization of multi-model flood forecasting[J]. Journal of Northwest A and F University: Natural Science Edition, 2013, 41 (9): 193- 199
[22]   石继海, 宋松柏, 李航 中长期径流预报模型优选研究[J]. 西北农林科技大学学报: 自然科学版, 2019, 47 (7): 174- 154
SHI Ji-hai, SONG Song-bai, LI Hang Optimization of mid-long term runoff forecasting models[J]. Journal of Northwest A and F University: Natural Science Edition, 2019, 47 (7): 174- 154
[23]   金菊良, 陈磊, 陈梦璐, 等 基于集对分析和风险矩阵的水资源承载力评价方法[J]. 人民长江, 2018, 49 (7): 35- 41
JIN Ju-liang, CHEN Lei, CHEN Meng-lu, et al Evaluation method of water resources carrying capacity based on set pair analysis and risk matrix[J]. Yangtze River, 2018, 49 (7): 35- 41
[24]   CLARK C O Flood-storage accounting[J]. Eos, Transactions American Geophysical Union, 1994, 25 (6): 1014- 1030
[25]   赵人俊, 王厥谋 论滞后演算法[J]. 水文, 1983, (5): 30- 34
ZHAO Ren-jun, WANG Jue-mou A discussion of the lag and route method[J]. Journal of China Hydrology, 1983, (5): 30- 34
[26]   刘苏宁, 甘泓, 魏国孝 粒子群算法在新安江模型参数率定中的应用[J]. 水利学报, 2010, 41 (5): 537- 544
LIU Su-ning, GAN Hong, WEI Guo-xiao Application of PSO algorithm to calibrate the Xin’anjiang hydrological model[J]. Journal of Hydraulic Engineering, 2010, 41 (5): 537- 544
[27]   赵克勤. 集对分析及其初步应用[M]. 杭州: 浙江科学技术出版社, 2000.
[28]   段新光, 栾芳芳 基于模糊综合评判的新疆水资源承载力评价[J]. 中国人口·资源与环境, 2014, 24 (Suppl. 1): 119- 122
DUAN Xin-guang, LUAN Fang-fang Evaluation of water resources carrying capacity in xinjiang based on fuzzy comprehensive model[J]. China Population, Resources and Environment, 2014, 24 (Suppl. 1): 119- 122
[29]   XIN G, YOU H E, XIAO Y I Attribute measure recognition approach and its applications to emitter recognition[J]. Science China Information Sciences, 2005, 48 (2): 225- 233
doi: 10.1360/122004-82
[30]   王光生, 夏士谆 SMAR模型及其改进[J]. 水文, 1998, (Suppl. 1): 28- 30
WANG Guang-sheng, XIA Shi-chun SMAR model and its improvement[J]. Journal of China Hydrology, 1998, (Suppl. 1): 28- 30
[31]   张建全. 基于人工神经网络的集合洪水预报研究[D]. 杭州: 浙江大学, 2016.
ZHANG Jian-quan. Ensemble flood forecasting by artificial neural network[D]. Hangzhou: Zhejiang University, 2016.
[1] CHIANG Yen ming, HAO Ruo nan, TSAI Wen ping. Progress and practical applications of flood forecasting systems in Taiwan[J]. Journal of ZheJiang University (Engineering Science), 2016, 50(9): 1784-1790.
[2] CHIANG Yen ming, ZHANG Jian quan, MING Yan. Flood forecasting by ensemble neural networks[J]. Journal of ZheJiang University (Engineering Science), 2016, 50(8): 1471-1478.
[3] ZHANG Hao, QIU Le-miao, ZHANG Shu-you, ZHANG Peng. Evaluation method based on set pair analysis for configuration scheme of reconfigurable machine tools[J]. Journal of ZheJiang University (Engineering Science), 2015, 49(7): 1232-1241.