基于EMD组合模型的径流多尺度预测

doi:10.13745/j.esf.sf.2020.10.22

摘要/Abstract

摘要：

受全球气候变化与人类活动影响,径流序列愈发呈现出非稳态与非线性特征,为降低由此而引发的预报误差,充分发挥不同模型对提高径流预测精度的优势,针对传统径流预报模型的单一性,以干旱区典型内陆河玛纳斯河为例,采用经验模态分解(EMD)提取径流序列中具有物理含义的信号,得到不同时间尺度的多个固有模态函数(IMF)及1个趋势项,利用 ARIMA模型与GRNN模型分别对不同时间尺度的IMF分量进行模拟,分析径流未来变化趋势。运用多元线性回归法、Spearman相关系数法、平均影响值法筛选大气环流因子作为神经网络模型的输入项,根据子序列的局部频率特点构建组合模型。最后将各IMF分量的预测结果重构,得到径流的最终预测值。单一评价指标无法全面评价模型精度,本文通过构建TOPSIS评价模型对径流预测模型进行定量评估,客观评价模型优度。结果表明:EMD分解能有效提取径流序列中隐含的多时间尺度信号,由趋势项可知玛纳斯河径流量总体呈上升趋势;EMD分解可提高ARIMA模型25%的合格率,但对于高频率分量IMF1、IMF2、IMF3,ARIMA模型的相对误差达到70%以上,预测结果不理想;经过筛选预报因子可有效提高GRNN模型精度,其中MIV法筛选的预报因子最适合玛纳斯河,与EMD-ARIMA组合后的GRNN模型的合格率最高,TOPSIS模型得分也最高。预测结果可作为水资源规划与调度的科学依据,建模思路也可为优化径流预测模型提供新途径。

关键词: EMD, 径流预测, GRNN模型, 组合模型, 模型评价

Abstract:

Affected by global climate change and human activities, runoff sequences increasingly show unsteady and non-linear characteristics. In order to reduce the forecast errors caused by these runoff characteristics, we took full advantages of different models to improve the accuracy of runoff prediction traditionally done by the single model approach. Taking the Manas River, a typical inland river in the arid area as an example, we used empirical mode decomposition (EMD) to extract physically meaningful signals from the runoff sequence to obtain multiple intrinsic mode functions (IMF) at different time scales and a trend indicator. We then used the ARIMA and GRNN models to simulate the IMF components at different time scales and analyze the future runoff changing trends. Next, we used the multiple linear regression, Spearman correlation coefficient and average influence value methods to screen the atmospheric circulation factors and them used as inputs to the neural network model; we then constructed the combined model according to the local frequency characteristics of the sub-sequences. Finally, we reconstructed the prediction results of each IMF component to obtain the runoff prediction. However, a single evaluation index cannot fully evaluate the accuracy of the prediction model. In this paper, we constructed the TOPSIS evaluation model to quantitatively evaluate the runoff prediction model and objectively evaluate the model’s superiority. The results show that using EMD can effectively extract the multi-timescale signals hidden in the runoff sequence, and the trend indicator indicated that the Manas River runoff is on the rise. Using EMD could improve the passing rate of the ARIMA model by 25%; but the relative errors of high frequency components IMF1, IMF2 and IMF3 in the ARIMA model were more than 70%, indicating the prediction results are not ideal. In the GRNN model, selected predictors could effectively improve the model accuracy, and the predictors selected by the MIV method were shown to be most suitable for the Manas River. Overall, the GRNN-EMD-ARIMA combination model had the highest passing rate, and the TOPSIS model had the highest score. The prediction results can be used as a scientific basis for water resources planning and dispatching, and the modeling ideas can also provide new ways to optimize runoff prediction models.

Key words: EMD, runoff prediction, GRNN model, coupled model, model evaluation

中图分类号:

X143
P333

李福兴, 陈伏龙, 蔡文静, 何朝飞, 龙爱华. 基于EMD组合模型的径流多尺度预测[J]. 地学前缘, 2021, 28(1): 428-437.

LI Fuxing, CHEN Fulong, CAI Wenjing, HE Chaofei, LONG Aihua. Multiscale runoff prediction based on the EMD combined model[J]. Earth Science Frontiers, 2021, 28(1): 428-437.

图/表 15

参考文献 27

[1]	IPCC. Climate change 2013: the physical science basis: contribution of working group I to the fifth assessment report of the intergovernmental panel on climate change[M]. New York: Cambridge University Press, 2013.
[2]	程国栋, 王根绪. 中国西北地区的干旱与旱灾: 变化趋势与对策[J]. 地学前缘, 2006, 13(1):3-14.
[3]	穆兴民, 李靖, 王飞, 等. 黄河天然径流量年际变化过程分析[J]. 干旱区资源与环境, 2003(2):1-5.
[4]	徐炜, 张弛, 彭勇, 等. 基于多模型预报信息融合的中长期径流预报研究[J]. 水力发电学报, 2013, 32(6):11-18.
[5]	陈守煜, 王大刚. 基于遗传算法的模糊优选BP网络模型及其应用[J]. 水利学报, 2003(5):116-121.
[6]	FENG Z K, NIU W J, TANG Z Y, et al. Monthly runoff time series prediction by variational mode decomposition and support vector machine based on quantum-behaved particle swarm optimization[J]. Journal of Hydrology, 2020, 583. DOI: 10.1016/jhydrol.2020.124627. DOI
[7]	崔庆. 最小二乘支持向量机模型的改进及其在径流预报中的应用[D]. 咸阳: 西北农林科技大学, 2013.
[8]	于瑞宏, 张宇瑾, 张笑欣, 等. 无测站流域径流预测区域化方法研究进展[J]. 水利学报, 2016, 47(12):1528-1539.
[9]	AYENU-PRAH A Y, ATTOH-OKINE N O. Comparative study of Hilbert-Huang transform[J]. Vehicle System Dynamics, 2009, 47(4):437-456. DOI URL
[10]	BESAW L E, RIZZO D M, BIERMAN P R, et al. Advances in ungauged streamflow prediction using artificial neural networks[J]. Journal of Hydrology, 2010, 386(1):27-37. DOI URL
[11]	BARGE J T, SHARIF H O. An ensemble empirical mode decomposition, self-organizing map, and linear genetic programming approach for forecasting river streamflow[J]. Water, 2016, 8(6):247. DOI URL
[12]	高祥涛, 李士进. 基于经验模态分解的水文时间序列预测研究[J]. 水文, 2009, 29(3):66-68.
[13]	梁浩, 黄生志, 孟二浩, 等. 基于多种混合模型的径流预测研究[J]. 水利学报, 2020, 51(1):112-125.
[14]	沈雪峰, 艾成. 新疆玛纳斯河径流时间变化特征及其趋势分析[J]. 干旱区资源与环境, 2012, 26(7):14-19.
[15]	凌红波, 徐海量, 张青青, 等. 1957—2007年新疆天山山区气候变化对径流的影响[J]. 自然资源学报, 2011, 26(11):1908-1917.
[16]	南峰, 李有利, 史兴民. 新疆玛纳斯河水量波动与气候变化之间的关系[J]. 水土保持研究, 2003(3):59-61.
[17]	凌红波, 徐海量, 张青青, 等. 1956—2007年新疆玛纳斯河流域气候变化趋势分析[J]. 冰川冻土, 2011, 33(1):64-71.
[18]	凌红波, 徐海量, 乔木, 等. 1958—2006年玛纳斯河流域水系结构时空演变及驱动机制分析[J]. 地理科学进展, 2010, 29(9):1129-1136.
[19]	赵雪花, 陈旭, 袁旭琦. 基于EMD的数据驱动模型在径流预测中的应用[J]. 系统工程, 2014, 32(9):150-154.
[20]	张建新, 鲁江, 邢旭东. 基于ARIMA模型的洞庭湖第四纪沉积物中镉演化及意义[J]. 地学前缘, 2008, 15(5):219-226.
[21]	DING H, WANG Z C, GUO Y C. Multi-objective optimization of fiber laser cutting based on generalized regression neural network and non-dominated sorting genetic algorithm[J]. Infrared Physics and Technology, 2020. DOI: 10.1016/j.infrared.2020.103337. DOI
[22]	DRISCOLL J M, HAY L E, VANDERHOOF M K, et al. Spatiotemporal variability of modeled watershed scale surface-Depression storage and runoff for the Conterminous United States[J]. Journal of the American Water Resources Association, 2020, 56(1):16-29. DOI URL
[23]	DOMBI G W, NANDI P, SAXE J M, et al. Prediction of rib fracture injury outcome by an artificial neural network, 1995, 39(5): 915-21.
[24]	ADAMU M, ADO A. Application of principal component analysis and multiple regression models in surface water quality assessment[J]. Journal of Environment and Earth Science, 2012, 2(2):16-23.
[25]	中华人民共和国国家质量监督检验检疫总局, 中国国家标准化管理委员会. 水文情报预报规范: GB/T 22482—2008[S]. 北京: 中国标准出版社, 2008.
[26]	齐文平, 王艳慧, 万圆, 等. 基于G-TOPSIS模型的贫困农户多目标发展评价: 以云南福贡县为例[J]. 地理科学进展, 2020, 39(6):1024-1036. DOI
[27]	杨莲梅, 关学锋, 张迎新. 亚洲中部干旱区降水异常的大气环流特征[J]. 干旱区研究, 2018, 35(2):249-259.

指标	各大气环流因子的相关系数值
指标	西藏高原 (30°~40°N, 75°~105°E)	北半球极涡中心强度 (JQ)	西藏高原 (25°~35°N, 80°~100°E)	太平洋副高北界 (110°E~115°W)	北美大西洋副高北界 (110°~20°W)		大西洋副高北界 (55°~25°W)	南海副高北界 (100°~120°E)
相关系数	0.885^**	0.866^**	0.863^**	0.821^**	0.764^**		0.753^**	0.732^**

指标	各大气环流因子的相关系数值
指标	西藏高原 (30°~40°N, 75°~105°E)	北半球极涡中心强度 (JQ)	西藏高原 (25°~35°N, 80°~100°E)	太平洋副高北界 (110°E~115°W)	北美大西洋副高北界 (110°~20°W)		大西洋副高北界 (55°~25°W)	南海副高北界 (100°~120°E)
相关系数	0.885^**	0.866^**	0.863^**	0.821^**	0.764^**		0.753^**	0.732^**

指标	各模型相关系数指标值
指标	1	2	3	4	5	6
R	0.820(a)	0.856(b)	0.869(c)	0.906(d)	0.912(e)	0.912(f)
R²	0.673	0.732	0.756	0.821	0.833	0.831
调整后R²	0.672	0.731	0.755	0.82	0.831	0.83
指标	各模型相关系数指标值
指标	7	8	9	10	11	12
R	0.914(g)	0.916(h)	0.918(i)	0.918(j)	0.920(k)	0.922(l)
R²	0.836	0.839	0.844	0.842	0.847	0.85
调整后R²	0.835	0.838	0.842	0.841	0.846	0.848

指标	各模型相关系数指标值
指标	1	2	3	4	5	6
R	0.820(a)	0.856(b)	0.869(c)	0.906(d)	0.912(e)	0.912(f)
R²	0.673	0.732	0.756	0.821	0.833	0.831
调整后R²	0.672	0.731	0.755	0.82	0.831	0.83
指标	各模型相关系数指标值
指标	7	8	9	10	11	12
R	0.914(g)	0.916(h)	0.918(i)	0.918(j)	0.920(k)	0.922(l)
R²	0.836	0.839	0.844	0.842	0.847	0.85
调整后R²	0.835	0.838	0.842	0.841	0.846	0.848

大气环流因子	MIV值
北美大西洋副高北界(110°~20°W)	0.011
印缅槽(15°~20°N,80°~100°E)	0.009 2
太平洋副高北界(110°E~115°W)	0.008 6
西藏高原(30°~40°N,75°~105°E	0.008 5
欧亚纬向环流指数(IZ,0°~150°E)	0.007 7
东亚槽强度(CQ)	0.007 6
大西洋欧洲环流型W	0.006 6
北美区极涡面积指数(3区,120°~30°W)	0.006 3