基于MODIS遥感产品和神经网络模拟太阳辐射

doi:10.13249/j.cnki.sgs.2017.06.013

摘要/Abstract

摘要：

现有的神经网络模拟太阳辐射的模型很少考虑云、气溶胶、水汽对太阳辐射的影响,采用MODIS提供的气溶胶、云、水汽高空大气遥感产品和常规气象数据,输入LM（Levenberg-Marquardt）算法优化后的BP（Back-Propagation）神经网络模型（简称LM-BP）模拟了和田、西宁、固原、延安4个辐射站点的太阳辐射月均值。验证结果表明：神经网络模型中加入气溶胶、云、水汽之后,4个辐射站点的 $R 2$ 均大于0.90,且各项误差指标均小于仅用常规气象站点数据模拟的太阳辐射结果。

关键词: 太阳辐射, MODIS, 神经网络, LM-BP, 云, 气溶胶, 水汽

Abstract:

Climate change is a major global issue of common concern of the international community, over the past century, the earth experienced a temperature rise, while solar radiation is an indicator of climate change. At the same time, solar radiation data is an important parameter about crop models, hydrological models and climate change models, many Artificial Neural Network ensemble models are developed to estimate solar radiation using routinely measured meteorolological variables, but it do not consider cloud, aerosol, and water vapor influence on solar radiation. In this article, we use cloud, aerosols, atmospheric precipitable water vapor from MODIS atmosphere remote sensing products and conventional meteorological data including air pressure, Air temperature, vapor pressure, relative humidity, and sunshine duration, and we analyze the relationship between solar radiation and meteorological data. In terms of conventional meteorological data, we make the selection of variables, the redundant variables are proposed. Then, BP artificial neural network model optimized by LM (Levenberg-Marquardt) algorithm (referred to as LM-BP) is used to stimulate solar radiation. This LM algorithm has fast local convergence feature about Gauss-Newton method, but also has global search feature about gradient descent method, which allows error along the direction of deterioration to search, and greatly improving the convergence rate and generalization ability of the network. Therefore, this article use LM-BP model to predict monthly mean daily global solar radiation from 2010 to 2013 about Hetian, Xining, Guyuan, Yan’an radiating station using only conventional meteorological data(referred to as A) and using MODIS atmosphere remote sensing products binding conventional meteorological data(referred to as $A +$ ) respectively. Then, we validate performance of the model with measured data about radiation station. The results show that cloud amount, cloud optical thickness, aerosol optical depth, and atmospheric precipitable water vapor these factors are added to the established model, the degree of matching simulated solar radiation and actual observations is more higher. And correlation determination (R²) for 4 radiation station are 0.90 or higher, while error indicators are small. This article showed that the use of LM-BP neural network model, combining with remote sensing data and conventional meteorological data to simulate solar radiation is a reasonable and effective way to simulate solar radiation.

Key words: solar radiation, MODIS, neural network, LM-BP, cloud, aerosol, water vapor

中图分类号:

P422.1

李净, 王丹, 冯姣姣. 基于MODIS遥感产品和神经网络模拟太阳辐射[J]. 地理科学, 2017, 37(6): 912-919.

Jing Li, Dan Wang, Jiaojiao Feng. Simulation of Solar Radiation Based on Neural Network and MODIS Remote Sensing Products[J]. SCIENTIA GEOGRAPHICA SINICA, 2017, 37(6): 912-919.

图/表 7

图1

表1

表2

表3

图2

图3

表4

误差指标统计表"

辐射站点	年份	$RMSE$ （MJ/m²）		MBE（MJ/m²）		MAPE（%）
辐射站点	年份	A	A+	A	A+	A	A+
西宁	2010	1.14	0.74	0.42	0.23	4.53	2.81
	2011	1.98	1.15	0.19	-0.27	10.64	5.94
	2012	1.13	0.82	0.8	0.44	5.81	4.42
	2013	1.08	1.04	0.55	0.47	5.17	5.09
和田	2010	1.51	1.37	-0.17	0.41	7	5.96
	2011	1.47	1.29	0.17	0.36	6.3	5.52
	2012	1.43	1.04	0.72	0.28	6.06	4.39
	2013	0.87	0.77	0.02	-0.11	3.69	0.3
固原	2010	1.3	0.78	-0.05	-0.2	6.77	3.92
	2011	1.21	1.04	-0.82	-0.27	6.71	5.74
	2012	1.68	1.3	1.17	0.21	8.27	7.31
	2013	1.15	1.07	0.13	-0.2	6.52	5.44
延安	2010	1.74	1.17	-1.41	-0.79	11.64	6.61
	2011	1.15	0.59	-0.88	-0.24	6.54	3.47
	2012	1.43	0.93	-0.88	-0.21	7.07	5.12
	2013	2.34	1.69	1.15	0.42	15.89	9.62

表4

参考文献 30

[1]	Lu Ning, Qin Jun, Yang Kun et al. A simple and efficient algorithm to estimate daily global solar radiation from geostationary satellite data[J]. Energy, 2011, 36(5): 3179-3188. doi: 10.1016/j.energy.2011.03.007
[2]	施国萍, 邱新法, 曾燕. 中国三种太阳辐射起始数据分布式模拟[J]. 地理科学, 2013, 33(4): 385-392.
	[Shi Guoping, Qiu Xinfa, Zeng Yan.Distributed Modeling of Three Kinds of Initial Data for Global Solar Radiation Simulation in China. Scientia Geographica Sinica, 2013, 33(4): 385-392.]
[3]	查良松. 我国地面太阳辐射量的时空变化研究[J]. 地理科学,1996, 16(3): 41-46.
	[Zha Liangsong.A Study on Spatial andTemporal Variation of Solar Radiation in China. Scientia Geographica Sinica, 1996, 16(3): 41-46.]
[4]	赵东, 罗勇, 高歌, 等. 我国近50年来太阳直接辐射资源基本特征及其变化[J]. 太阳能学报, 2009, 30(7): 946-952. doi: 10.3321/j.issn:0254-0096.2009.07.018
	[Zhao Dong, Luo Yong, Gao Ge et al.The basic characteristics and change of direct solar radiation resource in recent 50 years in China. Acta Energiae Solaris Sinica, 2009, 30(7): 946-952.] doi: 10.3321/j.issn:0254-0096.2009.07.018
[5]	曾燕, 邱新法, 刘昌明, 等. 起伏地形下黄河流域太阳直接辐射分布式模拟[J]. 地理学报, 2005, 60(4): 680-688. doi: 10.3321/j.issn:0375-5444.2005.04.017
	[Zeng Yan, Qiu Xinfa, Liu Changming et al. Distributed Model of Direct solar radiation of rugged terrain over Yellow River Basin. Acta Geographica Sinica, 2005, 60(4): 680-688.] doi: 10.3321/j.issn:0375-5444.2005.04.017
[6]	Chen Jilong, Xiao Beibei, Chen Chundi et al. Estimation of monthly-mean global solar radiation using MODIS atmospheric product over China[J]. Journal of Atmospheric and Solar-Terrestrial Physics, 2014, 110: 63-80. doi: 10.1016/j.jastp.2014.01.017
[7]	Sun Huaiwei, Zhao Na, Zeng Xiaofa et al. Study of solar radiation prediction and modeling of relationships between solar radiation and meteorological variables[J]. Energy Conversion and Management,2015, 105: 880-890. doi: 10.1016/j.enconman.2015.08.045
[8]	Teke A, Yıldırım H B, Çelik Ö.Evaluation and performance comparison of different models for the estimation of solar radiation[J]. Renewable and Sustainable Energy Reviews, 2015, 50: 1097-1107. doi: 10.1016/j.rser.2015.05.049
[9]	Wu Ji, Chan Chee keong, Zhang Yu et al. Prediction of solar radiation with genetic approach combing multi-model framework[J]. Renewable Energy, 2014, 66: 132-139. doi: 10.1016/j.renene.2013.11.064
[10]	Yao Wanxiang, Li Zhengrong, XiuTongbin et al. New decomposition models to estimate hourly global solar radiation from the daily value[J]. Solar Energy, 2015, 120: 87-99. doi: 10.1016/j.solener.2015.05.038
[11]	Manzano A, Martín M L, Valero F et al. A single method to estimate the daily global solar radiation from monthly data[J]. Atmospheric Research, 2015, 166: 70-82. doi: 10.1016/j.atmosres.2015.06.017
[12]	和清华,谢云. 我国太阳总辐射气候学计算方法研究[J]. 自然资源学报,2010,25(2):308-319.
	[He Qinghua, Xie Yun.Research on the Climatological Calculation Method of Solar Radiation in China. Journal of Natural Resources, 2010, 25(2): 308-319.]
[13]	Shamim M A, Remesan R, Bray M et al. An improved technique for global solar radiation estimation using numerical weather prediction[J]. Journal of Atmospheric and Solar-Terrestrial Physics, 2015, 129: 13-22. doi: 10.1016/j.jastp.2015.03.011
[14]	周开利, 康耀红. 神经网络模型及其MATLAB 仿真程序设计[M]. 北京: 清华大学出版社, 2005.
	[Zhou Kaili, Kang Yaohong.Neural Network Model and MATLAB programming. Beijing: Tsinghua University Press, 2005.]
[15]	Lam J C, Wan K K W, Yang L. Solar radiation modelling using ANNs for different climates in China[J]. Energy Conversion and Management, 2008, 49(5): 1080-1090. doi: 10.1016/j.enconman.2007.09.021
[16]	Wan K K W, Tang H L, Yang Let al. An analysis of thermal and solar zone radiation models using an Angstrom-Prescott equation and artificial neural networks[J]. Energy, 2008, 33(7): 1115-1127. doi: 10.1016/j.energy.2008.01.015
[17]	Ozgoren M, Bilgili M, Sahin B.Estimation of global solar radiation using ANN over Turkey[J]. Expert Systems with Applications, 2012, 39(5): 5043-5051. doi: 10.1016/j.eswa.2011.11.036
[18]	Şenkal O.Modeling of solar radiation using remote sensing and artificial neural network in Turkey[J]. Energy, 2010, 35(12): 4795-4801. doi: 10.1016/j.energy.2010.09.009
[19]	曹双华, 曹家枞, 刘凤强. 小波分析在太阳辐射神经网络预测中的应用研究[J]. 东华大学学报: 自然科学版, 2004, 30(6): 18-22. doi: 10.3969/j.issn.1671-0444.2004.06.005
	[Cao Shuanghua, CaoJiacong, Liu Fengqiang. Study of Application of Wavelet Analysis to Neural Networks for the Forecast of Solar Irradiance. Journal of Donghua University (Natural Science), 2004, 30(6):18-22.] doi: 10.3969/j.issn.1671-0444.2004.06.005
[20]	李净, 王卫东, 张福存. 利用LM-BP神经网络估算西北地区太阳辐射[J]. 干旱区地理, 2015, 38(3): 438-445.
	[Li Jing, Wang Weidong,Zhang Fucun.Simulation of solar radiation in northwest China based on LM-BP neural network. Arid Land Geography. 2015, 38(3): 438-445.]
[21]	李净, 冯姣姣, 王卫东, 等. 基于LM-BP神经网络的西北地区太阳辐射时空变化研究[J]. 地理科学, 2016, 36(5): 780-786.
	[Li Jing, Feng Jiaojiao, Wang Weidong et al. Spatial and Temporal Changes in Solar Radiation of Northwest China Based LM-BP Neural Network. Scientia Geographica Sinica, 2016, 36(5): 780-786.]
[22]	王卫东, 李净, 张福存, 等. 基于BP神经网络的太阳辐射预测——以兰州市为例[J]. 干旱区资源与环境, 2014, 28(2): 185-189. doi: 10.3969/j.issn.1003-7578.2014.02.032
	[ Wang Weidong, Li Jing, Zhang Fucun et al. Simulation of solar radiation in Lanzhou based on BP neural network. Journal of Arid Land Resource and Environment, 2014, 28(2): 185-189.] doi: 10.3969/j.issn.1003-7578.2014.02.032
[23]	高隽. 人工神经网络原理及仿真实例[M]. 北京: 机械工业出版社, 2003.
	[Gao Jun.Artificial neural network theory and simulation examples. Beijing: Machinery Industry Press, 2003.]
[24]	苏安玉, 李衡, 濮励杰, 等. 基于RAGA-BP神经网络模型的三江平原地下水资源预测研究[J]. 地理科学, 2009, 29(2): 283-287. doi: 10.3969/j.issn.1000-0690.2009.02.023
	[Su Anyu, Li Heng, Pu Lijie et al. Evaluation of Ground water Resources Based on RAGA-BP Neural Networks in the Sanjiang Plain. Scientia Geographica Sinica, 2009, 29(2): 283-287] doi: 10.3969/j.issn.1000-0690.2009.02.023
[25]	赵玲玲, 翁苏明, 曾华军, 等. 模式分析的核方法[M]. 北京: 机械工业出版社, 2006.
	[Zhao Lingling, Weng Suming, Zeng Huajun et al. Kernel methods for pattern analysis. Beijing: China Machine Press, 2006.]
[26]	宋宇辰, 甄莎. BP神经网络和时间序列模型在包头市空气质量预测中的应用[J]. 干旱区资源与环境, 2013, 27(7): 65-70.
	[Song Yuchen, Zhen Sha.The application of BP neural network and time series model to air quality forecast for Baotou. Journal of Arid Land Resource and Environment,2013, 27(7): 65-70.]
[27]	王海军, 蒲晓勇, 杨志彪, 等. 影响人工站与自动站水汽压差值原因探究[J]. 气象科技, 2012, 40(1): 25-30.
	[Wang Haijun, Pu Xiaoyong, Yang Zhibiao et al. Causal study of differences in water vapor pressure measurements between manual and automatic weather stations. Meteorological Science and Technology, 2012, 40(1): 25-30.]
[28]	汪凯, 叶红, 陈峰, 等. 中国东南部太阳辐射变化特征、影响因素及其对区域气候的影响[J]. 生态环境学报, 2010, 19(5): 1119-1124. doi: 10.3969/j.issn.1674-5906.2010.05.023
	[Wang Kai, Ye Hong, Chen Feng et al. Long-term change of solar radiation in southeastern China: Variation, factors,and climate forcing. Ecology and Environmental Sciences, 2010, 19(5): 1119-1124.] doi: 10.3969/j.issn.1674-5906.2010.05.023
[29]	López G, Batlles F J.Estimating solar radiation from MODIS data[J]. Energy Procedia, 2014, 49: 2362-2369. doi: 10.1016/j.egypro.2014.03.250
[30]	尹青, 张华, 何金海. 近 48 年华东地区地面太阳总辐射变化特征和影响因子分析[J]. 大气与环境光学学报, 2011, 6(1): 37-46.
	[Yin Qing, Zhang Hua, He Jinhai.Long-term change of surface total solar radiation and influencing factors over east China in recent 48 years. Journal of Atmospheric and Environmental Optics,2011, 6(1): 37-46.]

Rp	S（MJ/m²）	P（hPa）	T（℃）	P_w（hPa）	H_r（%）	S₀（h）	H₀（MJ/m²）
S（MJ/m²）	1.00	-0.11	0.80	0.48	-0.40	0.63	0.20
P（hPa）	-0.11	1.00	0.17	0.12	-0.14	0.06	-0.02
T₁（℃）	0.80	0.17	1.00	0.83	-0.07	0.42	0.14
P_w（hPa）	0.48	0.12	0.83	1.00	0.43	0.11	0.06
H_r（%）	-0.40	-0.14	-0.07	0.43	1.00	-0.53	-0.11
S₀（h）	0.63	0.06	0.42	0.11	-0.53	1.00	0.01
H₀（MJ/m²）	0.20	-0.02	0.14	0.06	0.11	0.01	1.00

站点	输入层	隐含层	输出层	误差限	训练函数	训练误差	迭代次数	实际迭代次数
和田	P、T₁、H_r、S₀	10	S	0.005	trainlm	0.0048	2000	102
和田	P、T₁、H_r、S₀、CF、COT、AOT、PWV	10	S	0.005	trainlm	0.0031	2000	106
西宁	P、T₁、H_r、S₀	10	S	0.005	trainlm	0.0044	2000	127
西宁	P、T₁、H_r、S₀、CF、COT、AOT、PWV	10	S	0.005	trainlm	0.0029	2000	126
固原	P、T₁、H_r、S₀	10	S	0.005	trainlm	0.0036	2000	118
固原	P、T₁、H_r、S₀、CF、COT、AOT、PWV	10	S	0.005	trainlm	0.0032	2000	125
延安	P、T₁、H_r、S₀	10	S	0.005	trainlm	0.0050	2000	156
延安	P、T₁、H_r、S₀、CF、COT、AOT、PWV	10	S	0.005	trainlm	0.0045	2000	180