基于深度学习的区域生态安全时空模拟与预测

doi:10.3724/SP.J.1047.2017.00915

摘要/Abstract

摘要：

区域作为人类、自然、社会共同作用和互相影响的复杂系统,对区域进行生态量化建模与模拟仿真,是实现区域可持续发展战略的关键。传统机器学习方法对区域生态系统建模取得了一定的成果,但难以确定学习特征和实现时空模拟。深度学习不需事先确定训练特征,具有优异的特征学习能力,能够提高模型预测精度,因此利用深度学习进行建模具有显著优势。本文使用植被净初级生产力（NPP）、气溶胶光学厚度（AOD）和人口格网数据,充分利用深度学习的优点,采用最优深度神经网络时空模拟,得到了河南省2007-2014年3 km分辨率的生态赤字空间分布图和河南省2015-2020年的生态赤字时间预测结果并进行分析,为区域生态的科学管理和建设供科学依据和参考。

关键词: 深度学习, 机器学习, 生态承载力, 生态赤字, 时空模拟

Abstract:

A region is a complex system of human, nature and society. The quantitative modeling and simulation of the ecology of the region are the key to realize the strategy of regional sustainable development. Traditional methods of machine learning have made some achievements in the modeling of regional ecosystems, but it is difficult to determine the learning characteristics and realize the simulation of time and space. Deep learning does not need to determine the training characteristics and has excellent feature learning ability and higher accuracy of model prediction. In this paper, we used the net primary productivity (NPP), aerosol optical thickness (AOD) and population grid data to make full use of the advantages of depth learning. The optimal deep neural network is used to simulate the spatial and temporal patterns of Henan Province. The spatial distribution map of ecological deficit and the forecast of ecological deficit in Henan province from 2015 to 2020 are generated and analyzed. Our work provides relevant basic scientific support and reference for the scientific management and construction of regional ecology.

Key words: deep learning, machine learning, ecological capacity, ecological deficit, spatio-temporal simulation

陈工, 李琦, 金玲艳, 梁贺明, 莫玉琴. 基于深度学习的区域生态安全时空模拟与预测[J]. 地球信息科学学报, 2017, 19(7): 915-923.DOI:10.3724/SP.J.1047.2017.00915

CHEN Gong,LI Qi,JIN Lingyan,LIANG Heming,Hamed Karimian,MO Yuqin. Spatio-temporal Simulation and Prediction of Regional Ecological Security Based on Deep Learning[J]. Journal of Geo-information Science, 2017, 19(7): 915-923.DOI:10.3724/SP.J.1047.2017.00915

图/表 14

图1

表1

表2

图2

表3

图3

图4

图5

图7

图6

图8

图9

图10

图11

参考文献 49

[1]	Our Common Future[R].Our Common Future[R]. Report of the World Commission on Environment and Development, 1987.
[2]	彭斯震,孙新章.全球可持续发展报告:背景、进展与有关建议[J].中国人口·资源与环境. 2014,24(12):1-5. doi: 10.3969/j.issn.1002-2104.2014.12.001
	[ Peng S Z, Sun X Z.Global sustainability report: Background, progress and related proposals[J]. China Population, Resources and Environment, 2014,24(12):1-5. ] doi: 10.3969/j.issn.1002-2104.2014.12.001
[3]	新华网.习近平:生态文明建设是十三五规划重点方向[Z]. 2015:2016.
	[ Xi J P: Ecological civilization construction is the key direction of the thirteen five planning[Z]. 2015:2016. ]
[4]	张志强,孙成权,程国栋,等.可持续发展研究:进展与趋向[J].地球科学进展,1999,14(6):58. doi: 10.3321/j.issn:1001-8166.1999.06.011
	[ Zhang Z Q, Sun C Q, Cheng G D, et al.Research on sustainable development: progress and trend[J]. Advances in Earth Science, 1999,14(6):58. ] doi: 10.3321/j.issn:1001-8166.1999.06.011
[5]	Castaneda B E.An index of sustainable economic welfare (ISEW) for Chile[J]. Ecological Econonics, 1999,28(2):231-244. doi: 10.1016/S0921-8009(98)00037-8
[6]	Hard P.Measuring sustainable development: review of current practices[R]. 1997:591-595.
[7]	徐中民,程国栋,张志强.生态足迹方法:可持续性定量研究的新方法——以张掖地区1995年的生态足迹计算为例[J].生态学报,2001,21(9):1484-1493. doi: 10.3321/j.issn:1000-0933.2001.09.013
	[ Xu Z M, Cheng G D, Zhang Z Q.Ecological tracing method: A new method for quantitative research on sustainability: Taking the ecological footprint of Zhangye prefecture in 1995 as an example[J]. Acta Ecologica Sinica, 2001,21(9):1484-1493. ] doi: 10.3321/j.issn:1000-0933.2001.09.013
[8]	徐中民,张志强,程国栋.可持续发展定量研究的几种新方法评介[J].中国人口·资源与环境. 2000,10(2):61-65.
	[ Xu Z M, Zhang Z Q, Cheng G D.Evaluation of several new methods for quantitative research on sustainable development[J]. Chinese Journal of Resources and Environment, 2000,10(2):61-65. ]
[9]	Castaneda B E.An index of sustainable economic welfare (ISEW) for Chile[J]. Ecological Economics, 1999,28(2):231-244. doi: 10.1016/S0921-8009(98)00037-8
[10]	卢祥之. 评《1999中国可持续发展战略报告》[J].出版发行研究,1999(10):52.
	[ Lu X Z. Comment on "1999 China's Sustainable Development Strategy Report"[J]. Publishing and Research, 1999,10:52. ]
[11]	张中旺,郑伯红.区域可持续发展与人地系统协调研究[J].襄樊学院学报,2000(5):85-88.
	[ Zhang Z W, Zheng B H.Study on regional sustainable development and man-land system coordination[J]. Journal of Xiangfan University, 2000,5:85-88. ]
[12]	Ayres R U.Commentary on the utility of the ecological footprint concept[J]. Ecological Economics, 2000,32(3):347-349. doi: 10.1016/S0921-8009(99)00151-2
[13]	孙志远,鲁成祥.深度学习研究与进展[J].计算机科学,2016,43(2):31-38.
	[ Sun Z Y, Lu C X.Study and progress of deep learning[J]. Computer Science, 2016,43(2):31-38. ]
[14]	杨娟,王昌全,李冰,等.基于遗传神经网络的成都市人均生态足迹预测[J].生态学报,2009,29(1):359-367.
	[ Yang J, Wang C Q, Li B, et al.Prediction of per capita ecological footprint of Chengdu based on genetic neural network[J]. Journal of Ecology, 2009,29(1):359-367. ]
[15]	吴明,姚尧,贾冯睿,等.基于径向基函数神经网络方法的城市生态压力预测[J].自然资源学报,2013,28(2):328-335. doi: 10.11849/zrzyxb.2013.02.014
	[ Wu M, Yao Y, Jia F R, et al.State prediction of urban ecological pressure based on radial basis function neural network[J]. Journal of Natural Resources, 2013,28(2):328-335. ] doi: 10.11849/zrzyxb.2013.02.014
[16]	孙浩然,杨维,郭亚伟,等.基于GIS-BP神经网络的生态地质环境敏感性评价——以千山景区及周边为例[J].地下水,2009,31(6):126-128.
	[ Sun H R, Yang W, Guo Y W, et al.Assessment of the sensitivity of eco-geological environment based on GIS-BP neural network - Taking Qianshan scenic area and surrounding area as an example[J]. Journal of Groundwater, 2009,31(6):126-128.]
[17]	金新,唐德善,陈丽夫.基于GRNN神经网络的生态足迹影响因素预测[J].水电能源科学,2014(5):137-139.
	[ Jin X, Tang D S, Chen L F.Study on influencing factors of ecological footprint based on GRNN neural network[J]. International Journal Hydroelectric Energy, 2014,5:137-139. ]
[18]	Stoica C, Camejo J, Banciu A.Water quality of Danube Delta systems: ecological status and prediction using machine-learning algorithms[J]. Water Sci Technol. 2016,73(10):2413-2421. doi: 10.2166/wst.2016.097 pmid: 27191562
[19]	李开菊,郑波尽.浅谈深度学习[J].软件,2016,37(5):81-83.
	[ Li K J, Zheng B J.Talk about deep learning[J]. Software, 2016,37(5):81-83. ]
[20]	孙志军,薛磊,许阳明,等.深度学习研究综述[J].计算机应用研究,2012,29(8):2806-2810.
	[ Sun Z J, Xue L, Xu Y M, et al.Review of deep learn[J] .Application Research of Computers, 2012,29(8):2806-2810. ]
[21]	Le Cun Y, Bottou L, Bengio Y, et al.Gradient-based learning applied to document recognition[C]. Proc. IEEE, 1998. USA: IEEE, 1998:2278-2324.
[22]	Mo D.. A survey on deep learning: one small step toward AI[R]. .
[23]	朱新玲,黎鹏.基于BP神经网络的湖北省生态足迹拟合与预测研究[J].武汉科技大学学报(社会科学版),2015,17(1):77-80.
	[ Zhu X L, Li P.Study on ecological footprint fitting and prediction of Hubei province based on BP Neural Network[J]. Journal of Wuhan University of Science and Technology (Social Science Edition), 2015,17(1):77-80. ]
[24]	郭荣中,申海建,杨敏华. 基于灰色模型的长沙市生态足迹与生态承载力预测分析[J].水土保持研究,2015,22(4):195-200.
	[ Guo R Z, Shen H J, Yang M H.Prediction analysis of ecological footprint and ecological carrying capacity of Changsha city based on gray model[J]. Research of Soil and Water Conservation, 2015,22(4):195-200. ]
[25]	陈成忠,林振山,陈玲玲. 生态足迹与生态承载力非线性动力学分析[J].生态学报,2006,26(11):3812-3816. doi: 10.3321/j.issn:1000-0933.2006.11.038
	[ Chen C Z, Lin Z S, Chen L L.Nonlinear dynamics analysis of ecological footprint and ecological carrying capacity[J]. Acta Ecologica Sinica, 2006,26(11):3812-3816. ] doi: 10.3321/j.issn:1000-0933.2006.11.038
[26]	杨娟,王昌全,李冰,等.基于遗传神经网络的成都市人均生态足迹预测[J].生态学报,2009,3(1):359-367.
	[ Yang J, Wang C Q, Li B, et al.Prediction of per capita ecological footprint in Chengdu based on genetic neural network[J]. Acta Ecologica Sinica, 2009,3(1):359-367. ]
[27]	吴明,姚尧,贾冯睿,等.基于径向基函数神经网络方法的城市生态压力预测[J].自然资源学报,2013,28(2):328-335. doi: 10.11849/zrzyxb.2013.02.014
	[ Wu M, Yao Y, Jia F R, et al.State prediction of urban ecological pressure based on radial basis function neural network[J]. Journal of Natural Resources, 2013,28(2):328-335. ] doi: 10.11849/zrzyxb.2013.02.014
[28]	孙浩然,杨维,郭亚伟,等.基于GIS-BP神经网络的生态地质环境敏感性评价——以千山景区及周边为例[J].地下水,2009,31(6):126-128.
	[ Sun H R, Yang W, Guo Y W, etc. Based on GIS-BP neural network of ecological geological environment sensitivity evaluation - to Qianshan scenic area and the surrounding as an example[J]. Groundwater, 2009,31(6):126-128. ]
[29]	金新,唐德善,陈丽夫.基于GRNN神经网络的生态足迹影响因素预测[J].水电能源科学,2014(5):137-139.
	[ Jin X, Tang D S, Chen L F. Study on influencing factors of ecological footprint based on GRNN neural network[J]. Hydropower and Energy Science, 2014(5):137-139. ]
[30]	刘永超,郭怀成,代超.生态足迹预测模型优选——以武汉市为例[J].北京大学学报(自然科学版),2015,51(5):897-904.
	[ Liu Y C, Guo H C, Dai C.Ecological footprint forecasting model optimization - taking Wuhan city as an example[J]. Journal of Peking University (Natural Science Version, 2015,51(5):897-904. ]
[31]	肖和善. 植被净初级生产力模型研究[J].亚热带水土保持,2007,19(4):24-28.
	[ Xiao H S.Study on net primary productivity model of vegetation[J]. Subtropical Soil and Water Conservation, 2007,19(4):24-28. ]
[32]	Alfö Ldy. Aerosol optical depth, aerosol composition and air pollution during summer and winter conditions in Budapest[J]. Science of the Total Environment, 2007(383):141. doi: 10.1016/j.scitotenv.2007.04.037 pmid: 17570465
[33]	Jordan M I.Serial order: a parallel distributed processing approach[J]. Advances in Psychology, 1997,121:471-495.
[34]	Elman J L.Finding structure in time[J]. Cognitive science, 1990,14(2):179-211.
[35]	Hochreiter S, Schmidhuber J.Long short-term memory[J]. Neural Computation, 1997,9(8):1735-1780.
[36]	田垚,蔡猛,何亮.基于深度神经网络和Bottleneck 特征的说话人识别系统[J].清华大学学报(自然科学版),2012,49(7):2036-2043.
	[ Tian Y, Cai M, He L. 2012. Speaker recognition system based on depth neural network and bottleneck feature[J]. Journal of Tsinghua University (Science and Technology), 2012,49(7):2036-2043. ]
[37]	裴新富. 关于黄土高原范围问题[J].中国水土保持,1991(12):37-42.
	[ Pei X F.On the range of the Loess Plateau[J]. Chinese Journal of Soil and Water Conservation, 1991,12:37-42. ]
[38]	李雅兰. 河南省生态足迹变化及其影响因素研究[D].南京:南京农业大学,2010.
	[ Li Y L.Henan Province ecological footprint changes and its influencing factors[D]. Nanjing: Nanjing Agricultural University, 2010. ]
[39]	河南“十二五”投入巨资保护环境[J].城市道桥与防洪,2012(4):224-225.
	[ Henan "12th Five-Year" invested heavily in the protection of the environment[J]. Urban Road Bridge and Flood Control, 2012(4):224. ]
[40]	金玲燕. 基于CyberGIS_R的区域生态建模与时空预测研究[D].北京:北京大学,2016.
	[ Jin L Y.Study on regional ecological modeling and spatio-temporal prediction based on CyberGIS_R [D]. Beijing: Peking University, 2016. ]
[41]	商丘市政府.商丘市城市总体规划说明书(2005-2020)[R].商丘:商丘市政府,2004.
	[ Shangqiu City Government.Shangqiu City Master Plan Manual (2005-2020)[R]. Shangqiu: Shangqiu Municipal Government, 2004. ]
[42]	周口市政府,2009年周口市人民政府工作报告[R].周口: 周口市政府,2009.
	[ Zhoukou City Government, 2009 Zhoukou City People's Government Work Report [R]. Zhoukou: Zhoukou City Government, 2009. ]
[43]	“十二五”时期河南人口发展报告[R]. 河南:河南统计局,2016.
	[ "12th Five-Year" period Henan population development report [R]. Henan: Henan Bureau of Statistics, 2016. ]
[44]	商丘市政府.商丘城乡总体规划(2015-2030)[R].商丘:商丘市政府,2014.
	[ Shangqiu city government. Shangqiu urban and rural master plan (2015-2030)[R]. Shangqiu: Shangqiu city government, 2014. ]
[45]	商丘市政府.2014年商丘市政府工作报告[R].商丘:商丘市政府,2014.
	[ Shangqiu municipal government. 2014 Shangqiu municipal government work report [R]. Shangqiu: Shangqiu city government, 2014. ]
[46]	何亚芬. 基于格网的南昌市土地生态承载力评价[D].江西:江西农业大学,2015.
	[ He Y F.Grid-based evaluation of land ecological carrying capacity in Nanchang[J]. Jiangxi Agricultural University, 2015. ]
[47]	程国建,颜宇甲.基于多Agent的生态复杂适应系统建模和仿真[J].西安石油大学学报(自然科学版),2011,26(2):134-148.
	[ Cheng G J, Yan Y J.Modeling and simulation of ecological complex adaptive system based on multi-Agent[J]. Journal of Xi'an Shiyou University (Natural Science Edition),2011,26(2):134-148. ]
[48]	岳大鹏,张露露.河南省2000-2007年人均生态足迹动态变化及其驱动力分析[J].资源开发与市场,2010,13(7): 211-216.
	[ Yue D P, Zhang L L.Dynamic change of per capita ecological footprint and its driving force analysis in Henan Province from 2000 to 2007[J]. Resource Development and Market, 2010,13(7):211-216. ]
[49]	郑林昌,齐蒙.河北省贸易隐含二氧化碳排放及其影响因素研究[J].气候变化研究进展,2017,13(12):201-205. doi: 10.12006/j.issn.1673-1719.2016.126
	[ Zheng L C, Qi M.Study on implied carbon dioxide emissions and its influencing factors in Hebei province[J]. Progress in Climate Change Research, 2017,13(12):201-205. ] doi: 10.12006/j.issn.1673-1719.2016.126

神经元数	1	3	10	30	50	100
T_Loss	0.1325	0.1239	0.0097	0.0298	0.0268	0.0268
VAL_LOSS	0.1861	0.1821	0.1821	0.0327	0.0265	0.0027