基于随机森林模型的珠江三角洲30 m格网人口空间化

doi:10.18306/dlkxjz.2017.10.012

摘要/Abstract

摘要：

人口空间化是实现人口统计数据与其他环境资源空间数据融合分析的有效途径。本文选取夜间灯光数据、道路网数据、水域分布数据、建成区数据、数字高程模型和地形坡度数据作为影响珠江三角洲人口分布的变量因子,利用随机森林模型对珠江三角洲2010年人口数据进行了30 m格网空间化,并将模拟结果与三个公开数据集作精度对比,最后基于随机森林模型的变量因子重要性分析珠江三角洲人口空间分布的影响因素。结果表明：本文模拟整体精度达到82.32%,均优于WorldPop数据集以及中国公里网格人口数据集,接近GPW数据集,而且在人口密度中等区域模拟精度最高;通过对变量因子重要性进行度量,发现夜间灯光强度是珠江三角洲人口分布的最重要指示性指标,到水域的距离、到建成区的距离和路网密度对珠江三角洲人口分布均具有重要作用。利用随机森林模型结合多源信息能够实现高空间分辨率的人口空间化,可为精细化城市管理提供重要数据源,也可为相关政策决策制定提供支持。

关键词: 人口空间化, 随机森林, 人口分布, 影响因素, 珠江三角洲

Abstract:

Grid population data can enable integrated analysis of population statistics with other spatial data on resources and the environment. Based on a Random Forest model and using nighttime lights, road network, surface water network, built-up area, slope, and DEM as control variables, the 2010 population data of the Pearl River Delta were distributed into 30 m grids. The estimation results were compared with three other public datasets. The importance of input variables was analyzed based on the results. The result shows that the accuracy of this simulation reached 83.32%, which is better than the WorldPop and the Population Grids of China datasets, and more close to the GPW dataset. Moreover, the 30 m resolution of our result furnishes detailed information of population density of the Pearl River Delta. According to the importance of covariates from the Random Forest model, strength of nighttime lights, distance to water, distance to built-up area, and density of roads are important factors in population distribution modeling in the Pearl River Delta. With the Random Forest model and multi-source data, high resolution population spatialization can be achieved. High spatial resolution grid data can provide important data source for high precision city management and policy making.

Key words: population spatialization, random forest, population distribution, impact factors, the Pearl River Delta

谭敏, 刘凯, 柳林, 朱远辉, 王大山. 基于随机森林模型的珠江三角洲30 m格网人口空间化[J]. 地理科学进展, 2017, 36(10): 1304-1312.

Min TAN, Kai LIU, Lin LIU, Yuanhui ZHU, Dashan WANG. Spatialization of population in the Pearl River Delta in 30 m grids using random forest model[J]. PROGRESS IN GEOGRAPHY, 2017, 36(10): 1304-1312.

图/表 9

图1

图2

图3

图4

图5

表1

图6

图7

图8

参考文献 29

[1]	柏中强, 王卷乐, 姜浩, 等. 2015. 基于多源信息的人口分布格网化方法研究[J]. 地球信息科学学报, 17(6): 653-660. doi: 10.3724/SP.J.1047.2015.00653
	[Bai Z Q, Wang J L, Jiang H, et al.2015. The gridding approach to redistribute population based on multi-source data[J]. Journal of Geo-Information Science, 17(6): 653-660.] doi: 10.3724/SP.J.1047.2015.00653
[2]	柏中强, 王卷乐, 杨雅萍, 等. 2015. 基于乡镇尺度的中国25省区人口分布特征及影响因素[J]. 地理学报, 70(8): 1229-1242.
	[Bai Z Q, Wang J L, Yang Y P, et al.2015. Characterizing spatial patterns of population distribution at township level across the 25 provinces in China[J]. Acta Geographica Sinica, 70(8): 1229-1242.]
[3]	陈凯, 刘凯, 柳林, 等. 2015. 基于随机森林的元胞自动机城市扩展模拟: 以佛山市为例[J]. 地理科学进展, 34(8): 937-946.
	[Chen K, Liu K, Liu L, et al.2015. Urban expansion simulation by random-forest-based cellular automata: A case study of Foshan City[J]. Progress in Geography, 34(8): 937-946.]
[4]	陈晴, 侯西勇, 吴莉. 2014. 基于土地利用数据和夜间灯光数据的人口空间化模型对比分析: 以黄河三角洲高效生态经济区为例[J]. 人文地理, 29(5): 94-100.
	[Chen Q, Hou X Y, Wu L.2014. Comparing of population spatialization models based on land use data and DMSP/OLS data respectively: A case study in the efficient ecological economic zone of the Yellow River Delta[J]. Human Geography, 29(5): 94-100.]
[5]	方匡南, 吴见彬, 朱建平, 等. 2011. 随机森林方法研究综述[J]. 统计与信息论坛, 26(3): 32-38. doi: 10.3969/j.issn.1007-3116.2011.03.006
	[Fang K N, Wu J B, Zhu J P, et al.2011. A review of technologies on random forests[J]. Statistics & Information Forum, 26(3): 32-38.] doi: 10.3969/j.issn.1007-3116.2011.03.006
[6]	方瑜, 欧阳志云, 郑华, 等. 2012. 中国人口分布的自然成因[J]. 应用生态学报, 23(12): 3488-3495.
	[Fang Y, Ouyang Z Y, Zheng H, et al.2012. Natural forming causes of China population distribution[J]. Chinese Journal of Applied Ecology, 23(12): 3488-3495.]
[7]	高义, 王辉, 王培涛, 等. 2013. 基于人口普查与多源夜间灯光数据的海岸带人口空间化分析[J]. 资源科学, 35(12): 2517-2523.
	[Gao Y, Wang H, Wang P T, et al.2013. Population spatial processing for Chinese coastal zones based on census and multiple night light data[J]. Resources Science, 35(12): 2517-2523.]
[8]	康停军, 张新长, 赵元, 等. 2012. 基于多智能体的城市人口分布模型[J]. 地理科学, 32(7): 790-797.
	[Kang T J, Zhang X C, Zhao Y, et al.2012. Agent-based urban population distribution model[J]. Scientia Geographica Sinica, 32(7): 790-797.]
[9]	李玲, 沈静, 袁媛. 2008. 人口发展与区域规划[M]. 北京: 科学出版社.
	[Li L, Shen J, Yuan Y.2008. Renkou fazhan yu quyu guihua[M]. Beijing, China: Science Press.]
[10]	廖顺宝, 孙九林. 2003. 基于GIS的青藏高原人口统计数据空间化[J]. 地理学报, 58(1): 25-33. doi: 10.3321/j.issn:0375-5444.2003.01.004
	[Liao S B, Sun J L.2003. GIS based spatialization of population census data in Qinghai-Tibet Plateau[J]. Acta Geographica Sinica, 58(1): 25-33.] doi: 10.3321/j.issn:0375-5444.2003.01.004
[11]	刘志佳, 黄河清. 2015. 珠三角地区建设用地扩张与经济、人口变化之间相互作用的时空演变特征分析[J]. 资源科学, 37(7): 1394-1402.
	[Liu Z J, Huang H Q.2015. Tempo-spatial characteristics of interactions among changes in built-up land, GDP and demography in the Pearl River Delta[J]. Resources Science, 37(7): 1394-1402.]
[12]	苏飞, 张平宇. 2010. 辽中南城市群人口分布的时空演变特征[J]. 地理科学进展, 29(1): 96-102. doi: 10.11820/dlkxjz.2010.01.013
	[Su F, Zhang P Y.2010. Spatio-temporal dynamics of population distribution in the middle and southern Liaoning Urban Agglomeration[J]. Progress in Geography, 29(1): 96-102.] doi: 10.11820/dlkxjz.2010.01.013
[13]	王鹤饶, 郑新奇, 袁涛. 2012. DMSP/OLS数据应用研究综述[J]. 地理科学进展, 31(1): 11-19. doi: 10.11820/dlkxjz.2012.01.002
	[Wang H R, Zheng X Q, Yuan T.2012. Overview of researches based on DMSP/OLS nighttime light data[J]. Progress in Geography, 31(1): 11-19.] doi: 10.11820/dlkxjz.2012.01.002
[14]	王静, 杨小唤, 石瑞香. 2012. 山东省人口空间分布格局的多尺度分析[J]. 地理科学进展, 31(2): 176-182. doi: 10.11820/dlkxjz.2012.02.006
	[Wang J, Yang X H, Shi R X.2012. Spatial distribution of the population in Shandong Province at multi-scales[J]. Progress in Geography, 31(2): 176-182.] doi: 10.11820/dlkxjz.2012.02.006
[15]	王珂靖, 蔡红艳, 杨小唤. 2016. 多元统计回归及地理加权回归方法在多尺度人口空间化研究中的应用[J]. 地理科学进展, 35(12): 1494-1505.
	[Wang K J, Cai H Y, Yang X H.2016. Multiple scale spatialization of demographic data with multi-factor linear regression and geographically weighted regression models[J]. Progress in Geography, 35(12): 1494-1505.]
[16]	肖荣波, 丁琛. 2011. 城市规划中人口空间分布模拟方法研究[J]. 中国人口·资源与环境, 21(6): 13-18.
	[Xiao R B, Ding C.2011. Modeling spatial distribution of population density for urban planning[J]. China Population Resources and Environment, 21(6): 13-18.]
[17]	张建辰, 王艳慧. 2014. 基于土地利用类型的村级人口空间分布模拟: 以湖北鹤峰县为例[J]. 地球信息科学学报, 16(3): 435-442. doi: 10.3724/SP.J.1047.2014.00435
	[Zhang J C, Wang Y H.2014. Simulation of village-level population distribution based on land use: A case study of Hefeng County in Hubei Province[J]. Journal of Geo-Information Science, 16(3): 435-442.] doi: 10.3724/SP.J.1047.2014.00435
[18]	张雷, 王琳琳, 张旭东, 等. 2014. 随机森林算法基本思想及其在生态学中的应用: 以云南松分布模拟为例[J]. 生态学报, 34(3): 650-659.
	[Zhang L, Wang L L, Zhang X D, et al.2014. The basic principle of random forest and its applications in ecology: A case study of Pinus yunnanensis[J]. Acta Ecologica Sinica, 34(3): 650-659.]
[19]	周成虎, 欧阳, 马廷. 2009. 地理格网模型研究进展[J]. 地理科学进展, 28(5): 657-662. doi: 10.11820/dlkxjz.2009.05.002
	[Zhou C H, Ou Y, Ma T.2009. Progresses of geographical grid systems researches[J]. Progress in Geography, 28(5): 657-662.] doi: 10.11820/dlkxjz.2009.05.002
[20]	周春山, 金万富, 史晨怡. 2015. 新时期珠江三角洲城市群发展战略的思考[J]. 地理科学进展, 34(3): 302-312.
	[Zhou C S, Jin W F, Shi C Y.2015. Development strategy of the Pearl River Delta Urban Agglomeration under the current socioeconomic situation[J]. Progress in Geography, 34(3): 302-312.]
[21]	珠江三角洲城市群年鉴编纂委员会. 2015. 珠江三角洲城市群年鉴2015[M]. 广州: 广东人民出版社.
	[Urban Agglomeration in the Pearl River Delta Yearbook Committee Compilation. 2015. Urban Agglomeration in the Pearl River Delta yearbook 2015[M]. Guangzhou, China: Guangdong People's Publishing House.]
[22]	卓莉, 黄信锐, 陶海燕, 等. 2014. 基于多智能体模型与建筑物信息的高空间分辨率人口分布模拟[J]. 地理研究, 33(3): 520-531. doi: 10.11821/dlyj201403011
	[Zhuo L, Huang X R, Tao H Y, et al.2014. High spatial resolution population distribution simulation based on building information and multi-agent[J]. Geographical Research, 33(3): 520-531.] doi: 10.11821/dlyj201403011
[23]	Breiman L.2001a. Random forests[J]. Machine Learning, 45(1): 5-32. doi: 10.1023/A:1010933404324
[24]	Breiman L.2001b. Statistical modeling: The two cultures[J]. Statistical Science, 16(3): 199-231.
[25]	Gaughan A E, Stevens F R, Huang Z J, et al.2016. Spatiotemporal patterns of population in mainland China, 1990 to 2010[J]. Scientific Data, 3: 160005. doi: 10.1038/sdata.2016.5 pmid: 26881418
[26]	Martin D.1989. Mapping population data from zone centroid locations[J]. Transactions of the Institute of British Geographers, 14(1): 90-97. doi: 10.2307/622344 pmid: 12281689
[27]	Qi W, Liu S H, Gao X L, et al.2015. Modeling the spatial distribution of urban population during the daytime and at night based on land use: A case study in Beijing, China[J]. Journal of Geographical Sciences, 25(6): 756-768. doi: 10.1007/s11442-015-1200-0
[28]	Stevens F R, Gaughan A E, Linard C, et al.2015. Disaggregating census data for population mapping using random forests with remotely-sensed and ancillary data[J]. PLoS One, 10(2): e0107042. doi: 10.1371/journal.pone.0107042 pmid: 25689585
[29]	Taylor K E.2001. Summarizing multiple aspects of model performance in a single diagram[J]. Journal of Geophysical Research, 106(D7): 7183-7192. doi: 10.1029/2000JD900719

	数据集	MAE	RMSE	%RMSE/%
广州区域	本文	7146.73	13530.23	17.68
	WorldPop	9599.64	17597.13	23.00
	GPW	7272.93	13198.29	17.25
	中国公里网格人口数据集	13396.08	22551.34	29.73
人口密度低	本文	1374.80	2815.25	151.98
	WorldPop	1300.97	2020.59	107.27
	GPW	753.38	1170.74	63.09
	中国公里网格人口数据集	767.42	1101.52	58.02
人口密度中等	本文	7221.09	10087.84	45.93
	WorldPop	10068.19	13889.55	63.26
	GPW	10663.92	15526.47	70.69
	中国公里网格人口数据集	18472.54	22538.66	104.95
人口密度高	本文	36823.59	39434.55	58.09
	WorldPop	49972.75	50613.87	74.56
	GPW	24609.52	28022.69	41.46
	中国公里网格人口数据集	61669.22	62285.80	93.20