Please wait a minute...
img

官方微信

遥感技术与应用  2020, Vol. 35 Issue (4): 943-949    DOI: 10.11873/j.issn.1004-0323.2020.4.0943
遥感应用     
数据驱动的植被总初级生产力估算方法研究
张坤1(),刘乃文2(),高帅3,赵书慧1
1.山东师范大学信息科学与工程学院,山东 济南 250358
2.山东管理学院信息工程学院省高校重点实验室,山东 济南 250357
3.中国科学院遥感与数字地球研究所 遥感科学国家重点实验室,北京 100101
Data-Driven Estimation of Gross Primary Production
Kun Zhang1(),Naiwen Liu2(),Shuai Gao3,Shuhui Zhao1
1.School of Information Science & Engineering,Shandong Normal University, Jinan 250358, China
2.Provincial Key Laboratory of Information Technology, School of Information Engineering, Shandong Management University, Jinan 250357, China
3.The State Key Laboratory of Remote Sensing Science, Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing 100101, China
 全文: PDF(1723 KB)   HTML
摘要:

植被总初级生产力(Gross Primary Production,GPP)是指在单位时间和单位面积上,绿色植物通过光合作用固定二氧化碳所产生的全部有机物同化量,对GPP的准确估算有助于碳循环的研究。为了提高GPP的估算精度,将机器学习技术与遥感技术相结合,首先利用GEE平台下的遥感数据以及中国陆地生态系统通量观测研究网络的通量塔实测GPP数据,建立数据集。然后使用随机森林作为估算模型,建模后根据数据特点对模型调参。最后获得模型的预测结果,决定系数R2为0.87,均方根误差RMSE的值为1.132 gC·m-2·d-1。这说明随机森林模型可以较为精确地估算GPP。结果发现,以大数据以及人工智能为代表的计算机技术飞速发展,将为遥感技术注入新的活力,使遥感技术走向更加成熟的发展应用阶段。

关键词: 随机森林模型碳循环GPP大数据GEE    
Abstract:

Gross Primary Production (GPP) of vegetation refers to the assimilation of all organic matter produced by green plants through photosynthesis and fixed carbon dioxide per unit time and unit area. Accurate estimation of GPP is helpful for the study of carbon cycle. In order to improve the estimation accuracy of GPP, this study combines machine learning technology and remote sensing technology. First, the remote sensing data under the GEE platform and the flux tower measurement data of the China Terrestrial Ecosystem Flux Observation Research Network are used to establish a data set. Then use random forest as the estimation model, and adjust the model according to the data characteristics after modeling. Finally, the prediction results of the model are obtained, the determination coefficient R2 is 0.87, and the root mean square error RMSE is 1.132 gC·m-2·d-1. This shows that the random forest model can estimate GPP more accurately.From the results of this study, we can see that the rapid development of computer technology represented by big data and artificial intelligence will inject new vitality into remote sensing technology and make remote sensing technology enter a more mature stage of development and application.

Key words: Random forest regression    The carbon cycle    GPP    Big data    GEE platform
收稿日期: 2018-06-30 出版日期: 2020-09-15
ZTFLH:  TP701  
基金资助: 国家重点研发计划项目(2017YFA0603004);国家自然科学基金项目(41730107);中国科学院百人计划项目(Y6YR0700QM);高分项目(30-Y20A34-9010-15/17)
通讯作者: 刘乃文     E-mail: 329937012@qq.com;sdnwliu@126.com
作者简介: 张坤(1993—),男,山东烟台人,硕士研究生,主要从事机器学习与空间数据挖掘研究。E?mail: 329937012@qq.com
服务  
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章  
张坤
刘乃文
高帅
赵书慧

引用本文:

张坤,刘乃文,高帅,赵书慧. 数据驱动的植被总初级生产力估算方法研究[J]. 遥感技术与应用, 2020, 35(4): 943-949.

Kun Zhang,Naiwen Liu,Shuai Gao,Shuhui Zhao. Data-Driven Estimation of Gross Primary Production. Remote Sensing Technology and Application, 2020, 35(4): 943-949.

链接本文:

http://www.rsta.ac.cn/CN/10.11873/j.issn.1004-0323.2020.4.0943        http://www.rsta.ac.cn/CN/Y2020/V35/I4/943

站点名称纬度/°N经度/°E植被类型
长白山42.403128.096混交林
千烟洲26.733115.067木本热带稀树草原
鼎湖山23.167112.530常绿阔叶林
西双版纳21.950101.200常绿阔叶林
锡林格勒44.130116.320草地
禹城36.833116.567农用地
拉萨当雄站30.41091.080草地
海北站37.660101.330草地
表1  研究区信息
图1  流程图
图2  对结果的影响力
图3  参数与准确率关系
图4  最大迭代次数与决定系数关系
最大迭代次数RMSE排名
103.0610
202.829
302.738
402.707
502.666
602.644
702.632
802.621
902.643
1002.645
表2  最大迭代次数调参表
图5  预测值与实测值对比
图6  MODIS数据与预测数据对比图
MOD17A3通量测量数据
R20.660. 87
RMSE1.921.132
表3  模型效果对比表
1 Solomon S, Qin D, Manning M, et al. Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to The Fourth Assessment Report of The Intergovernmental Panel on Climate Change. Summary for Policymakers[J]. Intergovernmental Panel on Climate Change Climate Change, 2007, 18(2):95-123.
2 Bonan G B. Forests and Climate Change: Forcings, Feedbacks, and The Climate Benefits of Forests[J]. Science, 2008, 320(5882): 1444-1449.
3 Cannell M G R, Milne R, Hargreaves K J, et al. National Inventories of Terrestrial Carbon Sources and Sinks: The U.K. Experience[J]. Climatic Change, 1999, 42(3):505-530.
4 Yu Guirui.Global Change and Terrestrial Ecosystems Carbon Cycle and Carbon Accumulation[M]. Beijing:Meteorology Press, 2003 [于贵瑞. 全球变化与陆地生态系统碳循环和碳蓄积[M]. 北京:气象出版社, 2003.]
5 Lieth H. Primary Production: Terrestrial Ecosystems[J]. Human Ecology, 1973, 1(4):303-332.
6 Li Wenhua, Zhou Xingmin. Ecological System and Optimal Utilization Mode of Qinghai-Tibet Plateau[M]. Guangzhou:Guangdong Science and Technology Press, 1998.
6 李文华, 周兴民. 青藏高原生态系统及优化利用模式[M]. 广州:广东科技出版社, 1998.
7 Mao J, Wang B, Dai Y, et al. Improvements of A Dynamic Global Vegetation Model and Simulations of Carbon and Water at An Upland-Oak Forest[J]. Advances in Atmospheric Sciences, 2007, 24(2): 311-322.
8 Wang J, Wu C, Zhang C, et al. Improved Modeling of Gross Primary Productivity (GPP) by Better Representation of Plant Phenological Indicators from Remote Sensing Using A Process Model[J]. Ecological Indicators, 2018, 88: 332-340.
9 Wu C, Peng D, Soudani K, et al. Land Surface Phenology Derived from Normalized Difference Vegetation Index (NDVI) at Global FLUXNET Sites[J]. Agricultural and Forest Meteorology, 2017, 233(Complete):171-182.
10 Tramontana G, Ichii K, Camps-Valls G, et al. Uncertainty Analysis of Gross Primary Production Upscaling Using Random Forests, Remote Sensing and Eddy Covariance Data[J]. Remote Sensing of Environment, 2015, 168: 360-373.
11 Chen B, Xiao X, Li X, et al. A Mangrove Forest Map of China in 2015: Analysis of Time Series Landsat 7/8 and Sentinel-1A Imagery in Google Earth Engine Cloud Computing Platform[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2017, 131: 104-120.
12 Alonso A, Muñoz-carpena R, Kennedy R E, et al. Wetland Landscape Spatio-temporal Degradation Dynamics Using The New Google Earth Engine Cloud-based Platform: Opportunities for Non-Specialists in Remote Sensing[J]. Transactions of The ASABE, 2016, 59(5): 1331-1342.
13 Gorelick N, Hancher M, Dixon M, et al. Google Earth Engine: Planetary-scale Geospatial Analysis for Everyone[J]. Remote Sensing of Environment, 2017,202: 18-27.
14 Dong J, Xiao X, Menarguez M A, et al. Mapping Paddy Rice Planting Area in Northeastern Asia with Landsat 8 Images, Phenology-based Algorithm and Google Earth Engine[J]. Remote Sensing of Environment, 2016,185:142-154..
15 Tian H F, Meng M, Wu M Q, et al. Mapping Spring Canola and Spring Wheat Using Radarsat-2 and Landsat 8 Images with Google Earth Engine[J]. Current Science, 2019, 116: 291-298.
16 Cheng Yanpei,Zhang Fawang,Dong Hua, et al. Dynamic Monitoring of Water Bodies in Central Asia based on Modis Satellite Data[J]. Hydrogeological Engineering Geology,2010, 37(5): 33-37.
16 程彦培, 张发旺, 董华, 等. 基于MODIS卫星数据的中亚地区水体动态监测研究[J]. 水文地质工程地质, 2010, 37(5):33-37.
17 Giglio L, Van d W G R, Randerson J T, et al. Global Estimation of Burned Area Using MODIS Active Fire Observations[J]. Atmospheric Chemistry and Physics Discussions, 2005, 5(6):11091-11141.
18 Reed B, Budde M, Spencer P, et al. Integration of MODIS-derived Metrics to Assess Interannual Variability in Snowpack, Lake Ice, and NDVI in Southwest Alaska[J]. Remote Sensing of Environment, 2009, 113(7):1443-1452.
19 Maurer E P, Rhoads J D, Dubayah R O, et al. Evaluation of The Snow-covered Area Data Product from MODIS[J]. Hydrological Processes, 2003, 17: 59-71.
20 Huete A, Didan K, Miura T, et al. Overview of The Radiometric and Biophysical Performance of The MODIS Vegetation Indices[J]. Remote Sensing of Environment, 2002, 83(1-2):195-213.
21 Ashouri H, Hsu K L, Sorooshian S, et al. PERSIANN-CDR: Daily Precipitation Climate Data Record from Multisatellite Observations for Hydrological and Climate Studies[J]. Bulletin of the American Meteorological Society, 2015, 96(1):69-83.
22 Yu Guirui, Sun Xiaomin. The Principle and Method of Flux Observation of Terrestrial Ecosystem[M]. Beijing: Advanced Education Publishing House,2006.
22 于贵瑞, 孙晓敏. 陆地生态系统通量观测的原理与方法[M].北京:高等教育出版社, 2006.
23 Leuning R, Yu G R. Carbon Exchange Research in ChinaFLUX[J]. Agricultural & Forest Meteorology, 2006, 137(3-4):0-124.
24 Baldocchi D.Breathing of The Terrestrial Biosphere: Lessons Learned from A Global Network of Carbon Dioxide Flux Measurement Systems[J].Australian Journal of Botany,2008,56(1):1-26.
25 Doherty S J, Bojinski S, Henderson-Sellers A, et al. Lessons Learned from IPCC AR4: Scientific Developments Needed to Understand, Predict, and Respond to Climate Change[J]. Bulletin of the American Meteorological Society, 2009, 90(4): 497-514.
26 Stoy P C, Mauder M, Foken T, et al. A Data-driven Analysis of Energy Balance Closure Across FLUXNET Research Sites: The Role of Landscape Scale Heterogeneity[J]. Agricultural and Forest Meteorology, 2013, 171: 137-152.
27 Liu X, Guanter L, Liu L, et al. Downscaling of Solar-induced Chlorophyll Fluorescence from Canopy Level to Photosystem Level Using A Random Forest Model[J]. Remote Sensing of Environment, 2019,231:110772. doi:10.1016/j.rse.2018.05.035.
doi: 10.1016/j.rse.2018.05.035
28 Yang Siqi, Zhao Lihua. Application of Random Forest Algorithm in Urban Air Quality Prediction[J]. Statistics and Decision Making,2017(20):83-86.
28 杨思琪, 赵丽华. 随机森林算法在城市空气质量预测中的应用[J]. 统计与决策, 2017(20):84-87.
29 Chen Y, Shen W, Gao S, et al. Estimating Deciduous Broadleaf Forest Gross Primary Productivity by Remote Sensing Data Using A Random Forest Regression Model[J]. Journal of Applied Remote Sensing, 2019, 13(3): 038502.
[1] 沈倩,周艳莲,单良. 基于遥感和站点观测数据的生态系统呼吸模型比较[J]. 遥感技术与应用, 2020, 35(2): 435-447.
[2] 沈明,丁云生,段洪涛. 基于地球大数据的“美丽湖泊” SDG 6.3.2综合评价体系构建[J]. 遥感技术与应用, 2020, 35(2): 295-301.
[3] 冯艾琳,武晋雯,孟莹,姜鹏,董巍,张璇,方缘,刘斌. 基于MODIS GPP数据产品的辽宁省碳源/汇空间格局分布研究[J]. 遥感技术与应用, 2019, 34(4): 857-864.
[4] 奚雷, 单楠, 杨沈斌, 张永光. 基于SCOPE模型的水稻不同生育期日光诱导叶绿素荧光及GPP模拟研究[J]. 遥感技术与应用, 2019, 34(3): 488-499.
[5] 王莹莹, 袁金国, 张莹, 吴朝阳. 中国温带地区植被物候期时空变化特征及对总初级生产力的影响[J]. 遥感技术与应用, 2019, 34(2): 377-388.
[6] 刘亲亲, 崔耀平, 刘素洁, 李楠. 中国不同土地利用类型分光辐射地表反照率研究[J]. 遥感技术与应用, 2019, 34(1): 46-56.
[7] 王卷乐, 程凯, 边玲玲, 韩雪华, 王明明. 面向SDGs和美丽中国评价的地球大数据集成框架与关键技术[J]. 遥感技术与应用, 2018, 33(5): 775-783.
[8] 郝斌飞,韩旭军,马明国,刘一韬,李世卫. Google Earth Engine在地球科学与环境科学中的应用研究进展[J]. 遥感技术与应用, 2018, 33(4): 600-611.
[9] 张滔,唐宏. 基于Google Earth Engine的京津冀2001~2015年植被覆盖变化与城镇扩张研究[J]. 遥感技术与应用, 2018, 33(4): 593-599.
[10] 刘畅,李震,张平,田帮森,周建民. 基于Google Earth Engine评估新疆西南部MODIS积雪产品[J]. 遥感技术与应用, 2018, 33(4): 584-592.
[11] 胡云锋,商令杰,张千力,王召海. 基于GEE平台的1990年以来北京市土地变化格局及驱动机制分析[J]. 遥感技术与应用, 2018, 33(4): 573-583.
[12] 马敏娜,袁文平. 青藏高原总初级生产力估算的模型差异[J]. 遥感技术与应用, 2017, 32(3): 406-418.
[13] 管小彬,沈焕锋,甘文霞,张良培. 基于Landsat TM/ETM+影像的武汉市冬季NPP估算及其时空变化分析[J]. 遥感技术与应用, 2015, 30(5): 884-890.
[14] 王海波,马明国,王旭峰. 冻融过程与陆地生态系统碳循环的相互关系研究进展[J]. 遥感技术与应用, 2014, 29(3): 369-377.
[15] 王让会,衣怀峰,宁虎森,吕妍,吉小敏,赵福生,徐德福,李琪. 基于NDVI的人工植被碳储量估算[J]. 遥感技术与应用, 2011, 26(4): 399-404.