基于随机森林回归的茶园扩张驱动机制分析

doi:10.19741/j.issn.1673-4831.2019.0334

摘要/Abstract

摘要： 茶园开发与扩张是土地利用/覆被变化（LUCC）的典型过程之一，与区域农业发展和土地可持续利用密切相关。以我国优质绿茶主产区——信阳市浉河区为例，结合遥感影像、土地利用等多源数据，探索1990-2015年浉河区茶园分布时空演变格局。以像元和乡镇尺度的自然和社会经济等26个因子为自变量，以茶园变动情况为因变量，采用随机森林回归方法，探讨1990-2000和2000-2015年茶园扩张的驱动机制。结果表明：（1）浉河区茶园主要分布于丘陵区和浅山区，1990-2000年茶园面积快速扩张，面积增加65.2%，2000-2015年茶园扩张速度放缓。茶园主要向西北和东南的水源和公路方向扩张。（2）1990-2000和2000-2015年2阶段，像元尺度的茶园距城区中心距离、距农村居民点距离和土壤特性以及乡镇尺度的坡度和平均高程等驱动因子重要性均较高。随着经济发展和技术进步，邻域土地利用情况、土壤条件等自然驱动因子对茶园扩张的限制性降低，空间社会经济条件对茶园开发与扩张越发重要。（3）茶园扩张对各主要驱动因子的边际依赖性强，在驱动因子不同梯度间差异明显，茶园扩张倾向于发生在距农村居民点距离2 km内，乡镇尺度平均坡度为16~18°的区域。（4）随机森林回归方法结合多源信息，能够较好地挖掘研究区茶园扩张驱动机制，为改善区域土地利用提供一定决策支持。

关键词: 土地利用/覆被变化, 茶园, 随机森林回归, 驱动因子, 浉河区

Abstract: Tea plantation and expansion, a typical process of regional land use/cover change (LUCC), is closely related to agricultural development and sustainable land use. Taking the Shihe District of Xinyang City, the main producing area of high quality green tea in China, as an example, the spatio-temporal evolution in the distribution pattern of tea plantations from 1990 to 2015 was explored by coupling with multi-source data such as remote sensing images and land use data, etc. Random forest (RF) regression was used to explore driving mechanism of tea production expansion over two periods (1990-2000 and 2000-2015). Twenty-six exploratory variables, such as topography, land use, soil types, etc. were classified into biophysical and socioeconomic categories at the levels of pixel and village; and treated as the factors driving tea plantation expansion. The results show that tea plantations in the Shihe District are mainly distributed in hilly areas and shallow mountainous areas. From 1990 to 2000, the area of tea plantation expanded rapidly with a growth rate of 65.2%, while the expansion rate slowed down during 2000-2015. Tea plantations mainly expanded in the northwest and southeast directions, growing towards water sources and roads. The results indicate that driving factors of high importance can be divided into two scales:distance between tea garden and urban center/rural residential area at pixel scale, and soil types, slope and mean elevation at rural scale. Economic and technological development reduced restrictions imposed by biophysical driving factors such as land use and soil conditions. Spatial socio-economic conditions are increasingly essential for tea plantation expansion. The tea plantation expansion has a strong marginal dependence on the main driving determinants as growth tends to occur within 2 km from the rural settlements, where the average slope is 16-18° at village pixel region. Random forest regression is an efficient and effective means to explore the mechanisms of tea plantation expansion, which could provide guidance for sustainable land use planning.

Key words: land use/cover change, tea plantation, random forest regression, driving factors, Shihe District

中图分类号:

李雪柔, 陈飞燕, 林爱文, 邹建成, 周志高. 基于随机森林回归的茶园扩张驱动机制分析[J]. 生态与农村环境学报, 2020, 36(1): 44-52.

LI Xue-rou, CHEN Fei-yan, LIN Ai-wen, ZOU Jian-cheng, ZHOU Zhi-gao. Driving Mechanism of Tea Plantation Expansion Using a Random Forest Regression Model[J]. Journal of Ecology and Rural Environment, 2020, 36(1): 44-52.

参考文献

[1] YAN P,SHEN C,FAN L C,et al.Tea Planting Affects Soil Acidification and Nitrogen and Phosphorus Distribution in Soil[J].Agriculture,Ecosystems&Environment,2018,254:20-25.
[2] ZHU T B,ZHANG J B,MENG T Z,et al.Tea Plantation Destroys Soil Retention of NO₃^- and Increases N₂O Emissions in Subtropical China[J].Soil Biology and Biochemistry,2014,73:106-114.
[3] MORTIMER P E,GUI H,XU J,et al.Alder Trees Enhance Crop Productivity and Soil Microbial Biomass in Tea Plantations[J].Applied Soil Ecology,2015,96:25-32.
[4] XIAO Z,HUANG X J,ZANG Z,et al.Spatio-temporal Variation and the Driving Forces of Tea Production in China Over the Last 30 Years[J].Journal of Geographical Sciences,2018,28(3):275-290.
[5] 刁亚芹,韩莹,李兆富.2000年以来天目湖流域茶园分布变化及趋势模拟[J].湖泊科学,2013,25(6):799-808.[DIAO Ya-qin,HAN Ying,LI Zhao-fu.Spatio-temporal Change of Tea Plantation Since 2000 and Model-based Prediction in the Tianmuhu Reservoir Watershed[J].Journal of Lake Sciences,2013,25(6):799-808.]
[6] 郑鹏,林海荣,占昕,等.基于CA-Markov模型的安溪县茶叶种植区动态变化[J].生态学杂志,2018(9):2776-2785.[ZHENG Peng,LIN Hai-rong,ZHAN Xin,et al.Dynamics of Tea Planting Areas in Anxi County Based on CA-Markov Model[J].Chinese Journal of Ecology,2018(9):2776-2785.]
[7] 傅超,刘彦随.城市扩张对西湖龙井茶园用地及发展影响分析[J].经济地理,2015,35(6):138-142.[FU Chao,LIU Yan-sui.The Urban Expansion Effects on Land Use of Xihu Longjing Tea and Its Development[J].Economic Geography,2015,35(6):138-142.]
[8] 吴克华,赵卫权,廖凤林,等.贵州省茶园生态适宜性研究[J].中国农业资源与区划,2014,35(2):108-116.[WU Ke-hua,ZHAO Wei-quan,LIAO Feng-lin,et al.Study on Ecological Suitability of Green Tea Garden in Guizhou Province[J].Journal of China Agricultural Resources and Regional,2014,35(2):108-116.]
[9] 马恩朴,蔡建明,林静,等.远程耦合视角下的土地利用/覆被变化解释[J].地理学报,2019,74(3):421-431.[MA En-pu,CAI Jian-ming,LIN Jing,et al.Explanation of Land Use/Cover Change From the Perspective of Tele-coupling[J].Acta Geographica Sinica,2019,74(3):421-431.]
[10] SU S L,SUN Y F,LEI C R,et al.Reorienting Paradoxical Land Use Policies Towards Coherence:A Self-adaptive Ensemble Learning Geo-simulation of Tea Expansion Under Different Scenarios in Subtropical China[J].Land Use Policy,2017,67:415-425.
[11] WANG Q,REN Q F,LIU J F.Identification and Apportionment of the Drivers of Land Use Change on a Regional Scale:Unbiased Recursive Partitioning-based Stochastic Model Application[J].Agriculture,Ecosystems&Environment,2016,217:99-110.
[12] 王银峰,张继红,翟海国.信阳茶区生态环境分析与评价[J].地理学报,1993,48(5):428-436.[WANG Yin-feng,ZHANG Ji-hong,ZHAI Hai-guo.An Analysis of the Ecological Conditions in the Xinyang Tea-producing Area[J].Acta Geographica Sinica,1993,48(5):428-436.]
[13] BREIMAN L.Random Forests[J].Machine Learning,2001,45(1):5-32.
[14] 方匡南,吴见彬,朱建平,等.随机森林方法研究综述[J].统计与信息论坛,2011,26(3):32-38.[FANG Kuang-nan,WU Jian-bin,ZHU Jian-ping,et al.A Review of Technologies on Random Forests[J].Statistics&Information Forum,2011,26(3):32-38.]
[15] 徐枫,王占岐,张红伟,等.随机森林算法在农村居民点适宜性评价中的应用[J].资源科学,2018,40(10):2085-2098.[XU Feng,WANG Zhan-qi,ZHANG Hong-wei,et al.Application of Random Forest Algorithm in Suitability Evaluation of Rural Residential Land[J].Resources Science,2018,40(10):2085-2098.]
[16] KUHN M.Building Predictive Models in R Using the Caret Package[J].Journal of Statistical Software,2008,28(5):1-26.
[17] FAWCETT T.An Introduction to ROC Analysis[J].Pattern Recognition Letters,2006,27(8):861-874.
[18] LESSCHEN J P,VERBURG P H,STAAL S J.Statistical Methods for Analysing the Spatial Dimension of Changes in Land Use and Farming Systems[M].Nairobi,Kenya and Wageningen,Netherlands:International Livestock Research Institute and LUCC Focus 3 Office,2005:56-62
[19] 邵景安,李阳兵,魏朝富,等.区域土地利用变化驱动力研究前景展望[J].地球科学进展,2007,22(8):798-809.[SHAO Jing-an,LI Yang-bing,WEI Chao-fu,et al.The Drivers of Land Use Change at Regional Scale Assessment and Prospects[J].Advances in Earth Science,2007,22(8):798-809.]
[20] 张运清,谢作轮,李秀珍,等.桥梁连陆对海岛土地利用变化的影响:以朱家尖岛为例[J].生态与农村环境学报,2017,33(7):607-614.[ZHANG Yun-qing,XIE Zuo-lun,LI Xiu-zhen,et al.Influence of Connection With Mainland via Bridge on Land Use of Islands:A Case Study of Zhujiajian Island[J].Journal of Ecology and Rural Environment,2017,33(7):607-614.]
[21] 魏士凯,范顺祥,张玉珍,等.塞罕坝自然保护区主要植被类型动态及其驱动力[J].应用生态学报,2018,29(4):1170-1178.[WEI Shi-kai,FAN Shun-xiang,ZHANG Yu-zhen,et al.Dynamics and Driving Forces of Main Vegetation Types in the Saihanba Nature Reserve,Hebei Province,China[J].Chinese Journal of Applied Ecology,2018,29(4):1170-1178.]
[22] CRK T,URIARTE M,CORSI F,et al.Forest Recovery in a Tropical Landscape:What Is the Relative Importance of Biophysical,Socioeconomic,and Landscape Variables?[J].Landscape Ecology,2009,24(5):629-642.
[23] WEHNER S,HERRMANN S,BERKHOFF K.CLUENaban:A Land Use Change Model Combining Social Factors With Physical Landscape Factors for a Mountainous Area in Southwest China[J].Ecological Indicators,2014,36:757-765.
[24] XIAO R,SU S L,MAI G C,et al.Quantifying Determinants of Cash Crop Expansion and Their Relative Effects Using Logistic Regression Modeling and Variance Partitioning[J].International Journal of Applied Earth Observation and Geoinformation,2015,34:258-263.
[25] SU S L,ZHOU X C,WAN C,et al.Land Use Changes to Cash Crop Plantations:Crop Types,Multilevel Determinants and Policy Implications[J].Land Use Policy,2016,50:379-389.
[26] MÜLLER D,LEITÃO P J,SIKOR T.Comparing the Determinants of Cropland Abandonment in Albania and Romania Using Boosted Regression Trees[J].Agricultural Systems,2013,117:66-77.
[27] 金玉香.临翔区和双江县茶园遥感信息提取及其生态适宜性评价[D].昆明:云南大学,2015.[JIN Yu-xiang.Linxiang District and Shuangjiang County Tea Grden Information Extracted Based on Remote Sense and Ecological Suitability Evaluation[D].Kunming:Yunnan University,2015.]
[28] 杨丽,傅春.赣南生态屏障区林地变化的空间驱动力分析[J].地理与地理信息科学,2018,34(6):58-62.[YANG Li,FU Chun.Spatial Driving Force Analysis of Forest Land Change in Ganan Ecological Security Zone[J].Geography and Geo-information Science,2018,34(6):58-62.]
[29] MARCOS-MARTINEZ R,BRYAN B A,CONNOR J D,et al.Agricultural Land-use Dynamics:Assessing the Relative Importance of Socioeconomic and Biophysical Drivers for More Targeted Policy[J].Land Use Policy,2017,63:53-66.
[30] DIHKAN M,GUNEROGLU N,KARSLI F,et al.Remote Sensing of Tea Plantations Using an SVM Classifier and Pattern-based Accuracy Assessment Technique[J].International Journal of Remote Sensing,2013,34(23):8549-8565.
[31] 许光明.基于面向对象和多源数据融合的茶园遥感提取研究[D].西安:陕西师范大学,2016:44-48.
[32] ZHANG Q W,GAO W J,SU S L,et al.Biophysical and Socioeconomic Determinants of Tea Expansion:Apportioning Their Relative Importance for Sustainable Land Use Policy[J].Land Use Policy,2017,68:438-447.
[33] FRIEDMAN J H.Greedy Function Approximation:A Gradient Boosting Machine[J].Annals of Statistics,2001,29(5):1189-1232.
[34] 刘桂林,张落成,张倩.1985-2010年南京市耕地变化轨迹及驱动力分析[J].生态与农村环境学报,2013,29(6):688-694.[LIU Gui-lin,ZHANG Luo-cheng,ZHANG Qian.Trajectory and Driving Forces of Change in Farmland in Nanjing During the Period From 1985 to 2010[J].Journal of Ecology and Rural Environment,2013,29(6):688-694.]