%0 Journal Article %T 2000年以来天目湖流域茶园分布变化及趋势模拟 %T Spatio-temporal change of tea plantation since 2000 and model-based prediction in the Tianmuhu Reservoir watershed %A 刁亚芹 %A 韩莹 %A 李兆富 %A DIAO,Yaqin %A HAN,Ying %A LI,Zhaofu %J 湖泊科学 %J Journal of Lake Sciences %@ 1003-5427 %V 25 %N 6 %D 2013 %P 799-808 %K 天目湖流域;土地利用;茶园;时空特征;神经网络模型;情景预测 %K Tianmuhu Reservoir watershed;land use;tea plantation;spatio-temporal distribution;ANN model;scenario prediction %X 土地利用/土地覆被变化(LUCC)与区域的可持续发展密切相关,分析土地利用变化及其空间变量的关系,建立模拟模型来认识和揭示土地利用变化的特征和时空分布规律,为未来规划和政策的制定提供支持是LUCC研究的重点问题之一.本研究以太湖流域上游典型丘陵山区茶园开发作为重点研究对象,结合中高分辨率遥感影像信息、1:50000国家标准地形图数据和野外土壤采集分析数据获取2000年以来茶园及其它土地利用信息、研究区地形地貌数据和土壤数据,通过对2000年以来茶园及其新开茶园的区位分布分析,发现茶园的变化呈现快速增长的趋势,11年研究区茶园面积扩大了5倍,其中一半以上的面积是最后3年增长的;茶园在流域上不同分区的分布差异比较大,这些差异与分区自然条件、区位条件和开发条件有一定关系.根据Cramer V卡方检验结果值,比较不同的茶园区位影响因子对茶园分布的影响力,确定高程、坡度、水库、河流、村落、主干道路和茶园规模对茶园区位影响比较显著.用茶园区位影响显著性的7个因子通过迭代优化建立MLP神经网络模型,对研究区的未来茶园的分布区位进行模拟,并且用已开发茶园对模型进行验证,达到84%的区位预测精度.基于模型用现有茶园开发模式、2000年左右茶园开发模式和折中模式3种情景模拟未来茶园的分布,发现按现有的茶园开发速度,2020年茶园总面积将达到流域总面积的1/4,任何一种情景都改变不了茶园在重要功能区的增长趋势,因此迫切需要政府采取相应的措施对茶园面积的增长和区位的分布进行控制. %X Land use/cover change (LUCC) is closely related to regional sustainable development.It is an important issue in LUCC investigations to analyze the relationship between land use change and its influencing variables on spatial scale,to reveal the change in land use and its spatio-temporal distribution characteristics by developing a simulation model,and to support the future decisionmaking.This study took the Tianmuhu Reservoir watershed,the land use of which has been changed rapidly with the dominant cash crop of tea plantation in upstream of the Lake Taihu basin,as the study area.Interpretation of the high-resolution remote sensing imagery provided the data of tea plantation and other land uses since 2000.Topography and soil data were obtained from the 1:50000 scale national standard relief map and soil sampling with laboratory analysis,respectively.Analyses of the spatial distribution and area change of existing and newly-exploited tea plantation since 2000 revealed that the tea plantation area was in a rapidly increasing trend.The tea plantation area expanded 5 times in the past 11 years and half of those increased in the last 3 years.There was a big difference in the spatial distribution of tea plantation among sub-regions,because of their different natural backgrounds, geographical factors and suitability conditions for tea plantation.The results of Cramer V chi-square test showed that elevation, slope,reservoirs,rivers,villages,roads and the scale of tea garden were the most significant factors for the spatial distribution of tea plantation.The MLP neural network model for the prediction of tea planation was developed through iterations based on the seven significant influencing factors.It was validated using the area of existing tea plantation,the performance of which reached to 84%.Application of the model in predicting the future spatial distribution of tea plantation under current,2000,and the compromised tea-exploitation mode displayed that the area of tea plantation in 2020 will account for 1/4 of the watershed area if it is exploited following the current pace.None of the three scenarios could change the growing trend of tea plantation.Therefore,it is urgent to push the local government to take corresponding strategies and measures to control the increasing area of tea planation and adjust its spatial distributions %R 10.18307/2013.0603 %U http://www.jlakes.org/ch/reader/view_abstract.aspx %1 JIS Version 3.0.0