中文摘要：

在3S技术平台支撑下,用典范对应分析排序方法对武夷山风景区景观格局与环境因子的关系及其尺度效应进行了分析。结果表明,在6个取样尺度上（粒度分别为30 m、90 m、150 m、210 m、300 m和450 m）,排序轴总特征值呈递减趋势,依次为9.626、7.814、5.708、4.019、3.386、2.310;前4轴累积景观环境解释量变化程度小（97.3%～98.8%）,前2轴各尺度平均累积解释量为89.4%,景观-环境的相关系数呈先减后增趋势。6个尺度中环境因子与景观格局的平均相关系数依次为郁闭度（0.7989＊＊）〉腐殖质层厚度（0.7248＊＊）〉海拔（0.6083＊＊）〉坡度（0.6132＊＊）〉经度（-0.4064＊＊）〉土层厚度（0.3696＊＊）〉纬度（0.0614）〉坡向（0.0295）。风景区景观类型可划分为3类,即与海拔因子密切相关的裸地、河流、建设用地、农田景观,与土壤因子密切相关的杉木林、马尾松林、阔叶林景观,以及受人类活动密切相关的茶园、经济林、竹林、灌草丛景观。因此,风景区景观格局与环境因子间的关系具有尺度效应,各环境因子对排序轴相关系数的影响规律在6个取样尺度内可尺度推绎。

英文摘要：

Based on 3S technology,relationships and muti-scale effects between environmental factors and landscape pattern in the Wuyishan Scenery District was studied by using CCA（Canonical Correspondence Analysis）.The results showed that the total eigenvalue of ordination axes increased with upscaling by 9.626,7.814,5.708,4.019,3.386 and 2.310 in six scales（30 m×30 m、 90 m×90 m、 150 m×150 m、 210 m×210 m、 300 m×300 m、 450 m×450 m）,respectively.The cumulative variance of landscape-environment relationship ranged from 97.3% to 98.8% in the first four ordination axes,and that of the first two ordination axes had an average of 89.4%.Landscape-environment correlation coefficients decreased at first and then increased.The average correlation coefficients between environment factors and landscape pattern were in order as canopy density（0.7989＊＊） humus thickness（0.7248＊＊） elevation（0.6083＊＊） slope（0.6132＊＊） longitude（-0.4064＊＊） soil thickness（0.3696＊＊） latitude（0.0614） aspect（0.0295）.Landscape types in Wuyishan Scenery District could be classified into three types,such as elevation related landscape,including bare land,water body,construction land,and farmland,soil related landscape,including Cunninghamia lanceolata forest,Pinus massoniana forest,broad-leaved forest,and human managed landscape,including Camellia sinensis plantation,economic forest,bamboo forest,shrub and grassland.Therefore,muti-scale effect existed between landscape pattern and environment factors,and the response law of environment factors to ordination axes could be scaling within six scales.