基于遥感生态指数的道路网络生态效应分析——以福州市为例

以2000年和2016年两期Landsat影像为基础数据源，借助遥感生态指数（RSEI）对福州市生态环境进行评价，在此基础上，从道路缓冲区、城乡梯度带、剖面线三种不同取样方法定量探讨RSEI对路网的响应机制；再以500 m×500 m、1000 m×1000 m、1500 m×1500 m、2000 m×2000 m、2500 m×2500 m、3000 m×3000 m不同尺度的网格划分空间单元，运用全局空间自相关、地理加权回归分析等方法分析了道路核密度（KDE）和RSEI及其之间关系的空间异质性。结果表明：从2000年到2016年，福州市生态环境好的区域面积增幅大于生态环境差的区域面积，生态环境质量向好的方向发展。各类型道路缓冲区的RSEI变化规律都是呈从0 m到3000 m逐渐上升的趋势，其中国道、省道、县道、乡镇道路影响的阈值分别在900、900、450、750 m左右。在城乡梯度分析中，RSEI曲线的变化规律都是随着与行政中心距离的增大而增大，到达一定阈值后趋于平缓，甚至还有小幅度的下降，区级的影响阈值在20 km左右，县级的影响阈值在12 km左右；而KDE曲线的变化规律与RSEI相反，其变化阈值与RSEI正好对应。剖面线所经过的行政中心处，其RSEI为低值，KDE为高值，西北方向的内陆地区RSEI高于东南方向的沿海地区。在多尺度的地理加权回归分析中，1500 m×1500 m和2000 m×2000 m这两个网格单元采样下的空间集聚性较强，空间异质性明显，总体上来看，RSEI与KDE呈现负相关关系，且相关关系存在空间分异，负回归系数主要分布在研究区的中心区域。研究结果可为福州市生态建设和路网规划提供参考依据。

The ecological effect of road network based on remote sensing ecological index: a case study of Fuzhou City, Fujian Province

Based on the Landsat images in 2000 and 2016, the remote sensing ecological index (RSEI) is employed to estimate the eco-environment state of Fuzhou City, then the response mechanism of the RSEI to the road network is quantitatively explored using three different sampling strategies, including road buffer zone, urban-rural gradient zone, and section line. Moreover, the spatial heterogeneity of road kernel density (KDE), RSEI and their relationships are analyzed using globally spatial autocorrelation analysis and geographically weighted regression analysis, respectively, in different spatial scales of sample units, including 500 m×500 m, 1000 m×1000 m, 1500 m×1500 m, 2000 m×2000 m, 2500 m×2500 m, and 3000 m×3000 m. The results show that the areas in fine state of eco-environment are greater than those in poor status, indicating an improvement of the eco-environment quality from 2000 to 2016 in the study area. The RSEIs in various types of road buffers are steadily increasing from 0 m to 3000 m distances to roads. Among which the influence thresholds of the national road, provincial road, county road and township road are about 900 m, 900 m, 450 m and 750 m, respectively. In urban-rural gradient analysis, the curve of RSEI increases with the increasing distances from the administrative center, and it tends to be gentle after reaching a certain threshold, which is around 20 km in the district-level, and 12 km in the county-level, respectively. However, the trend of KDE is exactly opposite of the RSEI''s, but with the similar thresholds. Along the section line, the RSEI is relatively low, while the KDE is relatively high in the administrative center; and the RSEI of the inland area in the northwest is higher than that of the coastal area in the southeast. Spatial autocorrelation analysis shows that the RSEI and the KDE have a higher spatial aggregation effect at the two sampling units of 1500 m×1500 m and 2000 m×2000 m than that of the other sampling units, therefore, 1500 m×1500 m and 2000 m×2000 m are then employed as spatial analysis units to explore the spatial variations in the relationships between RSEI and KDE using geographically weighted regression (GWR) model. The GWR outcomes overall indicate a negative correlation between RSEI and KDE, but also identify the spatial paradigms in their divergent correlations, with the negative associations mainly distributing in the central area of the study area. Our study can provide scientific basis for the ecological civilization construction and sustainable development of road network for the study area.