Abstract:

Optimizing spatial configuration (e.g., fragmentation) of greenspace is an effective approach to improve the urban thermal environment. However, the effect of greenspace fragmentation on the urban thermal environment may vary with changes in percent green cover, raising significant challenges for optimizing greenspace fragmentation to improve the urban thermal environment. Few studies tested this hypothesis. Taking Changsha as an example, we characterized the urban thermal environment based on land surface temperature (LST) observed by Landsat, and quantified the percent area of different land cover and greenspace fragmentation (measured by edge density of greenspace) based on a 1 m resolution land cover map derived from the Gaofen 2 satellite images. Using 419 grids of 1 km as the analytical unit, we applied piecewise linear regression to model the nonlinear relationship between percent green cover and edge density of greenspace. Subsequently, we identified the threshold of percent green cover, based on which the 419 grids were grouped into two categories. We built multiple linear regression models with LST as dependent variable, and percent cover of greenpace, water, and bare land and greenspace fragmentation as explanatory variable respectively, elucidating the effects of greenspace fragmentation on LST in the whole study regions, and regions with high and low percent green cover. We finally conducted the variance partitioning analysis to reveal the independent and joint effects of three groups of factors (i.e., percent green cover, greenspace fragmentation, and percent cover of other land covers) on LST. The results showed that 1) greenspace fragmentation first increased and then decreased with the increase of percent green cover, and the threshold was 44.92%. 2) The four variables explained 69.1% of the variation of LST and they all showed significant negative impact on LST. 3) Greenspace fragmentation showed a significant negative effect on LST in regions with low percent green cover, but showed an insignificant positive effect on LST in regions with high percent green cover. 4) Percent green cover had higher independent effect on LST than greenspace fragmentation in the study area and in regions of both high and low percent green cover. We suggested that Changsha can increase greenspace fragmentation in areas with low vegetation coverage, as well as increase vegetation coverage to improve the urban heat environment.

Key words: urban thermal environment, percent green cover, urban greenspace, landscape fragmentation, edge density

摘要：

优化绿地空间配置（如破碎度）是改善城市热环境的有效途径。然而绿地破碎度对城市热环境的影响可能随绿化覆盖率的变化而变化，对改善城市热环境提出了巨大挑战，但相关研究十分缺乏。以长沙市为例，利用Landsat地表温度表征城市热环境，使用解译于高分2号遥感影像的城市绿地图在1 m像元尺度量化绿化覆盖率和绿地破碎度（用绿地边界密度表征）。以419个1 km格网为分析单元，应用分段线性回归揭示绿化覆盖率与绿地破碎度间的非线性关系，识别绿化覆盖率阈值并以此为标准将419个1 km格网划分为高绿化覆盖率区和低绿化覆盖率区。以1 km格网平均地表温度为因变量，绿化覆盖率、绿地破碎度、水体覆盖率和裸地覆盖率为自变量从研究区、高绿化覆盖率区和低覆盖率区3个方面建立多元线性回归模型阐明绿地破碎度对地表温度的影响。最后进行方差分解分析绿化覆盖率、绿地破碎度和其他土地覆盖率对地表温度的独立和联合影响。结果显示，1）绿地破碎度随绿化覆盖率的增加先增加后降低，阈值为44.9%。2）整个研究区，4个指标可解释69.1%的地表温度变异，地表温度随绿化覆盖率、水体覆盖率、裸地覆盖率和绿地破碎度的增加显著降低。3）低绿化覆盖率下绿地破碎度显著影响地表温度，高绿化覆盖率下绿地破碎度对地表温度呈不显著正影响。4）在研究区和不同绿化覆盖率区域，绿化覆盖率对地表温度的独立影响均高于绿地破碎度的独立影响。建议长沙城市绿地规划管理在提高绿化覆盖率的同时可提高低绿化覆盖率区域绿地破碎度以改善城市热环境。

关键词: 城市热环境, 绿化覆盖率, 城市绿地, 破碎化, 边界密度