基于用户兴趣点数据与Landsat遥感影像的城市热场空间格局研究

地图用户兴趣点(POI)数据能够反映微观尺度上城市系统中的人类活动。利用2015年夏季Landsat 8遥感影像提取了南京市地表温度和主要土地覆盖类型,利用空间与非空间多元回归模型在2、5、10 km 3个尺度上研究了地表温度与同期POI密度及植被和水体盖度的相关性,并利用方差分解技术定量区分人类活动因子(POI密度)及生态基础设施(植被和水体盖度)对城市热场的相对重要性。结果表明,在3个观测尺度上,POI密度与地表温度均存在极显著的正相关(P0.001),且相关性随观测尺度的增大而升高。植被和水体均具有显著的降温效应,水体盖度与地表温度的相关性仅在2 km尺度上显著,在5 km和10km尺度上其降温效应不再显著。方差分解结果表明,人类活动因子和生态基础设施对地表温度的独立解释率为1.6%—15%,而二者共同解释率达到了40%—70%。研究表明POI作为城市功能节点可以综合反映城市中人类活动的热源强度,在城市热场空间格局研究中是一种可与遥感数据互补的有用数据源。

Quantifying the spatial pattern of urban thermal fields based on point of interest data and Landsat images

Point of interest (POI) in digital maps can effectively reflect human activities in urban systems at micro spatial scales. We retrieved land surface temperature (LST) in the Nanjing metropolitan region from a Landsat 8 image, and examined LST in relation to POI density, as well as vegetation and water cover at three spatial scales, namely, 2, 5, and 10 km. The relative importance of human factors (represented by POI density) and ecological facilities (represented by vegetation and water cover) on the thermal field patterns was quantitatively distinguished using simultaneous autoregressive models and the variation partitioning technique. The results showed that POI density and LST exhibit significantly positive correlations (P < 0.001) that are generally amplified with increasing observational scale. Vegetation and water cover played a significant role in reducing LST; however, this cooling effect from water cover was detected only at the 2-km scale. At all three studied scales, the results from variation partitioning showed that human factors shared a considerable proportion of explanatory power with ecological facilities (40%-70%), whereas the unique explanatory power of human factors and ecological facilities ranged between 1.6% and 15%. POIs characterize urban functional nodes and can thereby serve as an effective indicator of the intensity of anthropogenic heat sources. Our results suggested that POI could be a useful data source for the study of urban thermal fields, which is complementary to remotely sensed information.