兰州中心城区热环境季节动态遥感分析

Remote sensing analysis of the seasonal dynamics of Lanzhou's urban thermal environment

根据2009年1月、3月、6月、9月的Landsat TM热红外遥感影像反演兰州市中心城区地表真实温度,采用面向对象的分形网络演化算法对地温图进行分割,获取热场基元,通过G*i指数的空间聚集分析提取热岛范围;选取破碎度、优势度、多样性指数、斑块数目等表征城市热岛空间格局变化的景观指数,分析了不同季节城市热岛的空间变化特征。结果表明,兰州中心城区热岛效应具有明显的季节变化特征,以春季(3月)最强,其次是夏季和秋季,冬季(1月)的热岛效应最弱。景观级别上,城区热场景观破碎度指数呈现春季、夏季、秋季、冬季依次减小的趋势;优势度指数从春季到冬季持续增加;冬季的Shannon多样性指数最小,热力景观趋于多样化。类型级别上,夏季的热岛破碎程度最大,而冬、春季破碎度较小。

The present paper aims to introduce a remote sensing analysis of the seasonal dynamic variations of the urban thermal environment of Lanzhou city. As. is known, with the fast advancing pace of urbanization, the urban heat island （UHI） effect has become a conspicuous influential factor in deteriorating the urban eco-environment. As a typical valley city, Lanzhou has its own particular thermal environment since its evolution into a modcru metropolis with its own specific characteristic urban landscape. As a result, the fast urbanization of the city has made its UHI ever more intensified. To study the UHI effect of the city, it seems more realistic and instructive to quote four scenarios of the thermal infrared remote sensing data （Landsat TM images）, collected in January, March, June and September in 2009,respectively, so as to analyze the spatial pattern of UHI in different seasons. And, for our purpose, we would like first of all to quote the ＂6S＂ model to rectify atmosphere, and then apply the mono-window algorithm to illustrate the land surface temperature （LST） fluctuations, and, finally, a normalized method has been adopted to divide the temperature rating categories so as to obtain the LST maps and the LST rating maps. Trying to obtain the urban heat island ratio index, it is necessary to extract the urban heat field elements by using the image segmentation algorithms of the target-oriented fractal net evolution approach （FNEA）, with the heat island seasonal pattern distinguisbed by G＋ index spatial aggregation analysis. The landscape metrics were used to quantify the variations of the urban thermal landscape pattern. The following facts should be clarified via the appraisal system of temperature landscape ： the UHI in the city can help to show the remarkable differences in a seasonal pattern. Generally speaking, the UHI of the city is likely to be most intensive in spring ; it is getting moderate in summer and autumn, and becomes weakest in winter. So far as the landscape level is concerned, the most conspicuous landscape fragmentation of the urban thermal field appears in spring, whereas the second comes in summer, and, then, followed by autumn and winter. Thus, the thermal landscape pattern indicates significant diversity in their characteristic features with seasonal alteration. On the other hand, the dominant index of the urban thermal field has been found increasing from spring to winter, whereas the Shannon diversity index goes up from winter to spring. From the point of view of classification, it can be thought that the fragmentation of UHI tends to reach the maximal limit in summer, followed by winter and spring. Therefore, the fragmentation changing trend of cold is- lands turns out to be consistent with its heat islands, however, the degree of fragmentation on the whole proves less than the heat island. The second beat island tends to be continuous in winter and interrupted in spring. What is more, we have analyzed the correlation between NDVI and the land surface temperature. As a result, we have come to the conclusion that the heat island strength may imply a negative hnear correlation with the urban vegetation coverage both in summer and autumn. Thus, it can be thought that the climate change and human activities should account for the seasonal changes of surface heat regime, to which the increase of urban green land may contribute to the reduction of the urban heat island effects.