NDVI、NDMI与地表温度关系的对比研究

通过研究归一化植被指数(NDVI)、归一化水汽指数(NDMI)与地表温度(LST)的相关关系,对比NDVI和NDMI定量分析LST的适宜程度。以Landsat 8遥感影像为数据源,以郑汴都市区为例,反演LST,计算NDVI和NDMI。从整体、分区、像元等不同层面分析NDVI和NDMI与LST的相关关系,并利用GEO-Da软件,分析150 m、300 m、450 m三种不同采样间隔数据的NDVI、NDMI与LST的空间相关性。主要结论为：①NDVI与LST的线性拟合度较差,而NDMI与LST具有很强的线性关系,剖面分析显示NDMI与LST呈显著负相关关系。②缓冲区分析结果表明,随着距城市中心距离的增加,用地类型增多,LST和NDMI之间的相关性逐渐增强。③在Moran's I空间相关性分析中,不同采样间隔下两指数与LST的负相关关系均比较明显,但由于水体在2个指数中数值的差异,使NDMI与LST表现出更强的空间负相关性,而NDVI与LST空间相关性则相对较弱。因此,总的来说NDMI是热环境研究的有效指标,与NDVI相比,NDMI与LST的负相关关系更强、更稳定,也更适宜于对LST的定量分析。

Comparative study on the correlations between NDVI,NDMI and LST

Through the comparative study of the correlations between the normalized difference vegetation index (NDVI), normalized difference moisture index (NDMI) and land surface temperature (LST), the suitability to research LST using NDVI and NDMI was verified. Based on Landsat 8 remote sensing imagery in the Zhengzhou-Kaifeng metropolitan area, LST was retrieved, and both NDVI and NDMI were calculated. At the overall, regional, and pixel levels, the correlations between LST and NDVI, NDMI were analyzed. Using GEO-Da, the spatial correlations between LST and NDVI, NDMI were simulated with the data of three sampling intervals of 150 m, 300 m and 450 m. The conclusions are as follows. First, there is a stronger linear negative correlation between LST and NDMI, and sectional analysis shows that NDMI resembles a mirror image of LST, whereas the correlation between LST and NDVI is much weaker. Second, buffer analysis shows that with the increase of distance from the down town and increase of the number of land use types, the correlation between NDMI and LST also increases gradually. Third, in the analysis of Moran's I spatial correlation, the negative correlations between LST and the two indices were significant, but because of the great difference between the values of water body in the two indices, NDMI and LST show stronger negative correlation, while the correlation between NDVI and LST is relatively weak. In general, compared with NDVI, NDMI is more suitable as an effective indicator for quantitative analysis of LST.