Analyzing long-term spatio-temporal patterns of land surface temperature in response to rapid urbanization in the mega-city of Tehran

Abstract

Analyzing and monitoring spatial and temporal urbanization is vital for a better understanding of land surface temperature (LST) variations. Only a few studies have examined LST variations in mega-cities regarding long-term spatial and temporal scales. First, this study compared heat differences between urban and sub-urban areas (referred to hereafter as the regional scale) in the entire mega-city of Tehran and urban heat variation among 22 regions located within Tehran (referred to hereafter as the local scale) during last three decades. Second, we examined how land cover, climate and elevation data were correlated with urban heat for regional and local scales in the mega-city of Tehran. A series of six Landsat TM images of Tehran in 1986, 1991, 1994, 2001, 2008, and 2011 were classified into four land cover classes (built-up, crop, open space, and green lands). These images were also used to calculate land surface temperature (LST) and normalized difference vegetation index (NDVI). We used 30m elevation data from the Shuttle Radar Topography Mission. We also downloaded 1km climate data including temperature, precipitation, solar radiation, vapor pressure and wind speed from Global Climate. At the regional scale, we found that urbanization in the mega-city of Tehran doubled during last three decades increasing from 21% in 1986–43% in 2011. Moreover, the mean LST difference among various land cover classes was on average 5.70°C in each time point. Furthermore, the mean LST for various land cover classes increased on average 6.46°C from 1986 to 2011. We also found that mean NDVI, mean elevation and mean climate were more effective in reducing mean LST inside Tehran than outside Tehran. At the local scale, regions north of Tehran experienced the lowest mean LST values while regions south and southwest of Tehran experienced the highest mean LST values from 1986 to 2011. Moreover, the mean LST difference among the 22 regions in each time point observed was 6.77°C on average. Furthermore, the mean LSTs for all regions increased on average 3.75°C from 1986 to 2011. As the mean LST increased, the mean NDVI became a stronger factor in mitigating the mean LSTs in all 22 regions. Finally, mean NDVI, mean elevation and mean climate across 22 regions played significant role to control LST variations inside Tehran. We concluded that long-term spatial and temporal analysis can inform decision-makers for better planning to mitigate and control urban heat variations in the mega-city of Tehran.