Assessment of Impacts Controllable Factors on Urban Heat Islands Using Multiple Regression Analysis (Case Study: Tabriz Metropolis)

Urban heat islands have become one of the most challenging issues in urban studies. In addition to their primary impact on average temperature, heat islands can also have secondary effects on local meteorology. Urban heat islands are primarily caused by excessive surface heating and heat storage, pollution emissions, human heating, wind circulation barriers, and reduced evapotranspiration.

In this study, the split-window algorithm of Landsat 8 sensor with a combination of thermal bands 10 and 11 was used to obtain land surface temperature during the day, and the MODIS sensor was used to obtain land surface temperature at night. To determine the heat island region in the city, global and local Moran's I autocorrelation analyses and Getis-Ord Gi hotspot analysis, both performed in ArcMap software, were used. Multiple linear regression analysis in SPSS software was used to obtain the maximum impact of indices.

The results indicate that the hottest points during the day were in Region 6 around Tabriz Airport with a temperature of 27.27 degrees Celsius, and at night, the temperature was 17.07 degrees in Region 3. According to spatial Moran's I analysis, both heat islands had a clustered pattern. According to regression results, open spaces and soil texture, with a score of 0.700, had the greatest impact on urban heat islands due to their high heat capacity and were the coolest surfaces at night. The least impact was from built spaces with reflective roofs, with a score of -0.192, which had a cooling effect. Building height with shading and random density had a negative effect on increasing heat. However, the influential index on nighttime heat islands was in conditions where the building density was moderate and the orientation was appropriate, but the height-to-width ratio of the passageway was 1.31 and 0.825, respectively, which caused heat retention.