Spatio-temporal changes in the surface urban heat island extent between 1986-2021 in the polycentric agglomeration, southern Poland based on Landsat satellite images

<p>Currently, more than 50% of the world's population lives in cities, but it is estimated that this number will increase, directly affecting the development of urban infrastructure. An increase in impervious surfaces in the city causes the intensification of the surface urban heat island (SUHI), manifesting itself in an increased temperature in a city compared to the surroundings. In summer, the SUHI is considered a hazardous phenomenon on a local scale for city inhabitants. Although the Górnośląsko-Zagłębiowska Metropolis (GZM) is one of the most urbanised and populated areas in Poland, the issue of SUHI has not yet been addressed for the entire metropolis.</p><p>This study discusses spatial and temporal changes in the extent and the risk of SUHI in GZM between 1986 to 2021. Based on LANDSAT data recorded in summer, the land surface temperature was estimated (LST) for twenty-one satellite images. The extents of surface heat island (SHI) and SUHI were marked out based on mean LST and standard deviation. The SHI occupied from 11.5 to 18.5% of the GZM. Variability in the vegetation of the agricultural areas (before and after the harvest) significantly contributed to intraseasonal variability in SHI extent. In contrast to SHI, when determining the SUHI extent, only urban areas (artificial surfaces), distinguished based on the Corine Land Cover (CLC) classification, were considered. Therefore, SUHI extent (4.2 to 13.8%, depending on date) was generally smaller than SHI distinguished only based on the thermal criterion. However, in both cases, there was an upward trend in their extents between 1986 and 2021. Additionally, the SUHI risk indicator, distinguishing the areas most exposed to SUHI based on land cover and demographic data, revealed that the number of districts exposed to a high risk of SUHI increased in the research period. Since SUHI extent depends on the contribution of various types of land cover, temporal changes in the percentage of land cover types in the GZM were also analysed. For this purpose, the land cover types in CLC classification existing for 1990, 2000, 2006, 2012, and 2018  were reorganised into the following five main types: (i) impervious surfaces; (ii) mines, dumps, and construction sites; (iii) agricultural areas; (iv) vegetation areas; (v) water areas. In the research period,  the most significant changes concerned the area of impervious surfaces (increase by 4.8%) and agricultural areas (loss by - 3.7%) that were transformed into built-up areas. The urban area extent increased from 26.9% in 1990 to 30.6% in 2018. Moreover, significant relationships were found between the SUHI area and inhabitant number and population density in GZM districts.</p>