Land surface temperature in an arid city: assessing spatio-temporal changes

Abstract

In a city located in a desert environment (Tulum Valley, Argentina) we proposed to assess in the coldest and warmest periods (1) the urban cold/heat island (UCI/UHI) phenomenon; 2) which driving factors currently affect the Land Surface Temperature (LST). In the study area, we selected 50 points for the urban class and 49 points for the rural class. The LST data was obtained from Landsat 5 TM y Landsat 8 OLI/TIRS, for 1988, 2000, 2010, and 2021 years. As driver factors, we assessed the median (med) and standard deviation (sd) of NDVI (Normalized Difference Vegetation Index), NDBI (Normalized Differences Built-up Index), and BI (Brightness Index). The Tulum Valley behaved like a UCI during almost all studied years, except for 2010 and 2021. Probably because the city was undergoing reconstruction after a major earthquake. In the urban class, the LST were affected positively by NDVImed during the cold period, while in the rural class the LST were explained by a negative relationship with NDVImed, and positive with NDBImed. In the urban class during the warm periods, the LST were affected negatively by BIsd, BImed, NDVImed, and NDVIsd, which leads to a decrease in LST. Contrarily, an increase of NDBIsd produces an increment of LST. The only driver for LST in the rural class was NDBImed, which had a positive effect. Different drivers affected the LST behavior in Tulum Valley, moreover, these drivers explained more variability in rural than in urban class.