Land Surface Temperature and Urban Heat Island Dynamics from 2010 to 2018 Using Landsat Imagery in an Arid Environment: A Case Study in Sunland Park, New Mexico, USA

Abstract

Urbanization significantly impacts land surface temperature (LST) and the characteristics of urban heat islands (UHI). However, the relationship between these factors can be complex, particularly in arid regions where exposed soils and sparse vegetation can lead to exceptionally high daytime temperatures. This study explores how changes in land use and land cover (LULC) related to urbanization affect LST, vegetation conditions, and UHI characteristics in Sunland Park, New Mexico. We analyzed five cloud-free Landsat images from the Landsat-5 TM and Landsat-8 OLI/TIRS satellites, captured in May during the years 2010, 2011, 2013, 2016, and 2018. To map LULC changes, we employed supervised maximum likelihood classification, while the Normalized Difference Vegetation Index (NDVI), emissivity-corrected LST, and a normalized UHI index were calculated to assess daytime thermal patterns on the surface. The overall accuracy of the classification ranged from 85% to 87%, with Kappa coefficients between 0.74 and 0.78. Over the study period, we observed an increase in urban land from 7.53 km² to 12.23 km², alongside a decrease in rangeland from 17.82 km² to 12.30 km². Our analysis indicated that the highest daytime LST values and the most intense UHI effects were typically linked to exposed barren land and sparsely vegetated rangelands. Notably, certain urban zones experienced lower LST than adjacent barren areas. A significant inverse relationship between NDVI and LST (R² = 0.60–0.80) highlighted the cooling effects of vegetation, largely attributed to shading and evapotranspiration. These findings suggest that daytime thermal conditions in arid areas are shaped not only by the expansion of urban areas but also by the specific characteristics of the land being converted and the land management practices that follow. It is important to note that this analysis, based on five selected Landsat scenes from May, should be considered reflective of the specific acquisition dates rather than indicative of year-round thermal behavior or a general trend that urbanization decreases surface temperatures in arid environments.