Abu Dhabi has changed a lot over the past two decades: deserts have given way to neighborhoods, parks, and roads. A new study reveals that this transformation has made the area cooler. By analyzing 20 years of satellite images, researchers mapped three land typesβurban, vegetation, and sand. Using AI (machine learning) and GIS tools, they found that the average temperature dropped from 45.3β―°C in 2001 to 39.4β―°C in 2019. Vegetation is the main driver: each extra percent of green space cools the surface by about 7β―°C. Urban growth helps too, but less so, especially when it replaces sand with heatβtrapping asphalt and concrete. Reducing exposed desert surfaces also contributed to the cooling trend. The findings suggest that planting trees, creating green spaces, and managing urban expansion can be powerful tools for climate adaptation in arid regions like Abu Dhabi.
The research employed Landsat 7 imagery (2001β2019) within a hybrid QGISβGoogle Colab workflow. A supervised classification generated landβcover maps with four classes: urban, vegetation, sand/desert, and water. Classification accuracy was assessed via a confusion matrix, achieving an overall kappa >0.85. To link land cover to LST, simple linear regressions were run between the percent cover of each class and the annual mean LST. Quantitatively, a 1β―% rise in vegetation cover corresponded to a 6.92β―°C drop in LST, whereas urban expansion alone was not statistically significant after adjusting for vegetation and sand reduction. Spatial analysis revealed βcool islandsβ around new parks and green corridors, highlighting how landscape design can ease urban heat in hyperβarid settings. The findings advocate combining remote sensing, ML, and GIS for climateβresilient planning in dryβland cities.
