Assessment of Outdoor Thermal Comfort during the last decade Using Landsat 8 Imageries with Machine Learning Tools over the Three Metropolitan Cities of India

Since there is rapid expansion of urban areas and significant growth in urban populations, there is a pressing need to accurately assess changes in land use and land cover (LULC). Such changes play a pivotal role in predicting outdoor thermal comfort. Alterations in LULC can considerably impact local meteorological conditions, subsequently affecting thermal comfort. Thus, we endeavoured to investigate the spatial patterns in outdoor thermal comfort across the cities of Hyderabad, Bangalore, and Jaipur.

To achieve this, we utilized high-resolution imagery from Landsat 8 along with on-site meteorological data. The Support Vector Machine (SVM) incorporating principal component analysis (PCA) was used to estimate Thermal comfort. Seven environmental indices—namely, normalized difference vegetation index (NDVI), normalized difference water index (NDWI), new built-up index (NBI), land surface temperature (LST), brightness, greenness, and wetness—were used as independent variables. Addressing the issue of multi-collinearity among these variables was accomplished through PCA. LULC classification was performed by the decision tree method.

The LULC analysis unveiled intriguing trends. In Hyderabad, the proportion of built-up areas surged from 37% to 48% between 2009 and 2019, while barren lands notably diminished from 42% to 18%. Water bodies maintained a consistent coverage of around 3%. Notably, vegetation expanded from 20% to 30%, with the northwestern part becoming more verdant and the southern region becoming more urbanized. Similarly, in Bangalore, built-up areas escalated from 25% to 80%, resulting in a substantial loss of vegetation (25% to 2%) and a reduction in bare lands (50% to 18%). While water bodies experienced a minor decrease, the trend was noticeable. In Jaipur, built-up areas increased by approximately 12%, accompanied by a marginal uptick in greenery. Water bodies, however, remained almost negligible within the city.

The outcomes of the thermal comfort analysis demonstrated that the most pronounced urbanization transpired in Bangalore, whereas Jaipur exhibited the least urban expansion. Discomfort levels were highest in bare lands, followed by urban areas, vegetation zones, and water bodies. During the summers from 2009 to 2019, Hyderabad encountered varying degrees of discomfort, with Bangalore also witnessing similar conditions, and Jaipur experiencing discomfort most of the time. However, during the winter seasons, Bangalore transitioned from neutral to comfortable conditions, while Hyderabad and Jaipur predominantly maintained neutral levels of comfort.