Spatiotemporal evolution of winter urban heat and vegetation dynamics in Ahmedabad based on decadal land surface temperature and NDVI analysis from 2015 to 2024

This comprehensive study analyzes Land Surface Temperature (LST) and Normalized Difference Vegetation Index (NDVI) trends in Ahmedabad, India, from 2015 to 2024 (excluding 2017) using three LST retrieval methods: Split Window (SW), Single Channel (SC), and Mono Window algorithms applied to Landsat 8 and 9 thermal infrared data, alongside NDVI derived from reflective bands. Quality Assessment (QA) masks were applied, filtering out approximately 15% of pixels on average by excluding clouds, water bodies, non-land surfaces, bare soil, and sparse vegetation areas. A detailed analysis of masking impacts reveals that cloud masking increases LST by 1.19–2◦ C, water masking by 0.4–0.8◦ C, non-land surface masking decreases LST by 0.1–0.3◦ C, soil masking decreases LST by 0.19–0.5◦ C, and vegetation masking increases NDVI by 0.01–0.03, with combined effects increasing NDVI by 0.04–0.05. The SW algorithm, adjusted for this study, indicates a warming trend of 0.42◦ C/year, matching the SC algorithm, while the Mono Window algorithm shows 0.43◦ C/year. NDVI declines at -0.0036/year, reflecting urbanization. Urban areas are 8–10◦ C warmer than riverbank zones, where the Sabarmati River reduces temperatures by 2–4◦ C. A strong negative LST–NDVI correlation (-0.81) underscores vegetation’s cooling effect. This paper provides detailed methodologies, algorithm comparisons, spatial and temporal analyses, and actionable implications for urban planning and climate adaptation in Ahmedabad, supported by robust remote sensing techniques [1, 2].