Stronger Vegetation Cooling in a Warming City: A Google Earth Engine Analysis of NDVI–Land Surface Temperature Relations in Rajshahi, Bangladesh (2001–2025)

Abstract

Rapid urbanisation in mid-sized South Asian cities has intensified the surface urban heat island, yet satellite assessments of the vegetation–temperature relationship for Rajshahi, Bangladesh, remain scarce. This study quantifies it across 546 km² of Rajshahi City for four pre-monsoon periods (2001, 2009, 2017 and 2025) from Landsat 5, 8 and 9 imagery processed in Google Earth Engine at 30 m resolution. The Normalised Difference Vegetation Index (NDVI) and land surface temperature (LST) were derived per pixel and analysed on stratified random samples (n ≈ 9,400–9,600 per year). A highly significant negative NDVI–LST correlation was found in all four years: 2001 (r = −0.649), 2009 (r = −0.855), 2017 (r = −0.481) and 2025 (r = −0.721), all at p < 0.001. The strongest coupling occurred in the drought year 2009 and the weakest in the wetter 2017 pre-monsoon, indicating climatic modulation rather than breakdown. The absolute cooling slope steepened from 8.32 to 11.58 °C per NDVI unit between 2001 and 2025, a 39.1 % intensification that is highly significant (Fisher z = 9.40, p < 0.001). The climate-regulating value of vegetation is therefore rising as Rajshahi warms, strengthening the case for protecting its mango orchards and peri-urban cropland. This version supersedes the earlier deposit. The processing pipeline was rebuilt in full following an internal methodological audit, and all results were regenerated from the scripts and data included in this record. Earlier versions remain accessible through the version history.