Analyzing landscape changes and their relationship with land surface temperature and vegetation indices using remote sensing and AI techniques

Abstract Land use patterns and consumption occur widely due to fast industrialization and development in the previous several decades, which might lead to problems such as over-exploitation of land resources, food shortages, and pollution. Monitoring and subsequent modeling of land use land cover (LULC) changes has become critical. A study of the variations in the LULC pattern of the Baghpat District of Uttar Pradesh, India, was attempted. This study assessed spatial patterns and fluctuations in growth in the Baghpat District of Uttar Pradesh (India) from 1991 to 2021. The study also analyzed the land cover changes and their effects on land surface temperature (LST), normalized difference vegetation index (NDVI), and soil indices in the Baghpat district. Decadal land use and land cover (LULC) changes were analyzed using Multitemporal Landsat Imagery and applying the maximum likelihood classifier in ENVI (Image Processing Software). Post-classification spatial measures were used to examine changes in LULC and the spatial distribution of urban growth, as well as to identify changes using ArcMap (GIS Software) across the period. Various AI techniques were used to show the trend variation in NDVI, LST, and SAVI indices using IBM-SPSS, Microsoft Office, OriginLab, and MATLAB for the study area to comprehend the variation in the index within the given period. The findings indicated significant improvements in agriculture between 1991 and 2021 (from 58.94 to 84.79%), but significant declines in vegetation cover (from 29.53 to 1.14%). The yearly percentage growth of the parallel built-up area was 3.77%, 5.59%, 6.71%, and 6.90%, respectively. Approximately 43.85% of the increase in agricultural land between 1991 and 2021 came from the conversion of vegetation covers, which fell by 96.13%. The analysis of LST, NDVI, and SAVI data revealed a substantial negative association for all years, except a slight positive correlation. NDVI and SAVI values were highest in agricultural fields with the lowest LST values, whereas fallow land regions exhibited the reverse pattern. With the help of these findings, urban planners and designers may reduce various socio-economic and environmental consequences.