Satellite remote sensing for post-fire recovery: sustainable reforestation using NDVI, LST and SMI

Purpose This study aims to investigate the impact of wildfire on soil moisture, vegetation health and surface thermal conditions in a forested area east of Parnitha, Greece. It evaluates the ecological changes caused by the fire by analyzing key remote sensing indices. These include the normalized difference vegetation index (NDVI), soil moisture index (SMI) and the derived land surface temperature (LST), all of which support forest recovery efforts. Design/methodology/approach A quantitative approach was employed using open-source Landsat 8 collection 2 level 2 science product (L2SP) data from 2020 (pre-fire) and 2022 (post-fire), provided by the United States Geological Survey. The analysis focuses on a wildfire that started in 2021, leveraging the inverse relationship between land surface temperature and normalized difference vegetation index to calculate soil moisture index. Data preprocessing, including scaling and normalization, was conducted in ArcGIS Pro. Subsequently, LST equations for SMI estimation were derived using linear regression in MATLAB. To evaluate the reliability of the results, statistical analyses were performed. These included the coefficient of determination (R2), distribution analysis via histograms and calculation of the mean and standard deviation (Std). A correlation matrix was also used to assess variable interrelationships and temporal changes. Findings The results reveal substantial ecological changes post-fire: a marked increase in surface temperature, reduced soil moisture levels and significant degradation of vegetation health. Specifically, SMI values show a shift toward drier conditions, NDVI indicates a decline in vegetation cover and density and LST exhibits increased mean and variability. Originality/value This study employs a novel integration using remote sensing indices and the L2SP dataset, which provides preprocessed surface reflectance and LST products. This approach reduces computational effort while maintaining analytical effectiveness. It delivers more accurate and efficient insights to support sustainable forest management and reforestation strategies in fire-prone environments.