Evaluating the Applicability of Satellite Imagery Indices for Predicting Tree Mortality in Surface Fire Damaged Forests

This study aimed to develop a comprehensive model for predicting fire-damaged pine tree survival and assess the usefulness of satellite imagery techniques. Logistic regression models were used to create integrated survival prediction models at two sites three years after fire damage. Diameter at breast height, char height index, and aspect were identified as key factors for survival prediction. The integrated model achieved a high accuracy level, with a predictive power of 82%. Remote sensing techniques were employed to estimate mortality rates at the stand level and identify relevant factors for survival prediction. Initial changes in dNBR, which showed a strong correlation with mortality rates, were found to be influential factors. While these factors could potentially improve the model as new indicators, clear trends under legend criteria were challenging to establish. High spatial and spectral resolution are necessary to enhance accuracy, along with classification categories specifically designed for fire-damaged areas.