‘Before Los Angeles, there was South Wales’: Computing the Impact of the 2019-2020 New South Wales Bushfires

In response to the devastating 2019-2020 New South Wales bushfires, which caused unprecedented ecological and economic damage, this study seeks to prove the importance of computational tools by using python as a remote sensing tool to compute pre and post bushfire Landsat-8 images of New South Wales. It calculated the extent of the damages in terms of burnt areas and changes in vegetation during the incident using indices such as the Normalised Burnt Ratio (NBR) and the Normalised Difference Vegetation Index (NDVI). Results reveal a difference of 98.38% and 35.81% between the mean and standard deviations of NBR of the post and pre fire periods, respectively. Additionally, the study reveals the total burnt area in 2020 as 6412.67km2 (641267 ha) and the total high severity burnt area as 72.54km2 (7254 ha). This study demonstrates the effectiveness of python for geospatial analysis, offering a reproducible and adaptable approach for assessing wildfire impacts. However, limitations such as threshold-based classification and memory constraints draw attention to the need of advanced machine learning and statistical models to replace thresholding techniques in remote sensing, thus enabling consideration of local vegetation type and condition to improve classification accuracy.