Landsat 9 Satellite Images Potentiality in Extracting Land Cover Classes in GEE Environment Using an Index-Based Approach: The Case Study of Savona City

Land use and land cover modeling is an essential tool because it enables scientists and policymakers to foresee prospective changes in landscape heritage and examine trends to minimize potential dangers. To attain this purpose, a continuous stream of data needs be collected and examined. Landsat missions present a viable alternative since they have been collecting continuous data for five decades, and a new platform was launched at the end of September 2021 to avoid disrupting such a series. Consequently, the purpose of this research is to assess the quality of Landsat 9 data in extracting land use information using the index-based approach. Following the conclusion of the collection and pre-processing operations, two of the most often used vegetation indices, NDVI and MSAVI2, were derived from a Landsat 9 data covering Savona city, which was chosen as the pilot site due to its unique geomorphological characteristics. Lastly, maps accuracy was assessed by computing the confusion matrices, k estimator and overall, producer and users accuracy. The entire method was implemented in the free cloud computing platform Google Earth Engine by writing custom Java code. The generated land use/cover maps were both satisfactory, albeit the MSAVI2 had a greater overall accuracy (90.24% vs 79.60%) and K parameter (84.45% vs 71.70%) due to its ability to minimize soil spectral effect. Those findings are consistent with those derived from Landsat 8 images. This means that Landsat 9 is an excellent successor to Landsat 8.