Visualization of Snow Cover Changes in the Greater Caucasus Mountains Using Integrated Multi-Sensor Image Fusion

Abstract

The main goal of the current research was to visualize recent snow cover changes in the parts of the Greater Caucasus Mountains along Shahdag, Bazarduzu, and Tufandag months, using multi-sensor and multi- spectral satellite image processing. Accordingly, accessible satellite imagery from the valid sites was obtained, and a few necessary pre-processing operations, such as atmospheric and radiometric corrections, were subjected to available Sentinel-2 and Landsat (8 and 9) images from 2017 to 2023, adjusted to annual, spring-summer months. Next, we employed a harmonized method to combine the Harmonized Landsat and Sentinel (HLS) image bands using object-oriented functions within the eCognition software. This integration led to diverse products, especially High Optimized Natural Color (HONC) images and Scene Classification maps of the study area. Finally, it became possible to evaluate changes in the snow covers with higher accuracy in the annual and monthly time scales. Examining the amount of snow cover, represented by the Normalized Difference Snow Index (NDSI) in recent years shows a significant reduction in snow packs on an annual scale, with rapid melting from the middle of spring to the end of summer months. The final results indicate the fact that the methods of optimization and fusion of Landsat and Sentinel sensors can be very effective in extracting and identifying snow cover with high accuracy on a daily scale.

Keywords

• snow cover changes
• greater caucasus mountains
• azerbaijan
• hls images
• integrated processing methods