Document Type : Research Paper
Authors
1
Department of Physical Geography, Faculty of Social Sciences, University of Mohaghegh Ardabili, Ardabil, Iran.
2
Master's Student ، Remote Sensing and Geographic Information Systems (GIS) ، Department of Physical Geography ، Faculty of Social Sciences ، University of Mohaghegh Ardabili ، Ardabil ، Iran، sinakhonkham@gmail.com
3
Professor ، Department of Physical Geography ، Faculty of Social Sciences ، University of Mohaghegh Ardabili ، Ardabil ، Iran،. zeynali.b@uma.ac.ir.
4
Master's Student ، Remote Sensing and Geographic Information Systems (GIS) ، Department of Physical Geography ، Faculty of Social Sciences ، University of Mohaghegh Ardabili ، Ardabil ، Iran، samadi.f.2019@gmail.com.
5
PhD in Geomorphology, Department of Physical Geography ، Faculty of Social Sciences ، University of Mohaghegh Ardabili ، Ardabil ، Iran، Aboozarsadeghi@uma.ac.ir.
6
PhD Student in Climatology,Department of Physical Geography ,Faculty of Social Sciences, Mohaghegh Ardabili University, Ardabil, Iran, Azardokht.r208@gmail.com
10.22034/gp.2026.72905.3559
Abstract
Objective: Accurate monitoring of snow cover, a critical component of the hydrological cycle in mountainous regions, is essential for water resource management and climate change assessment. This study evaluates and compares the performance of five snow indices—Normalized Difference Snow Index (NDSI), S3, Snow Ratio Index (SRI), Normalized Difference Snow Water Index (NDSWIn), and Snow Surface Reflectance (SSR)—for snow cover extraction using Sentinel-2 and Landsat 8/9 imagery in the Ardabil and Marivan regions of northwestern Iran.
Methods: Six satellite images acquired between 2022 and 2025 were processed via Google Earth Engine. Snow cover maps were generated for each index, and classification performance was assessed using 200 reference points, employing Overall Accuracy (OA) and Kappa coefficient metrics.
Results: All indices detected snow-covered areas; however, S3 and NDSWIn consistently achieved the highest accuracy across both sensors and regions. In Ardabil, their mean OA exceeded 99% with Kappa ≈ 0.99. Landsat 8/9 generally outperformed Sentinel-2 in accuracy and temporal stability, while SSR showed the highest variability and lowest performance throughout the study period.
Conclusions: S3 and NDSWIn are the most reliable and effective indices for snow cover mapping in mountainous regions of northwestern Iran, supporting operational snow monitoring using multispectral satellite imagery.
Keywords
Main Subjects
)