EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Shoreline Change Detection in Bandar Anzali (2015–2025) Using Landsat-8 Spectral Indices and Machine Learning: Implications for Coastal Management and Port Sustainability

Kiana Kazari, Mahdi Hasanlou

Khazar Journal of Water and Environment · 2026

DOI: 10.68277/kjwe.2026.2 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Khazar Journal of Water and Environment, 2026. Verfasst von Kiana Kazari und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Bandar Anzali, situated on the southern Khazar (Caspian) coast, is one of Iran’s most significant maritime hubs for trade, tourism, and fisheries. In recent years, the area has experienced pronounced sedimentation and shoreline progradation, particularly during summer months when hydrodynamic activity declines and sediment delivery from upstream sources increases. These eomorphological shifts threaten port operations, navigability, and coastal infrastructure, underscoring the need for precise shoreline monitoring to support sustainable coastal management. This study applied satellite-derived spectral indices—NDWI, MNDWI, and NDVI—from Landsat 8 imagery to classify water and non-water surfaces in the Anzali coastal zone over 11 summers from 2015 to 2025. To improve classification accuracy, Random Forest and Support Vector Machine (SVM) algorithms were trained using representative land–water samples, surface reflectance bands, and spectral indices. The SVM model, implemented with an RBF kernel, demonstrated greater ability to detect subtle shoreline transitions and shallow, sediment-laden waters. Shoreline displacement was quantified using Net Shoreline Movement (NSM) and End Point Rate (EPR) derived from the classified water masks. Random Forest produced an NSM of 147.7 m and an EPR of 14.77 m/yr, while SVM yielded an NSM of 146.8 m and an EPR of 14.68 m/yr. The minimal discrepancy between the two models indicates high consistency in capturing shoreline advance, with SVM offering slightly finer delineation in transitional zones. These results confirm ongoing sedimentation and shoreline growth in the Anzali system, particularly near the port entrance and lagoon margins, with direct implications for dredging, port accessibility, and ecological stability. Integrating spectral indices, machine learning classification, and quantitative shoreline metrics provides a robust framework for monitoring sediment-prone coasts. It supports strategic planning for resilience in economically critical regions such as Bandar Anzali.

Zitieren

APA

Kazari, K., & Hasanlou, M. (2026). Shoreline Change Detection in Bandar Anzali (2015–2025) Using Landsat-8 Spectral Indices and Machine Learning: Implications for Coastal Management and Port Sustainability. Khazar Journal of Water and Environment. https://doi.org/10.68277/kjwe.2026.2

Vancouver

Kazari K, Hasanlou M. Shoreline Change Detection in Bandar Anzali (2015–2025) Using Landsat-8 Spectral Indices and Machine Learning: Implications for Coastal Management and Port Sustainability. Khazar Journal of Water and Environment. 2026. doi:10.68277/kjwe.2026.2

BibTeX
@article{Kazari2026,
  author = {Kazari, Kiana and Hasanlou, Mahdi},
  title = {Shoreline Change Detection in Bandar Anzali (2015–2025) Using Landsat-8 Spectral Indices and Machine Learning: Implications for Coastal Management and Port Sustainability},
  journal = {Khazar Journal of Water and Environment},
  year = {2026},
  doi = {10.68277/kjwe.2026.2}
}
Kostenlos starten →