DOI: 10.26650/ase.2026.1913971 ISSN: 2602-473X

Assessing the Effects of Mussel Aquaculture on Coastal Water Quality Using Sentinel-2 Imagery: An Installation-Based Approach in the Sea of Marmara

Gökhan Tunçelli
Show PDF Cite
Mussel aquaculture is often considered environmentally sustainable due to its potential to filter particles and recycle nutrients, yet its large-scale influence on coastal water quality remains difficult to quantify, especially in complex and highly variable systems such as the Sea of Marmara. Here, an installation-based analytical framework was implemented to evaluate potential mussel farm associations with turbidity and chlorophyll dynamics using Sentinel-2 satellite observations processed in Google Earth Engine. Twenty-one mussel farms in the southern Sea of Marmara were confirmed and digitized from high-resolution Google Earth imagery (most recent imagery year: 2025), and approximate installation years were assigned using time-series imagery interpretation supported by administrative information. Control areas were defined as a 1–3 km ring buffer around each farm to improve interpretability. Turbidity (B3/B2) and the Normalized Difference Chlorophyll Index (NDCI) were derived from annual and monthly median Sentinel-2 surface reflectance composites for 2018, 2020, 2022, and 2024. Annual-scale analyses (ANOVA and linear regression) detected no statistically significant pre- vs post-installation differences (p>0.05) and no significant relationships with time since installation. However, seasonal comparisons during the spring bloom period (March–May) revealed a recurrent negative chlorophyll difference in farm areas relative to controls, although the magnitude was limited and not consistently statistically strong. Overall, results indicate that mussel-related signals in the Sea of Marmara are scale- and season-dependent: patterns that may be consistent with filtration-related processes were more apparent under high-productivity conditions but are largely masked by strong regional variability and mixing, as well as known uncertainties in coastal optical remote sensing.

More from our Archive