Automated cloud-based monitoring of spatiotemporal water quality dynamics in a tropical coastal lagoon: Ciénaga de Mallorquín, Colombia
The Mallorquín Lagoon (Ciénaga de Mallorquín), a tropical coastal ecosystem in northern Colombia, has experienced severe environmental degradation due to urban expansion, wastewater discharges, and hydrological alterations, while existing water quality monitoring remains spatially and temporally...
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Dokumenttyp: | Article (Journal) |
| Sprache: | Englisch |
| Veröffentlicht: |
26 March 2026
|
| In: |
Regional studies in marine science
Year: 2026, Jahrgang: 96, Pages: 1-20 |
| ISSN: | 2352-4855 |
| DOI: | 10.1016/j.rsma.2026.104952 |
| Online-Zugang: | Verlag, kostenfrei, Volltext: https://doi.org/10.1016/j.rsma.2026.104952 Verlag, kostenfrei, Volltext: https://www.sciencedirect.com/science/article/pii/S2352485526002070 
|
| Verfasserangaben: | Estefany Villanueva-García, Juan Camilo Restrepo-López, Lihki Rubio, Sebastián Narváez-Salcedo, Jorge Omar Pierini |
| Zusammenfassung: | The Mallorquín Lagoon (Ciénaga de Mallorquín), a tropical coastal ecosystem in northern Colombia, has experienced severe environmental degradation due to urban expansion, wastewater discharges, and hydrological alterations, while existing water quality monitoring remains spatially and temporally limited. To address this monitoring gap, we developed and deployed an automated, scalable, cloud-based monitoring framework using Sentinel-2 MSI imagery and machine learning (ML) models calibrated and validated with in situ data to monitor spatiotemporal variations in the lagoon’s water quality from 2016 to 2023 under data-limited conditions. Support Vector Regression, Random Forest, and XGBoost algorithms were implemented within Google Earth Engine (GEE) to estimate surface temperature, salinity, suspended sediment concentration (SSC), and chlorophyll-a (Chl-a). Model validation achieved R² values between 0.63 and 0.75, indicating robust predictive performance for most variables. Salinity reached 40 PSU in the dry season and dropped to 25 PSU in the wet season, showing a strong inverse correlation with temperature (ρ = -0.68). SSC and Chl-a were positively correlated (ρ = 0.67), suggesting a potential link between wind-driven sediment resuspension and phytoplankton dynamics during dry months. However, Chl-a predictions were constrained by limited in situ data, highlighting the need for expanded field monitoring to improve model robustness. Spatiotemporal analyses revealed spatially salinity patterns and temporal changes consistent with an emerging hyper-salinization tendency in the lagoon. All outputs were integrated into a public cloud-based web platform that enables near-real-time visualization and analysis in collaboration with the regional environmental authority. This study provides a scalable and cost-effective framework for operational environmental monitoring in data-limited tropical coastal lagoons, supporting data-driven management and restoration strategies. |
|---|---|
| Beschreibung: | Gesehen am 02.06.2026 Online veröffentlicht: 26. März 2026 |
| Beschreibung: | Online Resource |
| ISSN: | 2352-4855 |
| DOI: | 10.1016/j.rsma.2026.104952 |