Satellite-derived bathymetry for the Maltese islands, and new insights for analysing salinity and temperature from multispectral sensors

Abstract

Given the dynamic nature of coastal zones, it is imperative to gain an understanding of their evolutionary patterns. Safeguarding these ecosystems necessitates the ability to observe their physical features and controlling processes with precision in both space and time. This demands the acquisition of precise and up-to-date information regarding several coastal parameters. Thus, to gain a comprehensive understanding of these ecosystems, this study employs remote sensing techniques, machine learning algorithms and traditional in-situ approaches. Together, these serve as valuable tools to help comprehend the characteristics and mechanisms occurring within these transitional regions of the Maltese archipelago. An empirical workflow was implemented to predict the spatial and temporal variations in bathymetry, sea surface salinity (SSS), and sea surface temperature (SST) from 2022 to 2024. This was achieved by leveraging Sentinel-2 satellite platforms, the Random Forest (RF), and Linear Regression (LR) machine learning algorithms, as well as in-situ data collected from multi-beam echo sounder surveys, sea gliders, and floats. Subsequently, the numerical data generated by the various machine learning algorithms were validated with an error metric and converted into visual representations to illustrate the parametric variations across the Maltese Islands. The RF algorithm outperformed the LR model in predicting accurate bathymetric information for all three years, yielding highly precise bathymetric data for the entire Maltese archipelago. Furthermore, the RF algorithm demonstrated strong performance in predicting SSS and SST, indicating its capability to handle more dynamic parameters effectively. Lastly, the parametric maps generated for all three years provided a clear understanding of both the spatial and temporal changes within these three coastal parameters. Therefore, this study effectively combined satellite data, in-situ measurements, and machine learning algorithms to accurately predict bathymetry, SSS, and SST across three years within the Maltese Islands.