The quality of drinking water is a critical factor for public health and the environment. Inland drinking water reservoirs are essential sources of freshwater supply for many communities around the world. However, these reservoirs are susceptible to various forms of contamination, including the presence of muddy water, which can pose significant challenges for water treatment facilities and lead to serious health risks for consumers. In addition, such reservoirs are also used for recreational purposes which supports the local economy. In this work, we show as a proof-of-concept that muddy water mapping can be accomplished with machine learning-based semantic segmentation constituting an extra source of sediment-laden water information. Among others, such an approach can solve issues including (i) the presence/absence, frequency and spatial extent of pollutants (ii) generalization and expansion to unknown reservoirs (assuming a curated global dataset) (iii) indications about the presence of other pollutants since it acts as their proxy. Our train/test approach is based on 13 Sentinel-2 (S-2) scenes from inland/coastal waters around Europe while treating the data as tabular. Atmospheric corrections are applied and compared based on spectral signatures. Muddy water and non-muddy water samples are taken according to expert knowledge, S-2 scene classification layer, and a combination of normalized difference indices (NDTI and MNDWI) and are evaluated based on their spectral signature statistics. Finally, a Random Forest model is trained, fine-tuned and evaluated using standard classification metrics. The experiments have shown that muddy water can be detected with high enough discrimination capacity, opening the door to more advanced image-based machine learning techniques.

Satellite-derived bathymetry (SDB) enables the efficient mapping of shallow waters such as coastal zones but typically requires extensive local ground truth data to achieve high accuracy. This study evaluates the effectiveness of transfer learning in reducing this requirement while keeping estimation accuracy at acceptable levels by adapting a deep learning model pretrained on data from Puck Lagoon (Poland) to a new coastal site in Agia Napa (Cyprus). Leveraging the open MagicBathyNet benchmark dataset and a lightweight U-Net architecture, three scenarios were studied and compared: direct inference to Cyprus, site-specific training in Cyprus, and fine-tuning from Poland to Cyprus with incrementally larger subsets of training data. Results demonstrate that fine-tuning with 15 samples reduces RMSE by over 50% relative to the direct inference baseline. In addition, the domain adaptation approach using 15 samples shows comparable performance to the site-specific model trained on all available data in Cyprus. Depth-stratified error analysis and paired statistical tests confirm that around 15 samples represent a practical lower bound for stable SDB, according to the MagicBathyNet benchmark. The findings of this work provide quantitative evidence on the effectiveness of deploying data-efficient SDB pipelines in settings of limited in situ surveys, as well as a practical lower bound for clear and shallow coastal waters.

This study investigates the spatiotemporal evolution of oil released during the Agia Zoni II shipwreck in the Saronic Gulf in 2017, employing the OpenOil module of the OpenDrift framework. The simulation integrates oceanographic and meteorological data to model the transport, weathering, and fate of spilled oil over a six-day period. Oil behavior is examined across key transformation processes, including dispersion, emulsification, evaporation, and biodegradation, using particle-based modeling and a comprehensive set of environmental inputs. The modeled results are validated against in situ observations and visual inspection data, focusing on four critical dates. The study demonstrates OpenOil’s potential for accurately simulating oil dispersion dynamics in semi-enclosed marine environments and highlights the significance of environmental forcing, vertical mixing, and shoreline interactions in determining oil fate. It concludes with recommendations for improving real-time response strategies in similar spill scenarios.

Foundation models for Earth observation are commonly transferred to downstream tasks without explicit attention to the preprocessing mismatch between pretraining and target distributions. This study isolates preprocessing mismatch as a controlled experimental factor in transferring the Hydro multispectral foundation model, a Swin Transformer V2 Base encoder pretrained with SimMIM on Sentinel-2 Level-2A water-body imagery, to the Marine Debris and Oil Spill (MADOS) marine pollution benchmark dataset, processed through ACOLITE Rayleigh reflectance and providing 11 of the 12 spectral bands used during pretraining. The two datasets are therefore produced by different atmospheric correction algorithms under different reflectance conventions, and the resulting per-band statistical discrepancy is quantified as the starting point of the analysis. Three preprocessing dimensions are then systematically varied while all other settings are held constant: input normalisation, spectral band adaptation for the missing B09, and encoder transfer mode. From this, four findings emerge. Normalisation mismatch between training and inference is the single largest source of performance degradation, reducing the mean Intersection over Union (mIoU) by 0.458, more than seven times the largest radiometric perturbation tested. A zero-parameter Frobenius-matched column crop of the patch embedding adapts the 12-band pretrained encoder to the 11-band target, at least as effectively as any learnt linear or nonlinear adapter, at a lower cross-seed variance. Under limited target supervision (1433 training patches against an 87.9 million-parameter encoder), freezing the encoder outperforms both fine-tuning in full and random initialisation training from scratch. The gains of partial unfreezing are attributable to augmented training (very simple copy–paste (VSCP) augmentation, exponential moving average (EMA), and test-time augmentation (TTA)) rather than to encoder adaptation. With matched preprocessing, the frozen encoder reaches 0.600 mIoU and matches the published MariNeXt baseline within seed variability. Mechanistic analysis via band-occlusion attribution and feature-space separability shows that input normalisation determines which spectral bands the encoder relies upon, with the magnitude of the shift correlated to the per-band gap between the source and target distributions. Operationally, preprocessing alignment, rather than architectural modification, carries most of the practical effort in transferring a multispectral foundation model to marine surface segmentation. These results are established for a single encoder–benchmark pair under limited target supervision. The mechanism they identify is more portable than the magnitude reported. A frozen encoder’s representations remain bound to the normalisation statistics of its pretraining dataset, so any transfer that departs from these statistics at inference is predicted to degrade sharply in proportion to the per-band distance between the two distributions.