High-resolution mapping of a mixed siliciclastic-carbonate benthic environment using a hybrid Satellite-Derived Bathymetry (SDB) approach integrated with sedimentological data: North Safaga Bay, Egyptian Red Sea

Modern mixed siliciclastic-carbonate benthic environments in the Egyptian northern Red Sea are shaped by continuous carbonate production and episodic siliciclastic delivery, yet their spatial patterns remain difficult to resolve where high-resolution bathymetry is lacking. To address these challenges and improve habitat classification, this study used a Satellite-Derived Bathymetry (SDB) model and explored its implications for understanding sedimentary processes in North Safaga Bay.

We compiled legacy ground truth datasets from previous studies and supplemented them with new in situ sampling, granulometric, and petrographic analyses. While SDB is limited by turbidity and substrate reflectance ambiguity, we mitigated these uncertainties using a physics-based workflow to isolate ICESat-2 (Ice, Cloud, and Land Elevation Satellite-2) signal photons to extract depth points and a Random Forest regressor (RFR) trained on Sentinel-2 imagery. The RFR model achieved high predictive accuracy (coefficient of determination; R^2 = 0.95) at a 10 m spatial resolution, with internal cross-validation and an 80:20 training-to-test ratio. The reliability of this model was further validated against a fully independent set of in situ samples (R^2 = 0.85).

Our results reveal that high-energy bathymetric highs are dominated by compound-grain and coralgal facies, while the more sheltered back reef lagoons favor widespread foraminiferal sands, including soritid-bearing assemblages. In contrast, bathymetric depressions at the outlets of major wadi systems act as sediment traps for terrigenous siliciclastic facies. The delivery of these clastics is driven by a spectrum of episodic gravity-driven flows, including hyperpycnal events during flash floods, aeolian input, and wave-induced reworking.

The integration of remote sensing and sedimentological analyses provides a significant advancement over existing datasets by allowing quantifiable spatial correlation between geomorphology and facies distribution across the bay, rather than relying on interpolated point data. By resolving these fine-scale bathymetric controls, our results revealed a mosaic setting in which seafloor bathymetry and energy, directly dictated by the underlying inherited rift topography, act as the primary control on sediment partitioning, providing a localized model for mixed deposits in active settings.