The Santos Basin (SB) is the main petroliferous basin in the Brazilian continental margin and one of the most studied marine areas in Brazil. However, historical data suggest that new efforts should be carried out to acquire quantitative biological data, especially in the deep sea, to establish the baseline of essential ocean variables in different ecosystems for future monitoring programs. The Brazilian energy company Petrobras planned and executed 24 oceanographic cruises over a period of 2 years to assess the benthic (SANSED cruise) and pelagic (SANAGU cruise) systems of the SB (356 days at sea in 2019 and 2021/2022). These efforts were part of the Santos Project, which comprised a comprehensive environmental study aimed at investigating benthic and pelagic variables to characterize ecology, biogeochemistry, thermohaline properties of water masses, and ocean circulation patterns, geomorphology, and sedimentology, as well as organic and inorganic chemistry. Here we present the detailed sampling designs and the field methods employed on board, during the SB scientific cruises. All sampling protocols were based on standardized approaches. For the benthos analyses, triplicate sediment samples were performed using a GOMEX-type box corer (0.25 m²) or a large modified Van Veen grab (0.75 m²) at 100 stations ranging from 25 to 2400 m depth. At each station, 25 geochemical and physico-chemical parameters were analyzed in addition to micro-, meio-, and macrofauna and living foraminifera samples. For the pelagic system, 60 stations were selected to investigate the plankton community, ranging in size from pico- to macroplankton, through vertical, horizontal, and oblique net hauls (20, 200, and 500 μm mesh size), as well as 25 biogeochemical parameters collected with an aid of a CTD-rosette sampler. Part of this scientific information also serves the Regional Environmental Characterization Project (PCR-BS) in support of Petrobras’ Santos Basin drilling licensing process led by the Brazilian Environmental Agency – IBAMA. This project contributes to the sustainable development of the SB, in line with the guidelines of the United Nations Decade of Ocean Science for Sustainable Development.

We report Trichodesmium slicks in the South Atlantic associated with fronts produced by the southward displacement of the Brazil Current. For the first time, a relatively high spatial sampling grid of trichomes density is shown over an undersampled region with regard to Trichodesmium. In this study, the distributions of Trichodesmium along the southeastern Brazilian continental shelf break were characterized during four oceanographic cruises and related to seawater temperature, upper layer nutrient concentrations and wind dynamics. Meridional changes in the densities of trichomes followed an expected trend with higher (lower) values found in warmer (colder) and more saline (less saline) water masses. Between 23° and 28°S, large accumulations of trichomes were associated with high temperature, increased phosphate concentrations (>0.08 μM), nitrogen below 2.4 μM, and low wind speeds (<11 m s⁻¹). Maximum densities of 212.6 × 10⁵ trichomes L⁻¹, which are considerably larger than those reported in other studies in subtropical oceans, occurred at stations showing weak water column stability and higher phosphate levels over the upper 25 m. We suggest that shelf break dynamics, which drive and modulate the intensity of the shelf break upwelling, as well as wind divergence at spatial scales of approximately 600 km², associated with mesoscale activities, may favor Trichodesmium growth and accumulation. On the other hand, Trichodesmium densities were low at latitudes between 28° and 33°S, likely due to either or both the lowest sea surface temperatures and/or out‐competition by the faster‐growing nondiazotrophs.

Chlorophyll-a concentration (Chl-a) satellite retrieval is an important proxy to assess the ocean trophic state and marine ecosystem responses to climate variability. However, uncertainties in water-leaving remote sensing reflectance spectra (Rrs()) retrieved from Ocean Color (OC) satellite sensors may affect Chl-a products depending on sensor performance, processing scheme, and environmental conditions. Chl-a algorithms also have their own uncertainty, particularly related to optical water type (OWT), phytoplankton composition, and physiology. Here, we compared simultaneous in-situ Rrs() and Chl-a measurements in the South Brazil Bight (SBB) with collocated satellite products from MODIS-Aqua, VIIRS-JPSS-1/SNPP, and OLCI–S3A/B, using empirical (OCx) and semi-analytical (GIOP) algorithms. All sensors showed similar Rrs() performance, with an overall negative bias, particularly for shorter wavelengths. These differences were minimized by maximum band ratios, especially for MODIS-Aqua (RPD = 16%; R2 = 0.90). For Chl-a, MODIS-Aqua showed the highest correlation using OCx (R2 = 0.95), while OLCI-OC4ME presented the lowest mean bias (0.13 mg·m-3) and RMSE (0.6 mg·m-3). In inner shelf waters, empirical algorithms mostly overestimated Chl-a, whereas VIIRS-GIOP provided the best performance (R2 = 0.78, Bias = 0.09 mg·m-3, RMSE = 0.2 mg·m-3). Overall, the three sensors provided reliable Rrs and Chl-a data across inner shelf and open-ocean waters of the SBB. Empirical OCx algorithms using MODIS-Aqua or VIIRS are suitable for spatiotemporal analyses of Chl-a patterns, whereas VIIRS-GIOP, followed by OLCI-OCx, is preferable for quantitative analyses, particularly in optically complex inner shelf waters. These findings support the continuity of long-term OC observations beyond MODIS-Aqua and their application to monitoring bio-optical and biogeochemical variability in the SBB.

Harmful algal blooms (HABs) severely affect aquatic ecosystems and coastal economies but remain difficult to monitor due to their diversity and complex environmental forcing. Satellite remote sensing has advanced HAB detection, as some algal taxa exhibit specific spectral signatures. However, most studies have relied on low-to-medium spatial resolution imagery and focused on few reported bloom events. This study aimed to detect and discriminate blooms of key toxic phytoplankton taxa in coastal and oceanic domains, using Sentinel-2 (S2) and Sentinel-3 (S3) data. A comprehensive, global in situ database of blooms of four potentially toxigenic taxa (Alexandrium spp., Dinophysis spp., Gymnodinium catenatum and Pseudo-nitzschia spp.) was compiled and matched with concurrent S2 and S3 images. Bio-optical parameters, including chlorophyll-a (Chl-a) and normalized difference indices, were tested for their ability to detect blooms, and an unsupervised classification approach was applied to potentially discern taxa-specific optical signatures. Normalized difference indices outperformed Chl-a in HAB detection and mapping, reinforcing that Chl-a alone is insufficient as a reliable bloom indicator. The inclusion of other optical variables (inherent optical properties) enhanced cluster performance and group differentiation for both satellite missions. S2 only detected differences in reflectance magnitude, whereas S3 discerned pigment-related features that may be indicative of two of the target taxa. Although future hyperspectral missions will enhance our understanding of HAB bio-optical diversity, this study highlights the importance of maintaining a consistent acquisition of multispectral data to support HAB operational forecasting systems and management strategies, tailored to specific target taxa and spatial domains.