Robert Frouin

Bio‐optical algorithms serve a critical role, enabling satellite‐based monitoring of the global ocean and its ecosystems. These largely empirical algorithms invert satellite measurements of water leaving reflectance (“ocean color”) to biogeochemical parameters such as chlorophyll a, but often fail to generalize beyond their training data, leading to increased uncertainty when extrapolated. We present a simple geometric approach for assessing bio‐optical model domain using a convex hull. Using NASA Aqua‐MODIS observations, and selected comparisons with PACE‐OCI, we find large swaths of the ocean are not within the domain of existing bio‐optical models, with size and location of gaps varying across seasons, and becoming more pronounced as input dimensionality increases. Understanding these limitations is critical in a changing ocean. Our work provides an interpretable metric that points to specific regions and times that should be flagged in the short‐term and sampled in the long‐term, to better understand the ocean's ecological and biogeochemical diversity.

Accurate characterization of Photosynthetically Available Radiation (PAR) is essential for quantifying marine primary production and carbon cycling. However, current satellite sensors suffer from persistent data gaps due to orbital limitations, sun glint, and large solar zenith angles. Such gaps bias bloom phenology, obscure long-term trends, and degrade biogeochemical models requiring continuous radiative forcing. To address this, we developed a deep convolutional neural network (CNN) to generate gap-free daily PAR fields at 0.5° resolution using MODIS-Aqua and MERRA-2 datasets. Unlike traditional methods that rely solely on spatiotemporal patterns or reanalysis alone, the CNN utilizes co-temporal MODIS-Aqua cloud fractional coverage and optical thickness, seamlessly substituted by MERRA-2 in gap regions, to account for cloud-modulated PAR variability. The model demonstrates high accuracy using MODIS-Aqua inputs (R² = 0.96, bias = -0.19 E/m²/day (-0.5%), RMSE = 3.20 E/m²/day (8.3%)), though performance degrades with MERRA-2 cloud data (R² = 0.84, bias =0.17 E/m²/day (-0.4%), RMSE = 6.47 E/m²/day (16.7%)). Comparisons with independent EPIC estimates (R² = 0.82, RMSE = 6.43 E/m²/day, bias = -0.85 E/m²/day) confirm the physical realism of the reconstructed fields. This gapless product provides a robust forcing field for ocean-biogeochemistry models and data assimilation, enabling more accurate assessments of long-term climate trends. Future work will integrate multi-sensor fusion and higher-resolution cloud inputs.

The influence of light availability and mixed layer depth (MLD) on phytoplankton bloom dynamics was examined across the Argentine Continental Shelf in the Southwest Atlantic Ocean (SWAO). Using satellite-derived chlorophyll-a concentration (Chl-a), photosynthetically available radiation (PAR), and euphotic depth (Z_eu, defined as the depth at which the irradiance is 1 % of its PAR value at surface) data, together with reanalysis products for MLD and wind fields, we analyzed the spatial and temporal variability of key phenological parameters computed from the Chl-a time series, including bloom initiation, peak timing, and bloom intensity, over the 1998–2019 period. Distinct mean spatial distribution patterns in bloom dynamics were observed. In the Central Shelf (CS), blooms typically initiate (May–August) and peak (September–November) relatively early which correlated with shallow MLDs and increasing light, while coastal areas showed even earlier initiation (April) due to highly variable environmental conditions. In turn, the Patagonian Shelf (PS) experienced delayed initiation (September onwards) and peaks (December–January) probably due to deeper MLDs as result of the colder Subantarctic waters. Bloom intensity also exhibited spatial variability, with the highest values observed in the southern PS and regions influenced by frontal systems, where nutrient-rich upwelling and favorable light conditions enhanced phytoplankton growth. Statistical modeling revealed that light penetration (Z_eu) and its interplay with vertical mixing (Z_eu:MLD ratio) were the strongest predictors of bloom anomalies at most sites. However, the predictive power of these relationships varied in regions influenced by local processes, like tidal mixing or frontal zones. Predictive models need to be integrated with regional oceanographic features to improve assessments of bloom phenology and primary production in such highly variable shelf ecosystems.

Bio-optical algorithms serve a critical role, enabling satellite-based monitoring of the global ocean and its ecosystems. These largely empirical algorithms invert satellite measurements of water leaving reflectance (“ocean color”) to biogeochemical parameters such as chlorophyll a, but often fail to generalize beyond their training data, leading to increased uncertainty when extrapolated. We present a simple geometric approach for assessing bio-optical model domain using a convex hull. We find that large swaths of the ocean are not within the domain of existing bio-optical models, particularly in oligotrophic gyres, but also across seasons and locations, becoming more pronounced as input dimensionality increases. Understanding those limitations and biases is critical in a changing ocean. Our work provides an interpretable metric that points to specific regions and times that should be flagged in the short-term and sampled in the long-term, with the hope of better understanding the ocean’s ecological and biogeochemical diversity.

The influence of light availability and mixed layer depth (MLD) on phytoplankton bloom dynamics was examined across the Argentine Continental Shelf in the Southwest Atlantic Ocean (SWAO). Using satellite-derived chlorophyll-a concentration, photosynthetically available radiation, and euphotic depth (Z_eu) data, combined with reanalysis products for MLD and wind fields, the spatial and temporal variability of key phenological parameters was analyzed, including bloom initiation, peak timing, and bloom intensity, over the 1998–2019 period. Distinct geographic trends in bloom dynamics were observed. In the Central Shelf (CS), blooms typically initiate (May–August) and peak (September–November) relatively early which correlated with shallow MLDs and increasing light, while coastal areas showed even earlier initiation (April) due to highly variable environmental conditions. In turn, the Patagonian Shelf (PS) experienced delayed initiation (September onwards) and peaks (December–January) due to deeper MLDs and colder Subantarctic waters. Bloom intensity also exhibited spatial variability, with the highest values observed in the southern PS and regions influenced by frontal systems, where nutrient-rich upwelling and favorable light conditions enhanced phytoplankton growth. Statistical modeling revealed that light penetration (Z_eu) and its interplay with mixing (Z_eu:MLD ratio) were the strongest predictors of bloom anomalies at most sites. However, the predictive power of these relationships varied in regions influenced by local processes, like tidal mixing or frontal zones. Predictive models need to be integrated with regional oceanographic features to improve assessments of bloom phenology and primary production in such highly variable shelf ecosystems.

The influence of light availability and mixed layer depth (MLD) on phytoplankton bloom dynamics was examined across the Argentine Continental Shelf in the Southwest Atlantic Ocean (SWAO). Using satellite-derived chlorophyll-a concentration, photosynthetically available radiation, and euphotic depth (Z_eu) data, combined with reanalysis products for MLD and wind fields, the spatial and temporal variability of key phenological parameters was analyzed, including bloom initiation, peak timing, and bloom intensity, over the 1998–2019 period. Distinct geographic trends in bloom dynamics were observed. In the Central Shelf (CS), blooms typically initiate (May–August) and peak (September–November) relatively early which correlated with shallow MLDs and increasing light, while coastal areas showed even earlier initiation (April) due to highly variable environmental conditions. In turn, the Patagonian Shelf (PS) experienced delayed initiation (September onwards) and peaks (December–January) due to deeper MLDs and colder Subantarctic waters. Bloom intensity also exhibited spatial variability, with the highest values observed in the southern PS and regions influenced by frontal systems, where nutrient-rich upwelling and favorable light conditions enhanced phytoplankton growth. Statistical modeling revealed that light penetration (Z_eu) and its interplay with mixing (Z_eu:MLD ratio) were the strongest predictors of bloom anomalies at most sites. However, the predictive power of these relationships varied in regions influenced by local processes, like tidal mixing or frontal zones. Predictive models need to be integrated with regional oceanographic features to improve assessments of bloom phenology and primary production in such highly variable shelf ecosystems.

The annual periodicity present in spectral remote sensing reflectances (R_rs(λ)) derived from ocean color satellite data processed by NASA’s Ocean Biology Processing Group has been shown to diverge from that observed in situ, with the source of the discrepancies remaining unresolved. Here, we investigate such periodicity in reflectances derived from four satellite sensors and seven in situ locations. Periodicity in the time series of R_rs(λ) spectra is examined without prior assumption of the main constituent frequency through application of the floating-mean Lomb–Scargle periodogram. We show that a one-term sinusoidal function specified at the annual frequency with a constant offset reproduces seasonal trends observed at various regional oceanic sites. Although unique regional features in mean, amplitude, and phase are generally well-captured by R_rs(λ) derived from satellite measurements, discrepancies perpetuate the annual frequency into the satellite-to-in situ R_rs(λ) matchup difference. Specifying the origin of the discrepancy is challenging because the annual frequency is significantly contained in time series of many physical measurements, corrections, and ancillary data required for R_rs(λ) retrievals. Until differences in amplitude and phase between satellite and in situ R_rs(λ) time series are reduced, improved uncertainty estimates should be incorporated into downstream product analyses.

Geostationary-orbit sensors provide high-frequency observations, offering advantages over sun-synchronous sensors for studying short-term ocean dynamics, particularly in regions such as the East Sea/Japan Sea and Yellow Sea, where daily variability is significant. Primary production (PP), crucial for analyzing the carbon cycle, is typically estimated on a global and daily scale from space, but understanding daily changes in PP at a regional scale is also important. Estimating instantaneous PP (iPP) requires knowledge of instantaneous photosynthetically available radiation (iPAR), which is the planar solar flux reaching the surface at wavelengths from 400 to 700 nm, a fundamental controlling variable. Here, using a plane-parallel theory-based PAR model, iPAR from GOCI-I was estimated for the ocean around the Korean Peninsula and evaluated against in-situ iPAR measurements from ECO-PAR sensors deployed at two ocean research stations, Socheong-cho and Ieodo, from 2015 to 2020. Data from 2015 to 2017 were used for the training set, and data from 2018 to 2020 for validation. In-situ measurements from 2015 to 2017 were checked against expected values from radiative transfer simulations, and a 2nd order polynomial regression was applied to correct the measurements. The GOCI-I iPAR estimates showed good agreement with the corrected data, with an RMSE of 10.36% and an MBE of 1.55% from 2015 to 2017. Evaluation against in-situ data from 2018 to 2020 also showed similar RMSE (10.04%) and MBE (0.74%). The accuracy of GOCI-I iPAR was further compared with that of iPAR data from the advanced Himawari imager (AHI) and the moderate resolution imaging spectrometer (MODIS). iPAR values from AHI and MODIS exhibited higher RMSE and MBE than GOCI-I iPAR and lower R². These findings demonstrate that GOCI-I iPAR is a valuable dataset for assessing diurnal variability in oceanic PP around the Korean Peninsula, with implications for improved regional carbon cycle studies and ecosystem monitoring.

The influence of light availability and mixed layer depth (MLD) on phytoplankton bloom dynamics was examined across the Argentine Continental Shelf in the Southwest Atlantic Ocean (SWAO). Using satellite-derived chlorophyll-a concentration, photosynthetically available radiation, and euphotic depth (Z_eu) data, combined with reanalysis products for MLD and wind fields, the spatial and temporal variability of key phenological parameters was analyzed, including bloom initiation, peak timing, and bloom intensity, over the 1998–2019 period. Distinct geographic trends in bloom dynamics were observed. In the Central Shelf (CS), blooms typically initiate (May–August) and peak (September–November) relatively early which correlated with shallow MLDs and increasing light, while coastal areas showed even earlier initiation (April) due to highly variable environmental conditions. In turn, the Patagonian Shelf (PS) experienced delayed initiation (September onwards) and peaks (December–January) due to deeper MLDs and colder Subantarctic waters. Bloom intensity also exhibited spatial variability, with the highest values observed in the southern PS and regions influenced by frontal systems, where nutrient-rich upwelling and favorable light conditions enhanced phytoplankton growth. Statistical modeling revealed that light penetration (Z_eu) and its interplay with mixing (Z_eu:MLD ratio) were the strongest predictors of bloom anomalies at most sites. However, the predictive power of these relationships varied in regions influenced by local processes, like tidal mixing or frontal zones. Predictive models need to be integrated with regional oceanographic features to improve assessments of bloom phenology and primary production in such highly variable shelf ecosystems.

The influence of light availability and mixed layer depth (MLD) on phytoplankton bloom dynamics was examined across the Argentine Continental Shelf in the Southwest Atlantic Ocean (SWAO). Using satellite-derived chlorophyll-a concentration, photosynthetically available radiation, and euphotic depth (Z_eu) data, combined with reanalysis products for MLD and wind fields, the spatial and temporal variability of key phenological parameters was analyzed, including bloom initiation, peak timing, and bloom intensity, over the 1998–2019 period. Distinct geographic trends in bloom dynamics were observed. In the Central Shelf (CS), blooms typically initiate (May–August) and peak (September–November) relatively early which correlated with shallow MLDs and increasing light, while coastal areas showed even earlier initiation (April) due to highly variable environmental conditions. In turn, the Patagonian Shelf (PS) experienced delayed initiation (September onwards) and peaks (December–January) due to deeper MLDs and colder Subantarctic waters. Bloom intensity also exhibited spatial variability, with the highest values observed in the southern PS and regions influenced by frontal systems, where nutrient-rich upwelling and favorable light conditions enhanced phytoplankton growth. Statistical modeling revealed that light penetration (Z_eu) and its interplay with mixing (Z_eu:MLD ratio) were the strongest predictors of bloom anomalies at most sites. However, the predictive power of these relationships varied in regions influenced by local processes, like tidal mixing or frontal zones. Predictive models need to be integrated with regional oceanographic features to improve assessments of bloom phenology and primary production in such highly variable shelf ecosystems.