Phytoplankton taxonomy, pigment composition and photo-physiological state were studied in Galveston Bay (GB), Texas (USA), following the extreme flooding associated with Hurricane Harvey (25–29 August 2017) using field and satellite ocean color observations. The percentage of chlorophyll a (Chl a) in different phytoplankton groups was determined from a semi-analytical IOP (inherent optical property) inversion algorithm. The IOP inversion algorithm revealed the dominance of freshwater species (diatom, cyanobacteria and green algae) in the bay following the hurricane passage (29 September 2017) under low salinity conditions associated with the discharge of floodwaters into GB. Two months after the hurricane (29–30 October 2017), under more seasonal salinity conditions, the phytoplankton community transitioned to an increase in small-sized groups such as haptophytes and prochlorophytes. Sentinel-3A Ocean and Land Colour Instrument (OLCI)-derived Chl a obtained using a red ∕ NIR (near-infrared) band ratio algorithm for the turbid estuarine waters was highly correlated (R²>0.90) to the (high-performance liquid chromatography) HPLC-derived Chl a. Long-term observations of OLCI-derived Chl a (August 2016–December 2017) in GB revealed that hurricane-induced Chl a declined to background mean state in late October 2017. A non-negative least squares (NNLS) inversion model was then applied to OLCI-derived Chl a maps of GB to investigate spatiotemporal variations of phytoplankton diagnostic pigments pre- and post-hurricane; results appeared consistent with extracted phytoplankton taxonomic composition derived from the IOP inversion algorithm and microplankton pictures obtained from an Imaging FlowCytobot (IFCB). OLCI-derived diagnostic pigment distributions also exhibited good agreement with HPLC measurements during both surveys, with R² ranging from 0.40 for diatoxanthin to 0.96 for Chl a. Environmental factors (e.g., floodwaters) combined with phytoplankton taxonomy also strongly modulated phytoplankton physiology in the bay as indicated by measurements of photosynthetic parameters with a fluorescence induction and relaxation (FIRe) system. Phytoplankton in well-mixed waters (mid-bay area) exhibited maximum PSII photochemical efficiency (F_v∕F_m) and a low effective absorption cross section (σ_PSII), while the areas adjacent to the shelf (likely nutrient-limited) showed low F_v∕F_m and elevated σ_PSII values. Overall, the approach using field and ocean color data combined with inversion models allowed, for the first time, an assessment of phytoplankton response to a large hurricane-related floodwater perturbation in a turbid estuarine environment based on its taxonomy, pigment composition and physiological state.

Dissolved and particulate organic carbon, suspended particulate matter concentrations, and their optical proxies colored dissolved organic matter absorption and backscattering coefficients were studied in Galveston Bay, Texas, following the extreme flooding of Houston and surrounding areas due to Hurricane Harvey (25–29 August 2017) using field and ocean color observations. A three‐step empirical‐semianalytic algorithm for determination of colored dissolved organic matter absorption and backscattering coefficients revealed the dynamics of dissolved organic carbon and particle distribution from Visible and Infrared Imaging Radiometric Suite ocean color. Environmental drivers, especially floodwater discharge and winds, strongly influenced the spatiotemporal distribution of dissolved/particulate material in the bay and shelf waters following the hurricane passage. Over 10 days during/following the hurricane, ~25.2 × 10⁶ kg C of total organic carbon and ~314.7 × 10⁶ kg of suspended particulate matter were rapidly exported from Galveston Bay (representing ~0.65% and 0.27% of respective annual Mississippi River fluxes to the Gulf of Mexico), with potential for ecological impacts to shelf waters.

HyperCoast is a Python package designed to provide an accessible and comprehensive set of tools for visualizing and analyzing hyperspectral data in coastal environments. Hyperspectral data refers to the information collected by sensors that capture light across a wide range of wavelengths, beyond what the human eye can see. This data allows scientists to detect and analyze various materials and conditions on the Earth's surface with great detail. Unlike multispectral data, which captures light in a limited number of broad wavelength bands (typically 3 to 10), hyperspectral data captures light in many narrow, contiguous wavelength bands, often numbering in the hundreds. This provides much more detailed spectral information. Leveraging the capabilities of popular packages like Leafmap (Wu, 2021) and PyVista (Sullivan & Kaszynski, 2019), HyperCoast streamlines the exploration and interpretation of complex hyperspectral remote sensing data from existing spaceborne and airborne missions. It is also poised to support future hyperspectral missions, such as NASA's SBG and GLIMR (Dierssen et al., 2021).

As a key component of inherent optical properties (IOPs) in ocean color remote sensing, phytoplankton absorption coefficient ($a_{phy}$), especially in hyperspectral, greatly enhances our understanding of phytoplankton community composition (PCC). The recent launches of NASA’s hyperspectral missions, such as EMIT and PACE, have generated an urgent need for hyperspectral algorithms for studying phytoplankton. Retrieving $a_{phy}$ from ocean color remote sensing in coastal waters has been extremely challenging due to complex optical properties. Traditional methods often fail under these circumstances, while improved machine-learning approaches are hindered by data scarcity, heterogeneity, and noise from data collection. In response, this study introduces a novel machine learning framework for hyperspectral retrievals of $a_{phy}$ based on the mixture-of-experts (MOEs), named PhA-MOE. Various preprocessing methods for hyperspectral training data are explored, with the combination of robust and logarithmic scalers identified as optimal. The proposed PhA-MOE for $a_{phy}$ prediction is tailored to both past and current hyperspectral missions, including EMIT and PACE. Extensive experiments reveal the importance of data preprocessing and improved performance of PhA-MOE in estimating $a_{phy}$ as well as in handling data heterogeneity. Notably, this study marks the first application of a machine learning–based MOE model to real PACE-OCI hyperspectral imagery, validated using match-up field data. This application enables the exploration of spatiotemporal variations in $a_{phy}$ within an optically complex estuarine environment.

Phytoplankton taxonomy, pigment composition and photo-physiological state were studied in Galveston Bay (GB), Texas (USA) following the extreme flooding associated with Hurricane Harvey (August 25–29, 2017) using field and satellite ocean color observations. Percentage of chlorophyll a (Chl a) in different phytoplankton groups were determined from a semi-analytical IOP (inherent optical property) inversion algorithm. The IOP inversion algorithm revealed the dominance of freshwater species (cyanobacteria and green algae) in the bay following the hurricane passage (September 29, 2017) under low salinity conditions associated with the discharge of floodwaters into GB; 2 months after the hurricane (October 29–30, 2017), under more seasonal salinity conditions, the phytoplankton community transitioned to an increase in small sized groups such as haptophyte and prochlorophyte. Sentinel-3A OLCI-derived Chl a obtained using a red/NIR band ratio algorithm for the turbid estuarine waters was highly correlated (R² > 0.90) to HPLC-derived Chl a concentrations. A Non-Negative Least Square (NNLS) inversion model was then applied to OLCI-derived Chl a maps of GB to obtain spatiotemporal distributions of phytoplankton diagnostic pigments; results appeared consistent with extracted phytoplankton taxonomic composition derived from the IOP inversion algorithm. OLCI-derived diagnostic pigment distributions also exhibited good agreement with HPLC measurements, with mean R² ranging from 0.39 for violaxanthin to 0.98 for Chl a. Environmental factors (e.g. floodwaters) combined with phytoplankton taxonomy also strongly modulated phytoplankton physiology in the bay as indicated by measurements of photosynthetic parameters with a Fluorescence Induction and Relaxation (FIRe) system. Phytoplankton in well-mixed waters (mid-bay area) exhibited maximum PSII photochemical efficiency (F_V/F_M) and low effective absorption cross section (δ_PSII), while the areas adjacent to the shelf (likely nutrient-limited) showed low F_V/F_M and elevated values. Overall, the approach using field and ocean color data combined with inversion models allowed, for the first time, an assessment of phytoplankton response to a large hurricane-related floodwater perturbation in a turbid estuarine environment based on its taxonomy, pigment composition and physiological state.