Coastal and inland waters ecosystems are ecologically, culturally, and economically important. Monitoring these environments is therefore essential to understand ecosystem functioning, how to ensure sustainable practices and assess the impact of human activities. Among the large diversity of measurement techniques, optical remote sensing presents some clear advantages. Indeed, earth observation satellites nowadays allow to monitor the spatial variability of water quality parameters over large areas and with relatively short revisiting times. In water and above water radiometers, have a great potential for ecosystem monitoring, especially if they are integrated into autonomous measurement systems providing high temporal resolution data, or if they have a high spectral resolution opening the door to new environmental products based on fine spectral features. However, retrieving relevant information on water constituents from radiometric data in optically complex waters is still challenging. Indeed, although in clear, case-1, waters most of the bio-optic parameters are dependent of the chlorophyll-a concentration, in coastal and inland waters (i.e. case-2 waters) light absorption and scattering is affected by terrestrial inputs of sediments and/or dissolved organic carbon which can make the retrieval of simple parameters such as the chlorophyll-a concentration very complicated. In addition, atmospheric correction algorithms are more challenging because of potentially extreme optical water properties and the proximity with the coast or surrounding land. This session is open to all contributions presenting novel applications of inland and coastal aquatic monitoring based on visible and NIR radiometric remote sensing data either from satellite or in situ sensors.
Lead Organizer: Héloïse Lavigne, Royal Belgium Intistute of Natural Sciences (hlavigne@naturalsciences.be)
Co-organizers:
Clémence Goyens, Royal Belgium Institute of Natural Sciences (cgoyens@naturalsciences.be)
Pierre Gernez, Nantes Université (pierre.gernez@univ-nantes.fr)
David Doxaran, Laboratoire D'Oceanographie de Villfrenche, CNRS (david.doxaran@imev-mer.fr)
Evangelos Spyrakos, University of Stirling (evangelos.spyrakos@stir.ac.uk)
Presentations
08:30 AM
AUTOMATED HYPERSPECTRAL RADIOMETRY IN THE PERIALPINE REGION FOR VALIDATION OF SATELLITE WATER QUALITY PRODUCTS (6511)
Primary Presenter: Abolfazl Irani Rahaghi, Eawag, Swiss Federal Institute of Aquatic Science & Technology (abolfazl.irani@eawag.ch)
Satellite remote sensing is a powerful tool for lake water quality monitoring. Validation of remote sensing products requires ground reference measurements, which are time-consuming and costly to obtain. With the recent advances in sensor technology, automated hyperspectral radiometric systems are increasingly used to acquire frequent near-real-time data. Such datasets do not only drastically increase the number of matchups, but also ensure a more balanced seasonal sampling compared to traditional field campaigns. In this study, we present optical properties (including remote sensing reflectance) and biogeochemical data from three research platforms in Lake Geneva (Switzerland/France), Lake Garda (Italy), and Greifensee (Switzerland). Remote sensing reflectance data were obtained through automated hyperspectral radiometers below (Wetlabs Thetis), at (JB Hyperspectral RoX), or above the water surface (WaterInsight WISPstation, Hypstar, and PML So-Rad). The cross-validated measurements were used to evaluate the performance of selected atmospheric correction algorithms, namely POLYMER, C2RCC, and Acolite applied to Sentinel-2/-3 products. We also performed an optical closure exercise to further assess the uncertainty of optical measurements in Lake Geneva, where simultaneous inherent optical property measurements were available. Our results suggest that automated spectroradiometers are becoming a reliable data source for validating satellite products and associated uncertainties at regional and global scales, with the added value of simultaneous biogeochemical measurements.
08:45 AM
HYPERNETS MEASUREMENTS IN CONTRASTED FRENCH COASTAL WATERS: CALIBRATION/VALIDATION OF MULTI-SENSOR SATELLITE PRODUCTS (5909)
Primary Presenter: David Doxaran, CNRS / Sorbonne University (david.doxaran@imev-mer.fr)
Since 2021, two autonomous HYPERNETS stations are operated in contrasted French coastal waters: one in the center of the optically complex Berre coastal lagoon by the Mediterranean Sea and one at the mouth of the highly turbid Gironde Estuary along the Atlantic coast. Optimized quality controls have been designed and applied to the hyperspectral radiometric measurements, namely the downwelling irradiance, sky and water radiance signals, recorded following a strict viewing geometry every 15 mn during daytime on the two stations. The resulting water reflectance spectra are proved to be of high quality (e.g., detailed saturations effects at short visible wavelengths, correlation with water turbidity). They allow assessing the performance of different atmospheric (and glint) correction algorithms applied to high (Sentinel2-MSI and Landsat8/9-OLI) and medium (Sentinel3-OLCI and MODIS) spatial resolution satellite data, based on numerous match-ups. A new method is proposed for the inversion of the measured hyperspectral water reflectance signal into concentration of suspended particulate matter (SPM) but also inherent optical properties and associated spectral slopes which contain useful information on SPM size distribution and composition.
09:00 AM
Aquaverse: An Aquatic Inversion Scheme for Remote Sensing of Fresh and Coastal Waters (6025)
Primary Presenter: Nima Pahlevan, SSAI / NASA GSFC (nima.pahlevan@nasa.gov)
A primary challenge in aquatic remote sensing is developing a robust atmospheric correction (AC) method to estimate remote sensing reflectance (Rrs) products. Here, we describe the utility of Mixture Density Networks (MDNs) in an AC framework, termed Aquaverse, for processing Landsat-8/-9 (L8/9) and Sentinel-2A/B (S2A/B) images over inland and nearshore coastal waters. First, a coupled ocean-atmosphere radiative transfer (RT) model was adapted to simulate top-of-atmosphere reflectance spectra (rhot) for various imaging geometries and environmental conditions. Second, the hyperspectral pairs of rhot- Rrs spectra (@ 5 nm) were resampled with sensor spectral response functions and used subsequently to train MDNs to transform rhot to Rrs with the associated uncertainties. The trained model was applied to L8/9 and S2A/B imagery for an extensive matchup assessment obtained from AERONET-OC products (N ~ 800) and other field-measured datasets (N ~ 700). Compared to the state-of-the-art AC processors, our matchup analyses showed performance improvements from 20 to 100% depending on the spectral band and statistical metrics. Median estimated errors for L8/9 and S2A/B matchups combined were ~ 45%, 35%, 20%, and 30% for the 443, 482, 560, and 655 nm spectral bands, respectively. Aquaverse-generated maps for several images exhibited similar spatial patterns as those produced by other processors, and uncertainty maps conformed to the model’s expected performance. We demonstrate that Aquaverse can be utilized as one alternative data processing technique in fresh and coastal waters.
09:15 AM
SHORTWAVE INFRARED RADIOMETRY OF AQUATIC ENVIRONMENTS (6193)
Primary Presenter: Henry Houskeeper, Woods Hole Oceanographic Institution (henry.houskeeper@whoi.edu)
Optical remote sensing has improved monitoring of aquatic systems by enabling remote or automated in situ characterization of biogeochemical parameters. Algorithms have primarily been based on empirical relationships---generally corresponding to oceanic (i.e., case-1) waters or else tuned to specific regions---derived using visible (VIS) wavelengths. Recent technological and algorithmic advances have revealed that aquatic remote sensing algorithms that use nonvisible wavelengths, including ultraviolet (UV) and near infrared (NIR), confer greater robustness to optical complexity, a concept termed End-Member Analysis (EMA). Longer wavelengths outside the NIR, i.e., shortwave infrared (SWIR), have not yet been evaluated for EMA due to high attenuation by water, and legacy observations of SWIR water-leaving signals often do not exceed noise. This presentation leverages recent technological advances to evaluate high signal-to-noise ratio SWIR water-leaving radiance observations that are compliant with an absolute radiometry perspective, i.e., individual waveband observations are absolute and do not require spectral normalization. Langley calibration results are shown for evaluation of radiometric performance, and SWIR water-leaving radiances from aquatic environments including the Southern Ocean, plus coastal and inland waters of California and Nevada (USA), are compared. Finally, the applicability of a SWIR black-pixel assumption for atmospheric correction is considered, along with potential future applications of SWIR radiometry for aquatic monitoring.
09:30 AM
SeaHawk Low-Cost Ocean Color CubeSat Produces High Spatial Resolution and High-Quality Data: A Comparison with NOAA-20 VIIRS, NASA MODIS-Terra and MODIS-Aqua (7198)
Primary Presenter: Md Masud-Ul-Alam, University of georgia (masudocndu@uga.edu)
SeaHawk, with its multispectral HawkEye sensor, is the first dedicated ocean color (OC) non-commercial CubeSat satellite mission in orbit. This mission was designed to demonstrate that a low-cost CubeSat could retrieve high quality, high spatial resolution data from around the world. Here we present the first in-depth assessment of SeaHawk’s performance with respect to 3 other operational OC missions: MODIS-Terra, MODIS-Aqua and VIIRS. We selected 11 locations that represent a variety of optical water types and compared data on days when there were data for all four sensors. For each image of a location, we extracted data from thousands of points selected at random. We selected bands present in all sensors (412nm, 447nm, 488nm, 555nm, 670nm) and compared the matchups at 3 different levels of processing (top of the atmosphere (TOA) reflectance, remote sensing reflectance (Rrs), and chlorophyll-a concentration). We then compared five statistical approaches: root mean square error (RMSE), index of agreement (d), Concordance Correlation Coefficient (CCC), linear regression (R2), and bias. Overall, TOA results show good agreement between HawkEye and all other sensors (irrespective of bands), particularly Terra. Rrs results show band-specific patterns instead of sensor-specific patterns; based on R2, CCC, and d, 555nm and 412nm seemed to be the best and worst performing bands. Not all locations show good matchups; discrepancies could be due to a combination of spatial, geophysical, temporal, and technical aspects, e.g. hydrodynamics, time difference, or solar/sensor zenith angle.
09:45 AM
USING HIGH RESOLUTION OCEAN COLOUR IMAGES TO DETECT BIOLOGICALLY ACTIVE FINE-SCALE FRONTS: A CASE STUDY FROM THE HAURAKI GULF, AOTEAROA NEW ZEALAND (4928)
Primary Presenter: Alexandre Lhériau-Nice, The University of Auckland (alhe551@aucklanduni.ac.nz)
Fronts, meeting points with water masses having different properties, are common occurrences in the ocean and cover a wide range of both temporal and spatial scales. Coastal areas such as the Hauraki Gulf (Aotearoa New Zealand) are characterised strong frontal heterogeneity that is challenging to monitor. Here, we combine fine resolution satellite imaging with a front detection algorithm optimised for the region to investigate the spatio-temporal variability of ocean color fronts. We adapted the Belkin and O’Reilly Algorithm (BOA) to remove the noise typically associated with Ocean and Land Colour Instrument (OLCI) images. By compiling high resolution detection over the 2016-2022 period, we show the distribution of frontal features varied across seasons, wind direction, and ENSO phase. Locations of fronts oscillated seasonally from the inner to the outer gulf, going far from land between winter and spring, reversing between spring and autumn. Similarly, El Niño periods correspond to fronts closer to the coast, while La Niña periods correspond to locations further offshore. These patterns relate to the predominant wind direction in the region of interest: El Niño is dominated by westerlies while La Niña is linked with prevalent easterlies. This flexible approach supports the study of the spatio-temporal variability of fronts and provides a near-real time tool to monitor change in coastal regions.
SS102A Inland and Coastal Aquatic Ecosystems Monitoring from In Situ and Satellite Radiometric Measurements
Description
Time: 8:30 AM
Date: 9/6/2023
Room: Sala Menorca A