The atmospheric contribution constitutes about 90 percent of the signal measured by satellite sensors over oceanic and inland waters. Over open ocean waters, the atmospheric contribution is relatively easy to correct as it can be assumed that water-leaving radiance in the near-infrared (NIR) is equal to zero and it can be performed by applying a relatively simple dark-pixel-correction-based type of algorithm. Over inland and coastal waters, this assumption cannot be made since the water-leaving radiance in the NIR is greater than zero due to the presence of water components like sediments and dissolved organic particles. The aim of this study is to determine the most appropriate atmospheric correction processor to be applied on Sentinel-2 MultiSpectral Imagery over several types of inland waters. Retrievals obtained from different atmospheric correction processors (i.e., Atmospheric correction for OLI ‘lite’ (ACOLITE), Case 2 Regional Coast Colour (here called C2RCC), Case 2 Regional Coast Colour for Complex waters (here called C2RCCCX), Image correction for atmospheric effects (iCOR), Polynomial-based algorithm applied to MERIS (Polymer) and Sen2Cor or Sentinel 2 Correction) are compared against in situ reflectance measured in lakes and reservoirs in the Valencia region (Spain). Polymer and C2RCC are the processors that give back the best statistics, with coefficients of determination higher than 0.83 and mean average errors less than 0.01. An evaluation of the performance based on water types and single bands–classification based on ranges of in situ chlorophyll-a concentration and Secchi disk depth values- showed that performance of these set of processors is better for relatively complex waters. ACOLITE, iCOR and Sen2Cor had a better performance when applied to meso- and hyper-eutrophic waters, compare with oligotrophic. However, other considerations should also be taken into account, like the elevation of the lakes above sea level, their distance from the sea and their morphology.

The colored dissolved organic matter (CDOM) variable is the standard measure of humic substance in waters optics. CDOM is optically characterized by its spectral absorption coefficient, $a_{CDOM}$ at at reference wavelength (e.g., ≈ 440 nm). Retrieval of CDOM is traditionally done using bio-optical models. As an alternative, this paper presents a comparison of five machine learning methods applied to Sentinel-2 and Sentinel-3 simulated reflectance ( $R_{rs}$ ) data for the retrieval of CDOM: regularized linear regression (RLR), random forest regression (RFR), kernel ridge regression (KRR), Gaussian process regression (GPR) and support vector machines (SVR). Two different datasets of radiative transfer simulations are used for the development and training of the machine learning regression approaches. Statistics comparison with well-established polynomial regression algorithms shows optimistic results for all models and band combinations, highlighting the good performance of the methods, especially the GPR approach, when all bands are used as input. Application to an atmospheric corrected OLCI image using the reflectance derived form the alternative neural network (Case 2 Regional) is also shown. Python scripts and notebooks are provided to interested users.

Shallow and deep lakes receive and recycle organic and inorganic substances from within the confines of these lakes, their watershed and beyond. Hence, a large range in absorption and scattering and extreme differences in optical variability can be found between and within global lakes. This poses a challenge for atmospheric correction and bio-optical algorithms applied to optical remote sensing for water quality monitoring applications. To optimize these applications for the wide variety of lake optical conditions, we adapted a spectral classification scheme based on the concept of optical water types. The optical water types were defined through a cluster analysis of in situ hyperspectral remote sensing reflectance spectra collected by partners and advisors of the European Union 7th Framework Programme (FP7) Global Lakes Sentinel Services (GLaSS) project. The method has been integrated in the Envisat-BEAM software and the Sentinel Application Platform (SNAP) and generates maps of water types from image data. Two variations of water type classification are provided: one based on area-normalized spectral reflectance focusing on spectral shape (6CN, six-class normalized) and one that retains magnitude with no modification to the reflectance signal (6C). This resulted in a protocol, or processing scheme, that can also be applied or adapted for Sentinel-3 Ocean and Land Colour Imager (OLCI) datasets. We apply both treatments to MERIS imagery of a variety of European lakes to demonstrate its applicability. The studied target lakes cover a range of biophysical types, from shallow turbid to deep and clear, as well as eutrophic and dark absorbing waters, rich in colored dissolved organic matter (CDOM). In shallow, high-reflecting Dutch and Estonian lakes with high sediment load, 6C performed better, while in deep, low-reflecting clear Italian and Swedish lakes, 6CN performed better. The 6CN classification of in situ data is promising for very dark, high CDOM, absorbing lakes, but we show that our atmospheric correction of the imagery was insufficient to corroborate this. We anticipate that the application of the protocol to other lakes with unknown in-water characterization, but with comparable biophysical properties will suggest similar atmospheric correction (AC) and in-water retrieval algorithms for global lakes.

Parameter retrieval and model inversion are key problems in remote sensing and Earth observation. Currently, different approximations exist: a direct, yet costly, inversion of radiative transfer models (RTMs); the statistical inversion with in situ data that often results in problems with extrapolation outside the study area; and the most widely adopted hybrid modeling by which statistical models, mostly nonlinear and non-parametric machine learning algorithms, are applied to invert RTM simulations. We will focus on the latter. Among the different existing algorithms, in the last decade kernel based methods, and Gaussian Processes (GPs) in particular, have provided useful and informative solutions to such RTM inversion problems. This is in large part due to the confidence intervals they provide, and their predictive accuracy. However, RTMs are very complex, highly nonlinear, and typically hierarchical models, so that very often a single (shallow) GP model cannot capture complex feature relations for inversion. This motivates the use of deeper hierarchical architectures, while still preserving the desirable properties of GPs. This paper introduces the use of deep Gaussian Processes (DGPs) for bio-geo-physical model inversion. Unlike shallow GP models, DGPs account for complicated (modular, hierarchical) processes, provide an efficient solution that scales well to big datasets, and improve prediction accuracy over their single layer counterpart. In the experimental section, we provide empirical evidence of performance for the estimation of surface temperature and dew point temperature from infrared sounding data, as well as for the prediction of chlorophyll content, inorganic suspended matter, and coloured dissolved matter from multispectral data acquired by the Sentinel-3 OLCI sensor. The presented methodology allows for more expressive forms of GPs in big remote sensing model inversion problems.

In the ever-evolving landscape of marine, oceanic, and climate change monitoring, the intersection of cutting-edge artificial intelligence (AI), machine learning (ML), and data analytics has emerged as a pivotal catalyst for transformative advancements

Land Surface Temperature (LST) is one of the key parameters in the physics of land-surface processes on regional and global scales, combining the results of all surface-atmosphere interactions and energy fluxes between the surface and the atmosphere. With the advent of the European Space Agency (ESA) Sentinel 3 (S3) satellite, accurate LST retrieval methodologies are being developed by exploiting the synergy between the Ocean and Land Colour Instrument (OLCI) and the Sea and Land Surface Temperature Radiometer (SLSTR). In this paper we explain the implementation in the Basic ENVISAT Toolbox for (A)ATSR and MERIS (BEAM) and the use of one LST algorithm developed in the framework of the Synergistic Use of The Sentinel Missions For Estimating And Monitoring Land Surface Temperature (SEN4LST) project. The LST algorithm is based on the split-window technique with an explicit dependence on the surface emissivity. Performance of the methodology is assessed by using MEdium Resolution Imaging Spectrometer/Advanced Along-Track Scanning Radiometer (MERIS/AATSR) pairs, instruments with similar characteristics than OLCI/ SLSTR, respectively. The LST retrievals were properly validated against in situ data measured along one year (2011) in three test sites, and inter-compared to the standard AATSR level-2 product with satisfactory results. The algorithm is implemented in BEAM using as a basis the MERIS/AATSR Synergy Toolbox. Specific details about the processor validation can be found in the validation report of the SEN4LST project.