This work represents a contribution to the CSIC Thematic Interdisciplinary Platform PTI-POLARCSIC and PTI-TELEDETECT.-- Special Issue Remote Sensing Techniques for Ocean Dynamics: State of the Art, Present and Future Applications.-- 29 pages, 13 figures, 2 tables, This study assesses the capability of Surface Quasi-Geostrophy (SQG) to reconstruct the three- dimensional (3D) dynamics in four critical areas of the Arctic Ocean: the Nordic, Barents, East Siberian, and Beaufort Seas. We first reconstruct the upper ocean dynamics from TOPAZ4 re- analysis of sea surface height (SSH), surface buoyancy (SSB), and surface velocities (SSV) and validate the results with the geostrophic and total TOPAZ4 velocities. The reconstruction of upper ocean dynamics using SSH fields is in high agreement with the geostrophic velocities, with correlation coefficients greater than 0.8 for the upper 400 m. SSH reconstructions outperform surface buoyancy reconstructions, even in places near freshwater inputs from river discharges, melting sea ice, and glaciers. Surface buoyancy fails due to the uncorrelation of SSB and subsurface potential vorticity (PV). Reconstruction from surface currents correlates to the total TOPAZ4 velocities with correlation coefficients greater than 0.6 up to 200 meters. Due to a lower stratification of the water column, the performance of the SQG approach is better in fall and winter than in the spring and summer. In the second part, we apply the SQG approach validated withthe reanalysis outputs to satellite-derived upper ocean currents and validate the results against in-situ measurements. Our results demonstrate that using surface information from SSH or surface velocities, combined with information on the stratification of the water column, it is possible to effectively reconstruct upper ocean dynamics in the Arctic and Subarctic Seas up to 400 meters. Future remote sensing missions in the Arctic Ocean, such as SWOT, Seastar, WaCM, CIMR, and CRISTAL, will produce enhanced SSH and surface velocities observations, allowing SQG schemes to characterize upper ocean 3D mesoscale dynamics up to 400 meters with higher resolution and lower uncertainties, M. Umbert received funding from H2020 Marie Skłodowska-Curie Actions under Grant Agreement No. 840374. E. De Andres received funding from Margarita Salas fellowship un-´ der Next Generation EU Grant Agreement No. UP2021-035. This work has also been funded by the AEI with the ARCTIC-MON project (PID2021-125324OB-I00). [...] We also received funding from the Spanish government through the ”Severo Ochoa Centre of Excellence” accreditation (CEX2019-000928-S) and from the ESA Arctic+Salinity project (AO/1-9158/18/I-BG), Peer reviewed

Level-3 daily merged Copernicus-GlobColour multi-sensor Remote Sensing Reflectance (Rrs(λ)) from the Copernicus Marine Environmental Monitoring Service (CMEMS) is used as data source. The aCDOM(443) was estimated using the gsmA parametrizations for the southern Beaufort Sea (Matsuoka et al., 2013) with some modifications; while DOC was retrieved using an empirical aCDOM(443)-DOC relationship developed by Juhls et al. (2022), valid across the fluvial-marine transition zone of the Mackenzie River–Beaufort Sea region, Level-4 daily maps of aCDOM(443) and DOC for the Mackenzie River-Beaufort Sea region (122-142ºW and 68-73ºN) are distributed in the regular latitude–longitude grid at a spatial resolution of 4 km. The data is available from 1st January 2002 to 31st October 2023. Additionally, 8-day means, and monthly mean data are also provided. The dataset consists of four variables:
(i) Chromophoric Dissolved Organic Matter at 443 nm (aCDOM(443) [m-1])
(ii) Dissolved Organic Carbon (DOC [µM])
(iii) water type mask (coastal/oceanic)
(iv) model relative error for Rrs(443) (∆Rrs(443) [%])
(v) aCDOM(443) uncertainty (σaCDOM [m-1])
(vi) DOC uncertainty (σDOC [µM]), YYYYMMDD_aCDOM443_DOC_l4_4km_P1D.nc (This dataset contains daily Level-4 aCDOM and DOC product from satellite observations for the Mackenzie River-Beaufort Sea region)
YYYYMMDD_YYYYMMDD_aCDOM443_DOC_l4_4km_P8D.nc (This dataset contains daily Level-4 aCDOM and DOC product from satellite observations for the Mackenzie River-Beaufort Sea region)
YYYYMMDD_YYYYMMDD_aCDOM443_DOC_l4_4km_P1M.nc (This dataset contains monthly Level-4 aCDOM and DOC product from satellite observations for the Mackenzie River-Beaufort Sea region), In the regime-shifting Arctic, organic carbon export from river watersheds is expected to increase due to changes in hydrological regimes and permafrost thawing. Given the impact these changes have on the biogeochemical cycles of coastal and shelf areas, robust monitoring of major Arctic rivers is needed. During ice-free season, ocean color remote sensing has shown to be an effective monitoring tool for remote, under-sampled areas with limited access such as the Beaufort Sea region. It offers synoptic spatio-temporal coverage, filling in the gaps among the scarce in situ data and improving our knowledge of land-to-ocean transport pathways and coastal ocean dynamics near riverine zones. Remote sensing of Chromophoric Dissolved Organic Matter (CDOM) and Dissolved Organic Carbon (DOC) has proven to be a valuable approach for assessing variations in terrestrial carbon export. However, seasonal fluctuations and diverse vegetation and soil conditions of each river watershed make it difficult to establish a single retrieval method for the entire Arctic. This complexity has led to a development of several regional algorithms, yet accessible long-term series remain a challenge. We present the first long-term satellite-derived dataset quantifying the absorption coefficient of CDOM at 443 nm and DOC concentrations in the Mackenzie River–Beaufort Sea region (122–142ºW and 68–73ºN) from 1st January 1998 to 31st October 2023. We employed a modified version of the GIOP (Generalized Inherent Optical Properties) algorithm to retrieve CDOM(443) for the southern Beaufort Sea; DOC was then derived using a region-specific relationship between aCDOM(443) and DOC. The product has been validated using in situ observations from several field campaigns conducted in the area. We provide a consistent dataset that can be used to assess DOM spatial and temporal variability, trends, and export to the coastal seas, This work has been carried out as part of the ARCTIC-MON project (Advances in monitoring sea ice thickness and snow depth, ocean dynamics, and freshwater content in the Arctic using remote sensing data), funded through the Consejo Superior de Investigaciones Científicas (CSIC) (grant no. PID2021-125324OB-I00), With the institutional support of the ‘Severo Ochoa Centre of Excellence’ accreditation (CEX2019-000928-S), Level-4 daily maps of aCDOM(443) and DOC for the Mackenzie River-Beaufort Sea region (122-142ºW and 68-73ºN) are distributed in the regular latitude–longitude grid at a spatial resolution of 4 km. The data is available from 1st January 2002 to 31st October 2023. Additionally, 8-day means, and monthly mean data are also provided. The dataset consists of four variables: (i) Chromophoric Dissolved Organic Matter at 443 nm (aCDOM(443) [m-1]) (ii) Dissolved Organic Carbon (DOC [µM]) (iii) water type mask (coastal/oceanic) (iv) model relative error for Rrs(443) (∆Rrs(443) [%]) (v) aCDOM(443) uncertainty (σaCDOM [m-1]) (vi) DOC uncertainty (σDOC [µM]), Peer reviewed

13 pages, 7 figures, 2 tables.-- Code availability: Underlying research code can be accessed at https://github.com/martaumbert/Ocean-Science-2024 (De Andrés et al., 2024).-- Data availability: Underlying research data are available at https://doi.org/10.6084/m9.figshare.c.7084813.v1 (Umbert, 2024), The hydrography of the Arctic Ocean has experienced profound changes over the last 2 decades. The sea ice extent has declined by more than 10 % per decade, and its liquid freshwater content has increased mainly due to glaciers and sea ice melting. Further, new satellite retrievals of sea surface salinity (SSS) in the Arctic might contribute to better characterizing the freshwater changes in cold regions. Ocean salinity and freshwater content are intimately related such that an increase (decrease) in one entails a decrease (increase) in the other. In this work, we evaluate the freshwater content in the Beaufort Gyre using surface salinity measurements from the satellite radiometric mission Soil Moisture and Ocean Salinity (SMOS) and TOPAZ4b reanalysis salinity at depth, estimating the freshwater content from 2011 to 2019 and validating the results with in situ measurements. The results highlight the underestimation of the freshwater content using reanalysis data in the Beaufort Sea and a clear improvement in the freshwater content estimation when adding satellite sea surface salinity measurements in the mixed layer. The improvements are significant, with up to a 70 % reduction in bias in areas near the ice melting. Our research demonstrates how remotely sensed salinity can assist us in better monitoring the changes in the Arctic freshwater content and understanding key processes related to salinity variations that cause density differences with potential to influence the global circulation system that regulates Earth's climate, This project was funded by Marie Skłodowska-Curie grant agreement no. 840374. Eva De Andrés is funded by Margarita Salas grant no. UP2021-035 under the NextGenerationEU program and supported by the MCIN/AEI project PID2020-113051RB-C31. We also received funding from the AEI with the ARCTIC-MON project (PID2021-125324OB-I00) and from the ESA Arctic+ Salinity project (AO/1-9158/18/I-BG) and Arctic+ SSS CCN (4000125590/18/I-BG). This research was supported by the European Union's Horizon 2020 research and innovation program under grant agreement no. 101003826 via the project “CRiceS – Climate Relevant interactions and feedbacks: the key role of sea ice and Snow in the polar and global climate system”, With the institutional support of the ‘Severo Ochoa Centre of Excellence’ accreditation (CEX2019-000928-S), Peer reviewed

14 pages, 14 figures, 1 table, The retrieval of sea ice thickness using L-band passive remote sensing requires robust models for emission from sea ice. In this work, measurements obtained from surface-based radiometers during the MOSAiC expedition are assessed with the Burke, Wilheit and SMRT radiative transfer models. These models encompass distinct methodologies: radiative transfer with/without wave coherence effects, and with/without scattering. Before running these emission models, the sea ice growth is simulated using the Cumulative Freezing Degree Days (CFDD) model to further compute the evolution of the ice structure during each period. Ice coring profiles done near the instruments are used to obtain the initial state of the computation, along with Digital Thermistor Chain (DTC) data to derive the sea ice temperature during the analyzed periods. The results suggest that the coherent approach used in the Wilheit model results in a better agreement with the horizontal polarization of the in situ measured brightness temperature. The Burke and SMRT incoherent models offer a more robust fit for the vertical component. These models are almost equivalent since the scattering considered in SMRT can be safely neglected at this low frequency, but the Burke model misses an important contribution from the snow layer above sea ice. The results also suggest that a more realistic permittivity falls between the spheres and random needles formulations, with potential for refinement, particularly for L-band applications, through future field measurements, This project is funded from the AEI with the ARCTIC-MON project (PID2021-125324OB-I00) and also with the Programación Conjunta Internacional project called ‘MEJORANDO LOS MODELOS DE EMISIVIDAD DEL HIELO MARINO EN LAS MICROONDAS DE BAJA FRECUENCIA’ (ICE-MOD), with reference PCI2019-111844-2. This work represents a contribution to the CSIC Thematic Interdisciplinary Platform PTI-POLARCSIC and PTI-TELEDETECT and is supported by the Spanish government through the ‘Severo Ochoa Centre of Excellence’ accreditation (CEX2019-000928-S), Peer reviewed

22 pages, 11 figures, supporting information https://doi.org/10.1029/2023JC020733.-- Data Availability Statement: All data used in this study are freely available through public repositories. SMOS SSS data (Martínez et al., 2020) are freely distributed from the Barcelona Expert Center https://doi.org/10.20350/DIGITALCSIC/12620. SIC data were obtained from OSI-450 (OSI SAF, 2017) and OSI-430b (OSI SAF, 2020) products from EUMETSAT. Data sets of OSTIA SST (Good et al., 2020) and TOPAZ4b SSS (Xie, 2022) are generated by E.U. Copernicus Marine Service Information. ORAS5 (Zuo et al., 2018) reanalysis data on SSS, SIT, wind stress, and sea ice velocity are delivered by the ECMWF, Arctic sea ice is retreating. In the Beaufort Gyre (BG), liquid freshwater content (FWC) has increased by 40% in the last two decades, with sea ice melting being one of the contributors. Combining satellite observations and reanalysis outputs, we study the sea ice meltwaters in the BG from 2011 to 2019. SMOS sea surface salinity (SSS) data were limited to the ice-free period, and reanalysis outputs showed limited accuracy in capturing the postmelt freshening observed in SMOS SSS. Applying criteria of SSS 25 psu and sea ice presence, we identified meltwater lenses (MWLs) across different years with higher frequency in 2011, 2015, and 2016, contrasting with scarcity in 2012, 2017, and 2019. MWLs showed lower SSS and temperature than river-influenced waters. Daily evolution of SSS and sea ice area within the MWLs exhibited similar patterns during both melting and formation stages. Furthermore, events of sea surface salinification following sea ice formation suggest that SMOS SSS might be capturing information on brine rejection. Atmospheric conditions suggested that the anticyclonic wind regime in 2011 contributed to meltwater retention, while the cyclonic wind pattern in 2016 favored outward meltwater dispersion. From our estimations, September's sea ice melting may contribute 10%–30% of the BG's annual FWC increase, though our method likely underestimates meltwater volume due to unaccounted advection processes. This study provides a validated method for MWL detection and underscores the importance of monitoring sea ice meltwaters to better understand regional variability in the BG within a changing climate, E. De Andrés is supported by Margarita Salas Grant UP2021-035 under the Next Generation EU program. M. Umbert is funded by Marie Skłodowska-Curie Grant Agreement No. 840374. This study was also financed by the MCIN/AEI projects DINGLAC (PID2020-113051RB-C31) and ARCTIC-MON (PID2021-125324OB-I00), and from the ESA Arctic+Salinity project (AO/1–9158/18/I-BG). This work represents a contribution to the CSIC Thematic Interdisciplinary Platform PTI-POLARCSIC, and is supported by the Spanish Government through the “Severo Ochoa Centre of Excellence” accreditation (CEX2019-000928-S), Peer reviewed

This in situ dataset collected during the ARCTIC-MON field campaign will be linked in the Barcelona Polar Lab (BPL) webpage: https://polar.icm.csic.es, The ARCTIC-MON campaign conducted at the Canadian High Arctic Research Station (CHARS) collected brightness temperatures from L-band radiometric measurements of sea ice, using the ARIEL portable L-band radiometer. This data is key to improve the emission modelling that is required for the development of the sea ice thickness products from remote sensing observations, This project is funded with the ARCTIC-MON project (PID2021-125324OB-I00) and also Programación Conjunta Internacional project called ICE-MOD, with reference PCI2019-111844-2, from the Agencia Estatal de Investigación in Spain. With the institutional support of the ‘Severo Ochoa Centre of Excellence’ accreditation (CEX2019-000928-S), Brightness temperature measured over sea ice at each stop done during each days transects. Each stop is labelled within each day’s measurements, Peer reviewed

X Simposio de Estudios Polares, 15-17 May 2024, Salamanca.-- This work represents a contribution to the CSIC Thematic Interdisciplinary Platform PTI-POLARCSIC, Arctic sea ice is retreating, thinning, and exhibiting increased mobility. In the Beaufort Gyre (BG), liquid freshwater content has increased by 40% in the last two decades, with sea ice melting being a primary contributor. This study utilizes satellite observations of sea surface salinity (SSS) and sea ice concentration, along with model-based sea ice thickness from 2011 to 2019. The aim is to investigate the sea ice-SSS relationship at different scales in the Arctic and understand the sea-ice meltwater dynamics in the BG. Our findings reveal a strong synchrony and positive correlation between sea ice area and SSS in the Arctic Ocean. In September, when the BG exhibits the largest ice-free ocean surface, a noticeable release of freshwater from sea ice melting occurs, a phenomenon not accurately reproduced by the models. The SMOS (Soil Moisture and Ocean Salinity) mission proves valuable in detecting meltwater lenses (MWL) originating from sea ice melting. These MWLs exhibit mean SSS ranging from 19 psu at the begining of sea ice retreat to 25 psu before sea ice formation. Wind-driven anticyclonic eddies can trap MWLs, preserving the freshest SSS imprints on the sea surface for up to 10 days. Furthermore, events of sea surface salinification following sea ice formation suggest that SMOS SSS might be capturing information on brine rejection. The daily evolution of sea ice SSS within the MWLs demonstrates a tight correlation between both variables after sea ice melting and just before sea ice formation, indicating a transient period in between, E. De Andrés is supported by Margarita-Salas Grant No. UP2021-035 under the Next Generation-EU program. M. Umbert is funded by Marie Sk lodowska-Curie Grant Agreement No. 840374. This study was also financed by the MCIN/AEI projects DINGLAC (PID2020-113051RB-C31) and ARCTIC-MON (PID2021-125324OB-I00), and from the ESA Arctic+Salinity project (AO/1-9158/18/I-BG), Peer reviewed

Special Issue Machine Learning and Statistical Learning with Applications 2025.-- 18 pages, 5 figures, 5 tables.-- The data supporting the conclusions of this article will be made available by the authors on request. These data were derived from the following resources available in the public domain: The production of SMOS sea ice thickness data was funded by the ESA project SMOS & CryoSat-2 Sea Ice Data Product Processing and Dissemination Service, and data from 15 October 2010 to 15 April 2021 were obtained from AWI. The BGEP data used in the validation were collected and made available by the Beaufort Gyre Exploration Program based at the Woods Hole Oceanographic Institution (https://www2.whoi.edu/site/beaufortgyre, accessed on 24 July 2025) in collaboration with researchers from Fisheries and Oceans Canada at the Institute of Ocean Sciences. Remote sensing data processing was executed at the Barcelona Expert Center on Remote Sensing (BEC-RS, https://bec.icm.csic.es, accessed on 24 July 2025) of the Institut de Ciències del Mar ICM-CSIC. The source codes are available for downloading at the following link: https://github.com/ferranhema/smos-mlsit (accessed on 24 July 2025), Data availability: Hernández-Macià, Ferran; Gabarró, Carolina; 2025; BEC SMOS Sea Ice Thickness [Dataset]; DIGITAL.CSIC; https://doi.org/10.20350/digitalCSIC/17749; http://hdl.handle.net/10261/407464, This study evaluates machine learning-based methods for retrieving thin Arctic sea ice thickness (SIT) from L-band radiometry, using data from the European Space Agency’s (ESA) Soil Moisture and Ocean Salinity (SMOS) satellite. In addition to the operational ESA product, three alternative approaches are assessed: a Random Forest (RF) algorithm, a Convolutional Neural Network (CNN) that incorporates spatial coherence, and a Long Short-Term Memory (LSTM) neural network designed to capture temporal coherence. Validation against in situ data from the Beaufort Gyre Exploration Project (BGEP) moorings and the ESA SMOSice campaign demonstrates that the RF algorithm achieves robust performance comparable to the ESA product, despite its simplicity and lack of explicit spatial or temporal modeling. The CNN exhibits a tendency to overestimate SIT and shows higher dispersion, suggesting limited added value when spatial coherence is already present in the input data. The LSTM approach does not improve retrieval accuracy, likely due to the mismatch between satellite resolution and the temporal variability of sea ice conditions. These results highlight the importance of L-band sea ice emission modeling over increasing algorithm complexity and suggest that simpler, adaptable methods such as RF offer a promising foundation for future SIT retrieval efforts. The findings are relevant for refining current methods used with SMOS and for developing upcoming satellite missions, such as ESA’s Copernicus Imaging Microwave Radiometer (CIMR), This project is funded from the AEI with the ARCTIC-MON project (PID2021-125324OB-I00), and from a Doctorat Industrial (AGAUR), with expedient number 2023 DI 0007., This work is supported by the Spanish government through the “Severo Ochoa Centre of Excellence” accreditation (Grant CEX2024-001494-S funded by AEI 10.13039/501100011033), Peer reviewed

Sea ice thickness from SMOS measurements. The retrieval is based in a hybrid approach, combining the Burke emission model (Burke et al., 1979) with a machine learning algorithm. The emission model is used to generate the dataset that is then used to train a Random Forest algorithm. The sea ice thickness is retrieved only up to 1 - 1.5 m of thickness due to the limited sensitivity at the L-band. This maximum retrievable thickness is not yet included in the product but will be in the future; for safety trust values only up to 1 m, ARCTIC-MON project (PID2021-125324OB-I00), from Agencia Estatal de Investigación (AEI), This work contributes to the Institut de Ciències del Mar "Severo Ochoa Centre of Excellence" accreditation CEX2024-001494-S funded by AEI 10.13039/501100011033 of the Spanish Ministry of Science and Innovation, Daily netCDF files containing the grided sea ice thickness estimations from SMOS, Peer reviewed

13 pages, 5 figures, 4 tables, supplementary material https://doi.org/10.1017/S095410202510031X, The Earth’s poles have always aroused great interest and fascination, first for explorers looking for the Northern Pass in the Arctic or for a new continent and natural resources to exploit, and today for scientists due to the significant role that the Arctic and Antarctic oceans play in the dynamics and future of our planet, especially in the current context of global change. To raise awareness of the importance and vulnerability of the polar oceans and to bring scientific advances in marine science to schools and the general public, we have worked from different approaches: 1) the online participation of students in polar oceanographic expeditions (through the ICM Divulga educational website), 2) face-to-face talks and workshops on polar ecosystems involving the observation of samples and/or videos, 3) the coordination and the edition of the book Observando los polos with a global and multidisciplinary vision of the state of scientific knowledge on polar areas, 4) the elaboration of the photographic exhibition ‘Una mirada polar’ and 5) the multi-institutional collaborative project surrounding the XIth Scientific Committee on Antarctic Research (SCAR) International Biology Symposium. These approaches, some of which involved Dr Andrés Barbosa’s collaboration, align with his objective of disseminating the results of research and scientific experience to the public. An evaluation of the strengths and weaknesses of the diverse strategies used to provide education about polar science is presented, This study is part of the POLARCSIC platform activities and had the institutional support of the ‘Severo Ochoa Centre of Excellence’ accreditation (Grant CEX2024-001494-S funded by AEI 10.13039/501100011033) to the ICM CSIC. [...] We sincerely appreciate the opportunities offered by the project leaders of PEGASO (CTM2012-37615, R. Simó), PI-ICE (CTM2017-89117-R, E. Berdalet and M. Dall’Osto), POLAR CHANGE (PID2019-110288RB-I00, M. Dall’Osto and R. Simó), SSSICE-SO (ARICE-H2020-PONANT ANTARCTIC SEASON 2023–2024, M. Umbert) and ARCTIC-MON (PID2021-125324OB-I00, C. Gabarró). ‘A Polar Gaze’ was partially funded by the Fundación General del CSIC Cuenta la Ciencia project. The book Observando los polos was an initiative of PTI (now Connection) POLARCSIC (https://www.polarcsic.es). The ‘Antarctica: Life on the Edge - Scientific Voyages Today’ exhibition was co-funded by the Maritime Museum of Barcelona and the Institut de Ciències del Mar of Barcelona (CSIC), Peer reviewed

27 pages, 10 figures, 3 tables, supplemental material https://doi.org/10.1109/JSTARS.2026.3654597/mm1; https://doi.org/10.1109/JSTARS.2026.3654597/mm2, In the Arctic, sea surface salinity (SSS) serves as a key tracer of freshwater variability and upper-ocean dynamics. While spaceborne SSS observations provide a unique capability for monitoring near-surface variability across undersampled polar regions, systematic quality assessments are essential to improve retrieval accuracy and ensure reliable interpretation. This study evaluates the spatio-temporal accuracy of five widely utilized satellite-derived products from the Soil Moisture and Ocean Salinity (SMOS), Aquarius, and Soil Moisture Active-Passive (SMAP) missions across nine pan-Arctic regions through comparative analysis with in situ and model reanalysis data. While satellite and reanalysis products generally agreed on large-scale September salinity patterns, major discrepancies emerged in river-influenced areas. Arctic-dedicated satellite products exhibited stronger temporal correlations with model reanalysis (r up to 0.95) and greater spectral consistency relative to global datasets. However, salinity overestimation near ice margins and meltwater outputs persisted across most observational records—particularly during summer months—highlighting the impact of meltwater stratification and unresolved contamination on SSS retrievals. Comparative analysis with in situ measurements showed that the CCI dataset demonstrated the closest agreement among the five SSS products in the Nordic Seas, with RMSD (R2) values ranging from 0.41 to 0.54 (0.21 to 0.69). Accurately capturing freshwater masses in the central Arctic remains a challenge (RMSD > 1.12), with BEC exhibiting smaller errors in most marginal seas but less variability in the eastern Arctic. These findings reinforce the added value of regionally tailored satellite products in extending the utility and improving the accuracy of SSS monitoring in the Arctic Ocean, The work of María Sánchez-Urrea was supported by the Spanish Ministry for Science and Innovation through Ph.D. under Grant PRE2021-099346. The work of Marta Umbert was supported by the ERC-2024-StG FRESH-CARE under Grant 101164517. Thework ofEva De Andréswas supported byARCTIC-MON project under Grant PID2021-125324OB-I00. This work was supported in part by the Agencia Estatal de Investigación through the ARCTIC-MON project (Advances in monitoring sea ice thickness and snow depth, ocean dynamics, and freshwater content in the Arctic using remote sensing data) under Grant PID2021-125324OB-I00 and in part by the Spanish government through the “Severo Ochoa Centre of Excellence” accreditation under Grant CEX2019-000928-S, Peer reviewed

24 pages, 10 figures, 1 table, supplementary material https://www.frontiersin.org/articles/10.3389/frsen.2026.1768049/full#supplementary-material.-- Data availability statement: The products are publicly available through the
Digital CSIC server: https://digital.csic.es/handle/10261/340924 (last accessed: 20
December 2023), with the DOI number: https://doi.org/10.20350/digitalCSIC/15721
(S´anchez-Urrea et al., 2023). All code associated with the algorithm presented in this study is openly
accessible under the Creative Commons 4.0 license via Zenodo (https://doi.org/10.5281/
zenodo.16603282; S´anchez-Urrea et al., 2025), In the regime-shifting Arctic, organic carbon export from river watersheds is expected to rise due to changes in hydrological regimes and permafrost thawing, affecting coastal and shelf biogeochemistry. Ocean color remote sensing enables monitoring inaccessible areas like the Beaufort Sea, improving our knowledge of coastal dynamics and land-to-ocean transport of Chromophoric Dissolved Organic Matter (CDOM) and Dissolved Organic Carbon (DOC) during the ice-free season. While multi-sensor datasets provide consistent monitoring over extended periods, the viability of using merged L3 remote sensing reflectance (Rrs) to derive carbon-based regional products remains unexplored. This study uses the Copernicus-GlobColour multi-sensor merged Rrs to generate a long-term dataset of CDOM absorption at 443 nm, aCDOM(443), and DOC concentrations in the Mackenzie River–Beaufort Sea region. We employ a modified GIOP (Generalized Inherent Optical Properties) inversion algorithm to retrieve aCDOM(443); then, derive DOC using a region-specific aCDOM(443)-DOC relationship. Validation against in situ observations shows performance comparable to previous studies, with a MdAPD (r2) of 56% (0.55) for aCDOM(443) and 29% (0.70) for DOC. This novel dataset allows for a detailed analysis of riverine plume dynamics, interannual variability, and trends. Average summer plume extent ranges from 43,068 to 92,388 km2 and is significantly correlated with annual discharge (r = 0.6, p< 0.001). When compared to independent in situ DOC estimates (1999-2017) at the river mouth, our satellite product shows consistent variability patterns. Contrary to initial expectations, a significant decline of −0.017 m−1yr−1 and −3.40 μM yr−1 is observed over a 26-year period for aCDOM(443) and DOC, respectively, at the Mackenzie delta outflow, resulting from a slight increase (1998–2015) followed by a decrease (2016–2023). These results suggest that trends in DOC fluxes need to be assessed using longer time series while considering significant uncertainties in both retrieval accuracy and variable spatiotemporal coverage over multidecadal periods, This work has been carried out as part of the ARCTIC-MON project (Advances in monitoring sea ice thickness and snow depth, ocean dynamics, and freshwater content in the Arctic using remote sensing data), funded through the Agencia Estatal de Investigación (AEI) (grant no. PID 2021-125324OB-I00). María Sánchez-Urrea gratefully acknowledges the financial support from the PhD Grant PRE 2021-099346 from Spanish Ministry for Science and Innovation. MU was funded by Marie Skłodowska-Curie Grant Agreement No. 840374. ED was funded by Margarita Salas Grant No. UP 2021-035 under the Next-Generation EU program and supported by the MCIN/AEI project PID 2020-113051RB-C31. RG-A was supported by the Independent Research Fund Denmark (4251-00058B). This work constitutes a contribution to the CSIC Thematic Interdisciplinary Platform PTI-POLARCSIC and PTI-TELEDETECT and is supported by the Spanish government through the “Severo Ochoa Centre of Excellence” accreditation (CEX2019–000928-S), Peer reviewed

Satellite observations are a key tool in this effort, employing various technologies to measure sea‑ice thickness. Altimetry satellites are sensitive to ice thicker than approximately 1 meter, while low‑frequency passive microwave sensors such as SMOS can retrieve thinner sea ice by detecting its natural microwave emission. These two approaches therefore provide complementary information. However, there is still room to improve the quality of remotely sensed sea‑ice thickness products by refining the retrieval methodologies. Progress is limited mainly by the scarcity of in situ measurements, which are essential for developing empirical models and performing robust validation.
The main objective of the project was therefore to measure sea‑ice thickness by continuously operating the SIMS instrument mounted on the icebreaker Le Commandant Charcot, and to collect in situ sea‑ice measurements during station stops. The in situ measurements were obtained through ice‑coring and drilling operations, which allowed us to validate the SIMS data and to characterize the physical properties of the sea ice. These include sea‑ice and snow thickness (measured with high precision), sea‑ice temperature, salinity, and dielectric constant—key parameters for improving the modeling required by the retrieval algorithms.
The SIMS instrument was deployed when possible given the meteorological conditions encountered when navigating through sea ice.
The figure in the attached file shows the SIMS sea ice thickness measurements during SIMSVAL project, both raw and with a median filter applied to smooth it. The stars indicate the in situ drillings done under the instrument when the ship stopped in the ice. As shown, the ground truth clearly matches the SIMS value at that point, highlighting its reliability despite the noise that appears when the ship is navigating and breaking the ice floes, We acknowledge the Spanish funding agency Agencia Estatal de Investigación (España) and Ministerio de Ciencia e Innovación (España) for its support through the ARCTIC-MON project (PID2021-125324OB-I00) that enabled us to cover travel expenses, shipping, as well as the rental and purchase of some of the required field work equipment.
We also acknowledge PONANT for providing us with berths on the icebreaker Le Commandant Charcot, for providing the SIMS and drilling instruments, and for the technical support from Daniel Cron, This work contributes to the Institut de Ciències del Mar "Severo Ochoa Centre of Excellence" accreditation CEX2024-001494-S funded by AEI 10.13039/501100011033 of the Spanish Ministry of Science and Innovation, There are three files, one for each cruise segment: - SIMSVAL: 21 March 2025 – 05 April 2025 - POLARIS: 05 April 2025 – 17 April 2025 - FANS: 17 April 2025 – 02 May 2025, Peer reviewed

Satellite observations are a key tool in this effort, employing various technologies to measure sea ice thickness. Altimetry satellites are sensitive to ice thicker than approximately 1 meter, while low-frequency passive microwave sensors such as SMOS can retrieve thinner sea ice by detecting its natural microwave emission. These two approaches therefore provide complementary information. However, there is still room to improve the quality of remotely sensed sea ice thickness products by refining the retrieval methodologies. Progress is limited mainly by the scarcity of in situ measurements, which are essential for developing empirical models and performing robust validation.

The main objective of the project was to collect brightness temperatures from L-band radiometric measurements of sea ice using the ARIEL portable L-band radiometer, and simultaneously measure sea ice thickness with the SIMS instrument, as well as sea ice conditions (salinity, temperature, and snow properties). This dataset is key to improving the emission modelling required for the development of sea ice thickness products derived from remote sensing observations, We acknowledge the Spanish funding agency Agencia Estatal de Investigación (España) and Ministerio de Ciencia e Innovación (España) for its support through the ARCTIC-MON project (PID2021-125324OB-I00) enabled us to cover travel expenses, shipping, as well as the rental and purchase of some of the required equipment.
We also acknowledge PONANT for providing us with berths on the icebreaker Le Commandant Charcot, for providing the SIMS and drilling instruments, and for the technical support from Daniel Cron, This work contributes to the Institut de Ciències del Mar "Severo Ochoa Centre of Excellence" accreditation CEX2024-001494-S funded by AEI 10.13039/501100011033 of the Spanish Ministry of Science and Innovation, There are three files, one for each cruise segment: - SIMSVAL: 21 March 2025 – 05 April 2025 - POLARIS: 05 April 2025 – 17 April 2025 - FANS: 17 April 2025 – 02 May 2025, Peer reviewed

This study evaluates machine learning-based methods for retrieving thin Arctic sea ice thickness (SIT) from L-band radiometry, using data from the European Space Agency's (ESA) Soil Moisture and Ocean Salinity (SMOS) satellite. In addition to the operational ESA product, three alternative approaches are assessed: a Random Forest (RF) algorithm, a Convolutional Neural Network (CNN) that incorporates spatial coherence, and a Long Short-Term Memory (LSTM) neural network designed to capture temporal coherence. Validation against in situ data from the Beaufort Gyre Exploration Project (BGEP) moorings and the ESA SMOSice campaign demonstrates that the RF algorithm achieves robust performance comparable to the ESA product, despite its simplicity and lack of explicit spatial or temporal modeling. The CNN exhibits a tendency to overestimate SIT and shows higher dispersion, suggesting limited added value when spatial coherence is already present in the input data. The LSTM approach does not improve retrieval accuracy, likely due to the mismatch between satellite resolution and the temporal variability of sea ice conditions. These results highlight the importance of L-band sea ice emission modeling over increasing algorithm complexity and suggest that simpler, adaptable methods such as RF offer a promising foundation for future SIT retrieval efforts. The findings are relevant for refining current methods used with SMOS and for developing upcoming satellite missions, such as ESA's Copernicus Imaging Microwave Radiometer (CIMR).