Research papers
The current filters are: Starting year = 2025, Ending year = 2026
0 Kb
Crespin J., Clavel-Henry M., Canals M., M. Thyng K., Ruiz-Xomchuk V., Solé J. (2025)
Ocean Modelling, 198 DOI: 10.1016/j.ocemod.2025.102596. (BibTeX: crespin.etal.2025a)
Abstract: See
Modeling the distribution of biogeochemical components in the ocean is essential for further understanding
climate change impacts and assess the functioning of marine ecosystems. This requires robust and efficient
physical-biological simulations of coupled ocean-ecosystem models, which are often hindered by limited data
availability and computational resources. The option of running biological tracer fields offline, independently
from the physical ocean simulation, is appealing due to increased computational efficiency. Here, we present
an assessment and implementation of an offline biogeochemical model — the Offline Fennel model — within
the Regional Ocean Modeling System (ROMS). Our methodology employs ROMS hydrodynamic outputs to run
the biogeochemical model offline. This work also includes the first ground-truthing exercise of the referred
offline biogeochemical model. We use a variety of skill metrics to compare the simulated surface chlorophyll
to an ocean color dataset (Copernicus Marine Service Mediterranean Ocean Color) and BGC-Argo floats for the
2015–2020 period. The model is able to reproduce the temporal and spatial structures of the main chlorophyll
fluctuation patterns in the study area, the Northwestern Mediterranean Sea. This area is of particular interest
as it is one of the most productive regions in the entire Mediterranean Basin, with open-ocean upwellings and
deep winter convection events occurring seasonally. The typical behavior of the region is likewise effectively
represented in the implementation, including offshore primary production, nutrient supplies from the Rhone
and Ebro rivers, and mesoscale hydrographic structures. This study provides a baseline for ROMS users in need
of executing more biogeochemical simulations independently from more computationally demanding physical
simulations.
0 Kb
Crespin J., Solé J., Canals M. (2025)
Geoscientific Model Development, 18, 5891–5912. DOI: https://doi.org/10.5194/gmd-18-5891-2025. (BibTeX: crespin.etal.2025b)
Abstract: See
Ocean biogeochemical models are essential for
advancing our understanding of oceanographic processes.
Here, we present the Offline Fennel model, a biogeochemi
cal model that relies on previously computed physical fields,
within the Regional Ocean Modeling System (ROMS). We
evaluated the model performance against a fully coupled
physical–biogeochemical online application in the northern
Gulf of Mexico, a region with intense biogeochemical ac
tivity, including rather frequent hypoxia events. By leverag
ing physical hydrodynamic outputs, we ran the Offline Fen
nel model using various time-step multiples from the cou
pled configuration, significantly enhancing computational eff
iciency and reducing simulation computational time by up
to 87%. The accuracy of the offline model was assessed
using three different mixing schemes: the generic length
scale (GLS), Large–McWilliams–Doney (LMD), and Mel
lor and Yamada 2.5 (MY25). The offline model achieved an
average skill score of 93%, with minimal impact on perfor
mance from the time-step choice. While the GLS configura
tion yielded the highest accuracy, all three mixing schemes
performed well. Although some discrepancies appeared be
tween offline and coupled simulation outputs, these were
smaller than those observed when using different mixing
schemes within the same model configuration. A significant
challenge identified was the simulation of ammonium (NH4),
which exhibited the largest discrepancies due to its rapid
turnover timescale compared to other tracers. The promising
results achieved so far validate the Offline Fennel model’s
capability and efficiency, thus offering a powerful tool for re
searchers aiming to conduct extensive biogeochemical simu
lations without rerunning the hydrodynamic component, thus
significantly reducing computational demands
0 Kb
Crespin J., Solé J., Canals M. (2025)
Progress in Oceanography, 235 DOI: 10.1016/j.pocean.2025.103494. (BibTeX: crespin.etal.2025)
Abstract: See
Jet Streams (JS) are powerful upper-tropospheric winds that significantly influence weather and climate. As
anthropogenic climate change alters temperature gradients, subtropical JS are expected to shift poleward, which
can have unforeseen consequences on midlatitude Earth systems. Here, we demonstrate, for the first time, the
impact of the steady poleward migration of the Northern Hemisphere subtropical JS on Marine Primary Pro
duction (MPP). Using over two decades of data (2000–2023), we establish a direct relationship between the JS
latitudinal position and MPP variability in the Northwestern Mediterranean Sea. The observed northward
migration of approximately 75 km over the study period aligns with a consistent decline in chlorophyll con
centrations, representing a 40 % reduction, with rates reaching up to 5% per year. This is attributed to the
steady northward seasonal shift of the JS position, which drives changes in northern wind-stress and Ekman
pumping, subsequently reducing upwelling occurrence and intensity. While the primary influence of JS position
on MPP is seasonal, we demonstrate that its impact extends to non-seasonal components as well. Unlike other
studies linking JS shifts to short-term wind stress variations and isolated upwelling events, our findings highlight
a long-term impact on MPP. Our findings suggest that JS dynamics is a dominant driver of MPP variability in the
Northwestern Mediterranean Sea and point to equivalent situations in other marine regions worldwide. The
cascading effects of reduced MPP have the potential to significantly impact marine ecosystems and resources,
with broader implications for fisheries and the carbon cycle.
Keywords: Jet Stream, Marine Primary Production,Atmosphere-Ocean Interactions,Climate Change,Climate Change Impacts
0 Kb
Crespin J., Clavel-Henry M., Canals M., M. Thyng K., Ruiz-Xomchuk V., Solé J. (2025)
Ocean Modelling, 198 DOI: 10.1016/j.ocemod.2025.102596. (BibTeX: crespin.etal.2025a)
Abstract: See
Modeling the distribution of biogeochemical components in the ocean is essential for further understanding
climate change impacts and assess the functioning of marine ecosystems. This requires robust and efficient
physical-biological simulations of coupled ocean-ecosystem models, which are often hindered by limited data
availability and computational resources. The option of running biological tracer fields offline, independently
from the physical ocean simulation, is appealing due to increased computational efficiency. Here, we present
an assessment and implementation of an offline biogeochemical model — the Offline Fennel model — within
the Regional Ocean Modeling System (ROMS). Our methodology employs ROMS hydrodynamic outputs to run
the biogeochemical model offline. This work also includes the first ground-truthing exercise of the referred
offline biogeochemical model. We use a variety of skill metrics to compare the simulated surface chlorophyll
to an ocean color dataset (Copernicus Marine Service Mediterranean Ocean Color) and BGC-Argo floats for the
2015–2020 period. The model is able to reproduce the temporal and spatial structures of the main chlorophyll
fluctuation patterns in the study area, the Northwestern Mediterranean Sea. This area is of particular interest
as it is one of the most productive regions in the entire Mediterranean Basin, with open-ocean upwellings and
deep winter convection events occurring seasonally. The typical behavior of the region is likewise effectively
represented in the implementation, including offshore primary production, nutrient supplies from the Rhone
and Ebro rivers, and mesoscale hydrographic structures. This study provides a baseline for ROMS users in need
of executing more biogeochemical simulations independently from more computationally demanding physical
simulations.
0 Kb
D´Alimonte D., Kajiyama T., Pitarch J., Brando V.E., Talone M., Mazeran C., Twardowski M., Kolluru S., Tonizzo A., Kwiatkowska E., Dessailly D., Gossn J.I. (2025)
Remote Sensing of Environment, 321, 11, 4606. DOI: 10.1016/j.rse.2025.114606. (BibTeX: d´alimonte.etal.2025)
Abstract: See
Several methods were developed in Ocean Colour remote sensing over the last 25 years to model the anisotropy
of the upwelling radiant field with respect to observation and solar-illumination geometries, also denoted as
bidirectional reflectance distribution function (BRDF). These methods are necessary to produce normalized, or
“BRDF-corrected,” marine reflectance representative of the seawater’s inherent optical properties (IOPs) independently of the measurement conditions. Each scheme relies on specific modeling assumptions and implementation solutions, which can lead to different results depending on the actual combination of the seawater
IOPs with the illumination and viewing geometry. The first aim of this study is to analyze the principles and
methods of the reference BRDF schemes presented by Morel et al. (denoted as M02), Park and Ruddick (P05), Lee
et al. (L11), He et al. (H17), and Twardowski and Tonizzo (T18). Acknowledging the direct applicability of M02,
P05, and L11, their performance has been verified under a variety of conditions, including in situ measurements,
matchup observations, and space-borne images. Comparisons between non-corrected and normalized data
clearly confirm the need to account for the BRDF effect. In particular, the analysis of the results indicates 1) a
substantial equivalence of M02, P05, and L11 in clear waters and 2) the tendency to obtain better results with
M02 and L11 as the optical complexity increases. Although M02 was conceived for Case 1 waters, the underlying
Chlorophyll-a overestimation tendency in some optically complex conditions is likely the reason for its extended
applicability. Since L11 is based on a more comprehensive and flexible framework for all water types, the design
of this method is suggested for revisions and BRDF correction improvements.
Keywords: Ocean colour remote sensing Bidirectional reflectance distribution function
0 Kb
Hernández-Macià F., Gabarro C., Sanjuan- Gomez G., J. Escorihuela M. (2025)
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 17, 10752–10758. DOI: 10.1109/JSTARS.2024.3406921. (BibTeX: hernandez-macia.etal.2025b)
Abstract: See
This study proposes a machine learning based
methodology for estimating Arctic thin sea ice thickness (up to
1 m) from brightness temperature measurements of SMOS. The
approach involves employing the so-called Burke model for sea ice
emission modeling, integrating a suitable permittivity model and
a radiative transfer equation. The training dataset is generated
through a model-based simulation, and is then used to train and
evaluatetwomachinelearningregressionalgorithms:RandomFor
est and Gradient Boosting. Overall, this machine learning method
ology results in great agreement with the ESA’s official sea ice
thickness product. Additionally, a validation performed by using
data from mooring measurements shows a subtle improvement
by the machine learning algorithms with respect to the ESA’s
official product.Theseresultsindicatetheirpotentialtosurpassthe
performance of the current SMOSthinseaice thickness retrievals.
0 Kb
Hernández-Macià F., Gabarro C., Huntemann M., Naderpour R., T. Johnson J., C. Jezek K. (2025)
Annals of Glaciology, 65, e37 DOI: 10.1017/aog.2024.38. (BibTeX: hernandez-macia.etal.2025a)
Abstract: See
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 dur
ing 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.
Keywords: Sea ice; sea-ice modeltice geophysics; remote sensing
0 Kb
Hernández-Macià F., Sanjuan G., Gabarro C., Escorihuela M.J. (2025)
Computers, 14, 8, 305. DOI: 10.3390/computers14080305. (BibTeX: hernandez-macia.etal.2025)
Abstract: See
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
ESAproduct, 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 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).
Keywords: machine learning; remote sensing; sea ice; cryosphere
0 Kb
Hernández-Macià F., Gabarro C., Huntemann M., Naderpour R., T. Johnson J., C. Jezek K. (2025)
Annals of Glaciology, 65, e37 DOI: 10.1017/aog.2024.38. (BibTeX: hernandez-macia.etal.2025a)
Abstract: See
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 dur
ing 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.
Keywords: Sea ice; sea-ice modeltice geophysics; remote sensing
0 Kb
Hernández-Macià F., Sanjuan G., Gabarro C., Escorihuela M.J. (2025)
Computers, 14, 8, 305. DOI: 10.3390/computers14080305. (BibTeX: hernandez-macia.etal.2025)
Abstract: See
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
ESAproduct, 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 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).
Keywords: machine learning; remote sensing; sea ice; cryosphere
0 Kb
Makarova E., Portabella M., Stoffelen A. (2025)
IEEE Transactions on Geoscience and Remote Sensing, 63 DOI: 0.1109/TGRS.2025.3586375. (BibTeX: makarova.etal.2025)
Abstract: See
The numerical weather prediction (NWP) stress-
equivalent 10-m wind (U10S) forecasts are used as a common
forcing for ocean models; however, these forecasts present local
and systematic biases when compared to the observational
data. The scatterometer wind observations are being assimilated
by European Centre for Medium-Range Weather Forecasts
(ECMWF) Integrated Forecasting System (IFS), but even after
the assimilation, the sea-surface wind biases are still present.
A previous approach to reduce such biases was based on correct-
ing the forecasts with the mean differences between scatterometer
observations and the NWP output accumulated over a certain
period of time. However, this approach shows performance degra-
dation for the periods when fewer scatterometers are available
and in the operational framework. To overcome these limitations,
we propose the use of machine learning (ML) to predict such
biases using other atmospheric and oceanic NWP variables as
inputs, so that the observational data are only required during
the training. In this work, we show the results for the preliminary
ML models trained on a small subset of data that use U10S
scatterometer–NWP differences as the target. The predicted
corrections applied to the ECMWF fifth reanalysis dataset ERA5
show error variance reduction up to 9.86% on a test subset
globally when compared to Advanced Scatterometer (ASCAT-A)
and up to 6.25% against independent scatterometer HSCAT-B,
thereby reducing the local biases. The best performance is seen
in the extra tropics with error variance reduction up to 10.6%
Keywords: ERA5 biases, machine learning (ML), neural networks, scatterometers, stress-equivalent winds
0 Kb
Olivé A., Pelegrí J.L., Claret M. (2025)
Journal of Geophysical Research: Oceans, 130, 8 DOI: 10.1029/2024JC021494. (BibTeX: olive.etal.2025)
Abstract: See
Lagrangian simulations based on 18 years (2002–2019) of high-resolution thermohaline and three-dimensional velocity fields allow revisiting the fate and thermohaline changes of the upper-ocean Antarctic Circumpolar Current (ACC) waters that enter directly the South Atlantic Ocean basin. An advection-diffusion scheme, applied to both climatological annual-mean and daily mean fields, allows estimating the mean pathways and seasonal variability, as well as recirculation volume transports, times, and depths in the South Atlantic subtropical gyre (SASG). The annual-mean diffusive simulation shows that 96.5 Sv of the upper-ocean waters (up to the 28.00 kg m−3) crossing the Drake Passage remain in the ACC, while 13.0 Sv join the eastern margin of the SASG. About 8.6 Sv of this eastern input, plus an additional 2.7 Sv that enter the SASG through the interior ocean, reach the North Brazil Current, yielding a total Drake contribution of 11.2 Sv to the upper returning-limb of the Atlantic Meridional Overturning Circulation. The upper-ocean waters that reach the eastern SASG undergo substantial water mass transformations, with a net transfer of 6.7 Sv from intermediate-deep to surface layers and an increase in heat transport by 0.39 PW and salt transport by 8.5 × 106 kg s−1, but remain largely unchanged as they drift westward toward the western boundary at 21°S. Most waters within the SASG (86%) recirculate once, taking a median of 9.1 years, although some complete as many as three loops after reaching 32°S-W. Regarding seasonality, the transit times and transport fraction of the upper-ocean waters flowing into the SASG show higher variability than those remaining in the ACC path.
0 Kb
Viudez A. (2025)
Physics of Fluids, 37, 11 DOI: 10.1063/5.0296879. (BibTeX: viudez.2025b)
Abstract: See
The multipolar spherical vortices of any degree ℓ, which are exact solutions to the classical nonlin
ear equation of motion for a perfect fluid, exhibit two possible polarizations determined by the sign
of the radial wavenumber k. We propose that the spin-up and spin-down states of spin-1/2 quantum
particles correspond to these two classical polarization states of ℓ = 1 vortices. In the presence of
a homogeneous background vorticity field 2ν, these vortices precess around the axis defined by ν
and propagate with a drift velocity equal to 2ν/k. This drift enables the experimental separation
of vortices with opposite polarizations. It is shown that the correlation between measurements of
the drift velocity 2ν/k for pairs of vortices, as observed by two independent observers, can lead to
violations of the Clauser-Horne-Shimony-Holt inequality—suggesting a classical physics explanation
of quantum entanglement.
0 Kb
Viúdez A. (2025)
European Journal of Mechanics B/Fluids, 111, 81-86. DOI: 10.1016/j.euromechflu.2024.12.004. (BibTeX: viudez.2025)
Abstract: See
The multipolar spherical vortex solutions to the Euler equations for Newtonian fluids in background cylindrical flow with swirl satisfy, once their three-dimensional Cartesian velocity components are mapped into the components of a four-component complex vector wave function, the relativistic Dirac equation for a free particle. It is suggested that the vertical component of the intrinsic spin angular momentum of the quantum mechanics particles is the azimuthal wavenumber of the angular phase of the oscillation modes in presence of the background rotation.
Keywords: Euler fluid equations Multipolar spherical vortex solutions Dirac quantum mechanics equation
0 Kb
Viúdez A. (2025)
Physics of Fluids, 37, 5 DOI: 10.1063/5.0261868. (BibTeX: viudez.2025a)
Abstract: See
A spin-1 vortex consists of the three time-dependent multipolar spherical modes associated
with the three spherical harmonic functions of degree ℓ = 1. The spin-1 vortex is basically
the spherical Hicks-Moffatt vortex with an arbitrary orientation in the three-dimensional
space. It is shown that, in the presence of a background flow with cylindrical swirl of
arbitrary orientation and a background time-dependent radial expansion/contraction flow,
the orientation of the spin-1 vortex precesses about an axis and with a frequency both
prescribed by the background cylindrical flow. The time dependence of the precession
frequency is prescribed by a background radial divergent flow. It is shown that this vortex
precession in presence of a constant background vorticity is analogous to the precession
of the magnetic moment of a body in presence of an external constant magnetic field, or
Larmor precession.
0 Kb
Wang S., Portabella M., Dong X., Lin W., Bao Q. (2025)
IEEE Transactions on Geoscience and Remote Sensing, 63 DOI: 10.1109/TGRS.2025.3625931. (BibTeX: wang.etal.2025b)
Abstract: See
Ocean currents and winds are crucial parame
ters to understand ocean–atmosphere interactions, while the
simultaneous retrievals using Doppler scatterometry can provide
direct observational support. To obtain high-quality airborne
Doppler scatterometer wind and current products, accurate
estimation of the observed backscatter coe cient and reliable
wind field inversion are essential. Based on the Ocean Surface
Current Observation Mission (OSCOM) airborne experiment
data collected using a Ka-band rotating pencil-beam Doppler
scatterometer, we propose two di erent calibration methods for
the backscatter coe cient to account for the larger-than-expected
azimuthal modulation of the backscatter signal, as predicted by
consolidated geophysical model functions (GMFs) used in Ka
band scatterometry. Both methods are based on the numerical
ocean calibration (NOC) approach, which is in turn based on
the estimation of the mean backscatter di erences between real
measurements and simulated ones with the use of the GMF
and reference winds. The first method employs an azimuth
dependent calibration, which can be implemented using either an
overall ratio or a ratio per flight leg. The second method involves
modifying the GMF to match the observed azimuthal modulation,
with options for one or two GMF coe cient adjustments. The
retrieved wind speeds range from 4 to 7 m/s, with wind directions
around 155 . In comparison with collocated European Centre for
Medium-Range Weather Forecasts (ECMWF) winds, the wind
speed biases of di erent methods are all lower than 1.2 m/s,
and the wind direction standard deviations (SDs) are lower than
93 . The azimuth-dependent calibration method yields smaller
wind speed biases but larger wind direction SDs compared to the
modified GMF method. The azimuth-dependent calibration using
leg-dependent ratios leads to the closest retrieved wind speeds
and directions to ECMWF. The calibration methods proposed
in this study provide data support for future simultaneous
retrieval studies of ocean winds and currents. Additionally, these
methods can be applied to other airborne Doppler scatterometer
experiments.
Keywords: Backscatter measurements, calibration, Doppler scatterometer, ocean winds
0 Kb
Werner-Pelletier N., Carrasco-Serra O., Umbert M., Hoareau N., Salat J., Reynaud T. (2025)
Frontiers in Marine Science (BibTeX: werner-pelletier.etal.2025)
Abstract: See
Ever more often, opportunity vessels are used to provide in-situ sea surface
temperature and salinity data. In particular, sailing vessels participating in oceanic
races are often utilized, as they usually cover remote areas not reached by
commercial vessels, such as the southern oceans. The received signal from
temperature and salinity sensors-especially the latter- is often disturbed either
by bubbles, due to strong turbulent flows, or by non-renewal of the water in
contact with the sensor. Until now, only manual methods have been successfully
usedtofilter this data, since no automated procedurehasbeendeveloped. Inthis
paper, we present (i) a sensor housing to be placed on the keel, designed to
reduce the aforementioned physical issues, and (ii) an automatic filtering method
to override the manual procedure. The physical system was mounted on the
historic sailboat Pen Duick VI and has served to collect data along the Ocean
Globe Race route (2023-2024). This initiative was a collaboration between the
crew of the boat, the Institute of Marine Sciences (ICM-CSIC) in Barcelona, and
the Laboratoire d’Oceanographie Physique et Spatiale (Ifremer). The housing for
sensors consisted of a 3D-printed hydrodynamic support, designed to reduce
drag. The automated filtering approach was based on wavelet denoising
techniques and simple moving averages. The results are presented in an open
dataset and show that procedure yielded good performance in identifying and
rejecting outliers, while operating with far greater speed than manual filtering.
The method is intended to become a standard procedure for similar in-situ
datasets, and an open-source software is provided for this purpose. This work is a
step forward in oceanographic data processing and aims to provide a tool with a
wide range of applications.
Keywords: sea surface temperature, sea surface salinity, vessels of opportunity, ocean racing, sensor housing, wavelet denoising, data filtering, automated quality control
