^{1}

^{*}

^{2}

^{3}

^{1}

^{2}

^{3}

Edited by: Gilles Reverdin, Centre National de la Recherche Scientifique (CNRS), France

Reviewed by: Isabelle Taupier-Letage, UMR7294 Institut Méditerranéen d'Océanographie (MIO), France; Marie Isabelle Pujol, Collecte Localisation Satellites (CLS), France

This article was submitted to Ocean Observation, a section of the journal Frontiers in Marine Science

This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

Mesoscale eddies are a key oceanic feature relevant to the transport of water properties and biological material. These structures, through their surface signature, have been characterized and widely investigated using sea level anomaly (SLA) maps retrieved from satellite altimetry observations. Gridded SLA fields are routinely computed combining SLA data along satellite tracks with an optimal interpolation algorithm using a selected set of parameters. These are crucial because they define and constrain the temporal and spatial scales of the structures resolved. Here, we investigate the impact of the choice of the interpolation parameters in global and in the Mediterranean SLA products. Our findings demonstrate that the number of eddies detected in SLA maps over the Mediterranean Sea regional products is significantly larger along the satellite tracks than between the tracks, irrespective of the number of satellites used to produce the gridded maps, of the time period considered and of the eddy size. We also show that this is not the case of the global SLA product. We attribute the anomalously high number of eddies detected along the tracks in the regional products to the correlation scales of the optimal interpolation algorithm, which are different from those in the global product due to a smaller Rossby radius of the Mediterranean structures. Among these, we find that the time correlation scale is more restrictive than the spatial one.

- Anomalous larger eddy number detected along Topex/Poseidon tracks in the Mediterranean SLA products.

- Result observed during 23 years of SLA maps for all eddy radius but not for the Global SLA products.

- The parameters used in the optimal interpolation algorithm have an important contribution to the observed result.

Mesoscale eddies are found across all oceans (Chelton et al.,

The paper is organized as follows: the SLA data used and the eddy identification and tracking algorithm are detailed in section 2. In section 3, the spatial distribution and temporal evolution of the density of eddies for the Mediterranean and global SLA satellite products is closely checked. We also shed some light about how the parameters chosen for the optimal interpolation algorithm can explain the anomalous observed eddy density distribution. Finally, the main conclusions are included in section 4.

The satellite sea level anomaly (SLA) maps and their associated errors used in this study to compute the different eddy datasets correspond to DT-2014 product version [this product is not longer available and has been recently substituted by the DT-2018 version (Taburet et al.,

Two different SLA datasets were used for each one of the regions analyzed (global ocean and Mediterranean sea). The first one is the so called

We applied the eddy identification and tracking algorithm developed by Faghmous et al. (

The maps of eddy density (number of eddy centers per square degree and day in a grid of 0.1 × 0.1°) in the Mediterranean sea computed from both SLA products (

Eddy density maps computed on a 0.1° grid for the Mediterranean Sea

The abnormal eddy density distribution in the Mediterranean sea is observed during the whole period covered by the SLA products. We have averaged the eddy density in a 0.5° strip along the Topex/Poseidon satellite tracks and in the gaps between tracks (thick lines) for each map (using 0.25° strip did not change the results significantly indicating that this is not a critical parameter). The 6 months-low-pass filtered time series for both computations is shown in

This difference in the density of eddies along the Topex/Poseidon tracks and in the gaps between tracks is found for all eddy sizes (eddy radius computed as the radius of a circle with the same area of the eddy) with significant eddy density in both Mediterranean SLA products (

It is worth mentioning that the aim of this study is not to determine the exact eddy density in a given region or to determine the exact value of the TG ratio, since their precise values may depend on the SLA product used and the details of the eddy detection and tracking algorithm. The real aim of this manuscript is to show that there is an anomalous result in the spatial distribution of the eddies detected in the SLA maps for the Mediterranean Sea and to analyze its possible origin. Pursuing the latter goal, it can be stated that the global and Mediterranean SLA products have only two main differences, namely: (1) the spatial resolution (1/4 or 1/8 of a degree, respectively) and (2) the parameters used in the optimal interpolation (OI) algorithm (AVISO,

The results show that the TG ratio strongly depends on the combination of correlation scales chosen (

Ratio of number of eddies along Topex/Poseidon tracks and number of eddies between tracks depending on the spatial (R) and temporal (T) correlation scales used by the optimal interpolation algorithm. The contour plot has been computed linearly interpolating the points indicated by black crosses. The black square (circle) indicates the combination of parameters that (Amores et al.,

We have shown that the eddy density maps (number of eddies per square degree and day) computed in the Mediterranean Sea from altimetry based SLA products (

Publicly available datasets were analyzed in this study. This data can be found here:

AA wrote the paper and did the computations. GJ helped writing the paper and contributed to the idea. SM helped improving the paper and suggested the study of the optimal interpolation parameters.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

GJ acknowledges a Ramón y Cajal contract (RYC-2013-14714) funded by the Spanish Ministry of Economy and the Regional Government of the Balearic Islands.