the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 21 Oct 2025
Polarity and direction dependence of energetic cross-frontal eddy transport in the Southern Ocean’s Pacific sector
Abstract. Cross-frontal mesoscale eddies mediate meridional heat and mass transport across the Antarctic Circumpolar Current fronts. Yet their spatiotemporal characteristics and dynamical impacts in the Southern Ocean’s Pacific sector remain inadequately quantified. Utilizing 23 years (2000–2022) of satellite altimetry and Argo float data, we reveal, for the first time, a pronounced polarity and direction dependence in cross-frontal eddy (CFE) abundance, energetics, and interactions with jets. Equatorward-propagating cyclonic eddies (CEs) dominate CFE activity (36 % of total), exhibiting higher eddy kinetic energy (EKE, in terms of total EKE in eddy interiors, EKET), longer propagation distances, and stronger nonlinearity than other types, followed by poleward-moving anticyclonic eddies (AEs, 28 %). These two dominant directional groups primarily drive the significant increase in the overall CFEs’ EKET: CEs at (1.98 ± 1.53) × 106 m4 s−2 yr−1 (excluding the anomalously low 2017 value) and AEs at (1.58 ± 0.74) × 106 m4 s−2 yr−1. Specifically, complete CFEs (experience pre-crossing, crossing, and post-crossing phases) are responsible for these trends, distinct from non-CFEs, partial or transient CFEs, which show no trend. During frontal crossing, EKET enhances in equatorward CEs and poleward AEs but diminishes in poleward CEs and equatorward AEs, explaining the two former types’ capacity for long-distance propagation and energetic behaviors. The intensified CEs carry cold, fresh southern waters equatorward, while AEs transport warm, salty northern waters poleward. These cross-frontal exchanges mitigate thermohaline gradients between interfrontal zones while enhancing local mesoscale available potential energy. We conclude that CFEs serve dual climatic roles, in mediating meridional energy transport while dynamically stabilizing the ACC against strengthening winds.
- Preprint
(7242 KB) - Metadata XML
-
Supplement
(937 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2025-5038', Igor Belkin, 14 Dec 2025
-
RC2: 'Comment on egusphere-2025-5038', Anonymous Referee #2, 14 Jan 2026
Summary:
The manuscript is novel and addresses an interesting topic. However, I recommend rejection at this stage due to methodological shortcomings, primarily related to the eddy detection and tracking approach and key parameter choices that are insufficiently justified and validated. I encourage the authors to reprocess the analyses by addressing the issues and requested revisions outlined below, and to consider resubmitting a substantially revised version thereafter.Major issues:
Altimetry Dataset and Eddy Detection and Tracking (EDT)
Comments: The manuscript uses the C3S DUACS two-satellite L4 SLA product. While a fixed two-altimeter constellation can improve long-term stability/homogeneity (important for climate metrics), the C3S QUID explicitly states that the all-satellite CMEMS products provide better spatial sampling and reduced mesoscale errors and “should be preferred for oceanic mesoscale applications,” whereas the C3S two-satellite products are intended for climate applications. The authors should justify this choice in the context of eddy detection/tracking, and ideally add a short sensitivity test (e.g., a representative sub-period) comparing C3S vs CMEMS all-satellite to show that eddy counts/sizes/intensities and the main conclusions are not materially affected.
Comments: Several recent eddy atlases and methodologies increasingly rely on ADT rather than SLA for eddy detection/tracking, as SLA removes the mean dynamic topography and may therefore suppress quasi-permanent features associated with standing meanders, jets, or long-lived structures (e.g., Laxenaire et al., 2018; Pegliasco et al., 2022). Please justify the use of SLA here and clarify whether using ADT would affect eddy detection and the inferred eddy–front relationships.
145-147: “A threshold W= −0.2σw was often chosen to identify the outer boundary of an eddy, with being the standard deviation of W over the entire region (e.g., Henson and Thomas, 2008; Isern-Fontanet et al., 2006).”
Comments: The definition of σw is unclear. Is it the standard deviation of W computed over the full spatial domain for each day separately, or over the full domain and the entire study period? Please clarify. In addition, the choice of the threshold needs stronger justification for the Southern Ocean/ACC context. The cited studies (e.g., Isern-Fontanet et al., 2006) applied this criterion in different dynamical regimes and data products, and such a fixed fraction of σw may be product and region dependent. A brief sensitivity analysis (e.g., varying the coefficient around 0.2) would help demonstrate that the results are robust.150-151: “SLA data with 0.25° spatial resolution were first linearly interpolated to 0.125° for better performance in eddy detection.”
Comments: I am concerned this is not appropriate, sampling to a finer grid does not add information and may introduce interpolation artefacts (artificial noise/small-scale structure) that can bias eddy detection and tracking. I recommend performing the analysis on the native-resolution SLA.152-156: “For eddy tracking, the algorithm identifies eddies at time t+1 that meet the following criteria relative to time t: (1) minimal centroid distance, (2) identical polarity (i.e., rotation direction), and (3) the minimum radius variation. If no eddy at t+1 satisfies these proximity thresholds for a given eddy at t, the eddy is considered dissipated. Conversely, if an eddy detected at t+1 does not match any eddy at t, it is classified as a newly generated eddy. Based on eddy identification and tracking results, this study focuses on eddies with lifespans exceeding 7 days and amplitudes greater than 2 cm.”
Comments: The proposed eddy detection/tracking workflow (OW thresholding combined with outermost closed SLA contours and a local nearest-neighbour tracking based on centroid distance/polarity/radius change) appears insufficiently robust, especially for the Southern Ocean, where strong jets, crowding and deformation make eddy identification and track continuity challenging. Such simplified schemes are prone to track switching, fragmentation and sensitivity to thresholds. In addition, the adopted selection criteria (lifetime > 7 days; amplitude > 2 cm) require justification. I recommend either adopting a validated state-of-the-art EDT framework (e.g., Chelton/Schlax-type contour methods, META3.1exp/TOEddies) or providing thorough validation and sensitivity analyses (including reasonable variations of tracking parameters and thresholds) demonstrating that eddy statistics and the derived CFE conclusions are not materially affected by the chosen assumptions.ACC frontal zone definition
169-170: “To account for topographically induced frontal displacements, we defined frontal zones as a strap ±15 km expanded in the normal directions from each climatological front, consistent with observed SO frontal oscillation area (Kim and Orsi, 2014).”
Comments: It is not clear how the ±15 km band is justified from Kim and Orsi (2014). In that paper, the reported meridional drift of the ACC fronts between 1993 and 2010 is on the order of ~28–46 km (their Fig. 8), and the estimated total drifts assuming a linear trend show sector-dependent values, with the Pacific sector spanning roughly −100 to +20 km depending on the front (their Fig. 9). This suggests that frontal variability/displacement can exceed the adopted 15 km half-width. Because this threshold likely influences eddy-front colocation statistics and therefore the main conclusions, it should be clarified and explicitly justified. A more defensible approach could be to define frontal zones based on the observed latitudinal variability of each front in the study sector (e.g., mean ±1 standard deviation of front latitude), rather than using a fixed ±15 km value. Therefore, this definition should be revisited.Minor issues:
40: Please briefly clarify in the text the distinction between front and jet
70-73: Sentence is too articulated. I suggest simplifying it, by splitting it into two shorter sentences to improve clarity and flow.
111: Please provide the DOI of the Copernicus Marine product used, and report the date of access/download
114-117: Since the horizontal velocities are provided directly by Copernicus, I suggest removing the detailed description of how they are computed. As written, it may imply you derived the velocities yourself. Readers interested in the processing can refer to the product documentation.
118-126: This paragraph could be shortened for readability. Also, it may help to more clearly separate what is taken from Park et al. versus what is done in this study. For instance, you could briefly state that you used the ACC front positions from Park et al. (altimetry-based), which were validated against available subsurface observations (e.g., Argo over the relevant period and CTD data in 2016–2017).
167: Replace “:” with “.”
217-219: Please clarify the percentage values mentioned in the text, as they are not directly reported in the table. As I understand it, they are obtained by summing the corresponding AE and CE percentages for that category and specific front, please state this explicitly.
251: “, p=0.08)” replace with “(p=0.08)”
272-273: Consider replacing “Total number of CFEs, divided into types of equatorward and poleward directions;” with “Total number of CFEs, divided into types of equatorward and poleward directions, shown as a function of the different ACC fronts;”
379: Do the water-mass core values reported in Table 3 have circumpolar validity for the Southern Ocean? Please clarify the geographic representativeness of the cited values. Since this work focuses on the Pacific sector, you may also consider or discuss Pacific-sector estimates of SAMW and AAIW core properties (e.g., Bostock et al., 2013; Li et al., 2021, 2022).
383: I suggest using a different colour bar for salinity rather than reusing the same colour bar as potential temperature, to avoid confusion and improve readability.
465-469: Please provide the DOI of the Copernicus Marine product used and report the date of access/download. Also, the Argo data link does not work, please check and update it.
Citation: https://doi.org/10.5194/egusphere-2025-5038-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 215 | 100 | 24 | 339 | 29 | 15 | 16 |
- HTML: 215
- PDF: 100
- XML: 24
- Total: 339
- Supplement: 29
- BibTeX: 15
- EndNote: 16
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Review of the manuscript egusphere-2025-5038 “Polarity and direction dependence of energetic cross-frontal eddy transport in the Southern Ocean’s Pacific sector” by Huimin WANG, Lingqiao CHENG, Erik BEHRENS, Zhuang CHEN, Jennifer DEVINE, and Guoping ZHU submitted to EGUsphere
Reviewer: Igor M. Belkin
Date: December 14, 2025
Summary: This study is methodologically flawed to the point where I must recommend rejection. I do not see how these flaws (enumerated below) can be fixed.
Major issues:
Frontal pattern of the SO:
Line 49: “In the SO, the transition from the warm subtropical waters to the cold Antarctic waters is not smooth but concentrated along a series of fronts (Deacon, 1937).”
Comments: Deacon (1933, 1937) missed the Subantarctic Front (SAF) largely due to the poor spatial resolution of oceanographic observations during RV Discovery cruises, with most stations placed three degrees of latitude apart. The SAF was discovered much later. The first definitive report of SAF was published by R.W. Burling in 1961. Thus, regarding the large-scale frontal pattern of the SO, Deacon’s works of 1933 and 1937 are outdated. The first modern circumpolar surveys of the SO fronts were published by Orsi et al. (1995) and Belkin and Gordon (1996).
ACC’s northern boundary (NB):
The authors use the northern boundary (NB) of ACC, which was introduced by Park et al. (2019, pp. 4515-4516): “The streamlines corresponding to the above five circumpolar fronts are determined as follows. First, the ACC northern and southern boundaries are unambiguously defined as those circumpolar streamlines passing through the northernmost and southernmost latitudes of Drake Passage, coinciding with the MDT contour of 0.30 m for the NB and −1.11 m for the SB. There is no great alternative for defining these boundaries because their streamlines are constrained by continental slopes at Drake Passage… …the definition of the NB from altimetry seems new. Note however that our NB runs close to the northern branch of the SAF (SAF‐N) in Sokolov and Rintoul (2009) and Barré et al. (2011). Indeed, it is shown in the following subsection that the newly defined NB from altimetry matches well with the SAF‐N identified from hydrography in the study area. Therefore, the altimetry derived NB and the hydrography‐derived SAF‐N are used interchangeably in this study.”
Also, L123: “Park et al. (2019) produced the most updated mapping of the ACC fronts. As shown in Figure 1, from north to south, the key fronts include the northern boundary of ACC (NB), the Subantarctic Front (SAF), the Polar Front (PF), the Southern Antarctic Circumpolar Current Front (SACCF), and the southern boundary of the ACC (SB).”
Comments: Fronts are commonly defined as narrow high-gradient zones. The ACC’s northern boundary (NB) defined by Park et al. (2019) is a streamline, not a front. Therefore, the authors should not use NB in their analysis. Traditionally, the northern boundary of ACC is the Subantarctic Front (SAF).
Eddy tracking:
L150: “SLA data with 0.25°spatial resolution were first linearly interpolated to 0.125° for better performance in eddy detection.”
Comments: The interpolation does not mitigate the poor spatial resolution (28 km) of SLA data. Such data is inadequate for studies of mesoscale eddies, whose diameters can be as small as 40-50 km.
L152: “For eddy tracking, the algorithm identifies eddies at time t+1 that meet the following criteria relative to time t: (1) minimal centroid distance, (2) identical polarity (i.e., rotation direction), and (3) the minimum radius variation.”
Comments: Traditionally, ocean eddies are tracked using their thermohaline (TS) signatures. This is a standard approach successfully utilized in hundreds of field campaigns and data analyses. The authors’ ignorance of TS signatures of individual eddies is a major flaw of this study.
L156: “Based on eddy identification and tracking results, this study focuses on eddies with lifespans exceeding 7 days and amplitudes greater than 2 cm.”
Comments: Mesoscale eddies have a typical lifespan ranging from several weeks up to a few years (rings of major currents). Their typical SSHA amplitudes range widely. In the Southern Ocean, the mean SSHA is 12 cm (Frenger et al., 2015, JGR). In the North Pacific, the SSHA ranges from 10 cm to 30 cm and higher (Ebuchi and Hanawa, 2001, Journal of Oceanography, Fig. 2 and Fig. 3). Rings spawned by major currents have SSHA>40 cm in the Gulf Stream area (Belkin et al., 2020, JPO, Fig. 11), 35-50 cm in young Agulhas rings (van Aken et al., 2003, DSR2; Schmid et al., 2003, DSR2; Wang et al., 2016, JGR) and up to 60 cm in the South Atlantic (Souza et al., 2011, Ocean Science, Table 2). Thus, the 7-day, 2-cm SSH anomalies are hardly qualified as mesoscale eddies. Such anomalies should be considered noise, especially given the accuracy of satellite altimetry of about 1-2 cm.
Frontal zones:
L165: “While climatological fronts define the ACC’s mean structure (Park et al., 2019), their positions exhibit meridional variability influenced by both bathymetry and eddy activity (Kim and Orsi, 2014; Thompson et al., 2010): Fronts stabilize over major bathymetric features (e.g., the Pacific-Antarctic Ridge) but show maximum variability in flat basins, with widened frontal zones developing downstream of obstacles like the Campbell Plateau. To account for topographically induced frontal displacements, we defined frontal zones as a strap ±15 km expanded in the normal directions from each climatological front, consistent with observed SO frontal oscillation area (Kim and Orsi, 2014).”
Comments: The 15 km threshold grossly underestimates the real observed shifts of SO fronts. Frontal meanders lead to frontal shifts of up to ~100 km in either direction. When a frontal meander pinches off and forms a ring, the ring diameter is typically on the order of 100 km. The 30-km-wide frontal zone definition thus makes little sense.
L433: “…this study defines each frontal zone as a 30 km-wide strip-shaped area but does not
account for potential interannual or seasonal variations that may extend beyond this range. Similarly, all qualified Argo profiles from 2000 to 2022 were used without considering interannual or seasonal variability in hydrographic properties. These limitations inevitably introduce certain uncertainties.”
Comments: In addition to seasonal and interannual shifts of individual fronts, these fronts experience so-called synoptic (or intra-seasonal) variability caused by meanders. Such meanders effectively lead to shifts of individual fronts on the order of 100 km in cross-frontal directions.
Cross-frontal eddies:
L193: “CFE occurrence peaks downstream of prominent topographic features, particularly near the Campbell Plateau (150°E–180°E; 39% of total CFEs) and downstream of the Udintsev Fracture Zone (125°W–160°W; 38%), where multiple fronts converge (Figure 2a). Eddies may cross multiple fronts sequentially at these frontal convergent regions. The majority of eddies cross a single front (Figure 2c). Double-frontal crossings (total 434) occur preferentially at southern fronts (SACCF/SB; > 50% of cases; Figure 2d). Triple-frontal crossings are rare and primarily limited to the PF/SACCF/ SB system (Figure 2e), and no instances of quadruple-frontal crossings were observed.”
Comments: The cross-frontal transport by rings spawned by fronts is well known. However, the cross-frontal transport by other types of mesoscale eddies is a totally different phenomenon, which is exceedingly rare and not well documented from in situ observations. Dufour et al. (2015, JPO) wrote: “However, meridional transport in the Southern Ocean requires crossing multiple intense jets that form the ACC fronts and act as natural barriers to tracer transport. Both observational and experimental studies have demonstrated that mixing across the core of jets is strongly inhibited due to the speed of the jet being higher than the propagation speed of eddies, hence reducing the time during which eddies can stir tracers (e.g., Bower et al. 1985; Lozier et al. 1997; Sommeria et al. 1989; Ferrari and Nikurashin 2010). Nonetheless, strong perturbations, such as rings detaching from the front, might provide a way for tracers to cross the jet cores (Samelson 1992; Wiggins 2005).” – Comments: The above excerpt from Dufour et al. (2015, JPO) makes clear that the cross-frontal eddy transport is likely effectuated by rings, while the cross-frontal transport by other types of mesoscale eddies is exceedingly small except for the cross-frontal transport by subsurface (intrathermocline) lenses at depth (e.g., Bower et al., 2013, DSR2). Thus, the probability of the same eddy crossing multiple fronts is infinitesimally small. Nonetheless, the authors claim that their algorithm detected numerous cases of multiple frontal crossings by the same eddies, with 434 cases of double-frontal crossings, and several cases of triple-frontal crossings (lines 197-198 and Figure 2e). It seems that these numbers are artifacts that resulted from a faulty eddy tracking algorithm.
Minor edits:
L55: “Practically” – Delete.
L73: “Electrona calserbgi” should be “Electrona carlsbergi”
L84: “mitigating the eastward flow in the ACC” – “Mitigating” is not the best term in the given context.
L108: “the Campbell Plateau, Pacific-Antarctic Ridge, and Udintsev Fracture Zone (Figure 1).” – Add the Eltanin Fracture Zone to the text and to the maps in Figure 1 and elsewhere.
L112: “during 2000 and 2022” should be “during 2000-2022”
L216: “and mitigate thermohaline gradients across frontal zones.” – Comments: Replace “mitigate” with a more appropriate term.”
L239: “display marked longer lifespans” should be “display markedly longer lifespans”
L284: “Note the x-axis in a–b are not equidistant at higher values.” – Rewrite.
L334: “Argo θ-S profiles (Figure 8) demonstrate that marked meridional thermohaline gradients exist between eddy intervals in different interfrontal zones, with colder and fresher water properties poleward. In the same zones, CEs consistently exhibit colder, fresher properties with shallower isopycnals (upper 1000 dbar), while AEs contain warmer, saltier waters with deeper isopycnals, reflecting their respective meridional origins.” – Comments: This is a trivial observation: From north to south, the SO temperature and salinity generally decrease across all fronts and interfrontal zones. This general ocean-wide trend has been known for a century.
L423: “CFEs mitigate cross-frontal water mass property gradients” – Comments: Avoid using “mitigate” regarding gradients.
L456: “This process directly mitigates the sharp meridional gradients.” – Comments: Avoid using “mitigate” regarding gradients.
===== END of REVIEW =====