the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 05 May 2026
When the fjords take a breath: Influence of wind forcing and Ekman dynamics on deep ventilation in the North Patagonian fjords
Abstract. Climate change directly alters the dissolved oxygen (DO) dynamics in the ocean. Ocean ventilation is critical for redistributing nutrients and gases throughout the water column, supporting essential biogeochemical cycles, and regulating the ocean capacity to absorb carbon dioxide and heat, making it essential for climate regulation. In the northern Patagonian fjords, ventilation events mitigate the impacts of deoxygenation and hypoxia, which are primarily driven by the poleward transport of Equatorial Subsurface Water. However, climate-driven changes such as surface warming and increased stratification may reduce the frequency and intensity of these ventilation events, thereby contributing to long-term deoxygenation. In this study, we examined the physical drivers, frequency, and variability of deep ventilation events in the northern Patagonian fjords from 2016 to 2024. ERA5 reanalysis data, from an oceanographic platform moored at 170 m depth, were used to evaluate wind stress, Ekman transport, pumping, and layer depth, alongside in situ time series of temperature, salinity, density, and DO. The results showed a seasonal increase in DO concentrations, classified as ventilation periods, whereas synoptic-scale oxygen events were identified and quantified as ventilation pulses. In total, 16 ventilation periods and 35 pulses were registered over the time series. Onshore Ekman transport favouring downwelling was identified as the main driver of ventilation periods from the austral winter to spring, accompanied by salinity decreases before and during oxygen increases. Pulses coincided with synoptic events dominated by negative total Ekman transport, driven by northern wind stress, along with concurrent thermohaline changes. Overall, these findings provide novel insights into the mechanisms governing oxygen renewal in the Patagonian fjords and underscore the importance of sustained observation systems in a changing ocean.
- Preprint
(3614 KB) - Metadata XML
-
Supplement
(715 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-2329', Anonymous Referee #1, 18 Jul 2026
-
RC2: 'Comment on egusphere-2026-2329', Anonymous Referee #2, 21 Jul 2026
Review of
"When the fjords take a breath: Influence of wind forcing and Ekman dynamics on deep ventilation in the North Patagonian Fjords"General remarks:
The article analyses a long term time series at the Guafo mouth in 170 m depth of temperature, salinity and dissolved oxygen together with ERA-5 wind data.
By defining their own, yearly changing, baselines of oxygen concentrations, the authors can define venticalation pulses and ventilation periods. They find that both occur regularly and analyse if the coastal Ekman transport can be attributed to the events. A seasonal analysis finds that the Ekman transport is on average onshore (with a negative τy, Fig. 5). The Ekman pumping changes over the year, being negative (downward pumping) over the austral winter (Fig. 7) and slightly positive during the austral summer. Interesting here is that the mooring is located close to the location within the fjord, where the Ekman pumping switches its sign (Fig. 6, color from red to blue). This suggests a high fluctuation, comapared to more on- and offshore locations.
The authors find that Ekman transport process are a main driver for the oxygenization events, which is an important result and worth of publication.
The authors are further investigating the processes for the ventilation pulses by using a principal component analysis (PCA) and wavelet coherences.While the paper is well written, using a beautiful long term dataset, to which I congratulate the authors, I have several remarks. The for me most major issue is a more generic description of the location and the drivers of oxygen. I miss a Figure with typical vertical profiles for the different seasons, understanding the vertical structure of T,S and oxygen. As a person not familiar with the system I do not understand the sources and sinks for oxygen, why is there hypoxia in 170 m? Is this water transported from oxygen minimum zones into the fjord, or is it local consumption, or is this water from within the fjord, that is transported into the ocean? Here a description in the introduction and/or discussion in the conclusions would be very helpful. Maybe even with a small comic-style sketch of the system. Similarly the authors do "only" discuss oxygenization events, and very shortly other processes. If there are oxygenization events, are there vice-versa also hypoxic events? Sec. 4.5, "Limitations and perspectives" or elsewhere could be extended, also in comparison with other systems. As far as I understood, the fjord is less of a basin/sill like structure like many other hypoxic/anoxic systems where intrusions of oxygen play a larger role (see 2 examples out of many of literature below for examples of intrusions from the Baltic Sea and the Clew Bay in Ireland) but this differences makes it maybe interesting to discuss this system in comparison with others, that expererience for the same, or different reason, hypoxia. I understand that this request can be extensive, but a short comparison would be helpful.
In terms of methodology it would be beneficial to really see the comparison of the measured oxygen with the winkler titration. Measuring oxygen is always tricky, especially long term deployments with biofouling etc.
Please also review the equations carefully, there are for example often subscripts missing. Please do also mark vector quantities and their components more clearly. Why do the authors use for a transport the notation "U", which is typically a velocity? That might confuse readers. What about M_x, M_y or M_coast as examples?
An analysis that I dont really understand is the PCA. What is the use to do it here? For me a "PC1: 48.3 %, PC2: 22.7 %" does not tell me a lot. Please elaborate what that means otherwise this section can be removed.
As a last criticism, the AOU is not well defined.Details:
Fig. 1: The scale is ranging between +2000m to -5000m. This is ok for the overview, but for Panel c a new scale is better. What is distance between the isobaths in Panel c? Depths are stated differently (sometimes 170m, sometimes 175 m).
Fig. 2d: Why is there a temperature peak during the austral winter?
Fig 4, caption: τy, subscripts
65: "(GM) (Linford et al., 2023)": Two brackets, maybe (GM, Linford et al., 2023)
182: "along the water column": This is unclear, are vertical CTD profiles with water samples meant?
193: "Major to 7d): Again unclear, that das Major mean here?
257: AOU not defined and abbraviation not explained.
284: \rho_a subscript
294: U_e, V_e subscript
Line 310 and Equation 4: Ue and Ve are not defined, components of M? M is also a vectorial component, this needs to be made more clear.
Line 465: (τᵥ): Typo, \tau_y
466: "The long-term monthly mean ranged from −0.056 N m⁻² in August to 0.026 N m⁻² in January, yielding a seasonal amplitude of 0.082 N m⁻².": I dont see this, January seems also to have a negative windstress.
865: Here water renewal processes by intrusions could be compared to other well known hypoxic systems as the Baltic Sea.
920: "Ultimately, the GM functions as a dynamic gateway, where episodic renewal intermittently counteracts the long-term influence of low-oxygen subsurface waters entering northern Patagonia." This is an important sentence and should be elaborated more. See comment above.Fig. 5: Title fonts are not consistently bold, (b) is bold, the other labels not.
Fig. 9: Mark the location of GM in the map with a dotLiterature:
Holtermann, P., Prien, R., Naumann, M. and Umlauf, L. (2020), Interleaving of oxygenized intrusions into the Baltic Sea redoxcline. Limnol Oceanogr, 65: 482-503. https://doi.org/10.1002/lno.11317
Kelly, Seán, et al. "Characterizing ventilation events in an anoxic coastal basin: Observed dynamics and the role of climatic drivers." Limnology and Oceanography 65.10 (2020): 2420-2442Citation: https://doi.org/10.5194/egusphere-2026-2329-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 157 | 240 | 14 | 411 | 33 | 26 | 30 |
- HTML: 157
- PDF: 240
- XML: 14
- Total: 411
- Supplement: 33
- BibTeX: 26
- EndNote: 30
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
My apologies to the authors and editor for the delay in submitting this review. The manuscript by Sola-Hidalgo and co-authors describes the dynamics of deep-water ventilation at the gateway connecting a complex fjord system in southern Chile and the adjacent ocean. The study is based on the analysis of a unique time series of dissolved oxygen, temperature, and salinity, collected at 170 m depth from 2016 to 2024. A somewhat shorter version of the data has been described in previous papers by the same group. In this paper the authors make a clear distinction between seasonally and short-term ventilation sequences, referred to as “periods” and “pulses”, respectively. Despite the differences in time scales, the observed oxygen variability appears to be closely associated with changes of along-shore (meridional) winds through Ekman dynamics.
Major comments:
Overall, the manuscript is very well written. The most important result is the clear association between the near-bottom oxygen variability and the changes in wind pattern, which seems to hold for the short-lived synoptic-scale changes and the more persistent seasonal changes. The ventilation events appear to be associated with onshore Ekman transport and downwelling as a response to northerly winds, and a contribution of Ekman pumping. These results are fairly well supported by the available evidence (particularly Figs. 2, 5, 7, and 8). The authors conclude that the near-bottom ventilation is associated with the replacement of equatorial subsurface water by the relatively warm, fresh, and well-oxygenated subantarctic water. This later argument is a lot weaker, simply because the time series observations are only available near the bottom. Also, one would expect a northerly (downwelling) wind to promote a stronger southward advection of EESW. Please address this in the Discussion. In any case, the authors need to further develop the description of the processes linking the Ekman layer and the changes in near-bottom water masses, and provide more evidence to support their conclusion. These issues are revisited in the Discussion “… salinity at 170 m emerged as an important local factor controlling the strength and persistence of the ventilated state, even if part of this salinity signal was itself linked to the same forcing that promoted renewal.”. I strongly suggest reconsidering the wording, there is no evidence that salinity controls the strength nor the persistence of the ventilation. Coexistence does not imply control. Moreover, with the available observations it is not possible to determine whether the ventilation is local or associated with a ventilated near-bottom inflow.
The association between near-bottom oxygen variability and atmospheric rivers (paragraph in line 539), summarized in Fig. 9, is highly speculative. Please clearly spell out what the relevance of water vapor is. Fig. 9 displays a tight relationship between wind magnitude, zonal Ekman transport and integrated water transport. However, the implied moisture impact is totally unclear, and, most important, the oxygen response to the strong wind events (12, 14, 17, and 19 July) is not obvious. Why does DO drop after the 25 July pulse? What makes the 19 July event distinct to justify it being highlighted in Fig 9?
It is not possible to conclude much from this unless the analysis period is extended at least a few weeks prior to 11 July and a few weeks after 27 July. If that extension shows much lower DO during periods of weak winds, then the evidence will be somewhat more robust. However, I don´t think this adds much to the otherwise pretty solid paper.
Likewise, I find the association with ENSO plausible, as shown in Linford et al (2023), but the way it is being presented here is somewhat speculative. First, late 2023 and early 2024 also presented El Nino conditions, and no apparent anomaly at the mooring location. Second, are the concurrent warming (beyond 10°C and freshening under 34) observed during the 2016 event compatible with the alleged advection of SAAW? In line 792 you state that the lower boundary of ESSW deepened in 2015-2016, I presume you mean the upper boundary of ESSW.
In contrast, section 4.4 is very good.
Overall, the Discussion is rather long, a bit wordy and repetitive. Shortening this section will help conveying a clear message. I also suggest downplaying the potential role of pulses in sustaining the duration of the longer-term periods. Note that, 3 out of the 4 longest periods (8, 9, and 16), all exceeding 100 days, present no pulses, indicating that pulses are not a necessary condition for the development of extended ventilation events.
Terminology and description of events:
The short-lived ventilation events are rather loosely referred to as “events” (l. 390, 391), pulses (pretty much elsewhere), please be consistent. In general, to avoid confusion, I suggest avoiding referring to “events” without stating “period” or “pulse”. This should be checked throughout the manuscript.
The text refers to specific events by numbering them sequentially. To aid the reader following up the arguments, each period and pulse should be labeled in Figure 3.
Description of data processing:
The data being reported are of great scientific value. I commend the authors and responsible institutions for maintaining such a long time series (2016-2024). However, maintaining science-quality data for such a long period poses numerous challenges due to bio-fouling and sensor drift. You provide some information about the oxygen calibration procedures. It would be useful to see some sensor-Winkler comparisons. Conductivity sensors are also quite sensitive. Inclusion of more calibration details is of fundamental importance. This can be added as supplementary material.
For most of the analyses you use filtered hourly data re-sampled at daily frequency, yet, the filtering information is unclear and is a bit scattered. You first state that data were collected hourly and daily averaged “… for reducing high-frequency variability …” (l. 166). Later on (l. 201) you state that all hourly data were low-passed filtered with a cutoff period of 48 hours “…to suppress diurnal and semidiurnal tidal variability…”. I presume these two lead to similar results, but which one was used? Then in l. 290 you state that hourly wind data were daily-averaged. Finally, in l. 324 you state that all data were also high-passed filtered with a cutoff frequency of 250 days to focus “on the synoptic-to-intraseasonal variability…”. Please provide all filtering and smoothing information in a single subsection within Section 2 and then always clearly state whether each of the analyses is carried out on the low-pass filtered daily-averaged data, the original hourly records, or the low-pass and high-pass filtered data.
You describe how data gaps were dealt with. Did you check how sensitive the period durations are to the choice of the 7-day gap filling maximum? Some information could be added as supplementary material.
Parts of the text appear to be written for experts in the region, please consider readers unfamiliar with the region (see minor comments).
Please check the mooring location provided in the caption of Figure 1 and also in line 162, it does not seem to match the position indicated by the yellow circle in Figure 1c. Note that most of the paper is based on the analysis of data collected at this site.
You find a fairly tight relationship between monthly DO and potential density variability (Fig. 2). It seems very useful to explore if this holds for the daily data and also to present DO vs Θ, SA, and σ scatter plots. These may also provide somewhat more robust evidence on the near-bottom water masses, which the manuscript lacks in its present form.
Minor points
Conservative Temperature and Absolute Salinity should be capitalized throughout the manuscript
Line 15, two “essential” in this sentence, consider rewriting
Line 22, confusing sentence, consider rewriting: “Data from ERA5 reanalysis and from an oceanographic platform…”
Line 29, for clarity please replace “northern wind stress” by “northerly wind stress”
Line 47, delete bracket before Hannah
Lines 59 and 61, state latitudes for the “Patagonian fjord system” and the “southern coasts”
Line 67, Guafo region, refer the reader to Fig 1
Line 128, clarify that the range 31-33 g kg-1 corresponds to Absolute Salinity
Line 297-298, no need to repeat the latitude dependency of the Coriolis parameter in the same sentence
Line 313-316, you used a 50 km cross-shore length-scale to derive the vertical transport per unit coastline based on the estimated Ekman pumping velocity. Please discuss how does this length scale compare with the characteristic width of the downwelling regions shown in Fig 6.
Line 343, “Environmental variables were averaged within a ±10 d window relative to the pulse onset …”. I first found this very confusing, how would averaging the data within a ±10 d window help understand the dynamics a short-lived (2-4 days) pulse? Please consider rewriting so that the reader understands that this is not a time-average within each pulse but rather a time-average of each daily value across all pulses.
Line 373-374, “The seasonal cycle was evident, with a stronger amplitude (~3 °C at 170 m depth), as observed in the DO time series.” I don´t understand this statement, the seasonal amplitude is larger than what? And what has the DO time series to do with this?
Line 379-380, As you point out, the lack of phase alignment between Conservative Temperature and oxygen variability indicates that the oxygen changes are not solubility driven, note that this is also evident from the nearly perfect match between DO and %DO saturation.
Line 383, 385, “occurred…”, not “was occurred”
Line 393, replace “… longer moments …” by “… longer duration events …”
Line 451, “In terms of wind stress intensity, a northward maximum area was observed during austral summer …”. Since you are referring here to wind stress I strongly suggest you stick to the meteorological convention, from where the wind blows, so this would be “… a southerly maximum area …”. Likewise, in the following line I suggest replacing “… southward wind stress …” by “… northerly wind stress …”. Similar changes needed in the caption of Figure 4. Please check throughout the manuscript as this can lead to misinterpretation of the results.
Line 493, “… utilized to obtain the Ekman time series of the transport …” replace by “… utilized to obtain the time series of Ekman transport …”
Line 494, “The zonal Ekman transport (Uₑ), projected onto the cross-shore direction …” Do you really project the zonal Ekman transport? It is unclear what would the cross-shore direction be, depending on whether you use the mainland coast or the southern coast of Chiloe Island. Please elaborate and clarify whether that direction falls close to the main cross-shore flow pathway
Line 587, unclear what is meant by “unfiltered DO”. Are you referring to the low-pass (2-day cutoff) or the high-pass (250-day cutoff)?
Line 681, “… upper annual media” (?)
Line 692, the low salinity may be “a result of" rather than "favor" ventilation
Line 758, although pulses “can occur at almost any time of the year”, Fig 3a shows that out of 35 pulses only about 10% occurred during periods of very low DO.
Lin 786, according to Table 1, period 8 lasted 135 days and was therefore longer than period 1.
Figures
Figure 1: bottom topography is of fundamental importance. A brief description is provided in section 2.1. Yet, it is difficult to determine the depths of the channels and slopes. Please clearly indicate and label a few relevant isobaths in panel b. Also please make sure that all (numerous) geographic features mentioned in the text are marked in Figure 1.
There is no need to repeat the observed variables in the Figure caption. Please explain the meaning of the red circle.
Figure 3, What exactly is it meant by “High-resolution time series”? Are these unfiltered hourly data. If these are filtered daily averaged data I suggest just stating it that way. I strongly suggest labeling each period and pulse with their respective numbers. It is otherwise very hard to follow parts of the text.
Figure 4, caption (line 458) replace southward by northerly and northward by southerly
Figure 9a: Please use either µM or µmol/l throughout