the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 27 Jan 2025
Characterization of Northwest African Coastal Upwelling Systems
Abstract. Coastal upwellings are critical for nutrient supply, biological productivity, socio-economic activities, local weather patterns and regional climate variability. This study investigated characteristics of the Northwest African coastal upwelling system, including the Senegal-Mauritania and Gulf of Guinea coastal upwelling regions. The spatial and temporal variability were analysed using physical and biogeochemical ocean variables from 1982 to 2022. The analysis focused on upwelling indices derived from sea surface temperatures (SSTs), wind stress and chlorophyll-a concentration during the boreal winterspring and summer, when upwelling peaks in the Senegal-Mauritania and Gulf of Guinea upwelling systems, respectively. Additional indicators such as the annual cycles of sea level anomaly, ocean surface current velocity, geostrophic balance, salinity, nitrate and dissolved oxygen in surface waters were also examined as upwelling indicators. Upwelling trends were investigated by comparing the results of different datasets. Results highlight the complex interplay of local and large-scale processes that influence the dynamics of coastal upwelling in Northwest Africa. The variability in these coastal upwelling regions was not only influenced by local and remote wind forcing, but also by large-scale climatic drivers such as El Niño Southern Oscillations (ENSO) and Atlantic Niño events. These climatic factors influence the intensity and frequency and duration of upwelling seasons. In the Senegal-Mauritania coastal upwelling, a strong relationship was observed between upwelling and the equatorial ENSO events from late winter to the early spring, with a notable reduction of upwelling intensity after the peak of El Niño in the equatorial Pacific. In the Gulf of Guinea, the summer upwelling intensity was affected by the equatorial Atlantic mode and often decoupled from local wind forcing, providing new insights into alternative upwelling drivers.
- Preprint
(5594 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2024-4175', Anonymous Referee #1, 24 Feb 2025
Review of “Characterization of Northwest African coastal Upwelling Systems” by Yamaloula et at.
Summary
In this study the seasonal and interannual variability of upwelling in the Senegal-Mauritania and Gulf of Guinea regions are investigated. Using multiple datasets, the study analyzes different upwelling indices based on multiple variables like SST, wind, and chlorophyll concentration. The study analyzes spatiotemporal dynamics of these systems and their connections to large-scale climatic drivers.
The paper is well-written, presenting interesting analyses that have the potential to offer valuable insights into the Northwest African upwelling systems. However, some aspects of the analyses could be expanded and would benefit from a more thorough discussion, particularly in relation to existing studies. Additionally, I think the novelty of the paper’s findings is not always clear. A more explicit emphasis on what sets this work apart from previous studies would improve its overall contribution. Given these concerns, I recommend major revisions.
Major points
- In the study analyze different upwelling indices. I am missing a discussion about the advantages/disadvantages of using the respective indices. Why do we need different indices to characterize the upwelling? And what dynamics do the different indices represent? This would be also important to clarify in the context of the interannual variability.
- The study presents multiple datasets for the same variables, offering an interesting comparison. However, the discussion of differences between datasets is relatively brief. I suggest either expanding this discussion to better highlight the discrepancies and their implications or selecting the most representative dataset for the main text while moving the expanded dataset comparison and justification to the supplementary material. This would also allow more space for in-depth analysis elsewhere in the manuscript.
- In the analysis of the cumulative upwelling index (Figs. 9–10), the choice of regional separation is unclear to me. Previous studies have identified distinct subregions within these upwelling systems, such as the permanent upwelling in northern SMUS and seasonal upwelling in the south (Wooster et al., 1976; Mittelstaedt 1991; Lathuilière, 2008). Similarly, GGUS consists of two upwelling cells with different governing mechanisms (Wiafe and Nyadjro, 2015; Brandt et al., 2023). I recommend aligning the regional separation with these distinctions. Additionally, introducing these regional differences more thoroughly in the introduction would provide valuable context for the reader.
- In the composite maps (Figs. 11–12), the indication of significant correlations is missing. Without this, it is difficult to assess the reliability of the observed patterns. I recommend adding significance markers to improve interpretability.
- The discussion of interannual variability does not address potential underlying mechanisms. How do large-scale climate drivers influence Northwest African upwelling systems? While analyzing these mechanisms in detail may be beyond the study’s scope, referencing previous studies and discussing relevant processes for the teleconnections would provide valuable context. Additionally, Dakar Niños is not mentioned in the study but should be discussed or at least mentioned in my opinion.
- To strengthen the analysis of large-scale climate influences on upwelling variability, I suggest including time series of the upwelling indices alongside key interannual climate modes (ENSO, Atlantic Niños, Atlantic Meridional Mode, etc.). Lead-lag correlations between upwelling indices and climate indices could provide insights into potential drivers and their temporal relationships. These are just a few ideas you do not have to follow these exact ones. However, I do think including a more thorough view on the temporal evolution of the indices would give a better sense on the temporal variability. I think this could be a great base to deepen the understanding of large-scale influences on Northwest African coastal Upwelling Systems and show long term trends of these systems.
Minor points
- I recommend summarizing the different indices calculated in this study in a table. This would provide a clear overview of which data were used for each index and could also include the definitions of the respective variables for better clarity.
- Line 151: What is meant by drained by local wind? Do you mean driven by wind stress curl?
- Line 170: Maybe I missed it but how do you calculate SSHal and SSHon?
- The contour lines in Fig. 2 are hardly visible. I would suggests improving their visibility.
- Line 215: What do you mean with downwelling? Isn’t it rather a relaxation period?
- Line 232: What is meant with “This highest SST drive the weakening of the surface wind stress pattern”? Do you mean that the SST influences the wind pattern? Can you clarify this? Other studies discuss an influence of advection of warm water by the NECC (Faye et al., 2015). This could be mentioned while discussion this.
- 5a-b & Fig 8. a-b: I do not really understand what is presented here. What surface velocity is this? Is it the alongshore or the cross-shore velocity? And how does the surface velocities contribute to upwelling? There are also studies who describe that the poleward undercurrent or also called the Mauritanian Current surfaces during the relaxation part (e.g. Klenz et al. 2018). How does that relate to your analyses of surface velocities?
- In Fig. 8 a & b there is no geostrophic velocities at 0E. What is the reason for that? And what is the reason for the reversal in the sign of the geostrophic velocities east and west of 0E?
- Line 284: How does the SLA influence the productivity? Are you talking about the effect of CTWs? That section might benefit from some clarification.
- Related to the last point: The study primarily discusses coastal-trapped waves (CTWs) in terms of their role in geostrophic transport. However, CTWs can influence ecosystems through both advection and the deformation of the density field. A brief mention of these mechanism would be great.
- In Fig. 5 and Fig. 6 the oxygen concentration in October is much lower than during the rest of the year. What is the reason for this uniform signal in both Figures?
- Line 340: It is mentioned that the peak of sea surface salinity is driven by upwelling events. How would you explain this connection?
- In the Figure caption of Fig. 9-10 the black arrow and the magenta cross is not explained. However, a red vector is mentioned that I do not see in the figures. I would suggest clarifying the figure captions.
- Line 383: Do you have references for the trend being connected to anthropogenic climate change? There have been several studies who show weakened variability in the equatorial Atlantic (Prigent et al. 2020). Might this be connected to the reduction in the GGUS upwelling?
References
Brandt, Peter, et al. "Physical processes and biological productivity in the upwelling regions of the tropical Atlantic." Ocean science 19.3 (2023): 581-601.
Faye, Saliou, et al. "A model study of the seasonality of sea surface temperature and circulation in the Atlantic North-eastern Tropical Upwelling System." Frontiers in Physics 3 (2015): 76.
Klenz, Thilo, Marcus Dengler, and Peter Brandt. "Seasonal variability of the Mauritania Current and hydrography at 18 N." Journal of Geophysical Research: Oceans 123.11 (2018): 8122-8137.
Lathuiliere, Cyril, Vincent Echevin, and Marina Lévy. "Seasonal and intraseasonal surface chlorophyll‐a variability along the northwest African coast." Journal of Geophysical Research: Oceans 113.C5 (2008).
Mittelstaedt, Ekkehard. "The ocean boundary along the northwest African coast: Circulation and oceanographic properties at the sea surface." Progress in Oceanography 26.4 (1991): 307-355.
Prigent, Arthur, et al. "Weakened SST variability in the tropical Atlantic Ocean since 2000." Climate Dynamics 54.5 (2020): 2731-2744.
Wiafe, George, and Ebenezer S. Nyadjro. "Satellite observations of upwelling in the Gulf of Guinea." IEEE Geoscience and Remote Sensing Letters 12.5 (2015): 1066-1070.
Wooster, Warren S., Andrew Bakun, and Douglas R. McLain. "The seasonal upwelling cycle along the eastern boundary of the North Atlantic." (1976).
Citation: https://doi.org/10.5194/egusphere-2024-4175-RC1 -
AC1: 'Reply on RC1', Dametoti Yamoula, 18 Apr 2025
Dear Reviewer,
I understand the Reviewer’ concerns and the majors and minor points including suggestions and recommendations. The manuscript has been revised according the Reviewer’s comments, and some questions are directly answered here. But those which required complementary analyses were discussed in the attached pdf file.
This manuscript characterizes the seasonal and interannual variability of coastal upwelling in the Northwest Africa by highlighting the dynamics and contribution of each potential coastal upwelling factor using different upwelling indices. This novelty which was not seem clear in the manuscript has been taken into account in this new version. As recommended, some aspects of the manuscript were expanded and discussed in relation to existing studies.
Major points
This study analyses different upwelling indices. In the revised version of the manuscript (attached in pdf file), the discussion of the advantages and disadvantages of each upwelling index has been added in the method (section 2.2) as requested.
Q1: “Why do we need different indices to characterize upwelling?”
R1: we need different upwelling indices because studies in the eastern Atlantic boundary upwelling systems have shown that coastal upwelling is a complex system that only the wind stress cannot explain. Several mechanisms and factors, including wind stress, eastern propagation of Kelvin waves, coastal trapped waves, eddies, capes effects, geostrophic transport and remote forcing (ENSO) were proposed to understand the coastal upwelling process in this region. Based on these previous studies, in this manuscript uses different upwelling indices to characterize the dynamics and the contribution of each potential coastal upwelling factor to the seasonal and interannual variability of coastal upwelling along the Northwest African coast.
Q2: “What dynamics do the different indices represent?”
R2: -Ekman transport and wind stress curl are used to analyse the wind stress dynamics (main driver of upwelling) over the coastal and offshore region of the upwelling system. Ekman transport provides estimates of the offshore transport water, offering a direct measure of upwelling strength. While the wind stress curl shows the coastal divergence driven by the wind stress. However, these two upwelling indices only show the dynamic of upwelling process related to the wind forcing which is not often correlated to SST anomalies and the intensity of upwelling.
-The SST based upwelling index represents the SST gradient dynamics (ocean response) and measure the intensity of upwelling based on SST anomalies in a given area. However, this upwelling index does not consider the other internal and external factors likely to influence upwelling variability.
-The cumulative upwelling index (CUI) obtained integrating Ekman transport or SST gradient, represents the annual amount of upwelling.
-Total upwelling magnitude (TUM) represent the amount of upwelling from the start to the end dates.
-Upwelling index based on the sea level anomaly (SLA) pressure gradient or sea surface height represents the dynamics of onshore geostrophic flow and coastal trapped waves generated by Kelvin and Rossby waves, their influence on coastal divergence and upwelling.
The context of the interannual variability was clarified and discussed in relation with large-scale drivers and the previous studies (see section 4.3 and discussion).
2-This study uses multiple datasets for the same variables. As suggested, we expanded and improved the discussion of our results, highlighting the discrepancies implications of the datasets (see results and discussion section in the attached pdf file: revised_manuscript).
3- As recommended, the cumulative upwelling index (CUI) was recalculated considering the permanent (21-25N) and seasonal (10-21ºN) upwelling regions previously defined in the SMUS (Wooster et al., 1976; Mittelstaedt, 1991; Lathuiliere et al., 2008; Benazouzz et al., 2014; ) and the two upwelling cells observed in the East of Cape Plamas (7-2ºW) and East of Cape Three Points (2W-2ºE) in the GGUS (Djakoure et al., 2014; Wiafe and Nyadjro., 2015; Brandt et al., 2023). This regional separation was introduced in the Introduction section based on the corresponding previous studies.
4- As recommended, the markers (black dots) were added on the Fig. 12 (current Fig 11) to indicate the significant correlation between Nino 3.4 index and upwelling favourable wind stress anomalies during the during the extreme Nino events.
5- The discussion of interannual variability has been improved referencing previous studies, including the potential underlying mechanisms and the influence of the large-scale climate drivers and regional climate variabilities such as Dakar Niños/Niñas, Atlantic Meridional Mode.
Q3: “How large-scale climate drivers influence the Northwest African coastal upwelling?”
R3: Our analyses show that the large-scale climate modes such as ENSO and Atlantic Niño/ Niña events modulate the variability of the Northwest African coastal upwellings by reducing or supressing upwelling during the El Niño and Atlantic Niño phases or by enhancing the upwelling favourable conditions (strengthening of upwelling favourable wind stress, and SST cooling) during the La Niña and Atlantic Niña phases. More details of this result will be found in section 4.3 of the revised manuscript.
6.The discussion of interannual variability has been expanded on other climate mode of variability: lag and lead correlation analyses between Northwest African coastal upwelling indices (Ekman transport, SST gradient) and climate indices, such as AMM, AMO, NAO, NOI, ONI, PDO, PNA, NTA and SOI have been used to stablish the connections and teleconnections between the events (see section 4.3, Tables S3, 4, 5 supplementary material). These correlation analyses have allowed to identify the potential indices of coastal upwelling prediction over Northwest Africa.
Minor points
2.Line 151: The local wind means wind stress curl along the coast
3. Line 170: The geostrophic transport was computed in the direction parallel to the coast in each upwelling system. SSHal and SSHon correspond to SSH pressure gradient in the offshore and coastal boxes defined in each upwelling system. We first computed the SSH pressure gradient in the longitudinal and latitudinal direction. Then, the components of geostrophic current used to compute the geostrophic transport was computed in the direction perpendicular to the coast (see Table S2, supplementary material)
4.The visibility of the contour lines is improved on Fig. 2
5. Line 215: The downwelling means the relaxation or non-upwelling period
6. Line 232: Yes, we mean that the SST influence wind patterns. The seasonal variability of upwelling has clearly shown that the upwelling seasons is characterized by strong wind stress with low SST while the relaxation phases are marked by the weakening of wind stress, reducing Ekman transport. The analysis of SST and wind stress anomalies during the ENSO and ATL3 vents have also show these interactions between SST and wind stress anomalies and their associated upwelling indices. The (Fig. 2) shown that the wind-driven upwelling index decrease when the SST anomalies increase and vis versa. For example, in the southern part of SMUS (10º-14ºN), the weakening of upwelling-favourable wind stress (equatorward wind) is associated to the advection of warm water around the Guinea Dome. This advection of warm water drives by the poleward migration of North Equatorial Counter-Current (NECC) (Faye et al., 2015) (see manuscript revised, line 351)
7. The surface velocity is a mistake. It is the ocean surface current velocity. Our Ocean surface velocity plays a crucial role in modulating the interactions coastal divergence and offshore Ekman transport. In this study, the analysis has shown the upwelling seasons are characterized by the shoaling of ocean while the relaxation periods are characterized by the surfacing of ocean geostrophic current with rich-nutrient water, such the Poleward current in the SMUS and Guinea current in the Gulf of Guinea. This result is consistent with the findings of Klenz et al., (2008) and Kounta et al., (2018). More details of this result can be found (lines 348, 670) that the surfacing of ocean curren current shows that strong ocean velocity is likely to enhance SST cooling and the vertical mixing by driving warm water current under or above the cold-water current.
8. The absence of onshore geostrophic transport around 0º E in the GGUS was related to the quality of the dataset used. This analysis was improved in the revised manuscript (see Fig. 7). This weakening or absence of the geostrophic (ageostrophic) and reversal signs can be by several oceanographical factors.
9. Line 284: Yes, we are talking about the effects of CTWs: The influence of eastward propagation of Kelvin waves transform to coastal trapped waves was analysed by analysed by using sea level anomaly (SLA) pressure gradient. The results have shown strong effects of coastal waves drive by the wind stress, in particular in the GGUS. The combination of both coastal waves and local wind stress forcing was suggested to be the primary driver of coastal upwelling in this region.
10. In both upwelling regions, the oxygen concentration in October is much lower than during the rest of the year. This lower oxygen can be explained by several oceanographic and atmospheric processes. The two upwelling systems show October as a transition month between upwelling and non-upwelling period (in the GGUS) and between non-upwelling and upwelling season (in the SMUS) that bring deeper, oxygen-poor water to the surface. It corresponds also the maximum warm water period that enhance stratification, vertical mixing and limiting the replenishment of oxygen in surface waters from deeper layers or atmospheric exchange. In addition, this transition period can be accompanied by strong consumption of oxygen via photosynthesis and chlorophyll-a concentration.
11. The peak in surface salinity typically aligns with the strongest upwelling months. This connection can be explained by considering the physical and oceanographic processes associated with upwelling. The deep ocean waters that rise to the surface are typical stable and more saline than the surface layer often diluted by precipitation and river runoff. In the SMUS where upwelling come from the intermediate water masses, the increase in surface salinity can influence by the southward water masses.
12.Line 383: Yes, some recent analyses have related this decrease trend observed in the EBUSs (Bograd et al., 2023, Brandt et al., 2024), in particular the in the eastern Atlantic boundary to the changes in atmospheric circulation and upwelling drivers due to global warming and climate change. This decrease of the interannual SST variability was related to the substantial weakening of the correlation equatorial zonal wind stress and the role play by the meridional warming in the west Africa coast (Prigent et al., (2020).
-
RC2: 'Comment on egusphere-2024-4175', Anonymous Referee #2, 24 Mar 2025
This paper compares various indices of coastal upwelling in two different regions of the eastern Atlantic, the Senegalese-Mauritano Upwelling System (SMUS) and the Gulf of Guinea Upwelling System (GGUS). Indices are based on wind stress (coastal divergence and curl) and SST gradient. Other indices are based on surface currents and sea level. These indices are computed from two ocean reanalyses (GLORYS, ORAS5), an atmospheric reanalysis (ERA5), and two SST products with different spatial resolution (OSTIA, Hadl). The authors first study the seasonal variations of upwelling and the latitudinal changes in the two regions. In the second part of the paper they describe the interannual variations using cumulative indices, composite and correlation analyses.
I found the purpose of the paper interesting and the quality of the figures rather good. However the writing is often poor and there are many unclear sections. First, the authors do not seem to fully understand what they are comparing. The wind stress from an ocean reanalysis results from a bulk formulae using a atmospheric forcing (possibly an atmospheric reanalysis) and the ocean model SST as inputs. The ERA5 wind stress results from bulk formulae (possibly different than that used in the ocean reanalysis) using ERA5 atmospheric forcing and observed SST as input. The authors do not explain why they compare these products and what to conclude from the comparison is unclear. Another issue is the use of sea level in the analysis. The authors seem to have understood that cross-shore geostrophic transport impacts coastal upwelling, but they do not compute this index in the paper. Instead they use sea level anomaly, “ocean surface current” (it is unclear if the latter refers to current speed or to a particular component) and a unknown “geostrophic balance” index in the analysis. This makes the paper rather difficult to follow. In the study of interannual variability, they misinterpret their own figures, as El Nino events seem to enhance southward wind off the SMUS but tend to produce warm SST anomalies. Last the discussion of the results is poor, it merely summerizes the findings and does not provide any discussion of the mecanisms at stake to explain some of the differences that have been identified.
To conclude, I also have had to write a very long list of specific minor and not-so-minor comments that the authors need to address. As there is a lot of work to be done to improve the paper, my decision is reject. I encourage the authors to revise their work thoroughly before considering resubmission.
Specific comments:
L18 : « remote wind forcing » can not influence the region. It can influence the ocean remotely and then ocean waves may propagate to the region.
L38 : Bograd et al., 2022 would be a more suited reference here.
Bograd, et al. (2023). Climate change impacts on eastern boundary upwelling systems. Annual Review of Marine Science, 15(1), 303-328.
L46 : using a regional ocean model
L47 : coastline geometry
L49 : Recent upwelling indices take into account the cross-shore geostrophic current which can enhance or reduce coastal upwelling (see Jacox,et al. (2018). Coastal upwelling revisited: Ekman, Bakun, and improved upwelling indices for the U.S. West Coast. Journal of Geophysical Research: Oceans, 123, 7332–7350. https://doi.org/ 10.1029/2018JC014187). This needs to be accounted for in this study.
L52 : these are physical, not biophysical processes
L53 you mean « eastward propagation of equatorial Kelvin waves » and « poleward propagation of coastal trapped waves » ?
L61 : what do you mean by « coastal variability » ?
L66 : where does this influence take place more specifically ?
L68 : Using a regional model
L69 : « .. highlighted the Guinea current » : not clear what is meant here. Can you elaborate a little ?
L73-81 : I find this paragraph not really useful as the paper focuses on the past period. It could be useful if Bakun’s alleged dynamical processes driving upwelling variability were mentioned but it is not the case.
L86 : what remote forcing ? You mean equatorial Kelvin waves ? This is unclear.
L88-89 : the goals of the study need to be described more precisely in the introduction.
L94 : Refer to the spatial resolutions of ORAS5 and GLORYS in the Table here.
L105 : Should be Table « 1 ». GLORYS does not deliver oxygen and nitrate, there must be another product with another spatial resolution.
L127 : « corresponds also to the ... »
L129 : UI needs to be defined.
L130 : not clear what you mean here.
L153 : not clear what you mean by « drained by the local wind » : if there is no coast, there is no divergence of the Ekman current and no upwelling. Do you mean « by the wind stress curl » ?
L168 : you must mention this process before (see my comment line 49).
L174 : I do not understand the definition of Tg : to compute the onshore geostrophic transport, you must compute the gradient of sea level in the direction parallel to the coast. What are SSHal and SSHon ?
L176 : nitrate
L179 : before CUI can be defined, a clear definition of UI is needed somewhere.
L182-188 : a schematic is needed to explain how the indices are defined.
L190 : it should be an integral, not a sum ? What is the temporal resolution ?
L194 : « based on the gradient descent fit » , « Pettitt statistical test »: references are needed
L196 : a word is missing here.
L201 : analysing
Figure 2 : what are ORAS5 and GLORYS wind stress fields ? ORAS5 and GLORYS are ocean reanalyses, meaning that wind products or reanalyses were used to compuate wind stress and force the ocean surface. These winds are possibly from era5, jra-55 or another reanalysis. This should be clearly explained in the methodology section.
I am puzzled by Figure 2 because coastal upwelling is negative (upward) in Fig 2a-c .However you mention earlier that Tg is negative eastward, which can not be the case because an eastward (negative) Tg produces a (negative) downwelling at the coast. I think that upwelling should always be positive as a convention. I also find it very confusing to have a different sign convention for coastal upwelling and wind stress curl upwelling. This needs to be changed.
Last, it would be nice to relax the saturation of the colorbar for positive values to see if there is coastal downwelling during some time periods.
L214 : the start of the upwelling seasons depends on latitude, so what is the latitude here ?
L214-216 : this is not easy to see this in Figure 2 and 3. Another figure should be shown.
L225 : What about the geostrophic transport Tg ? It does not seem to be taken into account in this analysis. It should be added (or at least compared) to the zonal Ekman transport (with the correct sign).
L127 : why is the maximum SST gradient observed before the maximum Ekman transport ?
L232 : « the highest SST drive... » : unclear, I do not understand what you mean ? The SST forces the wind patterns ? This should be demonstrated.
Figure 3 : several issued with this figure : the « THETAO » (e) label is awkward : GLORYS is an ocean reanalysis, same as ORAS5, so the title should be GLORYS ; in a-e, the color scale indicates SST where ? In the coastal box ?; the definition of the green line is strange : you mean during the month of minimum SST ? « Minimum SST » where ?; I do not see the point of the orange line and its link with upwelling as it lies south of the upwelling region.
L242-243 : I don’t see how you reach that conclusion as these are 3 different estimates of wind stress and we do not know which one is closest to reality.
L244 : « the good SST index... » : OSTIA and Hadl provide the observed SST indexes, not the good ones. How is the Hadl index poor ? You can just hypothesize that the OSTIA index is better because this product is more precise near the coast.
L246 : this is getting more and more confusing : ERA5 is not an SST dataset, it is an atmospheric reanalysis which is forced by an SST product (possibly OSTIA , to be checked). This needs to be clarified. Apparently the authors do not known well enough what they are comparing.
L247 : what are the wind forcing used to force ORAS5 and GLORYS ? This needs to be added in the table.
L248-249 : « The maximum magnitude.. » of what ?
Figure 4 : why didn’t you sum Ekman divergence (Mx), wind stress curl upwelling and Tg ? Just looking at Ekman transport does not make sense. What is SLA ? Is it Tg ? This is getting more and more confusing.
L260 : « geostrophic ... » : which figure are you refering to ? In Figure 4 SLA is shown, not geostrophic current. This is difficult to follow.
L262 : « positive WSCD generally enhance upwelling » : no, they contribute positively to upwelling, which is the sum of coastal upwelling (divergence of Ekman transport) and wind stress curl upwelling which occurs offshore. The rest of the sentence («negative WSCD ..suppress coastal upwelling ») should therefore be clarified.
Figure 5 : Legend is unclear : what is « ocean velocity » ? zonal, meridional, total ? What is « geostrophic balance” ? You mean geostrophic transport ? It is not the ocean velocity that matters for coastal upwelling but the surface cross-shore geostrophic velocity. Apparently the authors used ORAS5 and GLORYS surface velocity, which is useless.
What are the isolines in figure (e) ?
L273-274 : these two lines are meaningless.
L274 : what is the « Rossby circulation » ?
L284 : correlation does not indicate causality : how would SLA influence productivity ? I do not believe it is the case, or it needs to be better explained.
L295 : you mean « meridional SST gradient » ?
L295 : « the upwelling waters were estimated... » : not clear what this means. Rephrase. What is the justification for this SST threshold?
L300 : « integrating...” in time and space? Otherwise, what is the longitude refered to here? This figure is not shown.
Figure 7: repetition of text in legend.
L302-303: this is the maximum magnitude of the integrated transport or the maximum in time and space?
Line 306 and Fig.8: as for Fig. 5, I do not understand what is plotted when geostrophic balance is mentioned. Moreover, surface velocity is different from meridional geostrophic velocity, which may impact the GGUS coastal upwelling. It is impossible to interpret Figure 8a,b.
Line 306: “ocean velocity (contours)”
L320: Fig. 7 must be Fig.6 here. Keep in mind that ERA5 is not an ocean reanalysis, its wind stress is computed using a particular bulk formulae and a particular SST product (possibly OSTIA, to be verified). In the case of ocean reanalyses, wind stress is computed using using bulk formulae and the ocean model SST. Thus the wind stress differences in the different products may originate from the SST products and differences in bulk formulae. This should be discussed.
L321: I do not really see the point in using Hadl product since it is obvious that this product is not adapted for the upwelling index due to its coarse spatial resolution. It could be suppressed from the study to shorten the paper.
L327: “This indicates ...”: I strongly disagree. Correlation is not causality.
L328: I do not see two cells in Fig.7f-j., just one.
L330: What do you mean by “the western cell is linked to the SLA”: what is the mechanism? Does the geostrophic current driven by a SSH gradient play a role?
L330-332: Citing papers is not sufficient, how Kelvin waves and CTW drive this upwelling needs to be explained.
Figure 8: it would help to represent Figure 8 with the horizontal axis for longitude and vertical axis for time.
L347: “the maximum concentration of Chlorophyll ...”: not at all. Oxygen concentration does not drive chlorophyll concentration. I think the figures and text about the Chl, Nitrate, Oxygen and surface salinity should be suppressed, they do not bring any useful information to the paper.
Figure 9: which wind stress and SST product are used here?
L370-372: what was used for the upwelling indices in the linear regression? The cumulative index? The maximum for each year? This lacks details and makes it very difficult to follow.
L382-383: “This decreasing trend...”: you can not prove that this is linked to anthropogenic climate change with this analysis, this could also be decadal variability.
L400: A table should be added to list the climate events used to compute the composite with time periods, SST anomalies in the Pacific, Atlantic, SMUS and GGUS. The SST product used to compute the composite is not indicated.
Figure 11: I do not see the point in showing the global ocean in this figure. You could show only the Atlantic and Pacific ocean between 30°S-30°N, thus it would be easier to visualize the SST and wind stress anomalies in the SMUS and in the GGUS (it is impossible to see anything about the GGUS in the present figure).
Figure 11: how do you know that the composite is significant? A statistical test is needed.
L402-403: “extremely warm SST anomalies”: 0.6°C is not extreme in my opinion. I think the arrows indicate that coastal upwelling in the south part of (entire) SMUS is enhanced in DJF (MAM) , but the SST increases. Coastal upwelling does not explain this SST increase, apparently.
L405: References are needed here, since the authors have not shown any of this.
L407-408: I can not see the positive anomalies in the Gulf of Guinea druing La Nina events.
Figure 12: same comment as for Figure 11: only the regions of interest (Equatorial Pacific, tropical Atlantic) should be shown.
L425: how was the correlation computed? What was the time series used? Over which time period?
L430: I think the authors misinterpreted the result of the correlation: the correlation between El Nino index and meridional wind stress off the SMUS is negative, meaning that when the El Nino index is high (strong El Nino conditions), the meridional wind stress anomaly is negative (southward) thus enhances coastal upwelling in the SMUS. This can also be seen in Figs 11a,c, wind stress anomalies appear to be southward in SMUS.
L452: how mesoscale variability impacts upwelling needs to be explained in the discussion.
L453: “weak upwelling trend..”: positive or negative?
L455: I think this needs to be refined, the trends are different in the north and south of the SMUS and this line is not consistent with Sylla et al. (2019) findings.
L470: I disagree with the statement. The wind-driven upwelling increases and the SST is warmer, meaning that another process explains the surface warming.
Citation: https://doi.org/10.5194/egusphere-2024-4175-RC2 -
AC2: 'Reply on RC2', Dametoti Yamoula, 20 Apr 2025
Dear Reviewer,
I understand the Reviewer’ concerns and the majors and minor points including suggestions and recommendations. The manuscript has been revised according the Reviewer’s comments, and some questions are directly answered here. But those which required complementary analyses were discussed in the attached in the revised manuscript.
Some aspects of the manuscript were expanded and discussed in relation to existing studies.
Major and minor corrections
L18: We agree with you, it was a mistake. The « remote wind forcing » leads to eastward propagation of Kelvin waves, which are transformed into coastal trap waves when they reach the West African coast modulating coastal upwelling. Thank you very much for the correction.
L38: Bograd et al., 2023 was used as one of the appropriate references for the impacts of climate change on the eastern boundary upwelling systems.
L46 : using a regional ocean model (corrected)
L47 : coastline geometry (corrected)
L49 : This suggestion (Jacox,et al. 2018) was taken into account in the revised manuscript. The influence cross-shore geostrophic current was discussed in the results of the revised manuscript. The seasonality of onshore geostrophic transport was computed based on the sea level pressure gradient, and used to analyse the relative contribution of onshore geostrophic flow in each upwelling system.
L52 : these are physical instead of biophysical processes (corrected)
L53 Yes, we mean « eastward propagation of equatorial Kelvin waves » and « poleward propagation of coastal trapped waves » (corrected)
L61 : By « coastal variability » we mean the coastal SST variability
L66 : This influence take place more specifically under the capes
L68 : Using a regional model (corrected)
L73-81 : The paragraph was removed in the new version
L86 : The remote forcing here mean ENSO and other large-scales climate events which are not occur near to the Northwest African coastal upwelling regions.
L88-89 : the goal of the manuscript was clear describe in the introduction of the revised manuscript (Line 95).
L94 : Refer to the spatial resolution of each dataset was added in the Table S1 (supplementary material).
L105 : My apologies, the oxygen and nitrate data were provided by the global Biogeochemistry Analysis and Forecasts from the Copernicus Marine Environmental Monitoring Services (CMEMS)
L127 : « corresponds also to the ... » (corrected)
L129 : Upwelling index (UI) was defined in the method before using as the abbreviation.
L153 : not clear what you mean by « drained by the local wind » : if there is no coast, there is no divergence of the Ekman current and no upwelling. Do you mean « by the wind stress curl »?
L168 : you must mention this process before (see my comment line 49).
L174 : The onshore geostrophic transport (Tg) was computed using the sea level pressure gradients of the direction parallel to the coast. This sea level pressure gradient was derived from sea level anomaly (SLA) and sea surface height (SSH). We used both SLA and SSH for the comparison because the SSH include SLA and absolute dynamics of the topography (geoid, tidal). and are the alongshore and onshore pressure gradients, respectively. After compute the SSH gradient at the longitudinal and latitudinal directions, the appropriate SSH pressure gradient were selected in the corresponding upwelling system to compute the onshore geostrophic transport. Here, we used the difference of sea level pressure between offshore () and coastal () boxes to compute the onshore geostrophic transport (see method section and attached supplementary material for more details).
L176 : nitrate (corrected)
L179 : The cumulative upwelling index (CUI) and upwelling index (UI) were clearly defined in the method section .
L182-188 : Table S2 (supplementary material) summarizes the different upwelling indices used in this study.
L190 : Integrating (corrected). What is the temporal resolution is daily
L194 : « based on the gradient descent fit » , « Pettitt statistical test »: references are needed
L196 : a word is missing here (added): the sentence was reformulated
L201 : analysing (corrected)
Figure 2 : what are ORAS5 and GLORYS wind stress fields are ocean reanalyses forced by the Era interim products, the spatial resolution, levels and more characteristics of this data were added in the Table S2 (supplementary material).
Here I don’t well understand your worry but I think my revised manuscript can respond this. The update Figures present negative Ekman transport (Fig 2, blue colour shading) and onshore geostrophic flow (Fig. 4) and positive coastal wind stress curl (Fig. 2, contours) and SST gradient (Fig. 3, contours) during the upwelling season in both upwelling systems (SMUS, GGUS). The opposite sign of these indices has been observed during the relaxation or non-upwelling periods. What I mentioned earlier was to explain that the negative onshore geostrophic transport enhances the wind stress curl (positive). You are right upwelling should always be positive (positive wind stress curl), but negative onshore geostrophic transport (Fig. 4). This sign convention has been adjusted. Here, Ekman transport is driven by meridional wind stress (negative sign). The upwelling based on the Ekman transport will have a positive while the relaxation period will be positive sign.
We also take into account your suggestion to improve the visibility of the colorbar for positive values to see if there is coastal downwelling during some time periods.
L214 : Although the upwelling timing varies over the latitudes, the start and end dates of upwelling presented here is the average of the seasonal upwelling region of SMUS (10 – 20oN).
L214-216 : The starts and end dates were clearly shown on Figures 8 and 9 in the revised manuscript
L225 : The geostrophic transport (Tg) was taken into account by not seem to be taken into account in comparing to the zonal Ekman transport.
L227 : The earlier peak in SST gradient relative to Ekman transport in the SMUS can be explain by the interplay between atmospheric forcing, rapid oceanic response times, and preconditioning of shallow thermocline dynamics.
L232 : « the highest SST drive... » : we mean that the SST influence wind patterns. The seasonal analysis upwelling indices derived from SST and wind stress fields (Ekman transport) clearly shows that wind-driven upwelling decrease when the SST anomalies increase and vis-versa. The upwelling season is characterized by strong wind stress while the relaxation period is characterized by a weakening of wind stress associated with strong SST anomalies. The composite analysis of SST and wind stress anomalies during the extreme large-scale climate events have also shown that SST anomalies modulate the wind stress patterns (Section 4.3).
Figure 3 : was update taken to account all the remarks, such the label name, the definition of green (minimum temperature observed before the beginning of upwelling season in the seasonal upwelling region), the colour scale and the orange line (25 oC isotherm: Threshold of SST used to estimate the southern margin of upwelling region) lying south of the upwelling region.
L242-243: I don’t see how you reach that conclusion as these are 3 different estimates of wind stress and we do not know which one is closest to reality.
L244 : Your suggestion was taken into account here. OSTIA and Hadl are the reconstituted observational SST data but at different spatial resolution (0.05o and 1o , respectively). However, the best representation of upwelling index in this coastal region was OSTIA, because this product is more precise near the coast.
L246 : ERA5 is atmospheric reanalysis that includes SST and wind stress products used in this study. It has been forced by Era interim, OSTIA and HadI products. This information clearly appears in Table S1 (supplementary material). This needs to be clarified.
L247 : ORAS5 was forcing with Era interim HadI and OSTIA products while GLORYS was forcing Era-interim and real-time forecast data (see Table S1).
L248-249 : « The maximum magnitude.. » of upwelling-favourable
Figure 4 : This Figure and it analysis is removed in the revised version to avoid the confusion. It was showing the seasonal cycle of sea level anomaly (SLA) compared to upwelling indices derived from wind stress (Ekman transport) and SST in the coastal box.
L260 : this analysis based on Figure 4 has been removed to make it easier to understand.
L262 : this analysis based on Figure 4 has been removed to make it easier to understand.
Figure 5 : This Figure was updated (Figure 4 in the revised manuscript) taken to account all the suggestions. Only the onshore geostrophic transport of the coastal and offshore boxes was compared with ORAS5 and GLORYS datasets.
What are the isolines? The isolines represented the onshore geostrophic magnitude
L273-274 : these two lines were removed as suggested.
L274 : mistake, it is not Rossby circulation, but Rossby waves caused by the rotation of the Earth and the des effects of the latitudinal variation of the Coriolis force. The writing is adjusted
L284 : the correlation does not indicate causality. But influence of SLA on the upwelling productivity was estimated comparing the seasonal evolution of SLA and chlorophyll-a concentration. The upwelling season was characterized by negative SLA over the coastal box indicating strong influence of sea level pressure gradient on the coastal SST variability, while the relaxation periods were characterized by positive SLA indicating the supressing of upwelling season. This seasonal variability of SLA is likely influence chlorophyll-a productivity. The analysis based on Figure was removed as suggested, to avoid the confusion.
L295 : Yes, the « meridional SST gradient » was computed in the GGUS
L295 : This sentence was rephrase: The upwelling region were estimated using 25oC isotherm. The SST threshold corresponds to the reference temperature used in the equatorial region to estimate cold water anomalies (Caniaux et al., 2011). We mentioned it in the methodology section.
L300 : The upwelling indices used to compute the cumulative upwelling index are time series computed averaging the upwelling boxes (offshore box for the Ekman transport, difference offshore - coastal box for SST gradient).
Figure 7: repetition of text in legend (was corrected).
L302-303: this is the maximum magnitude in time and space over the climatological annual cycle.
Line 306 and Fig.8: was showing the ocean surface current velocity with the onshore geostrophic transport in contours. However, this Fig. 8 was update and corresponds to Fig. 7 in the revised version of the manuscript. The Fig .7 in the revised manuscript clearly shows the seasonal cycle of the onshore geostrophic transport, and the contour represent the magnitude of this geostrophic transport
L320: Fig. 7 must be Fig.6 here: This confusion was taken into account in the revised version of the manuscript. The different wind stress products used to compute the Ekman transport was discussed based on their characteristics: nature of the product, spatial and temporal resolution, inputs products, … Era5 is atmospheric reanalysis product including different variables such as SST and wind stress components used in this study.
L321: Hadl product was conserved in this study despite it is obvious that this product is not adapted for the upwelling index due to its coarse spatial resolution., for the comparisons. But it can be removed in the last version of the manuscript to shorten the paper (as you suggested).
L327: This result was revised based on the geostrophic flow analysis.
L328: The two upwelling cells were observed on the seasonal cycle plots (Fig. 7A-e). It is unclear on the Fig.7f-j because of the size of the plots.
L330: “the western cell is linked to the SLA”: we mean that the onshore geostrophic flow and the coastal waves drive by the seas level pressure gradient influence more the coastal SST variability at the western of the GGUS. The mechanism is based on the onshore geostrophic flow and poleward propagation of coastal trapped waves.
L330-332: The influence of Kelvin waves and coastal waves was discussed in the new version of the manuscript (445).
Figure 8: was update with the onshore geostrophic transport
L347: “the maximum concentration of Chlorophyll was observed during the upwelling peak and coincides with the peak of dissolved oxygen. This linear correlation let suggest that oxygen concentration (useful for the photosynthesis) drives the maximum concentration observed in this upwelling region. However, requested, the plots of surface salinity, chlorophyll-a concentration, dissolved oxygen and nitrate were removed in the main text from the supplementary material.
Figure 9: as ORAS5 does not delivery the daily product, the ERA5 wind stress and OSTIA SST were used to compute the cumulative upwelling indices
L370-372: to analysis the long-term trend, the Ekman transport and SST-based upwelling index, both derived during the upwelling seasons (winter-spring in the SMUS and summer in the GGUS) were used for the linear regression analysis, not the cumulative upwelling or the maximum of each year. However, in the revised version, these historical trends were computed with the wind stress and SST anomalies during the upwelling seasons.
L382-383: we suggested the linked of this long-term trend with previous studies on the impacts on climate change in the eastern boundary upwelling systems (Bograd et al., 2023).
L400: As recommended, the Tables (Tables S3, 4, 5) of the correlation analyses between upwelling indices (Ekman transport and SST gradient) were added in the supplementary material.
Figure 11: As recommended, the Figs. 11 and 12 (new Figs. 10 and 11 in the revised manuscript) have been displayed between 30S-30N, to improve the visibility of the upwelling regions on the Figures.
Figure 11: In the new version of the manuscript, the significant correlation was displayed with MonteCarlo statistical test.
L402-403: This result has been re-discussed: El Nino phases (DJF, MAM) are characterized by the weakening of equatorward trade winds, the reinforcing the northward winds and the reducing and supressing of upwelling along the northwest African coast, especially in the SMUS. While La Nina phases are characterized by the strengthening of the equatorward winds and enhancing Ekman transport and upwelling conditions.
Atlantic Nino/Nino phases present the same behaviour in the GGUS with less significant impacts on the SMUS, in particular in the northern of this upwelling region
L405: These results have been re-discussed with references (based on the previous studies on the Dakar Nino/Nina, Atlantic Meridional mode, Atlantic Nino/Nina).
L407-408: This discussion of the result has been improved with the update Figs. 10, 11 in the new version of the manuscript (section 4.3).
L425: how was the correlation was computed between the composite indices of the extreme events of ENSO and Atlantic Nino mentioned in the methodology section, and the spatial distribution of upwelling-favourable wind stress (meridional wind stress in the SMUS during winter and spring, zonal wind stress in the GGUS during summer) during the same period.
L430: The discussion of correlation result between El Nino index and meridional wind stress off the SMUS has been improved and expanded based on the update Figs. 10-11 (section 4.3).
L452: the impacts of mesoscale variability on upwelling needs was discussed based on the previous studies.
L453: upwelling trend is different according the upwelling region (L522, revised manuscript)
L455: The long-term trend of upwelling was re-discussed in the revised manuscript (L522) based on each section of upwelling (permanent and seasonal in the SMUS, eastern of the capes in the GGUS): sea Fig. S4 (supplementary material).
L470: I this line was rephrased in the revised manuscript.
In summary, the manuscript was completely revised based on reviewers’ comments and suggestions. The last version and supplementary material will be sent to all reviewer for the checking.
Thank you very much for your interesting and useful comment and suggestions for this manuscript, I really appreciated and hope that my revised version will satisfy your comments
-
AC2: 'Reply on RC2', Dametoti Yamoula, 20 Apr 2025
Status: closed
-
RC1: 'Comment on egusphere-2024-4175', Anonymous Referee #1, 24 Feb 2025
Review of “Characterization of Northwest African coastal Upwelling Systems” by Yamaloula et at.
Summary
In this study the seasonal and interannual variability of upwelling in the Senegal-Mauritania and Gulf of Guinea regions are investigated. Using multiple datasets, the study analyzes different upwelling indices based on multiple variables like SST, wind, and chlorophyll concentration. The study analyzes spatiotemporal dynamics of these systems and their connections to large-scale climatic drivers.
The paper is well-written, presenting interesting analyses that have the potential to offer valuable insights into the Northwest African upwelling systems. However, some aspects of the analyses could be expanded and would benefit from a more thorough discussion, particularly in relation to existing studies. Additionally, I think the novelty of the paper’s findings is not always clear. A more explicit emphasis on what sets this work apart from previous studies would improve its overall contribution. Given these concerns, I recommend major revisions.
Major points
- In the study analyze different upwelling indices. I am missing a discussion about the advantages/disadvantages of using the respective indices. Why do we need different indices to characterize the upwelling? And what dynamics do the different indices represent? This would be also important to clarify in the context of the interannual variability.
- The study presents multiple datasets for the same variables, offering an interesting comparison. However, the discussion of differences between datasets is relatively brief. I suggest either expanding this discussion to better highlight the discrepancies and their implications or selecting the most representative dataset for the main text while moving the expanded dataset comparison and justification to the supplementary material. This would also allow more space for in-depth analysis elsewhere in the manuscript.
- In the analysis of the cumulative upwelling index (Figs. 9–10), the choice of regional separation is unclear to me. Previous studies have identified distinct subregions within these upwelling systems, such as the permanent upwelling in northern SMUS and seasonal upwelling in the south (Wooster et al., 1976; Mittelstaedt 1991; Lathuilière, 2008). Similarly, GGUS consists of two upwelling cells with different governing mechanisms (Wiafe and Nyadjro, 2015; Brandt et al., 2023). I recommend aligning the regional separation with these distinctions. Additionally, introducing these regional differences more thoroughly in the introduction would provide valuable context for the reader.
- In the composite maps (Figs. 11–12), the indication of significant correlations is missing. Without this, it is difficult to assess the reliability of the observed patterns. I recommend adding significance markers to improve interpretability.
- The discussion of interannual variability does not address potential underlying mechanisms. How do large-scale climate drivers influence Northwest African upwelling systems? While analyzing these mechanisms in detail may be beyond the study’s scope, referencing previous studies and discussing relevant processes for the teleconnections would provide valuable context. Additionally, Dakar Niños is not mentioned in the study but should be discussed or at least mentioned in my opinion.
- To strengthen the analysis of large-scale climate influences on upwelling variability, I suggest including time series of the upwelling indices alongside key interannual climate modes (ENSO, Atlantic Niños, Atlantic Meridional Mode, etc.). Lead-lag correlations between upwelling indices and climate indices could provide insights into potential drivers and their temporal relationships. These are just a few ideas you do not have to follow these exact ones. However, I do think including a more thorough view on the temporal evolution of the indices would give a better sense on the temporal variability. I think this could be a great base to deepen the understanding of large-scale influences on Northwest African coastal Upwelling Systems and show long term trends of these systems.
Minor points
- I recommend summarizing the different indices calculated in this study in a table. This would provide a clear overview of which data were used for each index and could also include the definitions of the respective variables for better clarity.
- Line 151: What is meant by drained by local wind? Do you mean driven by wind stress curl?
- Line 170: Maybe I missed it but how do you calculate SSHal and SSHon?
- The contour lines in Fig. 2 are hardly visible. I would suggests improving their visibility.
- Line 215: What do you mean with downwelling? Isn’t it rather a relaxation period?
- Line 232: What is meant with “This highest SST drive the weakening of the surface wind stress pattern”? Do you mean that the SST influences the wind pattern? Can you clarify this? Other studies discuss an influence of advection of warm water by the NECC (Faye et al., 2015). This could be mentioned while discussion this.
- 5a-b & Fig 8. a-b: I do not really understand what is presented here. What surface velocity is this? Is it the alongshore or the cross-shore velocity? And how does the surface velocities contribute to upwelling? There are also studies who describe that the poleward undercurrent or also called the Mauritanian Current surfaces during the relaxation part (e.g. Klenz et al. 2018). How does that relate to your analyses of surface velocities?
- In Fig. 8 a & b there is no geostrophic velocities at 0E. What is the reason for that? And what is the reason for the reversal in the sign of the geostrophic velocities east and west of 0E?
- Line 284: How does the SLA influence the productivity? Are you talking about the effect of CTWs? That section might benefit from some clarification.
- Related to the last point: The study primarily discusses coastal-trapped waves (CTWs) in terms of their role in geostrophic transport. However, CTWs can influence ecosystems through both advection and the deformation of the density field. A brief mention of these mechanism would be great.
- In Fig. 5 and Fig. 6 the oxygen concentration in October is much lower than during the rest of the year. What is the reason for this uniform signal in both Figures?
- Line 340: It is mentioned that the peak of sea surface salinity is driven by upwelling events. How would you explain this connection?
- In the Figure caption of Fig. 9-10 the black arrow and the magenta cross is not explained. However, a red vector is mentioned that I do not see in the figures. I would suggest clarifying the figure captions.
- Line 383: Do you have references for the trend being connected to anthropogenic climate change? There have been several studies who show weakened variability in the equatorial Atlantic (Prigent et al. 2020). Might this be connected to the reduction in the GGUS upwelling?
References
Brandt, Peter, et al. "Physical processes and biological productivity in the upwelling regions of the tropical Atlantic." Ocean science 19.3 (2023): 581-601.
Faye, Saliou, et al. "A model study of the seasonality of sea surface temperature and circulation in the Atlantic North-eastern Tropical Upwelling System." Frontiers in Physics 3 (2015): 76.
Klenz, Thilo, Marcus Dengler, and Peter Brandt. "Seasonal variability of the Mauritania Current and hydrography at 18 N." Journal of Geophysical Research: Oceans 123.11 (2018): 8122-8137.
Lathuiliere, Cyril, Vincent Echevin, and Marina Lévy. "Seasonal and intraseasonal surface chlorophyll‐a variability along the northwest African coast." Journal of Geophysical Research: Oceans 113.C5 (2008).
Mittelstaedt, Ekkehard. "The ocean boundary along the northwest African coast: Circulation and oceanographic properties at the sea surface." Progress in Oceanography 26.4 (1991): 307-355.
Prigent, Arthur, et al. "Weakened SST variability in the tropical Atlantic Ocean since 2000." Climate Dynamics 54.5 (2020): 2731-2744.
Wiafe, George, and Ebenezer S. Nyadjro. "Satellite observations of upwelling in the Gulf of Guinea." IEEE Geoscience and Remote Sensing Letters 12.5 (2015): 1066-1070.
Wooster, Warren S., Andrew Bakun, and Douglas R. McLain. "The seasonal upwelling cycle along the eastern boundary of the North Atlantic." (1976).
Citation: https://doi.org/10.5194/egusphere-2024-4175-RC1 -
AC1: 'Reply on RC1', Dametoti Yamoula, 18 Apr 2025
Dear Reviewer,
I understand the Reviewer’ concerns and the majors and minor points including suggestions and recommendations. The manuscript has been revised according the Reviewer’s comments, and some questions are directly answered here. But those which required complementary analyses were discussed in the attached pdf file.
This manuscript characterizes the seasonal and interannual variability of coastal upwelling in the Northwest Africa by highlighting the dynamics and contribution of each potential coastal upwelling factor using different upwelling indices. This novelty which was not seem clear in the manuscript has been taken into account in this new version. As recommended, some aspects of the manuscript were expanded and discussed in relation to existing studies.
Major points
This study analyses different upwelling indices. In the revised version of the manuscript (attached in pdf file), the discussion of the advantages and disadvantages of each upwelling index has been added in the method (section 2.2) as requested.
Q1: “Why do we need different indices to characterize upwelling?”
R1: we need different upwelling indices because studies in the eastern Atlantic boundary upwelling systems have shown that coastal upwelling is a complex system that only the wind stress cannot explain. Several mechanisms and factors, including wind stress, eastern propagation of Kelvin waves, coastal trapped waves, eddies, capes effects, geostrophic transport and remote forcing (ENSO) were proposed to understand the coastal upwelling process in this region. Based on these previous studies, in this manuscript uses different upwelling indices to characterize the dynamics and the contribution of each potential coastal upwelling factor to the seasonal and interannual variability of coastal upwelling along the Northwest African coast.
Q2: “What dynamics do the different indices represent?”
R2: -Ekman transport and wind stress curl are used to analyse the wind stress dynamics (main driver of upwelling) over the coastal and offshore region of the upwelling system. Ekman transport provides estimates of the offshore transport water, offering a direct measure of upwelling strength. While the wind stress curl shows the coastal divergence driven by the wind stress. However, these two upwelling indices only show the dynamic of upwelling process related to the wind forcing which is not often correlated to SST anomalies and the intensity of upwelling.
-The SST based upwelling index represents the SST gradient dynamics (ocean response) and measure the intensity of upwelling based on SST anomalies in a given area. However, this upwelling index does not consider the other internal and external factors likely to influence upwelling variability.
-The cumulative upwelling index (CUI) obtained integrating Ekman transport or SST gradient, represents the annual amount of upwelling.
-Total upwelling magnitude (TUM) represent the amount of upwelling from the start to the end dates.
-Upwelling index based on the sea level anomaly (SLA) pressure gradient or sea surface height represents the dynamics of onshore geostrophic flow and coastal trapped waves generated by Kelvin and Rossby waves, their influence on coastal divergence and upwelling.
The context of the interannual variability was clarified and discussed in relation with large-scale drivers and the previous studies (see section 4.3 and discussion).
2-This study uses multiple datasets for the same variables. As suggested, we expanded and improved the discussion of our results, highlighting the discrepancies implications of the datasets (see results and discussion section in the attached pdf file: revised_manuscript).
3- As recommended, the cumulative upwelling index (CUI) was recalculated considering the permanent (21-25N) and seasonal (10-21ºN) upwelling regions previously defined in the SMUS (Wooster et al., 1976; Mittelstaedt, 1991; Lathuiliere et al., 2008; Benazouzz et al., 2014; ) and the two upwelling cells observed in the East of Cape Plamas (7-2ºW) and East of Cape Three Points (2W-2ºE) in the GGUS (Djakoure et al., 2014; Wiafe and Nyadjro., 2015; Brandt et al., 2023). This regional separation was introduced in the Introduction section based on the corresponding previous studies.
4- As recommended, the markers (black dots) were added on the Fig. 12 (current Fig 11) to indicate the significant correlation between Nino 3.4 index and upwelling favourable wind stress anomalies during the during the extreme Nino events.
5- The discussion of interannual variability has been improved referencing previous studies, including the potential underlying mechanisms and the influence of the large-scale climate drivers and regional climate variabilities such as Dakar Niños/Niñas, Atlantic Meridional Mode.
Q3: “How large-scale climate drivers influence the Northwest African coastal upwelling?”
R3: Our analyses show that the large-scale climate modes such as ENSO and Atlantic Niño/ Niña events modulate the variability of the Northwest African coastal upwellings by reducing or supressing upwelling during the El Niño and Atlantic Niño phases or by enhancing the upwelling favourable conditions (strengthening of upwelling favourable wind stress, and SST cooling) during the La Niña and Atlantic Niña phases. More details of this result will be found in section 4.3 of the revised manuscript.
6.The discussion of interannual variability has been expanded on other climate mode of variability: lag and lead correlation analyses between Northwest African coastal upwelling indices (Ekman transport, SST gradient) and climate indices, such as AMM, AMO, NAO, NOI, ONI, PDO, PNA, NTA and SOI have been used to stablish the connections and teleconnections between the events (see section 4.3, Tables S3, 4, 5 supplementary material). These correlation analyses have allowed to identify the potential indices of coastal upwelling prediction over Northwest Africa.
Minor points
2.Line 151: The local wind means wind stress curl along the coast
3. Line 170: The geostrophic transport was computed in the direction parallel to the coast in each upwelling system. SSHal and SSHon correspond to SSH pressure gradient in the offshore and coastal boxes defined in each upwelling system. We first computed the SSH pressure gradient in the longitudinal and latitudinal direction. Then, the components of geostrophic current used to compute the geostrophic transport was computed in the direction perpendicular to the coast (see Table S2, supplementary material)
4.The visibility of the contour lines is improved on Fig. 2
5. Line 215: The downwelling means the relaxation or non-upwelling period
6. Line 232: Yes, we mean that the SST influence wind patterns. The seasonal variability of upwelling has clearly shown that the upwelling seasons is characterized by strong wind stress with low SST while the relaxation phases are marked by the weakening of wind stress, reducing Ekman transport. The analysis of SST and wind stress anomalies during the ENSO and ATL3 vents have also show these interactions between SST and wind stress anomalies and their associated upwelling indices. The (Fig. 2) shown that the wind-driven upwelling index decrease when the SST anomalies increase and vis versa. For example, in the southern part of SMUS (10º-14ºN), the weakening of upwelling-favourable wind stress (equatorward wind) is associated to the advection of warm water around the Guinea Dome. This advection of warm water drives by the poleward migration of North Equatorial Counter-Current (NECC) (Faye et al., 2015) (see manuscript revised, line 351)
7. The surface velocity is a mistake. It is the ocean surface current velocity. Our Ocean surface velocity plays a crucial role in modulating the interactions coastal divergence and offshore Ekman transport. In this study, the analysis has shown the upwelling seasons are characterized by the shoaling of ocean while the relaxation periods are characterized by the surfacing of ocean geostrophic current with rich-nutrient water, such the Poleward current in the SMUS and Guinea current in the Gulf of Guinea. This result is consistent with the findings of Klenz et al., (2008) and Kounta et al., (2018). More details of this result can be found (lines 348, 670) that the surfacing of ocean curren current shows that strong ocean velocity is likely to enhance SST cooling and the vertical mixing by driving warm water current under or above the cold-water current.
8. The absence of onshore geostrophic transport around 0º E in the GGUS was related to the quality of the dataset used. This analysis was improved in the revised manuscript (see Fig. 7). This weakening or absence of the geostrophic (ageostrophic) and reversal signs can be by several oceanographical factors.
9. Line 284: Yes, we are talking about the effects of CTWs: The influence of eastward propagation of Kelvin waves transform to coastal trapped waves was analysed by analysed by using sea level anomaly (SLA) pressure gradient. The results have shown strong effects of coastal waves drive by the wind stress, in particular in the GGUS. The combination of both coastal waves and local wind stress forcing was suggested to be the primary driver of coastal upwelling in this region.
10. In both upwelling regions, the oxygen concentration in October is much lower than during the rest of the year. This lower oxygen can be explained by several oceanographic and atmospheric processes. The two upwelling systems show October as a transition month between upwelling and non-upwelling period (in the GGUS) and between non-upwelling and upwelling season (in the SMUS) that bring deeper, oxygen-poor water to the surface. It corresponds also the maximum warm water period that enhance stratification, vertical mixing and limiting the replenishment of oxygen in surface waters from deeper layers or atmospheric exchange. In addition, this transition period can be accompanied by strong consumption of oxygen via photosynthesis and chlorophyll-a concentration.
11. The peak in surface salinity typically aligns with the strongest upwelling months. This connection can be explained by considering the physical and oceanographic processes associated with upwelling. The deep ocean waters that rise to the surface are typical stable and more saline than the surface layer often diluted by precipitation and river runoff. In the SMUS where upwelling come from the intermediate water masses, the increase in surface salinity can influence by the southward water masses.
12.Line 383: Yes, some recent analyses have related this decrease trend observed in the EBUSs (Bograd et al., 2023, Brandt et al., 2024), in particular the in the eastern Atlantic boundary to the changes in atmospheric circulation and upwelling drivers due to global warming and climate change. This decrease of the interannual SST variability was related to the substantial weakening of the correlation equatorial zonal wind stress and the role play by the meridional warming in the west Africa coast (Prigent et al., (2020).
-
RC2: 'Comment on egusphere-2024-4175', Anonymous Referee #2, 24 Mar 2025
This paper compares various indices of coastal upwelling in two different regions of the eastern Atlantic, the Senegalese-Mauritano Upwelling System (SMUS) and the Gulf of Guinea Upwelling System (GGUS). Indices are based on wind stress (coastal divergence and curl) and SST gradient. Other indices are based on surface currents and sea level. These indices are computed from two ocean reanalyses (GLORYS, ORAS5), an atmospheric reanalysis (ERA5), and two SST products with different spatial resolution (OSTIA, Hadl). The authors first study the seasonal variations of upwelling and the latitudinal changes in the two regions. In the second part of the paper they describe the interannual variations using cumulative indices, composite and correlation analyses.
I found the purpose of the paper interesting and the quality of the figures rather good. However the writing is often poor and there are many unclear sections. First, the authors do not seem to fully understand what they are comparing. The wind stress from an ocean reanalysis results from a bulk formulae using a atmospheric forcing (possibly an atmospheric reanalysis) and the ocean model SST as inputs. The ERA5 wind stress results from bulk formulae (possibly different than that used in the ocean reanalysis) using ERA5 atmospheric forcing and observed SST as input. The authors do not explain why they compare these products and what to conclude from the comparison is unclear. Another issue is the use of sea level in the analysis. The authors seem to have understood that cross-shore geostrophic transport impacts coastal upwelling, but they do not compute this index in the paper. Instead they use sea level anomaly, “ocean surface current” (it is unclear if the latter refers to current speed or to a particular component) and a unknown “geostrophic balance” index in the analysis. This makes the paper rather difficult to follow. In the study of interannual variability, they misinterpret their own figures, as El Nino events seem to enhance southward wind off the SMUS but tend to produce warm SST anomalies. Last the discussion of the results is poor, it merely summerizes the findings and does not provide any discussion of the mecanisms at stake to explain some of the differences that have been identified.
To conclude, I also have had to write a very long list of specific minor and not-so-minor comments that the authors need to address. As there is a lot of work to be done to improve the paper, my decision is reject. I encourage the authors to revise their work thoroughly before considering resubmission.
Specific comments:
L18 : « remote wind forcing » can not influence the region. It can influence the ocean remotely and then ocean waves may propagate to the region.
L38 : Bograd et al., 2022 would be a more suited reference here.
Bograd, et al. (2023). Climate change impacts on eastern boundary upwelling systems. Annual Review of Marine Science, 15(1), 303-328.
L46 : using a regional ocean model
L47 : coastline geometry
L49 : Recent upwelling indices take into account the cross-shore geostrophic current which can enhance or reduce coastal upwelling (see Jacox,et al. (2018). Coastal upwelling revisited: Ekman, Bakun, and improved upwelling indices for the U.S. West Coast. Journal of Geophysical Research: Oceans, 123, 7332–7350. https://doi.org/ 10.1029/2018JC014187). This needs to be accounted for in this study.
L52 : these are physical, not biophysical processes
L53 you mean « eastward propagation of equatorial Kelvin waves » and « poleward propagation of coastal trapped waves » ?
L61 : what do you mean by « coastal variability » ?
L66 : where does this influence take place more specifically ?
L68 : Using a regional model
L69 : « .. highlighted the Guinea current » : not clear what is meant here. Can you elaborate a little ?
L73-81 : I find this paragraph not really useful as the paper focuses on the past period. It could be useful if Bakun’s alleged dynamical processes driving upwelling variability were mentioned but it is not the case.
L86 : what remote forcing ? You mean equatorial Kelvin waves ? This is unclear.
L88-89 : the goals of the study need to be described more precisely in the introduction.
L94 : Refer to the spatial resolutions of ORAS5 and GLORYS in the Table here.
L105 : Should be Table « 1 ». GLORYS does not deliver oxygen and nitrate, there must be another product with another spatial resolution.
L127 : « corresponds also to the ... »
L129 : UI needs to be defined.
L130 : not clear what you mean here.
L153 : not clear what you mean by « drained by the local wind » : if there is no coast, there is no divergence of the Ekman current and no upwelling. Do you mean « by the wind stress curl » ?
L168 : you must mention this process before (see my comment line 49).
L174 : I do not understand the definition of Tg : to compute the onshore geostrophic transport, you must compute the gradient of sea level in the direction parallel to the coast. What are SSHal and SSHon ?
L176 : nitrate
L179 : before CUI can be defined, a clear definition of UI is needed somewhere.
L182-188 : a schematic is needed to explain how the indices are defined.
L190 : it should be an integral, not a sum ? What is the temporal resolution ?
L194 : « based on the gradient descent fit » , « Pettitt statistical test »: references are needed
L196 : a word is missing here.
L201 : analysing
Figure 2 : what are ORAS5 and GLORYS wind stress fields ? ORAS5 and GLORYS are ocean reanalyses, meaning that wind products or reanalyses were used to compuate wind stress and force the ocean surface. These winds are possibly from era5, jra-55 or another reanalysis. This should be clearly explained in the methodology section.
I am puzzled by Figure 2 because coastal upwelling is negative (upward) in Fig 2a-c .However you mention earlier that Tg is negative eastward, which can not be the case because an eastward (negative) Tg produces a (negative) downwelling at the coast. I think that upwelling should always be positive as a convention. I also find it very confusing to have a different sign convention for coastal upwelling and wind stress curl upwelling. This needs to be changed.
Last, it would be nice to relax the saturation of the colorbar for positive values to see if there is coastal downwelling during some time periods.
L214 : the start of the upwelling seasons depends on latitude, so what is the latitude here ?
L214-216 : this is not easy to see this in Figure 2 and 3. Another figure should be shown.
L225 : What about the geostrophic transport Tg ? It does not seem to be taken into account in this analysis. It should be added (or at least compared) to the zonal Ekman transport (with the correct sign).
L127 : why is the maximum SST gradient observed before the maximum Ekman transport ?
L232 : « the highest SST drive... » : unclear, I do not understand what you mean ? The SST forces the wind patterns ? This should be demonstrated.
Figure 3 : several issued with this figure : the « THETAO » (e) label is awkward : GLORYS is an ocean reanalysis, same as ORAS5, so the title should be GLORYS ; in a-e, the color scale indicates SST where ? In the coastal box ?; the definition of the green line is strange : you mean during the month of minimum SST ? « Minimum SST » where ?; I do not see the point of the orange line and its link with upwelling as it lies south of the upwelling region.
L242-243 : I don’t see how you reach that conclusion as these are 3 different estimates of wind stress and we do not know which one is closest to reality.
L244 : « the good SST index... » : OSTIA and Hadl provide the observed SST indexes, not the good ones. How is the Hadl index poor ? You can just hypothesize that the OSTIA index is better because this product is more precise near the coast.
L246 : this is getting more and more confusing : ERA5 is not an SST dataset, it is an atmospheric reanalysis which is forced by an SST product (possibly OSTIA , to be checked). This needs to be clarified. Apparently the authors do not known well enough what they are comparing.
L247 : what are the wind forcing used to force ORAS5 and GLORYS ? This needs to be added in the table.
L248-249 : « The maximum magnitude.. » of what ?
Figure 4 : why didn’t you sum Ekman divergence (Mx), wind stress curl upwelling and Tg ? Just looking at Ekman transport does not make sense. What is SLA ? Is it Tg ? This is getting more and more confusing.
L260 : « geostrophic ... » : which figure are you refering to ? In Figure 4 SLA is shown, not geostrophic current. This is difficult to follow.
L262 : « positive WSCD generally enhance upwelling » : no, they contribute positively to upwelling, which is the sum of coastal upwelling (divergence of Ekman transport) and wind stress curl upwelling which occurs offshore. The rest of the sentence («negative WSCD ..suppress coastal upwelling ») should therefore be clarified.
Figure 5 : Legend is unclear : what is « ocean velocity » ? zonal, meridional, total ? What is « geostrophic balance” ? You mean geostrophic transport ? It is not the ocean velocity that matters for coastal upwelling but the surface cross-shore geostrophic velocity. Apparently the authors used ORAS5 and GLORYS surface velocity, which is useless.
What are the isolines in figure (e) ?
L273-274 : these two lines are meaningless.
L274 : what is the « Rossby circulation » ?
L284 : correlation does not indicate causality : how would SLA influence productivity ? I do not believe it is the case, or it needs to be better explained.
L295 : you mean « meridional SST gradient » ?
L295 : « the upwelling waters were estimated... » : not clear what this means. Rephrase. What is the justification for this SST threshold?
L300 : « integrating...” in time and space? Otherwise, what is the longitude refered to here? This figure is not shown.
Figure 7: repetition of text in legend.
L302-303: this is the maximum magnitude of the integrated transport or the maximum in time and space?
Line 306 and Fig.8: as for Fig. 5, I do not understand what is plotted when geostrophic balance is mentioned. Moreover, surface velocity is different from meridional geostrophic velocity, which may impact the GGUS coastal upwelling. It is impossible to interpret Figure 8a,b.
Line 306: “ocean velocity (contours)”
L320: Fig. 7 must be Fig.6 here. Keep in mind that ERA5 is not an ocean reanalysis, its wind stress is computed using a particular bulk formulae and a particular SST product (possibly OSTIA, to be verified). In the case of ocean reanalyses, wind stress is computed using using bulk formulae and the ocean model SST. Thus the wind stress differences in the different products may originate from the SST products and differences in bulk formulae. This should be discussed.
L321: I do not really see the point in using Hadl product since it is obvious that this product is not adapted for the upwelling index due to its coarse spatial resolution. It could be suppressed from the study to shorten the paper.
L327: “This indicates ...”: I strongly disagree. Correlation is not causality.
L328: I do not see two cells in Fig.7f-j., just one.
L330: What do you mean by “the western cell is linked to the SLA”: what is the mechanism? Does the geostrophic current driven by a SSH gradient play a role?
L330-332: Citing papers is not sufficient, how Kelvin waves and CTW drive this upwelling needs to be explained.
Figure 8: it would help to represent Figure 8 with the horizontal axis for longitude and vertical axis for time.
L347: “the maximum concentration of Chlorophyll ...”: not at all. Oxygen concentration does not drive chlorophyll concentration. I think the figures and text about the Chl, Nitrate, Oxygen and surface salinity should be suppressed, they do not bring any useful information to the paper.
Figure 9: which wind stress and SST product are used here?
L370-372: what was used for the upwelling indices in the linear regression? The cumulative index? The maximum for each year? This lacks details and makes it very difficult to follow.
L382-383: “This decreasing trend...”: you can not prove that this is linked to anthropogenic climate change with this analysis, this could also be decadal variability.
L400: A table should be added to list the climate events used to compute the composite with time periods, SST anomalies in the Pacific, Atlantic, SMUS and GGUS. The SST product used to compute the composite is not indicated.
Figure 11: I do not see the point in showing the global ocean in this figure. You could show only the Atlantic and Pacific ocean between 30°S-30°N, thus it would be easier to visualize the SST and wind stress anomalies in the SMUS and in the GGUS (it is impossible to see anything about the GGUS in the present figure).
Figure 11: how do you know that the composite is significant? A statistical test is needed.
L402-403: “extremely warm SST anomalies”: 0.6°C is not extreme in my opinion. I think the arrows indicate that coastal upwelling in the south part of (entire) SMUS is enhanced in DJF (MAM) , but the SST increases. Coastal upwelling does not explain this SST increase, apparently.
L405: References are needed here, since the authors have not shown any of this.
L407-408: I can not see the positive anomalies in the Gulf of Guinea druing La Nina events.
Figure 12: same comment as for Figure 11: only the regions of interest (Equatorial Pacific, tropical Atlantic) should be shown.
L425: how was the correlation computed? What was the time series used? Over which time period?
L430: I think the authors misinterpreted the result of the correlation: the correlation between El Nino index and meridional wind stress off the SMUS is negative, meaning that when the El Nino index is high (strong El Nino conditions), the meridional wind stress anomaly is negative (southward) thus enhances coastal upwelling in the SMUS. This can also be seen in Figs 11a,c, wind stress anomalies appear to be southward in SMUS.
L452: how mesoscale variability impacts upwelling needs to be explained in the discussion.
L453: “weak upwelling trend..”: positive or negative?
L455: I think this needs to be refined, the trends are different in the north and south of the SMUS and this line is not consistent with Sylla et al. (2019) findings.
L470: I disagree with the statement. The wind-driven upwelling increases and the SST is warmer, meaning that another process explains the surface warming.
Citation: https://doi.org/10.5194/egusphere-2024-4175-RC2 -
AC2: 'Reply on RC2', Dametoti Yamoula, 20 Apr 2025
Dear Reviewer,
I understand the Reviewer’ concerns and the majors and minor points including suggestions and recommendations. The manuscript has been revised according the Reviewer’s comments, and some questions are directly answered here. But those which required complementary analyses were discussed in the attached in the revised manuscript.
Some aspects of the manuscript were expanded and discussed in relation to existing studies.
Major and minor corrections
L18: We agree with you, it was a mistake. The « remote wind forcing » leads to eastward propagation of Kelvin waves, which are transformed into coastal trap waves when they reach the West African coast modulating coastal upwelling. Thank you very much for the correction.
L38: Bograd et al., 2023 was used as one of the appropriate references for the impacts of climate change on the eastern boundary upwelling systems.
L46 : using a regional ocean model (corrected)
L47 : coastline geometry (corrected)
L49 : This suggestion (Jacox,et al. 2018) was taken into account in the revised manuscript. The influence cross-shore geostrophic current was discussed in the results of the revised manuscript. The seasonality of onshore geostrophic transport was computed based on the sea level pressure gradient, and used to analyse the relative contribution of onshore geostrophic flow in each upwelling system.
L52 : these are physical instead of biophysical processes (corrected)
L53 Yes, we mean « eastward propagation of equatorial Kelvin waves » and « poleward propagation of coastal trapped waves » (corrected)
L61 : By « coastal variability » we mean the coastal SST variability
L66 : This influence take place more specifically under the capes
L68 : Using a regional model (corrected)
L73-81 : The paragraph was removed in the new version
L86 : The remote forcing here mean ENSO and other large-scales climate events which are not occur near to the Northwest African coastal upwelling regions.
L88-89 : the goal of the manuscript was clear describe in the introduction of the revised manuscript (Line 95).
L94 : Refer to the spatial resolution of each dataset was added in the Table S1 (supplementary material).
L105 : My apologies, the oxygen and nitrate data were provided by the global Biogeochemistry Analysis and Forecasts from the Copernicus Marine Environmental Monitoring Services (CMEMS)
L127 : « corresponds also to the ... » (corrected)
L129 : Upwelling index (UI) was defined in the method before using as the abbreviation.
L153 : not clear what you mean by « drained by the local wind » : if there is no coast, there is no divergence of the Ekman current and no upwelling. Do you mean « by the wind stress curl »?
L168 : you must mention this process before (see my comment line 49).
L174 : The onshore geostrophic transport (Tg) was computed using the sea level pressure gradients of the direction parallel to the coast. This sea level pressure gradient was derived from sea level anomaly (SLA) and sea surface height (SSH). We used both SLA and SSH for the comparison because the SSH include SLA and absolute dynamics of the topography (geoid, tidal). and are the alongshore and onshore pressure gradients, respectively. After compute the SSH gradient at the longitudinal and latitudinal directions, the appropriate SSH pressure gradient were selected in the corresponding upwelling system to compute the onshore geostrophic transport. Here, we used the difference of sea level pressure between offshore () and coastal () boxes to compute the onshore geostrophic transport (see method section and attached supplementary material for more details).
L176 : nitrate (corrected)
L179 : The cumulative upwelling index (CUI) and upwelling index (UI) were clearly defined in the method section .
L182-188 : Table S2 (supplementary material) summarizes the different upwelling indices used in this study.
L190 : Integrating (corrected). What is the temporal resolution is daily
L194 : « based on the gradient descent fit » , « Pettitt statistical test »: references are needed
L196 : a word is missing here (added): the sentence was reformulated
L201 : analysing (corrected)
Figure 2 : what are ORAS5 and GLORYS wind stress fields are ocean reanalyses forced by the Era interim products, the spatial resolution, levels and more characteristics of this data were added in the Table S2 (supplementary material).
Here I don’t well understand your worry but I think my revised manuscript can respond this. The update Figures present negative Ekman transport (Fig 2, blue colour shading) and onshore geostrophic flow (Fig. 4) and positive coastal wind stress curl (Fig. 2, contours) and SST gradient (Fig. 3, contours) during the upwelling season in both upwelling systems (SMUS, GGUS). The opposite sign of these indices has been observed during the relaxation or non-upwelling periods. What I mentioned earlier was to explain that the negative onshore geostrophic transport enhances the wind stress curl (positive). You are right upwelling should always be positive (positive wind stress curl), but negative onshore geostrophic transport (Fig. 4). This sign convention has been adjusted. Here, Ekman transport is driven by meridional wind stress (negative sign). The upwelling based on the Ekman transport will have a positive while the relaxation period will be positive sign.
We also take into account your suggestion to improve the visibility of the colorbar for positive values to see if there is coastal downwelling during some time periods.
L214 : Although the upwelling timing varies over the latitudes, the start and end dates of upwelling presented here is the average of the seasonal upwelling region of SMUS (10 – 20oN).
L214-216 : The starts and end dates were clearly shown on Figures 8 and 9 in the revised manuscript
L225 : The geostrophic transport (Tg) was taken into account by not seem to be taken into account in comparing to the zonal Ekman transport.
L227 : The earlier peak in SST gradient relative to Ekman transport in the SMUS can be explain by the interplay between atmospheric forcing, rapid oceanic response times, and preconditioning of shallow thermocline dynamics.
L232 : « the highest SST drive... » : we mean that the SST influence wind patterns. The seasonal analysis upwelling indices derived from SST and wind stress fields (Ekman transport) clearly shows that wind-driven upwelling decrease when the SST anomalies increase and vis-versa. The upwelling season is characterized by strong wind stress while the relaxation period is characterized by a weakening of wind stress associated with strong SST anomalies. The composite analysis of SST and wind stress anomalies during the extreme large-scale climate events have also shown that SST anomalies modulate the wind stress patterns (Section 4.3).
Figure 3 : was update taken to account all the remarks, such the label name, the definition of green (minimum temperature observed before the beginning of upwelling season in the seasonal upwelling region), the colour scale and the orange line (25 oC isotherm: Threshold of SST used to estimate the southern margin of upwelling region) lying south of the upwelling region.
L242-243: I don’t see how you reach that conclusion as these are 3 different estimates of wind stress and we do not know which one is closest to reality.
L244 : Your suggestion was taken into account here. OSTIA and Hadl are the reconstituted observational SST data but at different spatial resolution (0.05o and 1o , respectively). However, the best representation of upwelling index in this coastal region was OSTIA, because this product is more precise near the coast.
L246 : ERA5 is atmospheric reanalysis that includes SST and wind stress products used in this study. It has been forced by Era interim, OSTIA and HadI products. This information clearly appears in Table S1 (supplementary material). This needs to be clarified.
L247 : ORAS5 was forcing with Era interim HadI and OSTIA products while GLORYS was forcing Era-interim and real-time forecast data (see Table S1).
L248-249 : « The maximum magnitude.. » of upwelling-favourable
Figure 4 : This Figure and it analysis is removed in the revised version to avoid the confusion. It was showing the seasonal cycle of sea level anomaly (SLA) compared to upwelling indices derived from wind stress (Ekman transport) and SST in the coastal box.
L260 : this analysis based on Figure 4 has been removed to make it easier to understand.
L262 : this analysis based on Figure 4 has been removed to make it easier to understand.
Figure 5 : This Figure was updated (Figure 4 in the revised manuscript) taken to account all the suggestions. Only the onshore geostrophic transport of the coastal and offshore boxes was compared with ORAS5 and GLORYS datasets.
What are the isolines? The isolines represented the onshore geostrophic magnitude
L273-274 : these two lines were removed as suggested.
L274 : mistake, it is not Rossby circulation, but Rossby waves caused by the rotation of the Earth and the des effects of the latitudinal variation of the Coriolis force. The writing is adjusted
L284 : the correlation does not indicate causality. But influence of SLA on the upwelling productivity was estimated comparing the seasonal evolution of SLA and chlorophyll-a concentration. The upwelling season was characterized by negative SLA over the coastal box indicating strong influence of sea level pressure gradient on the coastal SST variability, while the relaxation periods were characterized by positive SLA indicating the supressing of upwelling season. This seasonal variability of SLA is likely influence chlorophyll-a productivity. The analysis based on Figure was removed as suggested, to avoid the confusion.
L295 : Yes, the « meridional SST gradient » was computed in the GGUS
L295 : This sentence was rephrase: The upwelling region were estimated using 25oC isotherm. The SST threshold corresponds to the reference temperature used in the equatorial region to estimate cold water anomalies (Caniaux et al., 2011). We mentioned it in the methodology section.
L300 : The upwelling indices used to compute the cumulative upwelling index are time series computed averaging the upwelling boxes (offshore box for the Ekman transport, difference offshore - coastal box for SST gradient).
Figure 7: repetition of text in legend (was corrected).
L302-303: this is the maximum magnitude in time and space over the climatological annual cycle.
Line 306 and Fig.8: was showing the ocean surface current velocity with the onshore geostrophic transport in contours. However, this Fig. 8 was update and corresponds to Fig. 7 in the revised version of the manuscript. The Fig .7 in the revised manuscript clearly shows the seasonal cycle of the onshore geostrophic transport, and the contour represent the magnitude of this geostrophic transport
L320: Fig. 7 must be Fig.6 here: This confusion was taken into account in the revised version of the manuscript. The different wind stress products used to compute the Ekman transport was discussed based on their characteristics: nature of the product, spatial and temporal resolution, inputs products, … Era5 is atmospheric reanalysis product including different variables such as SST and wind stress components used in this study.
L321: Hadl product was conserved in this study despite it is obvious that this product is not adapted for the upwelling index due to its coarse spatial resolution., for the comparisons. But it can be removed in the last version of the manuscript to shorten the paper (as you suggested).
L327: This result was revised based on the geostrophic flow analysis.
L328: The two upwelling cells were observed on the seasonal cycle plots (Fig. 7A-e). It is unclear on the Fig.7f-j because of the size of the plots.
L330: “the western cell is linked to the SLA”: we mean that the onshore geostrophic flow and the coastal waves drive by the seas level pressure gradient influence more the coastal SST variability at the western of the GGUS. The mechanism is based on the onshore geostrophic flow and poleward propagation of coastal trapped waves.
L330-332: The influence of Kelvin waves and coastal waves was discussed in the new version of the manuscript (445).
Figure 8: was update with the onshore geostrophic transport
L347: “the maximum concentration of Chlorophyll was observed during the upwelling peak and coincides with the peak of dissolved oxygen. This linear correlation let suggest that oxygen concentration (useful for the photosynthesis) drives the maximum concentration observed in this upwelling region. However, requested, the plots of surface salinity, chlorophyll-a concentration, dissolved oxygen and nitrate were removed in the main text from the supplementary material.
Figure 9: as ORAS5 does not delivery the daily product, the ERA5 wind stress and OSTIA SST were used to compute the cumulative upwelling indices
L370-372: to analysis the long-term trend, the Ekman transport and SST-based upwelling index, both derived during the upwelling seasons (winter-spring in the SMUS and summer in the GGUS) were used for the linear regression analysis, not the cumulative upwelling or the maximum of each year. However, in the revised version, these historical trends were computed with the wind stress and SST anomalies during the upwelling seasons.
L382-383: we suggested the linked of this long-term trend with previous studies on the impacts on climate change in the eastern boundary upwelling systems (Bograd et al., 2023).
L400: As recommended, the Tables (Tables S3, 4, 5) of the correlation analyses between upwelling indices (Ekman transport and SST gradient) were added in the supplementary material.
Figure 11: As recommended, the Figs. 11 and 12 (new Figs. 10 and 11 in the revised manuscript) have been displayed between 30S-30N, to improve the visibility of the upwelling regions on the Figures.
Figure 11: In the new version of the manuscript, the significant correlation was displayed with MonteCarlo statistical test.
L402-403: This result has been re-discussed: El Nino phases (DJF, MAM) are characterized by the weakening of equatorward trade winds, the reinforcing the northward winds and the reducing and supressing of upwelling along the northwest African coast, especially in the SMUS. While La Nina phases are characterized by the strengthening of the equatorward winds and enhancing Ekman transport and upwelling conditions.
Atlantic Nino/Nino phases present the same behaviour in the GGUS with less significant impacts on the SMUS, in particular in the northern of this upwelling region
L405: These results have been re-discussed with references (based on the previous studies on the Dakar Nino/Nina, Atlantic Meridional mode, Atlantic Nino/Nina).
L407-408: This discussion of the result has been improved with the update Figs. 10, 11 in the new version of the manuscript (section 4.3).
L425: how was the correlation was computed between the composite indices of the extreme events of ENSO and Atlantic Nino mentioned in the methodology section, and the spatial distribution of upwelling-favourable wind stress (meridional wind stress in the SMUS during winter and spring, zonal wind stress in the GGUS during summer) during the same period.
L430: The discussion of correlation result between El Nino index and meridional wind stress off the SMUS has been improved and expanded based on the update Figs. 10-11 (section 4.3).
L452: the impacts of mesoscale variability on upwelling needs was discussed based on the previous studies.
L453: upwelling trend is different according the upwelling region (L522, revised manuscript)
L455: The long-term trend of upwelling was re-discussed in the revised manuscript (L522) based on each section of upwelling (permanent and seasonal in the SMUS, eastern of the capes in the GGUS): sea Fig. S4 (supplementary material).
L470: I this line was rephrased in the revised manuscript.
In summary, the manuscript was completely revised based on reviewers’ comments and suggestions. The last version and supplementary material will be sent to all reviewer for the checking.
Thank you very much for your interesting and useful comment and suggestions for this manuscript, I really appreciated and hope that my revised version will satisfy your comments
-
AC2: 'Reply on RC2', Dametoti Yamoula, 20 Apr 2025
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 331 | 219 | 22 | 572 | 11 | 27 |
- HTML: 331
- PDF: 219
- XML: 22
- Total: 572
- BibTeX: 11
- EndNote: 27
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1