the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 25 Apr 2023
Deoxygenation of the Gulf of Mexico thermocline linked to a decrease in the detachment frequency of Loop Current Eddies
Abstract. This study presents an oxygen time series of the Gulf Mexico deep water region spanning from 2010 to 2019 using data from 6 oceanographic cruises and one ARGO buoy. The data suggest a deoxygenation trend in the thermocline of the Gulf of Mexico. This deoxygenation trend seems to be connected to a reduction in the number of eddies that detached from the Loop Current from 2010 to 2019 observed with altimetry data from the last two decades, and although the average size of the mesoscale structures shows a slight increase, the average detached area per year almost halved from the 2000–2010 decade to the 2010–2020 decade. Using the oxygen measurements and the altimetry data, a simple box model was formulated to reproduce the measured oxygen temporal variability in the GoM main thermocline from 2000 to 2020. The results from the box model suggest that an average detached Loop Current Eddy area of about 90 000 km2 per year is needed in order to maintain constant oxygen levels in the main thermocline waters. This threshold wasn’t reached during the 2010 to 2020 decade and if the LCE detachment area per year continues to decrease in the future, oxygen concentrations in the Gulf of Mexico thermocline might continue to fall with still unknown effects in the ecological web structure at these depths.
- Preprint
(12081 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2023-751', Anonymous Referee #1, 14 May 2023
Review comments on manuscript egusphere-2023-751, titled “Deoxygenation of the Gulf of Mexico thermocline linked to a decrease in the detachment frequency of Loop Current Eddies” submitted to EGUsphere by Quintanilla et al.
In this manuscript, the authors studied the deoxygenation in the thermocline of the Gulf of Mexico using data from research cruises and one Argo buoy. Using a simple box model, they came to the conclusion that deoxygenation between 2010 and 2020 is caused by the reduced Loop Current Eddy area based on altimetry data. There are several fundamental issues in this study, which must be addressed before the publication of the paper.
- Oxygen is not a conservative variable in the ocean. Oxygen concentration or oxygen content is strongly influenced by biological processes (production and consumption) and air-sea oxygen flux near the sea surface. Although these two processes mainly influence the oxygen concentration in the upper ocean, which causes strong seasonal variations in the surface mixed layer, they can also influence the oxygen concentration in the thermocline via vertical mixing/diffusion, entrainment, downwelling etc. Horizontal advection, horizontal diffusion and mixing can also change the oxygen concentration. These were all ignored in the simple box model. To use the simple box model, the authors need to prove Loop Current Eddies is the dominant process to supply oxygen in the thermocline and all other processes can be ignored. Otherwise, the proposed box model cannot be used.
- Even though the box model can be used, it takes time for the oxygen in the Loop Current eddies to be mixed into the common water. That time scale needs to be estimated, which should be a function of horizontal diffusion. If the time scale is longer than several years, the model cannot be used directly. A time lag needs to be considered. The deoxygenation may not be observed immediately after the reduction of Loop current Eddy annual area if there is a lag. This is also critical to determine whether the box model is reasonable or not.
- The Argo O2 profiles used in this study were not adjusted/QCed. Based on the Argo QC manual (https://archimer.ifremer.fr/doc/00354/46542/), it may be necessary to correct the oxygen profiles from Argo, at least some quality check/quality assurance should be conducted before those profiles can be used. Data quality assurance is critical to the detection of long-term trends. Non-QC’d data should not be used in the trend study because any shift/bias may lead to inaccurate conclusions.
- Stratification and warming influence on deoxygenation in the thermocline should be considered. There are several recent publications showed recent warming in the GOM (e.g. https://journals.ametsoc.org/view/journals/phoc/51/4/JPO-D-19-0295.1.xml; https://journals.ametsoc.org/view/journals/clim/36/8/JCLI-D-22-0409.1.xml). Warming could cause oxygen solubility changes and stratification changes, which may contribute to the deoxygenation in the Gulf of Mexico. To what extent the deoxygenation in the study region is controlled by warming should be investigated and discussed.
- Seasonal variations of dissoved oxygen in the thermocline needs to be considered. The thermocline can still be influenced by seasonal changes. To determine the long-term trend, the seasonal cycle typically should be removed first from the signal because seasonal variations can be orders of magnitude larger than the long-term trend.
- It seems like in the box model, oxygen in the thermocline is only supplied by the Loop Current Eddies. Is this a reasonable assumption? Let’s assume an extreme case that no Loop Current Eddies are generated in the Gulf of Mexico. Based on the box model, the oxygen concentration in the thermocline will decrease to zero due to oxygen consumption. That’s not true. Because oxygen was measured in other oceans without Loop Current eddies. Surface mixed layer is a major source for oxygen in the thermocline via ventilation, entrainment due to mixed layer seasonal cycle and mixing/diffusion etc. The authors have already assumed deoxygenation in the thermocline by ignoring oxygen exchange between the surface mixed layer and thermocline, which is invalid. At least a two layer model is needed to consider the oxygen exchange between the upper ocean and the thermocline. Maybe the oxygen exchange between the upper and lower thermoclines should be considered as well.
- The possible reasons for the reduction of LCE area after 2010 needs to be further discussed and investigated.
- The data shown in the paper demonstrated strong horizontal difference in dissloved oxygen concentration in the thermocline. Because the cruise data were not always measured in the same locations. Therefore, horizontal difference might cause year to year difference in the average oxygen values if the data are not collected at same stations. The influence of sampling locations should be considered in this study.
Citation: https://doi.org/10.5194/egusphere-2023-751-RC1 -
AC2: 'Reply to referee 1', Jose Quintanilla, 29 Jun 2023
In this manuscript, the authors studied the deoxygenation in the thermocline of the Gulf of Mexico using data from research cruises and one Argo buoy. Using a simple box model, they came to the conclusion that deoxygenation between 2010 and 2020 is caused by the reduced Loop Current Eddy area based on altimetry data. There are several fundamental issues in this study, which must be addressed before the publication of the paper.
Dear referee 1, we appreciate your comments and suggestions on our manuscript, they will make this a better work, and we apologize for the time taken to submit our response.
We understand that the main issue of the work as presented is that it doesn’t provide enough evidence of the LCEs detachments as a mechanism that ventilates the Gulf of Mexico (GoM) thermocline and therefore how a decrease of LCE detachments can lead to a deoxygenation of this water volume. We believe that the following additions and corrections can make this paper a solid contribution to ocean science.
(1) Oxygen is not a conservative variable in the ocean. Oxygen concentration or oxygen content is strongly influenced by biological processes (production and consumption) and air-sea oxygen flux near the sea surface. Although these two processes mainly influence the oxygen concentration in the upper ocean, which causes strong seasonal variations in the surface mixed layer, they can also influence the oxygen concentration in the thermocline via vertical mixing/diffusion, entrainment, downwelling etc. Horizontal advection, horizontal diffusion and mixing can also change the oxygen concentration. These were all ignored in the simple box model. To use the simple box model, the authors need to prove Loop Current Eddies is the dominant process to supply oxygen in the thermocline and all other processes can be ignored. Otherwise, the proposed box model cannot be used.
1- In order to show evidence of why the variability of the Loop Current Eddies (LCEs) detachments is the dominant process that affects the variability of the oxygen in the GoM thermocline, while the other processes can be considered constant (but not ignored), we propose the following changes:
- We will add a more complete comparison of the oceanographic conditions measured between cruises XIXIMI-4, 5 and 6 (proposed figures 7 and 8 and table 5 in attached pdf). These three cruises took measurements during the summer season in three consecutive years, august 2015, june 2016 and august 2017, and share 24 stations located in the exact same coordinates that will be used for this comparison in order to reduce the horizontal differences stated in issue 8. Between the three expeditions there is a decline in the oxygen levels, going from 2.87 ml l-1 in 2015 to 2.8 ml l-1 in 2016 to 2.65 ml l-1 in 2017.
- The oxygen decrease rate between those cruises increased from 0.04 ml l-1 y-1 to 0.16 ml l-1 y-1 in the same depths. In comparison, a change of such magnitude in the Oxygen Utilization Rate (OUR) is only expected comparing water from 200m to water from 600m (as measured by Jenkins, 1982). We didn’t find data that would lead us to believe that the oxygen utilization rate could have spiked in that time producing the observed variability, but the Particulate Organic Carbon measured during the same XIXIMI cruises (reported in Contreras-Pacheco et al., 2023) doesn’t show an important variability in the upper 200m that could link the strong oxygen reduction to a strong variability of the primary productivity that could have led to a strong increase of the biological respiration.
- We couldn’t find any evidence of change in the thermocline vertical mixing of the GoM during those years in the literature, so we propose to add a T-S diagram (that was clearly lacking before) to better show the thermocline range and how from 2015 to 2017, there wasn’t any substantial change in the thermohaline properties of the thermocline that shows evidence of strong variability in the vertical mixing. The upper thermocline temperature was even a bit colder during XIXIMI-6 than during XIXIMI-5, showing that there wasn’t a warming of this volume between both cruises that could explain the important oxygen decrease.
- Oxygen diffusion variability is hard to measure but it is usually accepted that is a proses less important than oxygen advection
- The only variable that we can see a clear change in, is the number of LCEs that detached one year prior to each expedition, being three, two and zero from XIXIMI 4 to 5 to 6 respectively. If, as has already been proposed (Maul, 1980), the LCEs detachments work as pulses that introduce large volumes of water from the Caribbean to the GoM, and, as we show here and has already been largely observed, the Caribbean thermocline is more oxygenated than the GoM thermocline, then these pulses should be introducing important volumes of water with more oxygen inside the GoM. After these observations we will add a paragraph to clarify that the scope of this work isn’t to say that the OUR, the oxygen vertical mixing and the oxygen diffusion can be neglected inside the GoM, but to show that these processes can be considered constant (we called the sum of those processes as oxygen change rate and it is a key assumption for the model)
- We will better clarify that the box model presented is intended as a method to illustrate how the oxygen variability observed can be approximated using the LCEs detachment volume as the only source of variability, with all other processes considered constant. And to show how a decrease of LCEs detachments can lead a decrease of the oxygen levels in the GoM. It is not intended to estimate the oxygen concentration in an exact location or moment but to contribute as an approximation tool of the GoM thermocline oxygen levels.
(2) Even though the box model can be used, it takes time for the oxygen in the Loop Current eddies to be mixed into the common water. That time scale needs to be estimated, which should be a function of horizontal diffusion. If the time scale is longer than several years, the model cannot be used directly. A time lag needs to be considered. The deoxygenation may not be observed immediately after the reduction of Loop current Eddy annual area if there is a lag. This is also critical to determine whether the box model is reasonable or not.
2- The immediate mixing after detachment of the LCEs volumes with the GoM water is an extreme simplification of the model that we pointed as clearly false. The box model has the main simplification of assuming the whole GoM upper and lower thermoclines as horizontally homogeneous, we will better clarify that the model isn’t proposed as a tool to determine a precise oxygen concentration in a given moment in a given point of the GoM but as a tool to understand and visualize how the LCE variability can produce important oxygen variability inside the thermocline of the GoM. We tried another model that implemented a constant oxygen mixing rate between the LCE and the GoM during the LCE lifetime, but it introduced more complexity than needed and the final oxygen budget was the same. Currently we are working with a more complex model trying to eventually add a much more precise understanding of the ventilation processes inside the GoM. For now, the simplicity of the model allowed us to quickly add the latest three LCEs detachments since 2020, and compare the predictions to newly found BioARGO measurements from 2021 to the present (https://biogeochemical-argo.org/data-access.php).
(3) The Argo O2 profiles used in this study were not adjusted/QCed. Based on the Argo QC manual (https://archimer.ifremer.fr/doc/00354/46542/), it may be necessary to correct the oxygen profiles from Argo, at least some quality check/quality assurance should be conducted before those profiles can be used. Data quality assurance is critical to the detection of long-term trends. Non-QC’d data should not be used in the trend study because any shift/bias may lead to inaccurate conclusions.
3-The first BioARGO data presented was used to show that the oxygen data measured with the buoy is consistent with the oxygen measured and calibrated during 6 cruises. If the lack of additional correction of its data is an issue, that data can be eliminated from this work, as it wouldn’t affect the main observations provided by the XIXIMI cruises. However, after remarks from referee 2, we decided to add newly obtained data from a BioARGO buoy deployed in in the center of the GoM in October 2021 and still in operation. This data serves as another time series that evidences the oxygen ventilation by LCEs. The internal calibration of the buoy should be enough us the data as a time series whose function is to compare the variability.
(4) Stratification and warming influence on deoxygenation in the thermocline should be considered. There are several recent publications showed recent warming in the GOM (e.g. https://journals.ametsoc.org/view/journals/phoc/51/4/JPO-D-19-0295.1.xml; https://journals.ametsoc.org/view/journals/clim/36/8/JCLI-D-22-0409.1.xml). Warming could cause oxygen solubility changes and stratification changes, which may contribute to the deoxygenation in the Gulf of Mexico. To what extent the deoxygenation in the study region is controlled by warming should be investigated and discussed.
4- Thanks for pointing out the paper by Wang et al., 2022, we will cite it as it is very relevant. As observed, the temperature of the thermocline didn’t show interannual variability during the XIXIMI cruises, but as a multiannual to decadal trend the effect of thermocline warming must be more clearly addressed. The reported warming of the upper GoM could be a part of the deoxygenation trend that is observed in the Caribbean lower thermocline and must be kept under observation.
(5) Seasonal variations of dissoved oxygen in the thermocline needs to be considered. The thermocline can still be influenced by seasonal changes. To determine the long-term trend, the seasonal cycle typically should be removed first from the signal because seasonal variations can be orders of magnitude larger than the long-term trend.
5- We haven't found evidence of important seasonal variability in the thermocline of the GoM as the mixed layer is always above the thermocline (Portela et al., 2018). The upper thermocline oxygen measured during the XIXIMI-3 cruises in winter 2013 doesn’t show evidence of vertical mixing with upper waters with a mean potential temperature of 17.07 ± 0.6 ˚C and a mean salinity of 36.26 0.091 .
(6) It seems like in the box model, oxygen in the thermocline is only supplied by the Loop Current Eddies. Is this a reasonable assumption? Let’s assume an extreme case that no Loop Current Eddies are generated in the Gulf of Mexico. Based on the box model, the oxygen concentration in the thermocline will decrease to zero due to oxygen consumption. That’s not true. Because oxygen was measured in other oceans without Loop Current eddies. Surface mixed layer is a major source for oxygen in the thermocline via ventilation, entrainment due to mixed layer seasonal cycle and mixing/diffusion etc. The authors have already assumed deoxygenation in the thermocline by ignoring oxygen exchange between the surface mixed layer and thermocline, which is invalid. At least a two layer model is needed to consider the oxygen exchange between the upper ocean and the thermocline. Maybe the oxygen exchange between the upper and lower thermoclines should be considered as well.
6- The GoM is different from other oceans as a semi enclosed basin where the LCE detachments introduce large volumes of water as pulses (Meunier et al., 2020). The thermoclines of other oceans have different ventilation mechanisms that we will detail more. The box model does indeed assume that without LCE detachment the oxygen concentration of the GoM thermocline will eventually get to 0, because the oxygen change rate of -0.16ml l-1 y-1 is considered constant (as obtained from the extreme scenario of no LCE detachments that happened between cruises XIMIMI-5 and 6, with 19 months without detachments). This is another simplification that we state as clearly false as all the mechanism you pointed have their own variability. We will add the explanation that the GoM, as any oceanic system tend to balance themselves, a reduction of water entrance via LCEs detachment might be balanced by other processes, as horizontal and vertical advection and diffusion and biologic production and consumption that can change the oxygen change rate. We shall make a better bibliographic revision on that aspect to add in the discussion.
(7) The possible reasons for the reduction of LCE area after 2010 needs to be further discussed and investigated.
7- We will look for more information about the possible reasons for a reduction of the LCEs detachment frequency. Our first bibliographic research produced little results, we found that it is a subject that is still under debate and not clear enough. The hypothesis that we though were more substantial were that the decrease of frequency of LCEs could be connected to ocean warming via stratification of the Loop Current (Moreles et al., 2021) and that a slowdown of the Atlantic meridional circulation could affect the Loop Current. There is a lot to study and understand on this subject, we will keep on looking and hopefully we will be able to add a little more information about it in this paper.
(8) The data shown in the paper demonstrated strong horizontal difference in dissloved oxygen concentration in the thermocline. Because the cruise data were not always measured in the same locations. Therefore, horizontal difference might cause year to year difference in the average oxygen values if the data are not collected at same stations. The influence of sampling locations should be considered in this study.
8- As addressed in point 1, we think that this issue can be corrected by including a cruise comparison using only the stations located in the exact same coordinates.
-
RC2: 'Comment on egusphere-2023-751', Anonymous Referee #2, 29 May 2023
The authors demonstrated that the Gulf of Mexico thermocline is encountering a deoxygenation trend during the past decade (2010-2019) which is attributed to Loop Current Eddies (LCEs) activities. The findings are derived from observations from 6 oceanographic cruises and one BioARGO float and from a simple box model. Although the studied topic is interesting, I don’t think the manuscript is ready for publication.
(1) The thermocline shall not be demonstrated by potential density anomaly which is used for pycnoclines. If the “thermocline” is considered a studied object, potential temperature anomaly shall be shown and discussed. Otherwise, the authors would like to use “pycnocline” throughout the manuscript.
(2) The findings lack statistical support. The authors separated the manuscript into two main sections. In the first part, they provided evidence of deoxygenation of the gulf waters and changes in LC or LCE activity, while a box model is developed for further mechanistic studies in the second part. However, statistical tools like simple linear regression are lacking. For example, the trend of deoxygenation shall be detected by linear regression (with p values) not by the naked eye as the shown temporal differences in Table 1 are minor than the based values and are closed for different layers and geographical locations. So, the tune “seem to have” (e.g., Lines 212, 217, …) is weak and not convincing at all. Further, statistical evidence or mechanistic analysis is also needed to investigate the relationship between LCE and deoxygenated waters. The authors need to dig deeper when “building” such a linkage.
(3) More data is needed. The authors are studying mesoscale phenomena over a vast ocean basin from 2000 to 2020. So, I don’t think measurements from 6 cruises and 1 ARGO float are enough. I am not sure if there are any other cruise measurements over the south of 25°N, but, at least, there are many BioARGO profiles around the path of LC provided by other archive systems like the world ocean database. The author should try to explore more data.
(4) The temporal scale of interest may need to be reconsidered and clarified. The lifespan of LCE is typically within a year, however, the authors focus more on the decadal changes of LCE and its relationship to the decadal changes in dissolved oxygen concentration in the gulf. I think the relationship between LCE and gulf water deoxygenation should be stronger on an interannual scale. In addition, 20-yr data is not enough to study a decadal phenomenon but is more suitable for interannual or even intra-annual scales.
(5) Even though the authors can statistically “prove” that changes in the gulf DO concentration are related to changes in LCE activity, the study is carried out under a major consumption that other factors that contribute to DO changes like air-sea oxygen exchanges, biological processes in the upper ocean, and DO transports due to advection and diffusion are minor or negligible. To my understanding, it is usually not the case. Contributions by all these factors should be quantified and discussed before the authors can “prove” that LCE activity outcompetes others.
(6) Again, as factors like air-sea oxygen exchanges are ignored in this study, changes in oxygen solubility due to temperature changes are ignored. Many studies indicate that the Gulf of Mexico is warming in recent decades. Decadal changes in water temperature are apparent which will lead to a decrease of oxygen solubility and further water deoxygenation. Thus, the effect of warming water should be discussed.
(7) A simple box model is not enough to reveal the mechanism of how LCE activity affects the water deoxygenation in the main thermoclines as the model neglects or simplifies too much critical processes like air-sea oxygen exchanges, advection, diffusion, and biogeochemical processes. The authors may need to seek another numerical tool like a coupled 3D model to reproduce the oxygen dynamics and LCE activity.
Citation: https://doi.org/10.5194/egusphere-2023-751-RC2 -
AC3: 'Reply ro referee 2', Jose Quintanilla, 29 Jun 2023
The authors demonstrated that the Gulf of Mexico thermocline is encountering a deoxygenation trend during the past decade (2010-2019) which is attributed to Loop Current Eddies (LCEs) activities. The findings are derived from observations from 6 oceanographic cruises and one BioARGO float and from a simple box model. Although the studied topic is interesting, I don’t think the manuscript is ready for publication.
Dear referee, we appreciate your comments and suggestions on our manuscript, they will make this a better work, and we apologize for the time taken to submit our response. We hope that the following additions and corrections can improve this paper.
(1) The thermocline shall not be demonstrated by potential density anomaly which is used for pycnoclines. If the “thermocline” is considered a studied object, potential temperature anomaly shall be shown and discussed. Otherwise, the authors would like to use “pycnocline” throughout the manuscript.
1 - We understand that the paper needs to present a T-S diagram to show the potential temperature and salinity values that defines the potential density range for the upper and lower thermocline. We will add a T-S figure of all the data used with the color code from the upper thermocline oxygen ranges to show how the stations with the highest oxygen show the subsurface salinity of the Caribbean water (proposed figure 7 in attached pdf) . Another T-S diagram will be included in a figure comparing the oxygen variability for three consecutive summer cruises, using only 24 stations that are located in the exact same coordinates, to show how there is an important oxygen decrease in the upper thermocline from june 2016 to august 2017, period without any LCE detachments, but that the temperature and salinity remain close to constant (figure 8).
(2) The findings lack statistical support. The authors separated the manuscript into two main sections. In the first part, they provided evidence of deoxygenation of the gulf waters and changes in LC or LCE activity, while a box model is developed for further mechanistic studies in the second part. However, statistical tools like simple linear regression are lacking. For example, the trend of deoxygenation shall be detected by linear regression (with p values) not by the naked eye as the shown temporal differences in Table 1 are minor than the based values and are closed for different layers and geographical locations. So, the tune “seem to have” (e.g., Lines 212, 217, …) is weak and not convincing at all. Further, statistical evidence or mechanistic analysis is also needed to investigate the relationship between LCE and deoxygenated waters. The authors need to dig deeper when “building” such a linkage.
2- We will add a linear regression using the XIXIMI data and historical data, a linear regression using XIXIMI data can be shown in figure 6, the r2 values are low at 0.36 and 0.66 in the upper and lower thermocline respectively. Those are low values but as we want to convey the thermocline oxygen shows high variability between years product of the high LCE detachment variability. Using the historic data, the r2 values increase to 0.6 and 0.9 in the upper and lower thermocline.
Although the biological oxygen utilization rate (OUR), and the oxygen vertical mixing and diffusion rates must vary from year to year the observable variable that can explain this variability is the LCEs detachments. We can add a statistical correlation analysis like a Pearson’s correlation or a principal component analysis but we believe that the box model proposed illustrates more clearly how using the LCEs detachments as the only source of variability the oxygen values can be approximated.
(3) More data is needed. The authors are studying mesoscale phenomena over a vast ocean basin from 2000 to 2020. So, I don’t think measurements from 6 cruises and 1 ARGO float are enough. I am not sure if there are any other cruise measurements over the south of 25°N, but, at least, there are many BioARGO profiles around the path of LC provided by other archive systems like the world ocean database. The author should try to explore more data.
3- We agree that more data is needed to make a more solid study, but there is a deficit of measurements in the Mexican area of the GoM, the historical data presented in this manuscript was intended as a recompilation of published measurements. We think that the data presented from 6 one-month cruises with more than 40 stations each, is a solid addition of data and a strong time series that can really useful. We will add new data from a BioARGO buoy in the center of the gulf that we just found, deployed in October 2021 and that is still taking data. The data from that buoy can be added to show how the box model still estimates the measured values adding the LCEs detachments observed after 2020 (as proposed in figures 9 and 6). According to our estimates in January 2020 there was one big LCE detachment, then nothing until another detachment in august 2021 followed by another detachment of a really big LCE in July 2022, there hasn’t been any other LCE detachments since. The BioARGO time series show a decrease in oxygen as predicted, and it will keep on adding data, we expect it to show further O2 decrease as long as there is not another LCE detachment and the buoy stays in the center of the GoM.
(4) The temporal scale of interest may need to be reconsidered and clarified. The lifespan of LCE is typically within a year, however, the authors focus more on the decadal changes of LCE and its relationship to the decadal changes in dissolved oxygen concentration in the gulf. I think the relationship between LCE and gulf water deoxygenation should be stronger on an interannual scale. In addition, 20-yr data is not enough to study a decadal phenomenon but is more suitable for interannual or even intra-annual scales.
4- Great point, the observed variability and deoxygenation is on a interannual scale and we must make emphasis on that point. The reason the decadal scale is mentioned is because comparing the 2000 -2009 decade to the 2010-2019 decade there is a difference in the detachment frequency and total LCE area that matches the oxygen differences observed between measurements made in the 2000s and measurements made in the 2010s.
(5) Even though the authors can statistically “prove” that changes in the gulf DO concentration are related to changes in LCE activity, the study is carried out under a major consumption that other factors that contribute to DO changes like air-sea oxygen exchanges, biological processes in the upper ocean, and DO transports due to advection and diffusion are minor or negligible. To my understanding, it is usually not the case. Contributions by all these factors should be quantified and discussed before the authors can “prove” that LCE activity outcompetes others.
5 – Those processes aren’t ignored but considered constant and are included as the constant value that we called oxygen change rate. We stated that it is a big simplification and that is clearly false, we will add more evidence in order to prove that the LCEs detachment variability “outcompetes” other processes of why the sum of those processes can be considered constant. With the T-S diagrams and thermohaline values we expect to show that there is not evidence of vertical mixing variability, with particulate organic carbon data obtained from the same XIXIMI cruises (Contreras-Pacheco et al., 2023) we can show that there is no evidence of an increase in the organic matter concentration that could explain an OUR increase as observed. And we are still looking for a way to find evidence of oxygen diffusion change but there is not enough data.
(6) Again, as factors like air-sea oxygen exchanges are ignored in this study, changes in oxygen solubility due to temperature changes are ignored. Many studies indicate that the Gulf of Mexico is warming in recent decades. Decadal changes in water temperature are apparent which will lead to a decrease of oxygen solubility and further water deoxygenation. Thus, the effect of warming water should be discussed.
6- We need to further discuss the effects of warming water, mainly to explain the observed oxygen decrease in the lower thermocline of the loop current water, as a decadal change that can be adding to the oxygen variability. But the interannual oxygen variability that we observe doesn’t seem to be product of a thermocline warming as observed.
(7) A simple box model is not enough to reveal the mechanism of how LCE activity affects the water deoxygenation in the main thermoclines as the model neglects or simplifies too much critical processes like air-sea oxygen exchanges, advection, diffusion, and biogeochemical processes. The authors may need to seek another numerical tool like a coupled 3D model to reproduce the oxygen dynamics and LCE activity.
7- Right now we are working with the NEMO-PISCES coupled model as a way to better understand the oxygen dynamics of the GoM. But for the scope of this work, as a more descriptive type, the simplest box model used to illustrate the process as dominating the oxygen variability is more functional than to introduce a more complex model. We expect that eventually more observations the use of complex models will validate or invalidate our observations.
-
AC3: 'Reply ro referee 2', Jose Quintanilla, 29 Jun 2023
-
AC1: 'Comment on egusphere-2023-751', Jose Quintanilla, 05 Jun 2023
Dear reviewer, we appreciate your time for reading and reviewing this work, we are working on addressing your comments to improve our paper and getting it to a publishable state. We will post our proposed corrections as a response this week. Many thanks for your help.
Citation: https://doi.org/10.5194/egusphere-2023-751-AC1
Status: closed
-
RC1: 'Comment on egusphere-2023-751', Anonymous Referee #1, 14 May 2023
Review comments on manuscript egusphere-2023-751, titled “Deoxygenation of the Gulf of Mexico thermocline linked to a decrease in the detachment frequency of Loop Current Eddies” submitted to EGUsphere by Quintanilla et al.
In this manuscript, the authors studied the deoxygenation in the thermocline of the Gulf of Mexico using data from research cruises and one Argo buoy. Using a simple box model, they came to the conclusion that deoxygenation between 2010 and 2020 is caused by the reduced Loop Current Eddy area based on altimetry data. There are several fundamental issues in this study, which must be addressed before the publication of the paper.
- Oxygen is not a conservative variable in the ocean. Oxygen concentration or oxygen content is strongly influenced by biological processes (production and consumption) and air-sea oxygen flux near the sea surface. Although these two processes mainly influence the oxygen concentration in the upper ocean, which causes strong seasonal variations in the surface mixed layer, they can also influence the oxygen concentration in the thermocline via vertical mixing/diffusion, entrainment, downwelling etc. Horizontal advection, horizontal diffusion and mixing can also change the oxygen concentration. These were all ignored in the simple box model. To use the simple box model, the authors need to prove Loop Current Eddies is the dominant process to supply oxygen in the thermocline and all other processes can be ignored. Otherwise, the proposed box model cannot be used.
- Even though the box model can be used, it takes time for the oxygen in the Loop Current eddies to be mixed into the common water. That time scale needs to be estimated, which should be a function of horizontal diffusion. If the time scale is longer than several years, the model cannot be used directly. A time lag needs to be considered. The deoxygenation may not be observed immediately after the reduction of Loop current Eddy annual area if there is a lag. This is also critical to determine whether the box model is reasonable or not.
- The Argo O2 profiles used in this study were not adjusted/QCed. Based on the Argo QC manual (https://archimer.ifremer.fr/doc/00354/46542/), it may be necessary to correct the oxygen profiles from Argo, at least some quality check/quality assurance should be conducted before those profiles can be used. Data quality assurance is critical to the detection of long-term trends. Non-QC’d data should not be used in the trend study because any shift/bias may lead to inaccurate conclusions.
- Stratification and warming influence on deoxygenation in the thermocline should be considered. There are several recent publications showed recent warming in the GOM (e.g. https://journals.ametsoc.org/view/journals/phoc/51/4/JPO-D-19-0295.1.xml; https://journals.ametsoc.org/view/journals/clim/36/8/JCLI-D-22-0409.1.xml). Warming could cause oxygen solubility changes and stratification changes, which may contribute to the deoxygenation in the Gulf of Mexico. To what extent the deoxygenation in the study region is controlled by warming should be investigated and discussed.
- Seasonal variations of dissoved oxygen in the thermocline needs to be considered. The thermocline can still be influenced by seasonal changes. To determine the long-term trend, the seasonal cycle typically should be removed first from the signal because seasonal variations can be orders of magnitude larger than the long-term trend.
- It seems like in the box model, oxygen in the thermocline is only supplied by the Loop Current Eddies. Is this a reasonable assumption? Let’s assume an extreme case that no Loop Current Eddies are generated in the Gulf of Mexico. Based on the box model, the oxygen concentration in the thermocline will decrease to zero due to oxygen consumption. That’s not true. Because oxygen was measured in other oceans without Loop Current eddies. Surface mixed layer is a major source for oxygen in the thermocline via ventilation, entrainment due to mixed layer seasonal cycle and mixing/diffusion etc. The authors have already assumed deoxygenation in the thermocline by ignoring oxygen exchange between the surface mixed layer and thermocline, which is invalid. At least a two layer model is needed to consider the oxygen exchange between the upper ocean and the thermocline. Maybe the oxygen exchange between the upper and lower thermoclines should be considered as well.
- The possible reasons for the reduction of LCE area after 2010 needs to be further discussed and investigated.
- The data shown in the paper demonstrated strong horizontal difference in dissloved oxygen concentration in the thermocline. Because the cruise data were not always measured in the same locations. Therefore, horizontal difference might cause year to year difference in the average oxygen values if the data are not collected at same stations. The influence of sampling locations should be considered in this study.
Citation: https://doi.org/10.5194/egusphere-2023-751-RC1 -
AC2: 'Reply to referee 1', Jose Quintanilla, 29 Jun 2023
In this manuscript, the authors studied the deoxygenation in the thermocline of the Gulf of Mexico using data from research cruises and one Argo buoy. Using a simple box model, they came to the conclusion that deoxygenation between 2010 and 2020 is caused by the reduced Loop Current Eddy area based on altimetry data. There are several fundamental issues in this study, which must be addressed before the publication of the paper.
Dear referee 1, we appreciate your comments and suggestions on our manuscript, they will make this a better work, and we apologize for the time taken to submit our response.
We understand that the main issue of the work as presented is that it doesn’t provide enough evidence of the LCEs detachments as a mechanism that ventilates the Gulf of Mexico (GoM) thermocline and therefore how a decrease of LCE detachments can lead to a deoxygenation of this water volume. We believe that the following additions and corrections can make this paper a solid contribution to ocean science.
(1) Oxygen is not a conservative variable in the ocean. Oxygen concentration or oxygen content is strongly influenced by biological processes (production and consumption) and air-sea oxygen flux near the sea surface. Although these two processes mainly influence the oxygen concentration in the upper ocean, which causes strong seasonal variations in the surface mixed layer, they can also influence the oxygen concentration in the thermocline via vertical mixing/diffusion, entrainment, downwelling etc. Horizontal advection, horizontal diffusion and mixing can also change the oxygen concentration. These were all ignored in the simple box model. To use the simple box model, the authors need to prove Loop Current Eddies is the dominant process to supply oxygen in the thermocline and all other processes can be ignored. Otherwise, the proposed box model cannot be used.
1- In order to show evidence of why the variability of the Loop Current Eddies (LCEs) detachments is the dominant process that affects the variability of the oxygen in the GoM thermocline, while the other processes can be considered constant (but not ignored), we propose the following changes:
- We will add a more complete comparison of the oceanographic conditions measured between cruises XIXIMI-4, 5 and 6 (proposed figures 7 and 8 and table 5 in attached pdf). These three cruises took measurements during the summer season in three consecutive years, august 2015, june 2016 and august 2017, and share 24 stations located in the exact same coordinates that will be used for this comparison in order to reduce the horizontal differences stated in issue 8. Between the three expeditions there is a decline in the oxygen levels, going from 2.87 ml l-1 in 2015 to 2.8 ml l-1 in 2016 to 2.65 ml l-1 in 2017.
- The oxygen decrease rate between those cruises increased from 0.04 ml l-1 y-1 to 0.16 ml l-1 y-1 in the same depths. In comparison, a change of such magnitude in the Oxygen Utilization Rate (OUR) is only expected comparing water from 200m to water from 600m (as measured by Jenkins, 1982). We didn’t find data that would lead us to believe that the oxygen utilization rate could have spiked in that time producing the observed variability, but the Particulate Organic Carbon measured during the same XIXIMI cruises (reported in Contreras-Pacheco et al., 2023) doesn’t show an important variability in the upper 200m that could link the strong oxygen reduction to a strong variability of the primary productivity that could have led to a strong increase of the biological respiration.
- We couldn’t find any evidence of change in the thermocline vertical mixing of the GoM during those years in the literature, so we propose to add a T-S diagram (that was clearly lacking before) to better show the thermocline range and how from 2015 to 2017, there wasn’t any substantial change in the thermohaline properties of the thermocline that shows evidence of strong variability in the vertical mixing. The upper thermocline temperature was even a bit colder during XIXIMI-6 than during XIXIMI-5, showing that there wasn’t a warming of this volume between both cruises that could explain the important oxygen decrease.
- Oxygen diffusion variability is hard to measure but it is usually accepted that is a proses less important than oxygen advection
- The only variable that we can see a clear change in, is the number of LCEs that detached one year prior to each expedition, being three, two and zero from XIXIMI 4 to 5 to 6 respectively. If, as has already been proposed (Maul, 1980), the LCEs detachments work as pulses that introduce large volumes of water from the Caribbean to the GoM, and, as we show here and has already been largely observed, the Caribbean thermocline is more oxygenated than the GoM thermocline, then these pulses should be introducing important volumes of water with more oxygen inside the GoM. After these observations we will add a paragraph to clarify that the scope of this work isn’t to say that the OUR, the oxygen vertical mixing and the oxygen diffusion can be neglected inside the GoM, but to show that these processes can be considered constant (we called the sum of those processes as oxygen change rate and it is a key assumption for the model)
- We will better clarify that the box model presented is intended as a method to illustrate how the oxygen variability observed can be approximated using the LCEs detachment volume as the only source of variability, with all other processes considered constant. And to show how a decrease of LCEs detachments can lead a decrease of the oxygen levels in the GoM. It is not intended to estimate the oxygen concentration in an exact location or moment but to contribute as an approximation tool of the GoM thermocline oxygen levels.
(2) Even though the box model can be used, it takes time for the oxygen in the Loop Current eddies to be mixed into the common water. That time scale needs to be estimated, which should be a function of horizontal diffusion. If the time scale is longer than several years, the model cannot be used directly. A time lag needs to be considered. The deoxygenation may not be observed immediately after the reduction of Loop current Eddy annual area if there is a lag. This is also critical to determine whether the box model is reasonable or not.
2- The immediate mixing after detachment of the LCEs volumes with the GoM water is an extreme simplification of the model that we pointed as clearly false. The box model has the main simplification of assuming the whole GoM upper and lower thermoclines as horizontally homogeneous, we will better clarify that the model isn’t proposed as a tool to determine a precise oxygen concentration in a given moment in a given point of the GoM but as a tool to understand and visualize how the LCE variability can produce important oxygen variability inside the thermocline of the GoM. We tried another model that implemented a constant oxygen mixing rate between the LCE and the GoM during the LCE lifetime, but it introduced more complexity than needed and the final oxygen budget was the same. Currently we are working with a more complex model trying to eventually add a much more precise understanding of the ventilation processes inside the GoM. For now, the simplicity of the model allowed us to quickly add the latest three LCEs detachments since 2020, and compare the predictions to newly found BioARGO measurements from 2021 to the present (https://biogeochemical-argo.org/data-access.php).
(3) The Argo O2 profiles used in this study were not adjusted/QCed. Based on the Argo QC manual (https://archimer.ifremer.fr/doc/00354/46542/), it may be necessary to correct the oxygen profiles from Argo, at least some quality check/quality assurance should be conducted before those profiles can be used. Data quality assurance is critical to the detection of long-term trends. Non-QC’d data should not be used in the trend study because any shift/bias may lead to inaccurate conclusions.
3-The first BioARGO data presented was used to show that the oxygen data measured with the buoy is consistent with the oxygen measured and calibrated during 6 cruises. If the lack of additional correction of its data is an issue, that data can be eliminated from this work, as it wouldn’t affect the main observations provided by the XIXIMI cruises. However, after remarks from referee 2, we decided to add newly obtained data from a BioARGO buoy deployed in in the center of the GoM in October 2021 and still in operation. This data serves as another time series that evidences the oxygen ventilation by LCEs. The internal calibration of the buoy should be enough us the data as a time series whose function is to compare the variability.
(4) Stratification and warming influence on deoxygenation in the thermocline should be considered. There are several recent publications showed recent warming in the GOM (e.g. https://journals.ametsoc.org/view/journals/phoc/51/4/JPO-D-19-0295.1.xml; https://journals.ametsoc.org/view/journals/clim/36/8/JCLI-D-22-0409.1.xml). Warming could cause oxygen solubility changes and stratification changes, which may contribute to the deoxygenation in the Gulf of Mexico. To what extent the deoxygenation in the study region is controlled by warming should be investigated and discussed.
4- Thanks for pointing out the paper by Wang et al., 2022, we will cite it as it is very relevant. As observed, the temperature of the thermocline didn’t show interannual variability during the XIXIMI cruises, but as a multiannual to decadal trend the effect of thermocline warming must be more clearly addressed. The reported warming of the upper GoM could be a part of the deoxygenation trend that is observed in the Caribbean lower thermocline and must be kept under observation.
(5) Seasonal variations of dissoved oxygen in the thermocline needs to be considered. The thermocline can still be influenced by seasonal changes. To determine the long-term trend, the seasonal cycle typically should be removed first from the signal because seasonal variations can be orders of magnitude larger than the long-term trend.
5- We haven't found evidence of important seasonal variability in the thermocline of the GoM as the mixed layer is always above the thermocline (Portela et al., 2018). The upper thermocline oxygen measured during the XIXIMI-3 cruises in winter 2013 doesn’t show evidence of vertical mixing with upper waters with a mean potential temperature of 17.07 ± 0.6 ˚C and a mean salinity of 36.26 0.091 .
(6) It seems like in the box model, oxygen in the thermocline is only supplied by the Loop Current Eddies. Is this a reasonable assumption? Let’s assume an extreme case that no Loop Current Eddies are generated in the Gulf of Mexico. Based on the box model, the oxygen concentration in the thermocline will decrease to zero due to oxygen consumption. That’s not true. Because oxygen was measured in other oceans without Loop Current eddies. Surface mixed layer is a major source for oxygen in the thermocline via ventilation, entrainment due to mixed layer seasonal cycle and mixing/diffusion etc. The authors have already assumed deoxygenation in the thermocline by ignoring oxygen exchange between the surface mixed layer and thermocline, which is invalid. At least a two layer model is needed to consider the oxygen exchange between the upper ocean and the thermocline. Maybe the oxygen exchange between the upper and lower thermoclines should be considered as well.
6- The GoM is different from other oceans as a semi enclosed basin where the LCE detachments introduce large volumes of water as pulses (Meunier et al., 2020). The thermoclines of other oceans have different ventilation mechanisms that we will detail more. The box model does indeed assume that without LCE detachment the oxygen concentration of the GoM thermocline will eventually get to 0, because the oxygen change rate of -0.16ml l-1 y-1 is considered constant (as obtained from the extreme scenario of no LCE detachments that happened between cruises XIMIMI-5 and 6, with 19 months without detachments). This is another simplification that we state as clearly false as all the mechanism you pointed have their own variability. We will add the explanation that the GoM, as any oceanic system tend to balance themselves, a reduction of water entrance via LCEs detachment might be balanced by other processes, as horizontal and vertical advection and diffusion and biologic production and consumption that can change the oxygen change rate. We shall make a better bibliographic revision on that aspect to add in the discussion.
(7) The possible reasons for the reduction of LCE area after 2010 needs to be further discussed and investigated.
7- We will look for more information about the possible reasons for a reduction of the LCEs detachment frequency. Our first bibliographic research produced little results, we found that it is a subject that is still under debate and not clear enough. The hypothesis that we though were more substantial were that the decrease of frequency of LCEs could be connected to ocean warming via stratification of the Loop Current (Moreles et al., 2021) and that a slowdown of the Atlantic meridional circulation could affect the Loop Current. There is a lot to study and understand on this subject, we will keep on looking and hopefully we will be able to add a little more information about it in this paper.
(8) The data shown in the paper demonstrated strong horizontal difference in dissloved oxygen concentration in the thermocline. Because the cruise data were not always measured in the same locations. Therefore, horizontal difference might cause year to year difference in the average oxygen values if the data are not collected at same stations. The influence of sampling locations should be considered in this study.
8- As addressed in point 1, we think that this issue can be corrected by including a cruise comparison using only the stations located in the exact same coordinates.
-
RC2: 'Comment on egusphere-2023-751', Anonymous Referee #2, 29 May 2023
The authors demonstrated that the Gulf of Mexico thermocline is encountering a deoxygenation trend during the past decade (2010-2019) which is attributed to Loop Current Eddies (LCEs) activities. The findings are derived from observations from 6 oceanographic cruises and one BioARGO float and from a simple box model. Although the studied topic is interesting, I don’t think the manuscript is ready for publication.
(1) The thermocline shall not be demonstrated by potential density anomaly which is used for pycnoclines. If the “thermocline” is considered a studied object, potential temperature anomaly shall be shown and discussed. Otherwise, the authors would like to use “pycnocline” throughout the manuscript.
(2) The findings lack statistical support. The authors separated the manuscript into two main sections. In the first part, they provided evidence of deoxygenation of the gulf waters and changes in LC or LCE activity, while a box model is developed for further mechanistic studies in the second part. However, statistical tools like simple linear regression are lacking. For example, the trend of deoxygenation shall be detected by linear regression (with p values) not by the naked eye as the shown temporal differences in Table 1 are minor than the based values and are closed for different layers and geographical locations. So, the tune “seem to have” (e.g., Lines 212, 217, …) is weak and not convincing at all. Further, statistical evidence or mechanistic analysis is also needed to investigate the relationship between LCE and deoxygenated waters. The authors need to dig deeper when “building” such a linkage.
(3) More data is needed. The authors are studying mesoscale phenomena over a vast ocean basin from 2000 to 2020. So, I don’t think measurements from 6 cruises and 1 ARGO float are enough. I am not sure if there are any other cruise measurements over the south of 25°N, but, at least, there are many BioARGO profiles around the path of LC provided by other archive systems like the world ocean database. The author should try to explore more data.
(4) The temporal scale of interest may need to be reconsidered and clarified. The lifespan of LCE is typically within a year, however, the authors focus more on the decadal changes of LCE and its relationship to the decadal changes in dissolved oxygen concentration in the gulf. I think the relationship between LCE and gulf water deoxygenation should be stronger on an interannual scale. In addition, 20-yr data is not enough to study a decadal phenomenon but is more suitable for interannual or even intra-annual scales.
(5) Even though the authors can statistically “prove” that changes in the gulf DO concentration are related to changes in LCE activity, the study is carried out under a major consumption that other factors that contribute to DO changes like air-sea oxygen exchanges, biological processes in the upper ocean, and DO transports due to advection and diffusion are minor or negligible. To my understanding, it is usually not the case. Contributions by all these factors should be quantified and discussed before the authors can “prove” that LCE activity outcompetes others.
(6) Again, as factors like air-sea oxygen exchanges are ignored in this study, changes in oxygen solubility due to temperature changes are ignored. Many studies indicate that the Gulf of Mexico is warming in recent decades. Decadal changes in water temperature are apparent which will lead to a decrease of oxygen solubility and further water deoxygenation. Thus, the effect of warming water should be discussed.
(7) A simple box model is not enough to reveal the mechanism of how LCE activity affects the water deoxygenation in the main thermoclines as the model neglects or simplifies too much critical processes like air-sea oxygen exchanges, advection, diffusion, and biogeochemical processes. The authors may need to seek another numerical tool like a coupled 3D model to reproduce the oxygen dynamics and LCE activity.
Citation: https://doi.org/10.5194/egusphere-2023-751-RC2 -
AC3: 'Reply ro referee 2', Jose Quintanilla, 29 Jun 2023
The authors demonstrated that the Gulf of Mexico thermocline is encountering a deoxygenation trend during the past decade (2010-2019) which is attributed to Loop Current Eddies (LCEs) activities. The findings are derived from observations from 6 oceanographic cruises and one BioARGO float and from a simple box model. Although the studied topic is interesting, I don’t think the manuscript is ready for publication.
Dear referee, we appreciate your comments and suggestions on our manuscript, they will make this a better work, and we apologize for the time taken to submit our response. We hope that the following additions and corrections can improve this paper.
(1) The thermocline shall not be demonstrated by potential density anomaly which is used for pycnoclines. If the “thermocline” is considered a studied object, potential temperature anomaly shall be shown and discussed. Otherwise, the authors would like to use “pycnocline” throughout the manuscript.
1 - We understand that the paper needs to present a T-S diagram to show the potential temperature and salinity values that defines the potential density range for the upper and lower thermocline. We will add a T-S figure of all the data used with the color code from the upper thermocline oxygen ranges to show how the stations with the highest oxygen show the subsurface salinity of the Caribbean water (proposed figure 7 in attached pdf) . Another T-S diagram will be included in a figure comparing the oxygen variability for three consecutive summer cruises, using only 24 stations that are located in the exact same coordinates, to show how there is an important oxygen decrease in the upper thermocline from june 2016 to august 2017, period without any LCE detachments, but that the temperature and salinity remain close to constant (figure 8).
(2) The findings lack statistical support. The authors separated the manuscript into two main sections. In the first part, they provided evidence of deoxygenation of the gulf waters and changes in LC or LCE activity, while a box model is developed for further mechanistic studies in the second part. However, statistical tools like simple linear regression are lacking. For example, the trend of deoxygenation shall be detected by linear regression (with p values) not by the naked eye as the shown temporal differences in Table 1 are minor than the based values and are closed for different layers and geographical locations. So, the tune “seem to have” (e.g., Lines 212, 217, …) is weak and not convincing at all. Further, statistical evidence or mechanistic analysis is also needed to investigate the relationship between LCE and deoxygenated waters. The authors need to dig deeper when “building” such a linkage.
2- We will add a linear regression using the XIXIMI data and historical data, a linear regression using XIXIMI data can be shown in figure 6, the r2 values are low at 0.36 and 0.66 in the upper and lower thermocline respectively. Those are low values but as we want to convey the thermocline oxygen shows high variability between years product of the high LCE detachment variability. Using the historic data, the r2 values increase to 0.6 and 0.9 in the upper and lower thermocline.
Although the biological oxygen utilization rate (OUR), and the oxygen vertical mixing and diffusion rates must vary from year to year the observable variable that can explain this variability is the LCEs detachments. We can add a statistical correlation analysis like a Pearson’s correlation or a principal component analysis but we believe that the box model proposed illustrates more clearly how using the LCEs detachments as the only source of variability the oxygen values can be approximated.
(3) More data is needed. The authors are studying mesoscale phenomena over a vast ocean basin from 2000 to 2020. So, I don’t think measurements from 6 cruises and 1 ARGO float are enough. I am not sure if there are any other cruise measurements over the south of 25°N, but, at least, there are many BioARGO profiles around the path of LC provided by other archive systems like the world ocean database. The author should try to explore more data.
3- We agree that more data is needed to make a more solid study, but there is a deficit of measurements in the Mexican area of the GoM, the historical data presented in this manuscript was intended as a recompilation of published measurements. We think that the data presented from 6 one-month cruises with more than 40 stations each, is a solid addition of data and a strong time series that can really useful. We will add new data from a BioARGO buoy in the center of the gulf that we just found, deployed in October 2021 and that is still taking data. The data from that buoy can be added to show how the box model still estimates the measured values adding the LCEs detachments observed after 2020 (as proposed in figures 9 and 6). According to our estimates in January 2020 there was one big LCE detachment, then nothing until another detachment in august 2021 followed by another detachment of a really big LCE in July 2022, there hasn’t been any other LCE detachments since. The BioARGO time series show a decrease in oxygen as predicted, and it will keep on adding data, we expect it to show further O2 decrease as long as there is not another LCE detachment and the buoy stays in the center of the GoM.
(4) The temporal scale of interest may need to be reconsidered and clarified. The lifespan of LCE is typically within a year, however, the authors focus more on the decadal changes of LCE and its relationship to the decadal changes in dissolved oxygen concentration in the gulf. I think the relationship between LCE and gulf water deoxygenation should be stronger on an interannual scale. In addition, 20-yr data is not enough to study a decadal phenomenon but is more suitable for interannual or even intra-annual scales.
4- Great point, the observed variability and deoxygenation is on a interannual scale and we must make emphasis on that point. The reason the decadal scale is mentioned is because comparing the 2000 -2009 decade to the 2010-2019 decade there is a difference in the detachment frequency and total LCE area that matches the oxygen differences observed between measurements made in the 2000s and measurements made in the 2010s.
(5) Even though the authors can statistically “prove” that changes in the gulf DO concentration are related to changes in LCE activity, the study is carried out under a major consumption that other factors that contribute to DO changes like air-sea oxygen exchanges, biological processes in the upper ocean, and DO transports due to advection and diffusion are minor or negligible. To my understanding, it is usually not the case. Contributions by all these factors should be quantified and discussed before the authors can “prove” that LCE activity outcompetes others.
5 – Those processes aren’t ignored but considered constant and are included as the constant value that we called oxygen change rate. We stated that it is a big simplification and that is clearly false, we will add more evidence in order to prove that the LCEs detachment variability “outcompetes” other processes of why the sum of those processes can be considered constant. With the T-S diagrams and thermohaline values we expect to show that there is not evidence of vertical mixing variability, with particulate organic carbon data obtained from the same XIXIMI cruises (Contreras-Pacheco et al., 2023) we can show that there is no evidence of an increase in the organic matter concentration that could explain an OUR increase as observed. And we are still looking for a way to find evidence of oxygen diffusion change but there is not enough data.
(6) Again, as factors like air-sea oxygen exchanges are ignored in this study, changes in oxygen solubility due to temperature changes are ignored. Many studies indicate that the Gulf of Mexico is warming in recent decades. Decadal changes in water temperature are apparent which will lead to a decrease of oxygen solubility and further water deoxygenation. Thus, the effect of warming water should be discussed.
6- We need to further discuss the effects of warming water, mainly to explain the observed oxygen decrease in the lower thermocline of the loop current water, as a decadal change that can be adding to the oxygen variability. But the interannual oxygen variability that we observe doesn’t seem to be product of a thermocline warming as observed.
(7) A simple box model is not enough to reveal the mechanism of how LCE activity affects the water deoxygenation in the main thermoclines as the model neglects or simplifies too much critical processes like air-sea oxygen exchanges, advection, diffusion, and biogeochemical processes. The authors may need to seek another numerical tool like a coupled 3D model to reproduce the oxygen dynamics and LCE activity.
7- Right now we are working with the NEMO-PISCES coupled model as a way to better understand the oxygen dynamics of the GoM. But for the scope of this work, as a more descriptive type, the simplest box model used to illustrate the process as dominating the oxygen variability is more functional than to introduce a more complex model. We expect that eventually more observations the use of complex models will validate or invalidate our observations.
-
AC3: 'Reply ro referee 2', Jose Quintanilla, 29 Jun 2023
-
AC1: 'Comment on egusphere-2023-751', Jose Quintanilla, 05 Jun 2023
Dear reviewer, we appreciate your time for reading and reviewing this work, we are working on addressing your comments to improve our paper and getting it to a publishable state. We will post our proposed corrections as a response this week. Many thanks for your help.
Citation: https://doi.org/10.5194/egusphere-2023-751-AC1
Data sets
Deoxygenation of the Gulf of Mexico thermocline linked to a decrease in the detachment frequency of Loop Current Eddies José Quintanilla, Juan Carlos Herguera, and Julio Sheinbaum https://doi.org/10.5281/zenodo.7830465
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 428 | 166 | 45 | 639 | 31 | 33 |
- HTML: 428
- PDF: 166
- XML: 45
- Total: 639
- BibTeX: 31
- EndNote: 33
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 309 | 47 |
| Mexico | 2 | 59 | 9 |
| China | 3 | 42 | 6 |
| Germany | 4 | 32 | 4 |
| Ireland | 5 | 21 | 3 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 309