the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Ideas and perspectives: The benthic iron flux from sandy advective bioturbated sediments
Laura M. Wehrmann
Darci A. Swenson Perger
Ian P. Dwyer
Nils Volkenborn
Christina Heilbrun
Robert C. Aller
Abstract. Multiple investigators have suggested that the benthic flux of dissolved iron (Fed) from continental shelf sediments represents an important source of this micronutrient to ocean waters. The magnitude, biogeochemical controls, and seasonal dynamics of Fed fluxes to date, however, have mostly been studied for muddy cohesive sediments dominated by molecular diffusion. Data from these studies have been included in global biogeochemical models to determine the contribution of this Fe source to the ocean. Fed fluxes from sandy advective sediments have received little consideration, although these sediments cover 50–60 % of the continental shelves. Sandy permeable deposits function as dynamic catalytic filters characterized by the rapid exchange of solutes and infiltration of particles —including labile Corg and reactive metal oxides— and high biogeochemical reaction rates. In this article, we discuss how the fundamentally different modes of solute and particle transport in sands affect the sedimentary Fe cycle and Fed flux. We present a case study in which we simulate bioirrigation in sands in summer and winter. In our experiments, Fed fluxes from non-irrigated sediments under diffusive conditions did not exceed 6 and 13 μmol Fe m-2 d-1 in winter and summer, respectively. Fluxes from irrigated cores reached values of 150 μmol Fe m-2 d-1 (winter) and 115 μmol Fe m-2 d-1 (summer). The results indicate that the pumping activity of the benthic macrofauna plays a key role in controlling the extent of the benthic Fed flux from permeable sediments, and that both biogenic and physical advection enhance fluxes. We argue that bioturbated sandy advective sediments constitute an important benthic Fe source to coastal waters and advocate for a more differentiated treatment of sediment type (muddy diffusive vs. sandy advective) and macrofaunal activity -reflecting different functional groups of the macrobenthos- in global biogeochemical Fe models. A better understanding of the benthic Fe cycle in sandy advective sediments is particularly important to help predict how anthropogenic effects such as changes in the deposition patterns of Corg and metals, the expansion of oxygen minimum zones, and changes in benthic biodiversity will affect the tightly coupled benthic-pelagic ecosystem along continental shelves.
- Preprint
(788 KB) - Metadata XML
- BibTeX
- EndNote
Laura M. Wehrmann et al.
Status: closed
-
RC1: 'Comment on bg-2022-247', Sebastiaan van de Velde, 14 Jan 2023
Wehrmann et al. argue in this MS that dissolved iron (Fed) fluxes from sandy sediments – which have received little attention – are an important source of iron for the ocean. They present a case study to illustrate that Fed fluxes from sandy sediments with simulated bioirrigation are several times higher than in the absence of irrigation, and that these fluxes are in the higher range of fluxes measured from muddy deposits.
While I agree with the overall message of the perspective paper, and do not question the quality of the presented results, I have one main comment on the experimental design and conclusions drawn from the results. Because the way the experiment is designed (and the authors acknowledge this during the discussion, e.g. L243), you end up with a sandy deposit that is diffusion dominated. In reality however, as discussed in the introduction, sandy deposits are advection dominated – irregardless of the presence or absence of bio-irrigators. As a results, when you compare your irrigated results with your non-irrigated results, you can say something about the the importance of irrigation relative to diffusion – but you cannot make an accurate assessment of how important irrigation would be under natural and advective conditions (which is what you do on L281 and L363), since you have no control that includes advection without irrigation. I would ask the authors to consider this in their discussion – and either provide a quantitative argument as to why irrigation is important in sandy deposits relative to advection, or rephrase the discussion so it focuses on advective sandy sediments in general, and does not go into too much detail with respect to the different types of irrigators or other benthic fauna. In essence, this would link better with the introduction, where you do explain how sandy deposits w/o bioturbators are distinctly different from muddy deposits.
A second concern is the comparison of Fed fluxes from this study with the literature (L280-283). Since you use a novel method, the discrepancy is likely a consequence of your new method to measure Fed fluxes. Dale et al. base themselves on the available data, which is almost exclusively collected by the traditional flux incubation methods. They then calibrate their model on the available fluxes, which are lower than what you would have measured with your extractor-method.
So you could also make the argument that both muddy as sandy deposits are roughly comparable as Fed sources, but our older methods underestimate the actual flux. Please also consider this difference in methods in the discussion.I am however confident that the authors can address these concerns and the smaller remarks listed below. The message of the manuscript is an important and valid one, and worthy to be published. I will happily read a revised version.
Kind regards
Sebastiaan van de Velde
Minor comments:L32: or through acidic dissolution of FeS in sediments colonized by cable bacteria (Rao et al., 2016; Seitaj et al., 2015; Sulu-Gambari et al., 2016; van de Velde et al., 2016)
L40: could use a reference
L59: I would perhaps also say that this is due to practical limitations regarding the in-situ sampling and processing of sandy sediments.
L60: and the fact that most of these sands are dominated by aerobic respiration and denitrification, thus little dissolved Fe(2+) is expected to accumulate and be released from these sediments.
L63: clarify if this is µmol cm-3 sediments or solid phase
L59-70: This argument is based on the extracteable Fe contents – but you would also need to take into account the oxygenation state of sandy versus muddy deposits (how deep is the sediment oxygenated?).
L101: or anthropogenic activities (trawling, dredging) (van de Velde et al., 2018)
L197: does this compare to the natural populations? (I now see you say this at L296, I would move this upward)
L301ff: Or if Corg gets too high, most iron will be rapdily precipitated to FeS(2) and not recycled to the overlying water (van de Velde et al., 2020a, 2020b)
Fig. 1: there is no label on the x-axis
Fig. 2: and what is happening in the rest of the sediment? I also assume this conceptuel model is based on actual measurements of the O2 distribution, which you might either refer to or show alongside the image.
References:
Rao, A. M. F., Malkin, S. Y., Hidalgo-Martinez, S. and Meysman, F. J. R.: The impact of electrogenic sulfide oxidation on elemental cycling and solute fluxes in coastal sediment, Geochim. Cosmochim. Acta, 172, 265–286, doi:10.1016/j.gca.2015.09.014, 2016.
Seitaj, D., Schauer, R., Sulu-gambari, F., Hidalgo-martinez, S., Malkin, S. Y., Burdorf, L. D. W., Slomp, C. P. and Meysman, F. J. R.: Cable bacteria generate a firewall against euxinia in seasonally hypoxic basins, Proc. Natl. Acad. Sci., 112(43), 13278–13283, doi:10.1073/pnas.1510152112, 2015.
Sulu-Gambari, F., Seitaj, D., Behrends, T., Banerjee, D., Meysman, F. J. R. and Slomp, C. P.: Impact of cable bacteria on sedimentary iron and manganese dynamics in a seasonally-hypoxic marine basin, Geochim. Cosmochim. Acta, 192(2016), 49–69, doi:10.1016/j.gca.2016.07.028, 2016.
van de Velde, S. J., Lesven, L., Burdorf, L. D. W., Hidalgo-Martinez, S., Geelhoed, J. S., Van Rijswijk, P., Gao, Y. and Meysman, F. J. R.: The impact of electrogenic sulfur oxidation on the biogeochemistry of coastal sediments: a field study, Geochim. Cosmochim. Acta, 194, 211–232, doi:10.1016/j.gca.2016.08.038, 2016.
van de Velde, S. J., Van Lancker, V., Hidalgo-Martinez, S., Berelson, W. M. and Meysman, F. J. R.: Anthropogenic disturbance keeps the coastal seafloor biogeochemistry in a transient state, Sci. Rep., 8(1), 1–10, doi:10.1038/s41598-018-23925-y, 2018.
van de Velde, S. J., Reinhard, C. T., Ridgwell, A. and Meysman, F. J. R.: Bistability in the redox chemistry of sediments and oceans, Proc. Natl. Acad. Sci., 117(52), 33043–33050, doi:10.1073/pnas.2008235117, 2020a.
van de Velde, S. J., Hidalgo-Martinez, S., Callebaut, I., Antler, G., James, R., Leermakers, M. and Meysman, F.: Burrowing fauna mediate alternative stable states in the redox cycling of salt marsh sediments, Geochim. Cosmochim. Acta, 276, 31–49, doi:10.1016/j.gca.2020.02.021, 2020b.
Citation: https://doi.org/10.5194/bg-2022-247-RC1 -
AC2: 'Reply on RC1', Laura Wehrmann, 03 Mar 2023
We thank the reviewer Dr. van de Velde for his insightful and constructive comments to our manuscript which helped to improve its clarity and focus. We hope that our revisions and answers to the comments are sufficient to accept this work for publication in Biogeosciences. Please find our responses to the individual comments below.
Wehrmann et al. argue in this MS that dissolved iron (Fed) fluxes from sandy sediments – which have received little attention – are an important source of iron for the ocean. They present a case study to illustrate that Fed fluxes from sandy sediments with simulated bioirrigation are several times higher than in the absence of irrigation, and that these fluxes are in the higher range of fluxes measured from muddy deposits.
While I agree with the overall message of the perspective paper, and do not question the quality of the presented results, I have one main comment on the experimental design and conclusions drawn from the results. Because the way the experiment is designed (and the authors acknowledge this during the discussion, e.g. L243), you end up with a sandy deposit that is diffusion dominated. In reality however, as discussed in the introduction, sandy deposits are advection dominated – irregardless of the presence or absence of bio-irrigators. As a results, when you compare your irrigated results with your non-irrigated results, you can say something about the the importance of irrigation relative to diffusion – but you cannot make an accurate assessment of how important irrigation would be under natural and advective conditions (which is what you do on L281 and L363), since you have no control that includes advection without irrigation. I would ask the authors to consider this in their discussion – and either provide a quantitative argument as to why irrigation is important in sandy deposits relative to advection, or rephrase the discussion so it focuses on advective sandy sediments in general, and does not go into too much detail with respect to the different types of irrigators or other benthic fauna. In essence, this would link better with the introduction, where you do explain how sandy deposits w/o bioturbators are distinctly different from muddy deposits.
As previously discussed in several publications, for example, by Aller, 2001 and Meysman et al. 2006, the burrow flushing induced by the pumping activity of macrofauna, i.e., bioirrigation, represents an advective mode of water transport. In our experiment, we use irrigation mimics to inject water into the sediment at depth following the typically activity of C. torquata. This injection drives the advective transport of pore water upward, forcing fluid out of the sediment. Thus, our irrigated cores must be described as advection-dominated with diffusion only controlling the transport regime during times of non-pumping/irrigation. What the reviewer refers to as advection is (non-biogenic) advective transport of pore water induced by pressure gradients, for example, due to the unidirectional flow of bottom currents across an uneven sediment surface or density gradients. The reviewer is correct that we do not simulate this type of advective transport in our experiments, as previously demonstrated using recirculating flumes or natural bedforms (e.g., Huettel et al, 1996; 1998). As outlined above, our focus is on biogenic advection induced by bioirrigation at depth, which is superimposed on any physically-driven advection. Bioirrigation enhances advection and can occur in the absence of physical forcing. We focused on isolating these effects using realistic irrigation volumes and patterns. We agree with the reviewer, that we should state this more carefully -and highlight that an additional physically induced advective transport component would likely alter the overall transport dynamics. The reviewer is correct that our “non-irrigated cores” are diffusion-dominated not physical advection-dominated as would likely be the case for sandy deposits in the absence of macrofauna. We used the diffusion case as a Fe flux end-member. We noted that the magnitude of Fe flux that we measured in this case closely corresponds to the values reported in the literature from benthic chamber measurements in which advection is suppressed. We have changed our manuscript discussion to clarify these points.
A second concern is the comparison of Fed fluxes from this study with the literature (L280-283). Since you use a novel method, the discrepancy is likely a consequence of your new method to measure Fed fluxes. Dale et al. base themselves on the available data, which is almost exclusively collected by the traditional flux incubation methods. They then calibrate their model on the available fluxes, which are lower than what you would have measured with your extractor-method.
This is an important point the reviewer brings up. We discuss the details of our new flux incubation method and provide a comprehensive comparison to other flux methods in a recent paper “Estimating benthic dissolved Fe and reactive solute fluxes” (Aller, Dwyer, Swenson, Heilbrun, Volkenborn, Wehrmann) published in Marine Chemistry (https://doi.org/10.1016/j.marchem.2023.104221). However, to guide the reader, we now explain the potential deviations in the approaches to estimating fluxes in our Introduction section.
So you could also make the argument that both muddy as sandy deposits are roughly comparable as Fed sources, but our older methods underestimate the actual flux. Please also consider this difference in methods in the discussion.
It is indeed possible that muds and sands produce roughly comparable Fed sources but for different reasons. We wrote this Perspectives paper to emphasize our lack of knowledge in this regard. We simply don’t have sufficient data on sandy deposits. Thus, we don’t know whether fluxes from muds or sands are in a similar range. It is also possible that because sands are open systems constantly extracting reactive particles from overlying water, that the Fed flux from bioturbated sands exceeds that from muds. Again, we don’t know.
I am however confident that the authors can address these concerns and the smaller remarks listed below. The message of the manuscript is an important and valid one, and worthy to be published. I will happily read a revised version.
Kind regards
Sebastiaan van de Velde
Minor comments:L32: or through acidic dissolution of FeS in sediments colonized by cable bacteria (Rao et al., 2016; Seitaj et al., 2015; Sulu-Gambari et al., 2016; van de Velde et al., 2016)
Added.
L40: could use a reference
Added.
L59: I would perhaps also say that this is due to practical limitations regarding the in-situ sampling and processing of sandy sediments.
Added.
L60: and the fact that most of these sands are dominated by aerobic respiration and denitrification, thus little dissolved Fe(2+) is expected to accumulate and be released from these sediments.
Again, we don’t always know that because Fe cycling in sandy deposits remains so little investigated. However, some recent studies, e.g., by Jahnke et al., 2005 and Zhou et al., 2023, indicate that DIR may also play an important role for OM remineralization in sandy deposits.
L63: clarify if this is µmol cm-3 sediments or solid phase
Done.
L59-70: This argument is based on the extracteable Fe contents – but you would also need to take into account the oxygenation state of sandy versus muddy deposits (how deep is the sediment oxygenated?).
Yes, here we highlight that there may be enough extractable Fe to support significant DIR (and ultimately benthic Fe fluxes). Overall, the presence of oxygen is indeed a key factor in the release of Fe from sediments. However, as we also point out, sandy deposits are often characterized by a very patchy distribution of redox zones due to their permeable nature and fast redox dynamics. Rather than a one-dimensional vertical view of O2 (as well as H2S) distribution a 3D-depiction of the redox distribution is more appropriate to characterize these deposits.
L101: or anthropogenic activities (trawling, dredging) (van de Velde et al., 2018)
Added.
L197: does this compare to the natural populations? (I now see you say this at L296, I would move this upward)
We added the information to the method section.
L301ff: Or if Corg gets too high, most iron will be rapdily precipitated to FeS(2) and not recycled to the overlying water (van de Velde et al., 2020a, 2020b)
Revised.
Fig. 1: there is no label on the x-axis
Done. Thanks!
Fig. 2: and what is happening in the rest of the sediment? I also assume this conceptuel model is based on actual measurements of the O2 distribution, which you might either refer to or show alongside the image.
Added.
References:
Rao, A. M. F., Malkin, S. Y., Hidalgo-Martinez, S. and Meysman, F. J. R.: The impact of electrogenic sulfide oxidation on elemental cycling and solute fluxes in coastal sediment, Geochim. Cosmochim. Acta, 172, 265–286, doi:10.1016/j.gca.2015.09.014, 2016.
Seitaj, D., Schauer, R., Sulu-gambari, F., Hidalgo-martinez, S., Malkin, S. Y., Burdorf, L. D. W., Slomp, C. P. and Meysman, F. J. R.: Cable bacteria generate a firewall against euxinia in seasonally hypoxic basins, Proc. Natl. Acad. Sci., 112(43), 13278–13283, doi:10.1073/pnas.1510152112, 2015.
Sulu-Gambari, F., Seitaj, D., Behrends, T., Banerjee, D., Meysman, F. J. R. and Slomp, C. P.: Impact of cable bacteria on sedimentary iron and manganese dynamics in a seasonally-hypoxic marine basin, Geochim. Cosmochim. Acta, 192(2016), 49–69, doi:10.1016/j.gca.2016.07.028, 2016.
van de Velde, S. J., Lesven, L., Burdorf, L. D. W., Hidalgo-Martinez, S., Geelhoed, J. S., Van Rijswijk, P., Gao, Y. and Meysman, F. J. R.: The impact of electrogenic sulfur oxidation on the biogeochemistry of coastal sediments: a field study, Geochim. Cosmochim. Acta, 194, 211–232, doi:10.1016/j.gca.2016.08.038, 2016.
van de Velde, S. J., Van Lancker, V., Hidalgo-Martinez, S., Berelson, W. M. and Meysman, F. J. R.: Anthropogenic disturbance keeps the coastal seafloor biogeochemistry in a transient state, Sci. Rep., 8(1), 1–10, doi:10.1038/s41598-018-23925-y, 2018.
van de Velde, S. J., Reinhard, C. T., Ridgwell, A. and Meysman, F. J. R.: Bistability in the redox chemistry of sediments and oceans, Proc. Natl. Acad. Sci., 117(52), 33043–33050, doi:10.1073/pnas.2008235117, 2020a.
van de Velde, S. J., Hidalgo-Martinez, S., Callebaut, I., Antler, G., James, R., Leermakers, M. and Meysman, F.: Burrowing fauna mediate alternative stable states in the redox cycling of salt marsh sediments, Geochim. Cosmochim. Acta, 276, 31–49, doi:10.1016/j.gca.2020.02.021, 2020b.
All added.
Citation: https://doi.org/10.5194/bg-2022-247-AC2
-
AC2: 'Reply on RC1', Laura Wehrmann, 03 Mar 2023
-
RC2: 'Comment on bg-2022-247', Anonymous Referee #2, 24 Jan 2023
The manuscript by Wehrmann et al. focuses on the problem of benthic iron fluxes on continental margins. The scientific background to this problem is well described in the introductory chapter. The text then describes experiments where measurements of dissolved iron fluxes at the water-sediment interface were performed with permeable sands. In one case, pore water does not flow through the sediment, in the other case, a pumping system forces water flow and simulates the bio-irrigation activity of a common annelid in West Atlantic margin sediments. The experiment was conducted under both winter and summer temperature conditions. The idea of experimentally examining iron flux from a bioturbated sediment is interesting, however, I do not understand why the manuscript is submitted to Biogeosciences under "ideas and perspectives", since the manuscript describes and discusses the results of an experiment.
The experiments here lasting 7 and 12 days show that benthic fluxes of dissolved iron are higher for permeable bio-irrigated sediments than for non-irrigated sediments. Oxygen levels remained high in the water column during the experiments. At the end of the experiments, the dissolved ammonium inventory in the pore water was measured. This was 4 to 9 times higher in the non-irrigated sediments. In view of the samples that were collected and the discussion, it is surprising not to see some additional results that would have strongly supported the conclusions: only iron fluxes were measured, only oxygen was monitored in the water column, and only ammonium concentrations were measured in the pore water. Thus many questions remained unanswered.
Since the samples were available, why were dissolved nitrogen compounds not monitored in the water column? Why were the fluxes of other biogenic compounds not measured? What was the inventory of dissolved iron in the pore water? What is the proportion of iron that crossed the water-sediment interface compared to the initial inventory of pore water iron? The experiments last 7 and 12 days, what justifies this duration?
The dimensions of the benthic chambers indicate that the volume of pore water was 1100 ml. The water injected daily into the sediment to simulate bioirrigation was 281 ml, suggesting a turnover time of about 4 days. Thus I wondered if the experiment simply flushed the pore water initially rich in products of benthic biogeochemical processes. The experiment seems to have mimicked an environment where one would start from an initial situation without bioturbation followed by a situation with bioturbating organisms. Figure 1 shows the evolution of iron fluxes as a function of time. It shows a decrease with time: this suggests that pore waters are progressively replaced by seawater. The duration of the experiment did not allow to reach a stationary state. The measured fluxes thus seem to be essentially dictated by the initial state where dissolved iron in the pore water was abundant before irrigation. If the experiment seeks to simulate continental shelf sands colonized by benthic organisms, I believe that water should first be circulated for a sufficient amount of time before benthic fluxes are considered. One suggestion is to allow time for three times the volume of pore water to be renewed before making flux measurements, to avoid simply measuring fluxes from flushing the water prior to bioirrigation. In other words, I would tend to think that the results of the experiment should have been considered from the moment it was stopped here, i.e. after a dozen days. Thus, the extrapolation of the data produced here for the entire continental shelf presented in Figure 3 seems to me insufficiently justified. I therefore suggest that the authors better justify the durations of the experiments and better describe the volumes of water involved and the initial state. I would be happy to review a corrected version of the manuscript in which my remarks have been considered.
Citation: https://doi.org/10.5194/bg-2022-247-RC2 -
AC1: 'Reply on RC2', Laura Wehrmann, 03 Mar 2023
We thank Reviewer #2 for the insightful and valuable comments on our manuscript that helped to improve our method description and discussion. We also added a figure that shows Fe inventories at the beginning and end of the experiment to address some of the reviewers’ concerns. We hope that our revisions and answers to the comments are sufficient to accept this work for publication in Biogeosciences. Please find our responses to the individual comments below.
The manuscript by Wehrmann et al. focuses on the problem of benthic iron fluxes on continental margins. The scientific background to this problem is well described in the introductory chapter. The text then describes experiments where measurements of dissolved iron fluxes at the water-sediment interface were performed with permeable sands. In one case, pore water does not flow through the sediment, in the other case, a pumping system forces water flow and simulates the bio-irrigation activity of a common annelid in West Atlantic margin sediments. The experiment was conducted under both winter and summer temperature conditions. The idea of experimentally examining iron flux from a bioturbated sediment is interesting, however, I do not understand why the manuscript is submitted to Biogeosciences under "ideas and perspectives", since the manuscript describes and discusses the results of an experiment.
The Ideas and perspectives manuscripts “report new ideas and novel aspects of scientific investigations” and was chosen by us to discuss the clear gaps in our knowledge of the benthic Fe flux from sandy permeable sediments which cover a large part of continental shelfs. We only present a small dataset as a case study to provide an initial idea of the expected scale of these fluxes from sandy deposits. As we highlight in our discussion, there is a clear need for further studies investigating this topic in detail. Our manuscript aims to spark further interest and encourage such investigations.
The experiments here lasting 7 and 12 days show that benthic fluxes of dissolved iron are higher for permeable bio-irrigated sediments than for non-irrigated sediments. Oxygen levels remained high in the water column during the experiments. At the end of the experiments, the dissolved ammonium inventory in the pore water was measured. This was 4 to 9 times higher in the non-irrigated sediments. In view of the samples that were collected and the discussion, it is surprising not to see some additional results that would have strongly supported the conclusions: only iron fluxes were measured, only oxygen was monitored in the water column, and only ammonium concentrations were measured in the pore water. Thus many questions remained unanswered.
Since the samples were available, why were dissolved nitrogen compounds not monitored in the water column? Why were the fluxes of other biogenic compounds not measured?
This article specifically focuses on the cycling and fluxes of iron from sandy permeable sediments and we thus did not analyze any nitrogen compounds other than ammonium concentrations in pore-waters which were used get insight into differences in the overall remineralization signals in the non-irrigated and irrigate cores. As pointed out in our Introduction section, the nitrogen cycle in these deposits has been the focus of previous studies (e.g., Rao et al., 2007; Marchant et al., 2016) while little is known about Fe fluxes. The Fe accumulator we developed is specific to the determination of Fe (and potentially P and other redox-sensitive trace metals; Aller et al., 2023) while it does not trap dissolved inorganic carbon, ammonium and other biogenic compounds and thus their fluxes cannot be estimated using this technique.
What was the inventory of dissolved iron in the pore water? What is the proportion of iron that crossed the water-sediment interface compared to the initial inventory of pore water iron? The experiments last 7 and 12 days, what justifies this duration?
These are important points. We now show our calculations of the dissolved Fe inventories in initial and end cores and show these data in Figure 2. The Fe inventory in the winter experiment increased ~15 fold in non-irrigated cores and ~8 fold in the irrigated cores compared to the initial cores, while they stayed in the same range throughout the experiment in the summer cores. These data indicate that the observed Fe flux does not just reflect a “flushing” of the sediment and transport of Fe produced prior to irrigation out of the sediment but reflect Fe(II) production (driven by OM remineralization) and a build-up in the inventory (especially in winter) despite the advective transport of Fe out of the sediment.
The timescale of these types of experiments is dictated by the characteristic reactivity of organic matter, usually 14 - 24 days for fresh plankton. Thus, the experiments have to remain sufficiently short to avoid reactive OM depletion and significant changes in dissimilatory Fe reduction. Given the higher temperatures in summer, we decided to keep the summer experiment shorter than the winter experiment. In nature, the system is open and reactive particles are resupplied efficiently to the surface sediments (as we discuss). This mechanism is unfortunately not mimicked in core incubation systems and we thus have to keep the experiment time short to address questions that rely on the characteristic reactivity of organic matter.
The dimensions of the benthic chambers indicate that the volume of pore water was 1100 ml. The water injected daily into the sediment to simulate bioirrigation was 281 ml, suggesting a turnover time of about 4 days. Thus I wondered if the experiment simply flushed the pore water initially rich in products of benthic biogeochemical processes. The experiment seems to have mimicked an environment where one would start from an initial situation without bioturbation followed by a situation with bioturbating organisms. Figure 1 shows the evolution of iron fluxes as a function of time. It shows a decrease with time: this suggests that pore waters are progressively replaced by seawater. The duration of the experiment did not allow to reach a stationary state. The measured fluxes thus seem to be essentially dictated by the initial state where dissolved iron in the pore water was abundant before irrigation. If the experiment seeks to simulate continental shelf sands colonized by benthic organisms, I believe that water should first be circulated for a sufficient amount of time before benthic fluxes are considered. One suggestion is to allow time for three times the volume of pore water to be renewed before making flux measurements, to avoid simply measuring fluxes from flushing the water prior to bioirrigation. In other words, I would tend to think that the results of the experiment should have been considered from the moment it was stopped here, i.e. after a dozen days. Thus, the extrapolation of the data produced here for the entire continental shelf presented in Figure 3 seems to me insufficiently justified. I therefore suggest that the authors better justify the durations of the experiments and better describe the volumes of water involved and the initial state. I would be happy to review a corrected version of the manuscript in which my remarks have been considered.
The reviewer is correct that the experiment “flushed” the pore-water that had Fe present from the pre-incubation time period. As pointed out above, during the experiment -while irrigation occurred-, however, Fe(II) production continued resulting in an overall higher (winter) or constant (summer) Fe inventory in the cores than was present in the initial cores. The elevated Fe fluxes during the 2-4 days after irrigation can likely be attributed to higher organic carbon turnover rates following sediment disturbance and the transport of some Fe that had build-up during the initial time prior to irrigation and these values were not considered in calculation of average flux values and in the discussion. Fe fluxes after the first four days (i.e., after one turnover time) remained high, however, indeed suggesting further Fe(II) production and advective transport (and not just flushing of Fe from prior to irrigation). Furthermore, if we applied a simple flush-turnover time model to the flux data, we would expect a decrease of fluxes with a half-life of ~2.7 days which is clearly not the case.
As pointed out above, the timescale of the experiment has to remain sufficiently short to not run into the problem of ending up with a strongly altered (degraded OM pool) as well as the depletion of reactive Fe oxides. This, however, would likely be the case if waiting three pore-water turnover times. This would skew the experiments to lower Fe production rates (and thus fluxes) and potentially a system with enhanced presence of hydrogen sulfide.
It seems like the reviewer misinterpreted Figure 3 (now Figure 4) which does not show an extrapolation of Fe fluxes to the entire shelf but the NY-New Jersey continental shelf sediment lithology/ composition which highlights that sandy sediments like the ones we studied are present in most of the shelf region.
Citation: https://doi.org/10.5194/bg-2022-247-AC1
-
AC1: 'Reply on RC2', Laura Wehrmann, 03 Mar 2023
-
RC3: 'Comment on bg-2022-247', Anonymous Referee #3, 04 Mar 2023
General comments:
Wehrmann et al argue that Fe_d flux from bioturbated, sandy sediments is several times greater than from muddy sediments, and that sandy sediments should thus be considered in estimates of shelf Fe_d flux. The paper provides data from an experiment in which Fe_d flux is compared between two treatments: sandy sediments with diffusion only, and sandy sediments with imposed irrigation.
First of all, I want to commend the authors for taking on the difficult task of studying permeable sediments, which are both methodologically challenging and difficult to interpret, given temporal and spatial variability. I fully agree that permeable sediments need further study, and may be disproportionately important in the global cycling of many elements. I found this paper very well written, and fun to read.
That said, I have three major concerns, related to: 1) the confusion between advection and bioirrigation in the framing, discussion and interpretation. The authors are testing and discussing two different variables -- sediment type and bioirrigation -- with only one experiment, so the true treatment is unclear; 2) the experimental data do not clearly support the study’s argument; and 3) the data interpretation that benthic Fe_d flux escapes oxidation in advective conditions is not clearly supported (while this point may potentially be true, it is not sufficiently explained by the proposed model). A few examples related to these three points are below in "specific comments".
In conclusion, don’t think the experimental data in the current version is well-constrained enough as-is to prove the argument (i.e., that bioturbated, sandy sediments are indeed significant sources of Fe_d). I was ultimately left wondering more about the Fe_d trap they developed, and wanted a more detailed accounting of the Fe in the system. If more robust experimental data (Fe and other supporting data) were added to this paper, I could see it being more convincing and very valuable. Similarly, or alternatively, a more comprehensive survey of existing Fe_d flux data from advective vs diffusive sediments (even if limited at this point) would strengthen the bigger picture argument. As is, however, I believe the current draft needs to be improved on both fronts, as it is too long for an Ideas and Perspectives piece, but does not contain a complete enough dataset for a research article. (The "sandy" vs "bioirrigated" issue may not be fixable, but I could still see the study being valuable if more data were included.)
Specific comments:
1a) Related to the first point, the paper initially seems to focus on sandy sediments (including the framing in the Abstract). Later, however, the experiment and parts of the discussion are clearly focused on bioirrigation, and the specificity of bioirrigation in sands vs bioirrigation in general is unclear. For example, Section 1.2 describes bioirrigation, although it is not entirely clear what information from this section is specific to sands, vs universal to sediments in general. (Similarly the Discussion starting at L296.) The abstract states that “the results indicate that… both biogenic and physical advection enhance fluxes”, although the simulated bioirrigation didn’t test both these variables.
1b) The authors do not offer an estimate of how important bioirrigation is in advective sediments, vs physically-driven processes. I would like to know how common bioirrigators / bioturbators are, at what density, in sandy sediments, and if they are significant enough to be a major control on fluxes. (Or conversely, if physically-driven advection is far more important.) (These processes are explored in Santos et al, 2012; The Driving Forces of…) I realize this may be difficult to generalize. However, the lack of clarity between advection and bioturbation. The cited macrofauna descriptions and densities are typically not specified between muddy or sandy environments, so it’s difficult to tell how common bioirrigation is in sands.
2a) The two treatments are really diffusive (“non-irrigated”) vs bioirrigated (“irrigated”), and thus do not provide a strong conclusion about the distinction the paper is trying to make between muddy and sandy sediments. For example, would you potentially see a similar treatment effect in diffusive vs bioirrigated treatments using nearby muddy sediments? As the authors say on L38, “Bioturbation by benthic macrofauna plays a key role for the benthic Fe cycle in both sediment types.”
2b) The higher flux in the irrigated treatments may simply be Fe_d being washed out of the sediments by a change in advective regime (moving from the non-irrigated, equilibrium period where porewater was likely anoxic and rich with Fe_d, to the well-flushed, irrigated period during which Fe_d was potentially flushed out of the porewater). The decreasing flux of Fe_d (Fig. 1) over the course of the experiment suggests the observed fluxes have not reached “steady state” for the imposed advection (in quotes, since steady state is relative for advective systems). I see the authors responded to Reviewer 2 on this topic, but I do not see the updated manuscript. I do not follow the authors’ rebuttal that the flushing would yield a flux decrease with a 2.7 day half life.
2c) The low Fe_d fluxes observed in the papers cited on L244 suggest that high Fe_d fluxes in sands are not universal. These observations were explained to be due to low sediment Fe(III) content. The current paper, however, does not present sediment Fe(III) content for comparison.
3a) The model proposed to explain high Fe_d flux in the bioirrigated sediments (Fig 2) is not well supported by data. The authors claim that since bioirrigation “results in a narrowing of the oxygenated surface zone, a large amount of Fe_d is not oxidized here but instead remains in solution.” First, no porewater O2 data are presented. Second, Fe(III) need not be oxidized in oxic sediment, but can be rapidly oxidized in the water column. Thus, an adequate mechanism for higher Fe_d flux in bioirrigated sediments is not proposed. (E.g., perhaps higher rates of OM respiration and biological activity produce more organic Fe, or Fe-binding ligands that retain Fe in solution?) The authors mention (L283) that a large fraction of the Fe_d is likely re-deposited, but could explain a bit more why they think they observe higher flux then.
3b) The porewater ammonium data are used to strengthen the Fe_d argument, but the interpretation is strange (and seemingly irrelevant to the rest of the data set). The authors claim the “the lower NH4 of the irrigated cores – assuming similar NH4 production rates in both core types – reflect additional, advective transport of solutes out of the sediment”. This is not a realistic assumption: advection fuels OM respiration by providing O2 and OM to sediment-bound microbes, so NH4 production is likely higher in the irrigated sediments. Subsequently, the authors compare ratios of NH4 *stocks* to the Fe_d *fluxes*, and claim they are “remarkably close”. (The flux ratios are not clearly presented to verify.) Why wouldn’t you present the NH4 flux data instead? Quantitatively comparing a stock to a flux makes little sense. Finally, the authors use this ratio similarity to conclude that, “when advective Fe_d transport occurs, Fe2+ escapes the sediment with little precipitation occurring.” We have no idea how much Fe reduction occurs prior to the flux measurement, so this statement is unfounded.
Minor comments:
A figure of the irrigated chamber may be helpful to visualize the set up. (Not totally necessary, but I found the description initially hard to imagine.)
L275: The authors say the Fe_d fluxes are higher than would be predicted based on bottom water O2 concentrations. Can you provide a calculation / number? What are the predicted fluxes based on O2?
L302: A word is missing here (”that maximizes… “) ?
Citation: https://doi.org/10.5194/bg-2022-247-RC3 -
AC3: 'Reply on RC3', Laura Wehrmann, 12 Mar 2023
We thank Reviewer #3 constructive comments to our manuscript which led us to further refine our manuscript introduction, discussion an conclusions. We hope that our revisions and answers to the comments are sufficient to accept this work for publication in Biogeosciences. Please find our responses to the individual comments below.
General comments:
Wehrmann et al argue that Fe_d flux from bioturbated, sandy sediments is several times greater than from muddy sediments, and that sandy sediments should thus be considered in estimates of shelf Fe_d flux. The paper provides data from an experiment in which Fe_d flux is compared between two treatments: sandy sediments with diffusion only, and sandy sediments with imposed irrigation.
First of all, I want to commend the authors for taking on the difficult task of studying permeable sediments, which are both methodologically challenging and difficult to interpret, given temporal and spatial variability. I fully agree that permeable sediments need further study, and may be disproportionately important in the global cycling of many elements. I found this paper very well written, and fun to read.
That said, I have three major concerns, related to: 1) the confusion between advection and bioirrigation in the framing, discussion and interpretation. The authors are testing and discussing two different variables -- sediment type and bioirrigation -- with only one experiment, so the true treatment is unclear; 2) the experimental data do not clearly support the study’s argument; and 3) the data interpretation that benthic Fe_d flux escapes oxidation in advective conditions is not clearly supported (while this point may potentially be true, it is not sufficiently explained by the proposed model). A few examples related to these three points are below in "specific comments".
1) In sandy sediments, bioirrigation induces the advective transport of pore-water. This process adds to the “physical” (non-biogenic) advective transport processes that commonly characterize sandy deposits along continental margins. In response to this comment and a similar one by Reviewer#1, we now phrase this more clearly throughout the manuscript. We are not conducting a comparison of sediment types, i.e., we are not testing fluxes from muds in our core experiments but solely focus on the effect of bioirrigation on the benthic Fe flux from sands. Muddy diffusive sediments have been studied extensively over the last 4 – 5 decades, and we have a better idea of the size, scaling factors and spatial variability than for permeable sands. We now more specifically describe that we are focused on sands in the introduction and discussion and also discuss that the effect of bioirrigation is superimposed on other advective processes common in sandy deposits.
2) The main message we aim to convey is that we do not have a good understanding of the size and controlling mechanisms of the Fed flux from sandy deposits but that our case study, as well as previous studies (e.g., Jahnke, et al., 2005), data point to the potential for sandy sediments to be an important source of Fe to the coastal ocean. In the context of what is known about biogeochemical processes and solute and particulate transport in sandy sediments, we highlight that the mechanisms driving the cycling and transport of iron in sandy sediments are significantly different from muddy diffusive sediments, and thus values and scaling parameters determined for muds cannot simply extrapolated to sands. The case study presented in our paper provides a glimpse of the Fe fluxes that can be expected from sandy deposits, however, our example is not intended to be all encompassing or exhaustive. The idea behind this “Ideas and Perspectives paper” is to highlight unknowns, to make the case that overlooked Fe flux sources can be significant, and to encourage further studies to enhance the understanding of the mechanisms and controlling factors driving the benthic Fe flux from sands.
3) We determined Fed fluxes using Fe accumulators connected to the experiment chambers which implies that the Fe we detected from the irrigated cores reflects the benthic Fe flux, i.e., Fe that has “escaped oxidation” within the sediment under advective conditions. Clearly, Fed is oxidized in the water column within a few minutes, but this does not appear to be as strict of an inhibitor to flux as previously thought. We now additionally show pore-water dissolved Fe inventories of initial and final irrigated and non-irrigated cores for the two (summer and winter) experiments. These data show that pore-water Fe inventories increase further during the duration of the winter experiment and remain the same in summer experiment despite a significant flux of Fed out of the irrigated cores, which supports our model that Fed is continuously released into the pore-water and subsequently transported into the overlying water column from the irrigated cores. We further address this comment in the “specific comments” section below.
In conclusion, don’t think the experimental data in the current version is well-constrained enough as-is to prove the argument (i.e., that bioturbated, sandy sediments are indeed significant sources of Fe_d). I was ultimately left wondering more about the Fe_d trap they developed, and wanted a more detailed accounting of the Fe in the system.
The traps (and a detailed accounting of Fe in our test chambers) is provided in Aller et al. 2023 which is now available online at https://www.sciencedirect.com/science/article/pii/S0304420323000178
If more robust experimental data (Fe and other supporting data) were added to this paper, I could see it being more convincing and very valuable. Similarly, or alternatively, a more comprehensive survey of existing Fe_d flux data from advective vs diffusive sediments (even if limited at this point) would strengthen the bigger picture argument. As is, however, I believe the current draft needs to be improved on both fronts, as it is too long for an Ideas and Perspectives piece, but does not contain a complete enough dataset for a research article. (The "sandy" vs "bioirrigated" issue may not be fixable, but I could still see the study being valuable if more data were included.)
Our manuscript is intended to be a “Ideas and Perspectives” contribution with the goal of emphasizing that sandy deposits function differently than muds both when it comes to solute and particulate transport, and -linked to this- the distribution and roles of different biogeochemical processes. Our “case study” along with our discussion of the biogeochemical functioning of sandy advective sediments is meant to provide a first idea of what we can expect from these deposits and to encourage further research. We have now added pore-water Fe inventories to further strengthen our experimental dataset. Unfortunately, at this point there are very little other data of Fed fluxes from sandy advective continental margin sites available that would allow for a more comprehensive survey, and any such data sets would likely have unintentionally excluded a large fraction of the iron flux in addition to having other procedural artifacts of benthic chamber work (Aller, et al. 2023). Again, this is why we wrote this Ideas and Perspectives piece!
We are not testing “sandy” vs. “bioirrigated” but aim to gain a better understanding of the benthic Fe flux from sandy bioirrigated sediments.
Specific comments:
1a) Related to the first point, the paper initially seems to focus on sandy sediments (including the framing in the Abstract). Later, however, the experiment and parts of the discussion are clearly focused on bioirrigation, and the specificity of bioirrigation in sands vs bioirrigation in general is unclear. For example, Section 1.2 describes bioirrigation, although it is not entirely clear what information from this section is specific to sands, vs universal to sediments in general. (Similarly the Discussion starting at L296.) The abstract states that “the results indicate that… both biogenic and physical advection enhance fluxes”, although the simulated bioirrigation didn’t test both these variables.
The manuscript indeed focuses on sandy sediments, and we only mention muds to emphasize some of the important differences between these deposits (and because muds have received most of the attention from researchers to date, while permeable sediments remain relatively overlooked). We have revised the Introduction section to make it clearer that we are addressing sands. The reviewer is correct in that we are specifically interested in studying the effect of bioirrigation on Fe fluxes in sandy deposits, and we tried to minimize any text that discusses bioirrigation more broadly. We have now also revised the statement in the abstract to make it clearly that we only tested the effect of (simulated) biogenic advection.
1b) The authors do not offer an estimate of how important bioirrigation is in advective sediments, vs physically-driven processes. I would like to know how common bioirrigators / bioturbators are, at what density, in sandy sediments, and if they are significant enough to be a major control on fluxes. (Or conversely, if physically-driven advection is far more important.) (These processes are explored in Santos et al, 2012; The Driving Forces of…) I realize this may be difficult to generalize. However, the lack of clarity between advection and bioturbation. The cited macrofauna descriptions and densities are typically not specified between muddy or sandy environments, so it’s difficult to tell how common bioirrigation is in sands.
Sandy deposits along continental shelves, are an incredibly important habitat for benthic macrofauna and one will not find a square meter that is not teeming with life. In our discussion section, we review data for the density of macroorganisms along the Mid Atlantic Bight ranging in numbers from 250 individuals m-2 to 18,000 individuals m-2 -as seen in our map, most of this area is sandy. Benthic macrofauna are without a doubt significant enough to be a major control on element fluxes. In densely populated intertidal or subtidal areas, bioirrigation is thought to dominate porewater exchange (Volkenborn et al., 2007). In low-biomass North Sea sand flats, about half of the advection-driven particle flux into the sediment was also attributed to bioirrigation (Rusch et al., 2000). Based on the estimates in Santos et al. 2012, the spatial and temporal scales and flow rate of bioirrigation are in a similar range to those of wave pumping and flow- and topography-induced pressure gradients. We have now added this information to the introduction section. We emphasize that our bioirrigation mimics were specifically scaled to the magnitudes and frequency patterns observed in natural patches of maldanids (Clymenella torquata) at our primary field site.
2a) The two treatments are really diffusive (“non-irrigated”) vs bioirrigated (“irrigated”), and thus do not provide a strong conclusion about the distinction the paper is trying to make between muddy and sandy sediments. For example, would you potentially see a similar treatment effect in diffusive vs bioirrigated treatments using nearby muddy sediments? As the authors say on L38, “Bioturbation by benthic macrofauna plays a key role for the benthic Fe cycle in both sediment types.”
The two treatments are diffusive and bioirrigated (biotic advection) in SANDY sediments. As pointed out in our Introduction, the upward percolation and advective transport of pore-water induced by bioirrigation and the infiltration of particles into deeper sediment layers are important mechanisms in sands but largely absent in muds, which are impermeable. Although mobile burrowers in muds can enhance advective exchange during movement, bioirrigation within fixed tubes and burrow structures does not promote interstitial advection. We realize that our previous statement “Bioturbation by benthic macrofauna plays a key role for the benthic Fe cycle in both sediment types” (although technically true) was very broad and have removed it. We are discussing particle and solute transport in sandy advective sediments in more detail in Section 1.1.
2b) The higher flux in the irrigated treatments may simply be Fe_d being washed out of the sediments by a change in advective regime (moving from the non-irrigated, equilibrium period where porewater was likely anoxic and rich with Fe_d, to the well-flushed, irrigated period during which Fe_d was potentially flushed out of the porewater). The decreasing flux of Fe_d (Fig. 1) over the course of the experiment suggests the observed fluxes have not reached “steady state” for the imposed advection (in quotes, since steady state is relative for advective systems). I see the authors responded to Reviewer 2 on this topic, but I do not see the updated manuscript. I do not follow the authors’ rebuttal that the flushing would yield a flux decrease with a 2.7 day half life.
We now show the pore-water Fed inventories to highlight that concentration of Fed remains high in the pore-water -even further increasing in the winter- over the course of the experiment. This shows that the measured Fed flux of the bioirrigated sediment is not just a “washing out” of accumulated Fed. As indicated in our response to Reviewer #2, there is no “steady-state” in these systems due to the decrease in organic matter reactivity.
In a system where production is absent (as Reviewer#2) suggested, the flux of Fed out of the sediment will decay as 1/t where tis the pore-water residence time (turnover time of the pore fluid). In this case, t = total volume of pore fluid (V)/ irrigation input (v) (note the irrigation volume input must equal the volume output across the sediment-water interface). If there is no Fed production (or consumption) within the sediment, the time dependent, volume-averaged pore water concentration is: C(t) = C0e-t/t (assuming the inflowing fluid has a Fe concentration of 0). Ignoring diffusive flux as minor relative to advection, the advective flux across the sediment- water interface is: J(t) = vC(t)/A (where A = area of the sediment –water interface). Given the dimensions and pumping rate of our system, J(t), the flux would then decay with a 2.7 day half life, which it clearly does not. Note that the relatively small diffusive flux would also decay with the same time-scale as concentration gradients become zero in the absence of production.
2c) The low Fe_d fluxes observed in the papers cited on L244 suggest that high Fe_d fluxes in sands are not universal. These observations were explained to be due to low sediment Fe(III) content. The current paper, however, does not present sediment Fe(III) content for comparison.
We double checked, and the Eitel et al. 2020 paper does not provide any sedimentary Fe(III) content values for the sandy shelf sites to compare data against. Slomp et al 1997 proposes that low OC deposition was responsible for the low Fe fluxes at the sandy sites they investigated as these were dominated by erosion. We have now updated our manuscript conclusion section to indicate that the flux of Fed from sands is likely highly variable (ie., “not universal”).
3a) The model proposed to explain high Fe_d flux in the bioirrigated sediments (Fig 2) is not well supported by data. The authors claim that since bioirrigation “results in a narrowing of the oxygenated surface zone, a large amount of Fe_d is not oxidized here but instead remains in solution.” First, no porewater O2 data are presented. Second, Fe(III) need not be oxidized in oxic sediment, but can be rapidly oxidized in the water column. Thus, an adequate mechanism for higher Fe_d flux in bioirrigated sediments is not proposed. (E.g., perhaps higher rates of OM respiration and biological activity produce more organic Fe, or Fe-binding ligands that retain Fe in solution?) The authors mention (L283) that a large fraction of the Fe_d is likely re-deposited, but could explain a bit more why they think they observe higher flux then.
We did not measure O2 in the surface sediment. However, a narrowing of the surface oxygenated zones during the pumping activity of benthic macrofauna in sandy sediment is a common observation as, for example, previously made by Volkenborn et al., 2010, 2012, Kristensen et al., 2011, and Quintana et al., 2011 for Arenicola marina, Marenzelleria viridis and Neotrypaea californiensis using oxygen imaging. We also visually observed that the oxidized sediment zone, as reflected by light – yellow color, was clearly thinner than in the non-irrigated cores, consistent with zonal compression.
We would like to point out that the benthic Fe flux is defined as the Fed moving across the sediment-water interface over a specific area and time. The reviewer is correct that a fraction of it will likely be oxidized in the water column, however, this does not affect the benthic flux but is a post-flux process. Our experimental set-ups with the Fe accumulators quantitively extract all, or virtually all, Fed released from the nonirrigated and irrigated cores (Aller et al., 2023) so differences in the effectiveness of this method cannot explain the higher flux from the irrigated cores. Overall, our results -along with the results from previous studies- suggest a higher flux of Fed from irrigated cores due to the effective advective transport of Fed out of the sediment across a narrowed oxic zone which does not allow for the quantitative reoxidation of Fed before its escape. The enhanced mobilization of iron at depth in the sediment due to intermittently oxic conditions (driving iron sulfide oxidation) likely further increased Fe release from the irrigated cores. The reviewer, nonetheless brings up an important point which is that permeable sandy sediments can be characterized by high OM respiration rates (reflected in higher oxic respiration and denitrification rates). This has mainly been explained by the 1) enhanced availability of oxygen for aerobic respiration and nitrogen compounds for coupled nitrification-denitrification, as well as the increased supply of degradable DOM into the sediment (e.g., D’Andrea et al., 2002; Marchant et al., 2016; Rusch et al., 2006). In our experiments, the enhanced inflow of oxygen and nitrogen compounds are unlikely to fuel higher dissimilatory iron reduction rates (which would be needed to explain a higher Fed flux driven by enhanced OM remineralization). We did not measure the concentration of DOM in the overlying water in our experiment which we pumped into irrigation mimics but given our experimental set-up featuring a recirculation system (and no “fresh” seawater input), we do not believe that this substrate source can explain large differences in DIR rates (and Fed fluxes) between the irrigated and nonirrigated cores as postulated by the reviewer. Importantly, as we point out in our manuscript, in nature, the advection-driven infiltration of (nano)particulate Fe oxides into the sediment is an important process that helps sustain elevated DIR rates in sandy sediments in general (e.g., Huettel, et al., 1996; D’Andrea, et al., 2002; Jahnke, et al., 2005). However, in our experiment, the use of the Fe accumulators that removed iron from the recirculating setup meant that there was minimal Fe in the overlying water (comparable to what it would be in an open system with no outside iron supply) that was re-injected using the irrigation mimics.
3b) The porewater ammonium data are used to strengthen the Fe_d argument, but the interpretation is strange (and seemingly irrelevant to the rest of the data set). The authors claim the “the lower NH4 of the irrigated cores – assuming similar NH4 production rates in both core types – reflect additional, advective transport of solutes out of the sediment”. This is not a realistic assumption: advection fuels OM respiration by providing O2 and OM to sediment-bound microbes, so NH4 production is likely higher in the irrigated sediments. Subsequently, the authors compare ratios of NH4 *stocks* to the Fe_d *fluxes*, and claim they are “remarkably close”. (The flux ratios are not clearly presented to verify.) Why wouldn’t you present the NH4 flux data instead? Quantitatively comparing a stock to a flux makes little sense. Finally, the authors use this ratio similarity to conclude that, “when advective Fe_d transport occurs, Fe2+ escapes the sediment with little precipitation occurring.” We have no idea how much Fe reduction occurs prior to the flux measurement, so this statement is unfounded.
We agree with the reviewer that without additional justification, this section did not significantly strengthen our main discussion points. Because it is not critical to our fundamental thesis, we decided that rather than expanding and defending our argument further, we would remove the section (calculation) in the revised manuscript version. Doing so helped to shorten the manuscript, partly addressing the comment by the reviewer that the paper is too long.
Minor comments:
A figure of the irrigated chamber may be helpful to visualize the set up. (Not totally necessary, but I found the description initially hard to imagine.)
A nice picture of the set-up can be found in Aller et al. 2023 (see Supplemental Materials 1).
L275: The authors say the Fe_d fluxes are higher than would be predicted based on bottom water O2 concentrations. Can you provide a calculation / number? What are the predicted fluxes based on O2?
We have added an estimated flux based on Dale et al. 2015.
L302: A word is missing here (”that maximizes… “) ?
We moved the “that” to make it clearer.
Citation: https://doi.org/10.5194/bg-2022-247-AC3
-
AC3: 'Reply on RC3', Laura Wehrmann, 12 Mar 2023
Status: closed
-
RC1: 'Comment on bg-2022-247', Sebastiaan van de Velde, 14 Jan 2023
Wehrmann et al. argue in this MS that dissolved iron (Fed) fluxes from sandy sediments – which have received little attention – are an important source of iron for the ocean. They present a case study to illustrate that Fed fluxes from sandy sediments with simulated bioirrigation are several times higher than in the absence of irrigation, and that these fluxes are in the higher range of fluxes measured from muddy deposits.
While I agree with the overall message of the perspective paper, and do not question the quality of the presented results, I have one main comment on the experimental design and conclusions drawn from the results. Because the way the experiment is designed (and the authors acknowledge this during the discussion, e.g. L243), you end up with a sandy deposit that is diffusion dominated. In reality however, as discussed in the introduction, sandy deposits are advection dominated – irregardless of the presence or absence of bio-irrigators. As a results, when you compare your irrigated results with your non-irrigated results, you can say something about the the importance of irrigation relative to diffusion – but you cannot make an accurate assessment of how important irrigation would be under natural and advective conditions (which is what you do on L281 and L363), since you have no control that includes advection without irrigation. I would ask the authors to consider this in their discussion – and either provide a quantitative argument as to why irrigation is important in sandy deposits relative to advection, or rephrase the discussion so it focuses on advective sandy sediments in general, and does not go into too much detail with respect to the different types of irrigators or other benthic fauna. In essence, this would link better with the introduction, where you do explain how sandy deposits w/o bioturbators are distinctly different from muddy deposits.
A second concern is the comparison of Fed fluxes from this study with the literature (L280-283). Since you use a novel method, the discrepancy is likely a consequence of your new method to measure Fed fluxes. Dale et al. base themselves on the available data, which is almost exclusively collected by the traditional flux incubation methods. They then calibrate their model on the available fluxes, which are lower than what you would have measured with your extractor-method.
So you could also make the argument that both muddy as sandy deposits are roughly comparable as Fed sources, but our older methods underestimate the actual flux. Please also consider this difference in methods in the discussion.I am however confident that the authors can address these concerns and the smaller remarks listed below. The message of the manuscript is an important and valid one, and worthy to be published. I will happily read a revised version.
Kind regards
Sebastiaan van de Velde
Minor comments:L32: or through acidic dissolution of FeS in sediments colonized by cable bacteria (Rao et al., 2016; Seitaj et al., 2015; Sulu-Gambari et al., 2016; van de Velde et al., 2016)
L40: could use a reference
L59: I would perhaps also say that this is due to practical limitations regarding the in-situ sampling and processing of sandy sediments.
L60: and the fact that most of these sands are dominated by aerobic respiration and denitrification, thus little dissolved Fe(2+) is expected to accumulate and be released from these sediments.
L63: clarify if this is µmol cm-3 sediments or solid phase
L59-70: This argument is based on the extracteable Fe contents – but you would also need to take into account the oxygenation state of sandy versus muddy deposits (how deep is the sediment oxygenated?).
L101: or anthropogenic activities (trawling, dredging) (van de Velde et al., 2018)
L197: does this compare to the natural populations? (I now see you say this at L296, I would move this upward)
L301ff: Or if Corg gets too high, most iron will be rapdily precipitated to FeS(2) and not recycled to the overlying water (van de Velde et al., 2020a, 2020b)
Fig. 1: there is no label on the x-axis
Fig. 2: and what is happening in the rest of the sediment? I also assume this conceptuel model is based on actual measurements of the O2 distribution, which you might either refer to or show alongside the image.
References:
Rao, A. M. F., Malkin, S. Y., Hidalgo-Martinez, S. and Meysman, F. J. R.: The impact of electrogenic sulfide oxidation on elemental cycling and solute fluxes in coastal sediment, Geochim. Cosmochim. Acta, 172, 265–286, doi:10.1016/j.gca.2015.09.014, 2016.
Seitaj, D., Schauer, R., Sulu-gambari, F., Hidalgo-martinez, S., Malkin, S. Y., Burdorf, L. D. W., Slomp, C. P. and Meysman, F. J. R.: Cable bacteria generate a firewall against euxinia in seasonally hypoxic basins, Proc. Natl. Acad. Sci., 112(43), 13278–13283, doi:10.1073/pnas.1510152112, 2015.
Sulu-Gambari, F., Seitaj, D., Behrends, T., Banerjee, D., Meysman, F. J. R. and Slomp, C. P.: Impact of cable bacteria on sedimentary iron and manganese dynamics in a seasonally-hypoxic marine basin, Geochim. Cosmochim. Acta, 192(2016), 49–69, doi:10.1016/j.gca.2016.07.028, 2016.
van de Velde, S. J., Lesven, L., Burdorf, L. D. W., Hidalgo-Martinez, S., Geelhoed, J. S., Van Rijswijk, P., Gao, Y. and Meysman, F. J. R.: The impact of electrogenic sulfur oxidation on the biogeochemistry of coastal sediments: a field study, Geochim. Cosmochim. Acta, 194, 211–232, doi:10.1016/j.gca.2016.08.038, 2016.
van de Velde, S. J., Van Lancker, V., Hidalgo-Martinez, S., Berelson, W. M. and Meysman, F. J. R.: Anthropogenic disturbance keeps the coastal seafloor biogeochemistry in a transient state, Sci. Rep., 8(1), 1–10, doi:10.1038/s41598-018-23925-y, 2018.
van de Velde, S. J., Reinhard, C. T., Ridgwell, A. and Meysman, F. J. R.: Bistability in the redox chemistry of sediments and oceans, Proc. Natl. Acad. Sci., 117(52), 33043–33050, doi:10.1073/pnas.2008235117, 2020a.
van de Velde, S. J., Hidalgo-Martinez, S., Callebaut, I., Antler, G., James, R., Leermakers, M. and Meysman, F.: Burrowing fauna mediate alternative stable states in the redox cycling of salt marsh sediments, Geochim. Cosmochim. Acta, 276, 31–49, doi:10.1016/j.gca.2020.02.021, 2020b.
Citation: https://doi.org/10.5194/bg-2022-247-RC1 -
AC2: 'Reply on RC1', Laura Wehrmann, 03 Mar 2023
We thank the reviewer Dr. van de Velde for his insightful and constructive comments to our manuscript which helped to improve its clarity and focus. We hope that our revisions and answers to the comments are sufficient to accept this work for publication in Biogeosciences. Please find our responses to the individual comments below.
Wehrmann et al. argue in this MS that dissolved iron (Fed) fluxes from sandy sediments – which have received little attention – are an important source of iron for the ocean. They present a case study to illustrate that Fed fluxes from sandy sediments with simulated bioirrigation are several times higher than in the absence of irrigation, and that these fluxes are in the higher range of fluxes measured from muddy deposits.
While I agree with the overall message of the perspective paper, and do not question the quality of the presented results, I have one main comment on the experimental design and conclusions drawn from the results. Because the way the experiment is designed (and the authors acknowledge this during the discussion, e.g. L243), you end up with a sandy deposit that is diffusion dominated. In reality however, as discussed in the introduction, sandy deposits are advection dominated – irregardless of the presence or absence of bio-irrigators. As a results, when you compare your irrigated results with your non-irrigated results, you can say something about the the importance of irrigation relative to diffusion – but you cannot make an accurate assessment of how important irrigation would be under natural and advective conditions (which is what you do on L281 and L363), since you have no control that includes advection without irrigation. I would ask the authors to consider this in their discussion – and either provide a quantitative argument as to why irrigation is important in sandy deposits relative to advection, or rephrase the discussion so it focuses on advective sandy sediments in general, and does not go into too much detail with respect to the different types of irrigators or other benthic fauna. In essence, this would link better with the introduction, where you do explain how sandy deposits w/o bioturbators are distinctly different from muddy deposits.
As previously discussed in several publications, for example, by Aller, 2001 and Meysman et al. 2006, the burrow flushing induced by the pumping activity of macrofauna, i.e., bioirrigation, represents an advective mode of water transport. In our experiment, we use irrigation mimics to inject water into the sediment at depth following the typically activity of C. torquata. This injection drives the advective transport of pore water upward, forcing fluid out of the sediment. Thus, our irrigated cores must be described as advection-dominated with diffusion only controlling the transport regime during times of non-pumping/irrigation. What the reviewer refers to as advection is (non-biogenic) advective transport of pore water induced by pressure gradients, for example, due to the unidirectional flow of bottom currents across an uneven sediment surface or density gradients. The reviewer is correct that we do not simulate this type of advective transport in our experiments, as previously demonstrated using recirculating flumes or natural bedforms (e.g., Huettel et al, 1996; 1998). As outlined above, our focus is on biogenic advection induced by bioirrigation at depth, which is superimposed on any physically-driven advection. Bioirrigation enhances advection and can occur in the absence of physical forcing. We focused on isolating these effects using realistic irrigation volumes and patterns. We agree with the reviewer, that we should state this more carefully -and highlight that an additional physically induced advective transport component would likely alter the overall transport dynamics. The reviewer is correct that our “non-irrigated cores” are diffusion-dominated not physical advection-dominated as would likely be the case for sandy deposits in the absence of macrofauna. We used the diffusion case as a Fe flux end-member. We noted that the magnitude of Fe flux that we measured in this case closely corresponds to the values reported in the literature from benthic chamber measurements in which advection is suppressed. We have changed our manuscript discussion to clarify these points.
A second concern is the comparison of Fed fluxes from this study with the literature (L280-283). Since you use a novel method, the discrepancy is likely a consequence of your new method to measure Fed fluxes. Dale et al. base themselves on the available data, which is almost exclusively collected by the traditional flux incubation methods. They then calibrate their model on the available fluxes, which are lower than what you would have measured with your extractor-method.
This is an important point the reviewer brings up. We discuss the details of our new flux incubation method and provide a comprehensive comparison to other flux methods in a recent paper “Estimating benthic dissolved Fe and reactive solute fluxes” (Aller, Dwyer, Swenson, Heilbrun, Volkenborn, Wehrmann) published in Marine Chemistry (https://doi.org/10.1016/j.marchem.2023.104221). However, to guide the reader, we now explain the potential deviations in the approaches to estimating fluxes in our Introduction section.
So you could also make the argument that both muddy as sandy deposits are roughly comparable as Fed sources, but our older methods underestimate the actual flux. Please also consider this difference in methods in the discussion.
It is indeed possible that muds and sands produce roughly comparable Fed sources but for different reasons. We wrote this Perspectives paper to emphasize our lack of knowledge in this regard. We simply don’t have sufficient data on sandy deposits. Thus, we don’t know whether fluxes from muds or sands are in a similar range. It is also possible that because sands are open systems constantly extracting reactive particles from overlying water, that the Fed flux from bioturbated sands exceeds that from muds. Again, we don’t know.
I am however confident that the authors can address these concerns and the smaller remarks listed below. The message of the manuscript is an important and valid one, and worthy to be published. I will happily read a revised version.
Kind regards
Sebastiaan van de Velde
Minor comments:L32: or through acidic dissolution of FeS in sediments colonized by cable bacteria (Rao et al., 2016; Seitaj et al., 2015; Sulu-Gambari et al., 2016; van de Velde et al., 2016)
Added.
L40: could use a reference
Added.
L59: I would perhaps also say that this is due to practical limitations regarding the in-situ sampling and processing of sandy sediments.
Added.
L60: and the fact that most of these sands are dominated by aerobic respiration and denitrification, thus little dissolved Fe(2+) is expected to accumulate and be released from these sediments.
Again, we don’t always know that because Fe cycling in sandy deposits remains so little investigated. However, some recent studies, e.g., by Jahnke et al., 2005 and Zhou et al., 2023, indicate that DIR may also play an important role for OM remineralization in sandy deposits.
L63: clarify if this is µmol cm-3 sediments or solid phase
Done.
L59-70: This argument is based on the extracteable Fe contents – but you would also need to take into account the oxygenation state of sandy versus muddy deposits (how deep is the sediment oxygenated?).
Yes, here we highlight that there may be enough extractable Fe to support significant DIR (and ultimately benthic Fe fluxes). Overall, the presence of oxygen is indeed a key factor in the release of Fe from sediments. However, as we also point out, sandy deposits are often characterized by a very patchy distribution of redox zones due to their permeable nature and fast redox dynamics. Rather than a one-dimensional vertical view of O2 (as well as H2S) distribution a 3D-depiction of the redox distribution is more appropriate to characterize these deposits.
L101: or anthropogenic activities (trawling, dredging) (van de Velde et al., 2018)
Added.
L197: does this compare to the natural populations? (I now see you say this at L296, I would move this upward)
We added the information to the method section.
L301ff: Or if Corg gets too high, most iron will be rapdily precipitated to FeS(2) and not recycled to the overlying water (van de Velde et al., 2020a, 2020b)
Revised.
Fig. 1: there is no label on the x-axis
Done. Thanks!
Fig. 2: and what is happening in the rest of the sediment? I also assume this conceptuel model is based on actual measurements of the O2 distribution, which you might either refer to or show alongside the image.
Added.
References:
Rao, A. M. F., Malkin, S. Y., Hidalgo-Martinez, S. and Meysman, F. J. R.: The impact of electrogenic sulfide oxidation on elemental cycling and solute fluxes in coastal sediment, Geochim. Cosmochim. Acta, 172, 265–286, doi:10.1016/j.gca.2015.09.014, 2016.
Seitaj, D., Schauer, R., Sulu-gambari, F., Hidalgo-martinez, S., Malkin, S. Y., Burdorf, L. D. W., Slomp, C. P. and Meysman, F. J. R.: Cable bacteria generate a firewall against euxinia in seasonally hypoxic basins, Proc. Natl. Acad. Sci., 112(43), 13278–13283, doi:10.1073/pnas.1510152112, 2015.
Sulu-Gambari, F., Seitaj, D., Behrends, T., Banerjee, D., Meysman, F. J. R. and Slomp, C. P.: Impact of cable bacteria on sedimentary iron and manganese dynamics in a seasonally-hypoxic marine basin, Geochim. Cosmochim. Acta, 192(2016), 49–69, doi:10.1016/j.gca.2016.07.028, 2016.
van de Velde, S. J., Lesven, L., Burdorf, L. D. W., Hidalgo-Martinez, S., Geelhoed, J. S., Van Rijswijk, P., Gao, Y. and Meysman, F. J. R.: The impact of electrogenic sulfur oxidation on the biogeochemistry of coastal sediments: a field study, Geochim. Cosmochim. Acta, 194, 211–232, doi:10.1016/j.gca.2016.08.038, 2016.
van de Velde, S. J., Van Lancker, V., Hidalgo-Martinez, S., Berelson, W. M. and Meysman, F. J. R.: Anthropogenic disturbance keeps the coastal seafloor biogeochemistry in a transient state, Sci. Rep., 8(1), 1–10, doi:10.1038/s41598-018-23925-y, 2018.
van de Velde, S. J., Reinhard, C. T., Ridgwell, A. and Meysman, F. J. R.: Bistability in the redox chemistry of sediments and oceans, Proc. Natl. Acad. Sci., 117(52), 33043–33050, doi:10.1073/pnas.2008235117, 2020a.
van de Velde, S. J., Hidalgo-Martinez, S., Callebaut, I., Antler, G., James, R., Leermakers, M. and Meysman, F.: Burrowing fauna mediate alternative stable states in the redox cycling of salt marsh sediments, Geochim. Cosmochim. Acta, 276, 31–49, doi:10.1016/j.gca.2020.02.021, 2020b.
All added.
Citation: https://doi.org/10.5194/bg-2022-247-AC2
-
AC2: 'Reply on RC1', Laura Wehrmann, 03 Mar 2023
-
RC2: 'Comment on bg-2022-247', Anonymous Referee #2, 24 Jan 2023
The manuscript by Wehrmann et al. focuses on the problem of benthic iron fluxes on continental margins. The scientific background to this problem is well described in the introductory chapter. The text then describes experiments where measurements of dissolved iron fluxes at the water-sediment interface were performed with permeable sands. In one case, pore water does not flow through the sediment, in the other case, a pumping system forces water flow and simulates the bio-irrigation activity of a common annelid in West Atlantic margin sediments. The experiment was conducted under both winter and summer temperature conditions. The idea of experimentally examining iron flux from a bioturbated sediment is interesting, however, I do not understand why the manuscript is submitted to Biogeosciences under "ideas and perspectives", since the manuscript describes and discusses the results of an experiment.
The experiments here lasting 7 and 12 days show that benthic fluxes of dissolved iron are higher for permeable bio-irrigated sediments than for non-irrigated sediments. Oxygen levels remained high in the water column during the experiments. At the end of the experiments, the dissolved ammonium inventory in the pore water was measured. This was 4 to 9 times higher in the non-irrigated sediments. In view of the samples that were collected and the discussion, it is surprising not to see some additional results that would have strongly supported the conclusions: only iron fluxes were measured, only oxygen was monitored in the water column, and only ammonium concentrations were measured in the pore water. Thus many questions remained unanswered.
Since the samples were available, why were dissolved nitrogen compounds not monitored in the water column? Why were the fluxes of other biogenic compounds not measured? What was the inventory of dissolved iron in the pore water? What is the proportion of iron that crossed the water-sediment interface compared to the initial inventory of pore water iron? The experiments last 7 and 12 days, what justifies this duration?
The dimensions of the benthic chambers indicate that the volume of pore water was 1100 ml. The water injected daily into the sediment to simulate bioirrigation was 281 ml, suggesting a turnover time of about 4 days. Thus I wondered if the experiment simply flushed the pore water initially rich in products of benthic biogeochemical processes. The experiment seems to have mimicked an environment where one would start from an initial situation without bioturbation followed by a situation with bioturbating organisms. Figure 1 shows the evolution of iron fluxes as a function of time. It shows a decrease with time: this suggests that pore waters are progressively replaced by seawater. The duration of the experiment did not allow to reach a stationary state. The measured fluxes thus seem to be essentially dictated by the initial state where dissolved iron in the pore water was abundant before irrigation. If the experiment seeks to simulate continental shelf sands colonized by benthic organisms, I believe that water should first be circulated for a sufficient amount of time before benthic fluxes are considered. One suggestion is to allow time for three times the volume of pore water to be renewed before making flux measurements, to avoid simply measuring fluxes from flushing the water prior to bioirrigation. In other words, I would tend to think that the results of the experiment should have been considered from the moment it was stopped here, i.e. after a dozen days. Thus, the extrapolation of the data produced here for the entire continental shelf presented in Figure 3 seems to me insufficiently justified. I therefore suggest that the authors better justify the durations of the experiments and better describe the volumes of water involved and the initial state. I would be happy to review a corrected version of the manuscript in which my remarks have been considered.
Citation: https://doi.org/10.5194/bg-2022-247-RC2 -
AC1: 'Reply on RC2', Laura Wehrmann, 03 Mar 2023
We thank Reviewer #2 for the insightful and valuable comments on our manuscript that helped to improve our method description and discussion. We also added a figure that shows Fe inventories at the beginning and end of the experiment to address some of the reviewers’ concerns. We hope that our revisions and answers to the comments are sufficient to accept this work for publication in Biogeosciences. Please find our responses to the individual comments below.
The manuscript by Wehrmann et al. focuses on the problem of benthic iron fluxes on continental margins. The scientific background to this problem is well described in the introductory chapter. The text then describes experiments where measurements of dissolved iron fluxes at the water-sediment interface were performed with permeable sands. In one case, pore water does not flow through the sediment, in the other case, a pumping system forces water flow and simulates the bio-irrigation activity of a common annelid in West Atlantic margin sediments. The experiment was conducted under both winter and summer temperature conditions. The idea of experimentally examining iron flux from a bioturbated sediment is interesting, however, I do not understand why the manuscript is submitted to Biogeosciences under "ideas and perspectives", since the manuscript describes and discusses the results of an experiment.
The Ideas and perspectives manuscripts “report new ideas and novel aspects of scientific investigations” and was chosen by us to discuss the clear gaps in our knowledge of the benthic Fe flux from sandy permeable sediments which cover a large part of continental shelfs. We only present a small dataset as a case study to provide an initial idea of the expected scale of these fluxes from sandy deposits. As we highlight in our discussion, there is a clear need for further studies investigating this topic in detail. Our manuscript aims to spark further interest and encourage such investigations.
The experiments here lasting 7 and 12 days show that benthic fluxes of dissolved iron are higher for permeable bio-irrigated sediments than for non-irrigated sediments. Oxygen levels remained high in the water column during the experiments. At the end of the experiments, the dissolved ammonium inventory in the pore water was measured. This was 4 to 9 times higher in the non-irrigated sediments. In view of the samples that were collected and the discussion, it is surprising not to see some additional results that would have strongly supported the conclusions: only iron fluxes were measured, only oxygen was monitored in the water column, and only ammonium concentrations were measured in the pore water. Thus many questions remained unanswered.
Since the samples were available, why were dissolved nitrogen compounds not monitored in the water column? Why were the fluxes of other biogenic compounds not measured?
This article specifically focuses on the cycling and fluxes of iron from sandy permeable sediments and we thus did not analyze any nitrogen compounds other than ammonium concentrations in pore-waters which were used get insight into differences in the overall remineralization signals in the non-irrigated and irrigate cores. As pointed out in our Introduction section, the nitrogen cycle in these deposits has been the focus of previous studies (e.g., Rao et al., 2007; Marchant et al., 2016) while little is known about Fe fluxes. The Fe accumulator we developed is specific to the determination of Fe (and potentially P and other redox-sensitive trace metals; Aller et al., 2023) while it does not trap dissolved inorganic carbon, ammonium and other biogenic compounds and thus their fluxes cannot be estimated using this technique.
What was the inventory of dissolved iron in the pore water? What is the proportion of iron that crossed the water-sediment interface compared to the initial inventory of pore water iron? The experiments last 7 and 12 days, what justifies this duration?
These are important points. We now show our calculations of the dissolved Fe inventories in initial and end cores and show these data in Figure 2. The Fe inventory in the winter experiment increased ~15 fold in non-irrigated cores and ~8 fold in the irrigated cores compared to the initial cores, while they stayed in the same range throughout the experiment in the summer cores. These data indicate that the observed Fe flux does not just reflect a “flushing” of the sediment and transport of Fe produced prior to irrigation out of the sediment but reflect Fe(II) production (driven by OM remineralization) and a build-up in the inventory (especially in winter) despite the advective transport of Fe out of the sediment.
The timescale of these types of experiments is dictated by the characteristic reactivity of organic matter, usually 14 - 24 days for fresh plankton. Thus, the experiments have to remain sufficiently short to avoid reactive OM depletion and significant changes in dissimilatory Fe reduction. Given the higher temperatures in summer, we decided to keep the summer experiment shorter than the winter experiment. In nature, the system is open and reactive particles are resupplied efficiently to the surface sediments (as we discuss). This mechanism is unfortunately not mimicked in core incubation systems and we thus have to keep the experiment time short to address questions that rely on the characteristic reactivity of organic matter.
The dimensions of the benthic chambers indicate that the volume of pore water was 1100 ml. The water injected daily into the sediment to simulate bioirrigation was 281 ml, suggesting a turnover time of about 4 days. Thus I wondered if the experiment simply flushed the pore water initially rich in products of benthic biogeochemical processes. The experiment seems to have mimicked an environment where one would start from an initial situation without bioturbation followed by a situation with bioturbating organisms. Figure 1 shows the evolution of iron fluxes as a function of time. It shows a decrease with time: this suggests that pore waters are progressively replaced by seawater. The duration of the experiment did not allow to reach a stationary state. The measured fluxes thus seem to be essentially dictated by the initial state where dissolved iron in the pore water was abundant before irrigation. If the experiment seeks to simulate continental shelf sands colonized by benthic organisms, I believe that water should first be circulated for a sufficient amount of time before benthic fluxes are considered. One suggestion is to allow time for three times the volume of pore water to be renewed before making flux measurements, to avoid simply measuring fluxes from flushing the water prior to bioirrigation. In other words, I would tend to think that the results of the experiment should have been considered from the moment it was stopped here, i.e. after a dozen days. Thus, the extrapolation of the data produced here for the entire continental shelf presented in Figure 3 seems to me insufficiently justified. I therefore suggest that the authors better justify the durations of the experiments and better describe the volumes of water involved and the initial state. I would be happy to review a corrected version of the manuscript in which my remarks have been considered.
The reviewer is correct that the experiment “flushed” the pore-water that had Fe present from the pre-incubation time period. As pointed out above, during the experiment -while irrigation occurred-, however, Fe(II) production continued resulting in an overall higher (winter) or constant (summer) Fe inventory in the cores than was present in the initial cores. The elevated Fe fluxes during the 2-4 days after irrigation can likely be attributed to higher organic carbon turnover rates following sediment disturbance and the transport of some Fe that had build-up during the initial time prior to irrigation and these values were not considered in calculation of average flux values and in the discussion. Fe fluxes after the first four days (i.e., after one turnover time) remained high, however, indeed suggesting further Fe(II) production and advective transport (and not just flushing of Fe from prior to irrigation). Furthermore, if we applied a simple flush-turnover time model to the flux data, we would expect a decrease of fluxes with a half-life of ~2.7 days which is clearly not the case.
As pointed out above, the timescale of the experiment has to remain sufficiently short to not run into the problem of ending up with a strongly altered (degraded OM pool) as well as the depletion of reactive Fe oxides. This, however, would likely be the case if waiting three pore-water turnover times. This would skew the experiments to lower Fe production rates (and thus fluxes) and potentially a system with enhanced presence of hydrogen sulfide.
It seems like the reviewer misinterpreted Figure 3 (now Figure 4) which does not show an extrapolation of Fe fluxes to the entire shelf but the NY-New Jersey continental shelf sediment lithology/ composition which highlights that sandy sediments like the ones we studied are present in most of the shelf region.
Citation: https://doi.org/10.5194/bg-2022-247-AC1
-
AC1: 'Reply on RC2', Laura Wehrmann, 03 Mar 2023
-
RC3: 'Comment on bg-2022-247', Anonymous Referee #3, 04 Mar 2023
General comments:
Wehrmann et al argue that Fe_d flux from bioturbated, sandy sediments is several times greater than from muddy sediments, and that sandy sediments should thus be considered in estimates of shelf Fe_d flux. The paper provides data from an experiment in which Fe_d flux is compared between two treatments: sandy sediments with diffusion only, and sandy sediments with imposed irrigation.
First of all, I want to commend the authors for taking on the difficult task of studying permeable sediments, which are both methodologically challenging and difficult to interpret, given temporal and spatial variability. I fully agree that permeable sediments need further study, and may be disproportionately important in the global cycling of many elements. I found this paper very well written, and fun to read.
That said, I have three major concerns, related to: 1) the confusion between advection and bioirrigation in the framing, discussion and interpretation. The authors are testing and discussing two different variables -- sediment type and bioirrigation -- with only one experiment, so the true treatment is unclear; 2) the experimental data do not clearly support the study’s argument; and 3) the data interpretation that benthic Fe_d flux escapes oxidation in advective conditions is not clearly supported (while this point may potentially be true, it is not sufficiently explained by the proposed model). A few examples related to these three points are below in "specific comments".
In conclusion, don’t think the experimental data in the current version is well-constrained enough as-is to prove the argument (i.e., that bioturbated, sandy sediments are indeed significant sources of Fe_d). I was ultimately left wondering more about the Fe_d trap they developed, and wanted a more detailed accounting of the Fe in the system. If more robust experimental data (Fe and other supporting data) were added to this paper, I could see it being more convincing and very valuable. Similarly, or alternatively, a more comprehensive survey of existing Fe_d flux data from advective vs diffusive sediments (even if limited at this point) would strengthen the bigger picture argument. As is, however, I believe the current draft needs to be improved on both fronts, as it is too long for an Ideas and Perspectives piece, but does not contain a complete enough dataset for a research article. (The "sandy" vs "bioirrigated" issue may not be fixable, but I could still see the study being valuable if more data were included.)
Specific comments:
1a) Related to the first point, the paper initially seems to focus on sandy sediments (including the framing in the Abstract). Later, however, the experiment and parts of the discussion are clearly focused on bioirrigation, and the specificity of bioirrigation in sands vs bioirrigation in general is unclear. For example, Section 1.2 describes bioirrigation, although it is not entirely clear what information from this section is specific to sands, vs universal to sediments in general. (Similarly the Discussion starting at L296.) The abstract states that “the results indicate that… both biogenic and physical advection enhance fluxes”, although the simulated bioirrigation didn’t test both these variables.
1b) The authors do not offer an estimate of how important bioirrigation is in advective sediments, vs physically-driven processes. I would like to know how common bioirrigators / bioturbators are, at what density, in sandy sediments, and if they are significant enough to be a major control on fluxes. (Or conversely, if physically-driven advection is far more important.) (These processes are explored in Santos et al, 2012; The Driving Forces of…) I realize this may be difficult to generalize. However, the lack of clarity between advection and bioturbation. The cited macrofauna descriptions and densities are typically not specified between muddy or sandy environments, so it’s difficult to tell how common bioirrigation is in sands.
2a) The two treatments are really diffusive (“non-irrigated”) vs bioirrigated (“irrigated”), and thus do not provide a strong conclusion about the distinction the paper is trying to make between muddy and sandy sediments. For example, would you potentially see a similar treatment effect in diffusive vs bioirrigated treatments using nearby muddy sediments? As the authors say on L38, “Bioturbation by benthic macrofauna plays a key role for the benthic Fe cycle in both sediment types.”
2b) The higher flux in the irrigated treatments may simply be Fe_d being washed out of the sediments by a change in advective regime (moving from the non-irrigated, equilibrium period where porewater was likely anoxic and rich with Fe_d, to the well-flushed, irrigated period during which Fe_d was potentially flushed out of the porewater). The decreasing flux of Fe_d (Fig. 1) over the course of the experiment suggests the observed fluxes have not reached “steady state” for the imposed advection (in quotes, since steady state is relative for advective systems). I see the authors responded to Reviewer 2 on this topic, but I do not see the updated manuscript. I do not follow the authors’ rebuttal that the flushing would yield a flux decrease with a 2.7 day half life.
2c) The low Fe_d fluxes observed in the papers cited on L244 suggest that high Fe_d fluxes in sands are not universal. These observations were explained to be due to low sediment Fe(III) content. The current paper, however, does not present sediment Fe(III) content for comparison.
3a) The model proposed to explain high Fe_d flux in the bioirrigated sediments (Fig 2) is not well supported by data. The authors claim that since bioirrigation “results in a narrowing of the oxygenated surface zone, a large amount of Fe_d is not oxidized here but instead remains in solution.” First, no porewater O2 data are presented. Second, Fe(III) need not be oxidized in oxic sediment, but can be rapidly oxidized in the water column. Thus, an adequate mechanism for higher Fe_d flux in bioirrigated sediments is not proposed. (E.g., perhaps higher rates of OM respiration and biological activity produce more organic Fe, or Fe-binding ligands that retain Fe in solution?) The authors mention (L283) that a large fraction of the Fe_d is likely re-deposited, but could explain a bit more why they think they observe higher flux then.
3b) The porewater ammonium data are used to strengthen the Fe_d argument, but the interpretation is strange (and seemingly irrelevant to the rest of the data set). The authors claim the “the lower NH4 of the irrigated cores – assuming similar NH4 production rates in both core types – reflect additional, advective transport of solutes out of the sediment”. This is not a realistic assumption: advection fuels OM respiration by providing O2 and OM to sediment-bound microbes, so NH4 production is likely higher in the irrigated sediments. Subsequently, the authors compare ratios of NH4 *stocks* to the Fe_d *fluxes*, and claim they are “remarkably close”. (The flux ratios are not clearly presented to verify.) Why wouldn’t you present the NH4 flux data instead? Quantitatively comparing a stock to a flux makes little sense. Finally, the authors use this ratio similarity to conclude that, “when advective Fe_d transport occurs, Fe2+ escapes the sediment with little precipitation occurring.” We have no idea how much Fe reduction occurs prior to the flux measurement, so this statement is unfounded.
Minor comments:
A figure of the irrigated chamber may be helpful to visualize the set up. (Not totally necessary, but I found the description initially hard to imagine.)
L275: The authors say the Fe_d fluxes are higher than would be predicted based on bottom water O2 concentrations. Can you provide a calculation / number? What are the predicted fluxes based on O2?
L302: A word is missing here (”that maximizes… “) ?
Citation: https://doi.org/10.5194/bg-2022-247-RC3 -
AC3: 'Reply on RC3', Laura Wehrmann, 12 Mar 2023
We thank Reviewer #3 constructive comments to our manuscript which led us to further refine our manuscript introduction, discussion an conclusions. We hope that our revisions and answers to the comments are sufficient to accept this work for publication in Biogeosciences. Please find our responses to the individual comments below.
General comments:
Wehrmann et al argue that Fe_d flux from bioturbated, sandy sediments is several times greater than from muddy sediments, and that sandy sediments should thus be considered in estimates of shelf Fe_d flux. The paper provides data from an experiment in which Fe_d flux is compared between two treatments: sandy sediments with diffusion only, and sandy sediments with imposed irrigation.
First of all, I want to commend the authors for taking on the difficult task of studying permeable sediments, which are both methodologically challenging and difficult to interpret, given temporal and spatial variability. I fully agree that permeable sediments need further study, and may be disproportionately important in the global cycling of many elements. I found this paper very well written, and fun to read.
That said, I have three major concerns, related to: 1) the confusion between advection and bioirrigation in the framing, discussion and interpretation. The authors are testing and discussing two different variables -- sediment type and bioirrigation -- with only one experiment, so the true treatment is unclear; 2) the experimental data do not clearly support the study’s argument; and 3) the data interpretation that benthic Fe_d flux escapes oxidation in advective conditions is not clearly supported (while this point may potentially be true, it is not sufficiently explained by the proposed model). A few examples related to these three points are below in "specific comments".
1) In sandy sediments, bioirrigation induces the advective transport of pore-water. This process adds to the “physical” (non-biogenic) advective transport processes that commonly characterize sandy deposits along continental margins. In response to this comment and a similar one by Reviewer#1, we now phrase this more clearly throughout the manuscript. We are not conducting a comparison of sediment types, i.e., we are not testing fluxes from muds in our core experiments but solely focus on the effect of bioirrigation on the benthic Fe flux from sands. Muddy diffusive sediments have been studied extensively over the last 4 – 5 decades, and we have a better idea of the size, scaling factors and spatial variability than for permeable sands. We now more specifically describe that we are focused on sands in the introduction and discussion and also discuss that the effect of bioirrigation is superimposed on other advective processes common in sandy deposits.
2) The main message we aim to convey is that we do not have a good understanding of the size and controlling mechanisms of the Fed flux from sandy deposits but that our case study, as well as previous studies (e.g., Jahnke, et al., 2005), data point to the potential for sandy sediments to be an important source of Fe to the coastal ocean. In the context of what is known about biogeochemical processes and solute and particulate transport in sandy sediments, we highlight that the mechanisms driving the cycling and transport of iron in sandy sediments are significantly different from muddy diffusive sediments, and thus values and scaling parameters determined for muds cannot simply extrapolated to sands. The case study presented in our paper provides a glimpse of the Fe fluxes that can be expected from sandy deposits, however, our example is not intended to be all encompassing or exhaustive. The idea behind this “Ideas and Perspectives paper” is to highlight unknowns, to make the case that overlooked Fe flux sources can be significant, and to encourage further studies to enhance the understanding of the mechanisms and controlling factors driving the benthic Fe flux from sands.
3) We determined Fed fluxes using Fe accumulators connected to the experiment chambers which implies that the Fe we detected from the irrigated cores reflects the benthic Fe flux, i.e., Fe that has “escaped oxidation” within the sediment under advective conditions. Clearly, Fed is oxidized in the water column within a few minutes, but this does not appear to be as strict of an inhibitor to flux as previously thought. We now additionally show pore-water dissolved Fe inventories of initial and final irrigated and non-irrigated cores for the two (summer and winter) experiments. These data show that pore-water Fe inventories increase further during the duration of the winter experiment and remain the same in summer experiment despite a significant flux of Fed out of the irrigated cores, which supports our model that Fed is continuously released into the pore-water and subsequently transported into the overlying water column from the irrigated cores. We further address this comment in the “specific comments” section below.
In conclusion, don’t think the experimental data in the current version is well-constrained enough as-is to prove the argument (i.e., that bioturbated, sandy sediments are indeed significant sources of Fe_d). I was ultimately left wondering more about the Fe_d trap they developed, and wanted a more detailed accounting of the Fe in the system.
The traps (and a detailed accounting of Fe in our test chambers) is provided in Aller et al. 2023 which is now available online at https://www.sciencedirect.com/science/article/pii/S0304420323000178
If more robust experimental data (Fe and other supporting data) were added to this paper, I could see it being more convincing and very valuable. Similarly, or alternatively, a more comprehensive survey of existing Fe_d flux data from advective vs diffusive sediments (even if limited at this point) would strengthen the bigger picture argument. As is, however, I believe the current draft needs to be improved on both fronts, as it is too long for an Ideas and Perspectives piece, but does not contain a complete enough dataset for a research article. (The "sandy" vs "bioirrigated" issue may not be fixable, but I could still see the study being valuable if more data were included.)
Our manuscript is intended to be a “Ideas and Perspectives” contribution with the goal of emphasizing that sandy deposits function differently than muds both when it comes to solute and particulate transport, and -linked to this- the distribution and roles of different biogeochemical processes. Our “case study” along with our discussion of the biogeochemical functioning of sandy advective sediments is meant to provide a first idea of what we can expect from these deposits and to encourage further research. We have now added pore-water Fe inventories to further strengthen our experimental dataset. Unfortunately, at this point there are very little other data of Fed fluxes from sandy advective continental margin sites available that would allow for a more comprehensive survey, and any such data sets would likely have unintentionally excluded a large fraction of the iron flux in addition to having other procedural artifacts of benthic chamber work (Aller, et al. 2023). Again, this is why we wrote this Ideas and Perspectives piece!
We are not testing “sandy” vs. “bioirrigated” but aim to gain a better understanding of the benthic Fe flux from sandy bioirrigated sediments.
Specific comments:
1a) Related to the first point, the paper initially seems to focus on sandy sediments (including the framing in the Abstract). Later, however, the experiment and parts of the discussion are clearly focused on bioirrigation, and the specificity of bioirrigation in sands vs bioirrigation in general is unclear. For example, Section 1.2 describes bioirrigation, although it is not entirely clear what information from this section is specific to sands, vs universal to sediments in general. (Similarly the Discussion starting at L296.) The abstract states that “the results indicate that… both biogenic and physical advection enhance fluxes”, although the simulated bioirrigation didn’t test both these variables.
The manuscript indeed focuses on sandy sediments, and we only mention muds to emphasize some of the important differences between these deposits (and because muds have received most of the attention from researchers to date, while permeable sediments remain relatively overlooked). We have revised the Introduction section to make it clearer that we are addressing sands. The reviewer is correct in that we are specifically interested in studying the effect of bioirrigation on Fe fluxes in sandy deposits, and we tried to minimize any text that discusses bioirrigation more broadly. We have now also revised the statement in the abstract to make it clearly that we only tested the effect of (simulated) biogenic advection.
1b) The authors do not offer an estimate of how important bioirrigation is in advective sediments, vs physically-driven processes. I would like to know how common bioirrigators / bioturbators are, at what density, in sandy sediments, and if they are significant enough to be a major control on fluxes. (Or conversely, if physically-driven advection is far more important.) (These processes are explored in Santos et al, 2012; The Driving Forces of…) I realize this may be difficult to generalize. However, the lack of clarity between advection and bioturbation. The cited macrofauna descriptions and densities are typically not specified between muddy or sandy environments, so it’s difficult to tell how common bioirrigation is in sands.
Sandy deposits along continental shelves, are an incredibly important habitat for benthic macrofauna and one will not find a square meter that is not teeming with life. In our discussion section, we review data for the density of macroorganisms along the Mid Atlantic Bight ranging in numbers from 250 individuals m-2 to 18,000 individuals m-2 -as seen in our map, most of this area is sandy. Benthic macrofauna are without a doubt significant enough to be a major control on element fluxes. In densely populated intertidal or subtidal areas, bioirrigation is thought to dominate porewater exchange (Volkenborn et al., 2007). In low-biomass North Sea sand flats, about half of the advection-driven particle flux into the sediment was also attributed to bioirrigation (Rusch et al., 2000). Based on the estimates in Santos et al. 2012, the spatial and temporal scales and flow rate of bioirrigation are in a similar range to those of wave pumping and flow- and topography-induced pressure gradients. We have now added this information to the introduction section. We emphasize that our bioirrigation mimics were specifically scaled to the magnitudes and frequency patterns observed in natural patches of maldanids (Clymenella torquata) at our primary field site.
2a) The two treatments are really diffusive (“non-irrigated”) vs bioirrigated (“irrigated”), and thus do not provide a strong conclusion about the distinction the paper is trying to make between muddy and sandy sediments. For example, would you potentially see a similar treatment effect in diffusive vs bioirrigated treatments using nearby muddy sediments? As the authors say on L38, “Bioturbation by benthic macrofauna plays a key role for the benthic Fe cycle in both sediment types.”
The two treatments are diffusive and bioirrigated (biotic advection) in SANDY sediments. As pointed out in our Introduction, the upward percolation and advective transport of pore-water induced by bioirrigation and the infiltration of particles into deeper sediment layers are important mechanisms in sands but largely absent in muds, which are impermeable. Although mobile burrowers in muds can enhance advective exchange during movement, bioirrigation within fixed tubes and burrow structures does not promote interstitial advection. We realize that our previous statement “Bioturbation by benthic macrofauna plays a key role for the benthic Fe cycle in both sediment types” (although technically true) was very broad and have removed it. We are discussing particle and solute transport in sandy advective sediments in more detail in Section 1.1.
2b) The higher flux in the irrigated treatments may simply be Fe_d being washed out of the sediments by a change in advective regime (moving from the non-irrigated, equilibrium period where porewater was likely anoxic and rich with Fe_d, to the well-flushed, irrigated period during which Fe_d was potentially flushed out of the porewater). The decreasing flux of Fe_d (Fig. 1) over the course of the experiment suggests the observed fluxes have not reached “steady state” for the imposed advection (in quotes, since steady state is relative for advective systems). I see the authors responded to Reviewer 2 on this topic, but I do not see the updated manuscript. I do not follow the authors’ rebuttal that the flushing would yield a flux decrease with a 2.7 day half life.
We now show the pore-water Fed inventories to highlight that concentration of Fed remains high in the pore-water -even further increasing in the winter- over the course of the experiment. This shows that the measured Fed flux of the bioirrigated sediment is not just a “washing out” of accumulated Fed. As indicated in our response to Reviewer #2, there is no “steady-state” in these systems due to the decrease in organic matter reactivity.
In a system where production is absent (as Reviewer#2) suggested, the flux of Fed out of the sediment will decay as 1/t where tis the pore-water residence time (turnover time of the pore fluid). In this case, t = total volume of pore fluid (V)/ irrigation input (v) (note the irrigation volume input must equal the volume output across the sediment-water interface). If there is no Fed production (or consumption) within the sediment, the time dependent, volume-averaged pore water concentration is: C(t) = C0e-t/t (assuming the inflowing fluid has a Fe concentration of 0). Ignoring diffusive flux as minor relative to advection, the advective flux across the sediment- water interface is: J(t) = vC(t)/A (where A = area of the sediment –water interface). Given the dimensions and pumping rate of our system, J(t), the flux would then decay with a 2.7 day half life, which it clearly does not. Note that the relatively small diffusive flux would also decay with the same time-scale as concentration gradients become zero in the absence of production.
2c) The low Fe_d fluxes observed in the papers cited on L244 suggest that high Fe_d fluxes in sands are not universal. These observations were explained to be due to low sediment Fe(III) content. The current paper, however, does not present sediment Fe(III) content for comparison.
We double checked, and the Eitel et al. 2020 paper does not provide any sedimentary Fe(III) content values for the sandy shelf sites to compare data against. Slomp et al 1997 proposes that low OC deposition was responsible for the low Fe fluxes at the sandy sites they investigated as these were dominated by erosion. We have now updated our manuscript conclusion section to indicate that the flux of Fed from sands is likely highly variable (ie., “not universal”).
3a) The model proposed to explain high Fe_d flux in the bioirrigated sediments (Fig 2) is not well supported by data. The authors claim that since bioirrigation “results in a narrowing of the oxygenated surface zone, a large amount of Fe_d is not oxidized here but instead remains in solution.” First, no porewater O2 data are presented. Second, Fe(III) need not be oxidized in oxic sediment, but can be rapidly oxidized in the water column. Thus, an adequate mechanism for higher Fe_d flux in bioirrigated sediments is not proposed. (E.g., perhaps higher rates of OM respiration and biological activity produce more organic Fe, or Fe-binding ligands that retain Fe in solution?) The authors mention (L283) that a large fraction of the Fe_d is likely re-deposited, but could explain a bit more why they think they observe higher flux then.
We did not measure O2 in the surface sediment. However, a narrowing of the surface oxygenated zones during the pumping activity of benthic macrofauna in sandy sediment is a common observation as, for example, previously made by Volkenborn et al., 2010, 2012, Kristensen et al., 2011, and Quintana et al., 2011 for Arenicola marina, Marenzelleria viridis and Neotrypaea californiensis using oxygen imaging. We also visually observed that the oxidized sediment zone, as reflected by light – yellow color, was clearly thinner than in the non-irrigated cores, consistent with zonal compression.
We would like to point out that the benthic Fe flux is defined as the Fed moving across the sediment-water interface over a specific area and time. The reviewer is correct that a fraction of it will likely be oxidized in the water column, however, this does not affect the benthic flux but is a post-flux process. Our experimental set-ups with the Fe accumulators quantitively extract all, or virtually all, Fed released from the nonirrigated and irrigated cores (Aller et al., 2023) so differences in the effectiveness of this method cannot explain the higher flux from the irrigated cores. Overall, our results -along with the results from previous studies- suggest a higher flux of Fed from irrigated cores due to the effective advective transport of Fed out of the sediment across a narrowed oxic zone which does not allow for the quantitative reoxidation of Fed before its escape. The enhanced mobilization of iron at depth in the sediment due to intermittently oxic conditions (driving iron sulfide oxidation) likely further increased Fe release from the irrigated cores. The reviewer, nonetheless brings up an important point which is that permeable sandy sediments can be characterized by high OM respiration rates (reflected in higher oxic respiration and denitrification rates). This has mainly been explained by the 1) enhanced availability of oxygen for aerobic respiration and nitrogen compounds for coupled nitrification-denitrification, as well as the increased supply of degradable DOM into the sediment (e.g., D’Andrea et al., 2002; Marchant et al., 2016; Rusch et al., 2006). In our experiments, the enhanced inflow of oxygen and nitrogen compounds are unlikely to fuel higher dissimilatory iron reduction rates (which would be needed to explain a higher Fed flux driven by enhanced OM remineralization). We did not measure the concentration of DOM in the overlying water in our experiment which we pumped into irrigation mimics but given our experimental set-up featuring a recirculation system (and no “fresh” seawater input), we do not believe that this substrate source can explain large differences in DIR rates (and Fed fluxes) between the irrigated and nonirrigated cores as postulated by the reviewer. Importantly, as we point out in our manuscript, in nature, the advection-driven infiltration of (nano)particulate Fe oxides into the sediment is an important process that helps sustain elevated DIR rates in sandy sediments in general (e.g., Huettel, et al., 1996; D’Andrea, et al., 2002; Jahnke, et al., 2005). However, in our experiment, the use of the Fe accumulators that removed iron from the recirculating setup meant that there was minimal Fe in the overlying water (comparable to what it would be in an open system with no outside iron supply) that was re-injected using the irrigation mimics.
3b) The porewater ammonium data are used to strengthen the Fe_d argument, but the interpretation is strange (and seemingly irrelevant to the rest of the data set). The authors claim the “the lower NH4 of the irrigated cores – assuming similar NH4 production rates in both core types – reflect additional, advective transport of solutes out of the sediment”. This is not a realistic assumption: advection fuels OM respiration by providing O2 and OM to sediment-bound microbes, so NH4 production is likely higher in the irrigated sediments. Subsequently, the authors compare ratios of NH4 *stocks* to the Fe_d *fluxes*, and claim they are “remarkably close”. (The flux ratios are not clearly presented to verify.) Why wouldn’t you present the NH4 flux data instead? Quantitatively comparing a stock to a flux makes little sense. Finally, the authors use this ratio similarity to conclude that, “when advective Fe_d transport occurs, Fe2+ escapes the sediment with little precipitation occurring.” We have no idea how much Fe reduction occurs prior to the flux measurement, so this statement is unfounded.
We agree with the reviewer that without additional justification, this section did not significantly strengthen our main discussion points. Because it is not critical to our fundamental thesis, we decided that rather than expanding and defending our argument further, we would remove the section (calculation) in the revised manuscript version. Doing so helped to shorten the manuscript, partly addressing the comment by the reviewer that the paper is too long.
Minor comments:
A figure of the irrigated chamber may be helpful to visualize the set up. (Not totally necessary, but I found the description initially hard to imagine.)
A nice picture of the set-up can be found in Aller et al. 2023 (see Supplemental Materials 1).
L275: The authors say the Fe_d fluxes are higher than would be predicted based on bottom water O2 concentrations. Can you provide a calculation / number? What are the predicted fluxes based on O2?
We have added an estimated flux based on Dale et al. 2015.
L302: A word is missing here (”that maximizes… “) ?
We moved the “that” to make it clearer.
Citation: https://doi.org/10.5194/bg-2022-247-AC3
-
AC3: 'Reply on RC3', Laura Wehrmann, 12 Mar 2023
Laura M. Wehrmann et al.
Laura M. Wehrmann et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 419 | 164 | 27 | 610 | 12 | 14 |
- HTML: 419
- PDF: 164
- XML: 27
- Total: 610
- BibTeX: 12
- EndNote: 14
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1