Review of Carbon sequestration along a gradient of tidal marsh degradation in response to sea level rise

The authors present a paper describing variation in sedimentation and organic carbon accumulation between levee and basin position and along a gradient of degradation in a tidal marsh of the Chesapeake Bay ecosystem. The authors collected data from eight study sites representing levee or basin geomorphic positions and in different plant communities from the least degraded to most degraded portion of the marsh with degradation being caused by sea level rise and eventual reclamation of these wetlands by the bay. Degradation is determined by the ratio of vegetated to non-vegetated area. The authors conclude there are substantial differences in sediment deposition and carbon accumulation between levee and basin position and more modest differences along the degradation gradient. Overall the paper is well written, clear, and easy to follow although I have some concerns I would like to see addressed.

1. The authors should emphasize what is novel about the work they have conducted. Much of the paper summarizes basic physical geography of tidal wetlands. That levee positions receive more sediment is not novel – in fact, its why they are levees. Similarly, that they receive more OCAR input is simply because they receive more sediment. Plant community differences similarly are not novel – the zonation of tidal wetland communities associated with deposition and salinity are well understood. What seems most interesting is the degradation aspect of the study and how these systems change with sea-level rise. I think this theme could better come forward in the paper overall.
2. I am somewhat confused by the design of the sampling. As written in the text, the sampling appears quite limited. Table 1 and the text around 145-155 suggests 8 sampling sites and 4 soil cores per site. This would make n=32. However, the figures showing data points such as Figs. 2 and 6 suggest many many more data points. The sampling needs to be clarified. Moreover, please describe how the sampling withing a site is independent. Were samples collected along multiple transects? Minimum distance between soil cores? Overall how the soils were collected needs to be better described.
3. Similarly, I am concerned about the sampling of the gradient in degradation. The gradient is described/quantified as the ratio of vegetated to non-vegetated surface and while this ratio is reported in table 1, the sampling as I understand it is somewhat misleading since only the vegetated portions of the marsh were sampled, regardless of gradient position. Clearly the vegetated and non-vegetated portions of the marsh would experience differences in OCAR input so the decision to only sample vegetated – i.e., least degraded regardless of the degradation gradient needs to be justified and the implication of this choice clearly described.
4. The description of the statistical analysis is too limited for the statistical procedure to be evaluated. Please expand the analysis section to indicate if fixed or mixed models were used and any random effects, any data transformations, selection of post-hoc tests (the results of which are show in the figures).
5. I find several inconsistencies in the arguments surrounding the differences between the levee and basin communities. Line 380 suggests that high accretion rates in levees may be due to rapid burial of organic matter and low O2 availability leading to lower decomposition. However, on line 339, there is the suggestion that deep-soil pore water drainage on levees promotes oxygenation and more rapid plant growth. While perhaps these can both be true depending on the precise depth of anoxia, its reads as inconsistent. Similarly, on line 394, the packing of high-density mineral matter on the levees is used as a justification for the greater bulk density of levee soils would further argue against rapid drainage and oxygenation.
6. I have concerns with the interpretation of the 13C data as presented here. The sediment varies considerable in 13C suggesting different sources of OCAR input as the authors indicate. However, the endpoints of the carbon is somewhat ambiguous. The argument is made that 13C can determine the difference between autochthonous C and allochthonous C. However, autochthonous C can come from two sources – the C4 grass Spartina and C3 rush Schoenoplectus while allochthonous C is assumed to be C3 (presumably phytoplankton and other algae). Therefore, while seeing a highly C4 signature in sediment is good indication of local C in a Spartina zone the opposite is not necessarily true since the deposition could be from OCAR input from outside the wetland as well as OCAR input from remobilized sediment with a local source. Please address this concern in interpreting these data. Figure 6 I think shows the community shift happening with the Schoenoplectus OCAR being mostly C4-derived in the least degraded and intermediate sites and mostly C3 derived in the most degraded. Since this is a C3 plant, the data suggest a recent conversion (and the large error bars support this) in the least and intermediate sites but a long-term history of the C3 rush in the most degraded. Combined with the assertion that basins are sediment starved, the data argue for local carbon inputs dominating the basin system.

General Comments

Thank you for the opportunity to review the manuscript titled “Carbon sequestration along a gradient of tidal marsh degradation in response to sea level rise” by Mona Huyzentruyt and colleagues. This paper reports on the differences in organic carbon accumulation rates among levees and basins along a marsh degradation gradient within a microtidal wetland. The results of the paper indicate that carbon accumulation rates are much greater on the levees than within the basins. Some differences in organic carbon accumulation rates were detected among the different marsh degradation zones which may suggest accumulation rates tend to be greater in more degraded zones.

The manuscript represents a substantial contribution to scientific progress, and important new data, in the Biogeosciences. The paper is well organized and written without errors but could be improved by editing to reduce text, and re-arranging where some information is introduced within the text.

There are a few issues within the statistical approach that can be addressed and will likely not have a large effect on the main results reported by the paper.  Specifically, there are several analyses (dry bulk density and C-13) where it appears all of the subsamples from all cores were included as individual observations and run in an ANOVA. If this is the case then this method would artificially inflate the sample size and therefore artificially decrease the error term, as each sample is being treated as independent even though multiple samples came from the same core. To deal with this, the authors could consider a repeated measures ANOVA or linear mixed modeling approach where “depth” can be nested within “core”. It could be that the authors want to consider taking this approach in other analyses as well to potentially gain a little more insight from the data they have, as it could help take advantage of all the depth data rather than just averaging it all to one value. But the other analyses are ok as is if the authors don’t want to make that change.

The authors also make the claim that the degraded marsh zones are experiencing the same rate of relative sea level rise as the other zones, but this should be better discussed and documented as they also make statements that suggest the degraded zone may be decreasing in elevation (more ponds, etc) which would mean that the rate of relative sea level rise may be greater in the degraded marsh zones.  

This paper is sound, and the results are interesting. I believe this research will be of interest to the readers of Biogeosciences.

Specific Comments:

Graphical abstract

In the graphical abstract – suggest indicating the direction of the coast is given that the setting is describing tidal marsh ecosystems.

Abstract

In line 24 the authors state: “Additionally, OCAR was observed to increase with increasing degree of marsh degradation in response to sea level rise” but really there is just one difference in the basin rates (‘most degraded’ is higher) and one difference in the levee rates (‘least degraded’ is lower).

Introduction

Does ‘marsh degradation’ specifically mean less vegetation? If so, make sure this is clearly defined in the Introduction.

How wide-spread is the phenomenon of decreasing vegetation in tidal marshes globally?

How were the degradation zones (least degraded, intermediately degraded, most degraded) determined? Now I see this is answered in the supplement, but a brief explanation should be included in the paper text (the material in the supplement can remain the same).

Introduce the difference between C3 and C4 vegetation and why it matters in this context in the introduction (as they were investigated separately in this work).

Methods

It seems that four cores were taken at within each zone, but at the same site. Why weren’t cores taken from multiple sites throughout the zone… that seems like it would better represent the carbon dynamics of that zone.

Explain why two vegetation types were sampled. Perhaps the C3 vs C4 difference should be introduced in the Introduction if the authors think it is important.

Why were water samples taken at just one location and how is this information used? Was the site inundated, and this was the water present above the soil surface?

Statistical Analyses

The description of the stats leaves some questions. Were levee and basin sites from all of the degradation zones all analyzed against each other, or were levee and basin nested within zone?

Figure 2.

It is not statistically appropriate to pool values from different depths within one core as they are not independent. However, the authors could address this by adopting a linear mixed effects modeling framework that nests “depth” within “core”. Such an approach may actually be useful for some of the subsequent analyses as well because it may help the authors determine more about the differences in carbon density, for example, in the basin sites along the degradation gradient that currently are all the same using the ANOVA analyses, but the differences may be parsed out if all of the measurements from each core were included. It may be overkill as the authors do not specifically have ‘depth’ questions – but as the analyses are set up right now a lot of information (and work!!) is being ‘tossed out’ as the cores are averaged to just one value.

Lines 266-270, as the authors point out, there is no statistical difference in the mean carbon density among the basins of different degradation zones, so that cannot be reported as a finding.

Introduce the C3 vs C4 difference among the two vegetation types earlier as the reasoning for separating them and explain why this is important in introduction.

Why are the individual depth measurements used for the C-13 plot and not for the others, and how did the authors avoid pseudo-replication? (Does the statistical model include depth nested within core to avoid inflating the sample size and artificially shrinking the error term?) Remember that multiple depths are essentially ‘repeated measures’ within a core.

Discussion

Figure 7 is unnecessary.

Line 312 states that the study examines: “accumulation rates (OCAR) in response to gradients in marsh degradation and levee-basin gradients.” It seems there is an important distinction between examining accumulation rates across gradients of degradation, and “in response” to degradation. It seems the authors are doing the former and therefore should use that language here, i.e. change to “accumulation rates across gradients in marsh degradation.”

Line 313, what is the relative area of levees to basins in this wetland, and in most tidal wetlands? This will help provide context on the relative importance of these ‘hotspots’

Line 315: when the authors state that levees are “among the fastest carbon accumulating environments on Earth” they are talking specifically about soil organic carbon accumulation, right?

Figure 8. Clear and relatively easy to understand but at first glance the relative size of the arrows among the levee and the two vegetation communities appears to be the same. It is difficult to determine the ‘point’ of the conceptual figure – are there differences in the relative strength of these processes among the different locations? It seems there must be if the accumulation rates are so different, but it is difficult to see this from the figure.

Line 410: Be specific that one basin rate differs from the other two and one levee rate differs from the other two. Especially given that these rates were only sampled in one location per zone (via three cores), it seems to be overstating the results a bit to claim that there is an increase in OCAR with increasing marsh degradation.

Line 419: This is interesting - do the authors know that the degraded march experiences the same rate of relative sea level rise? It seems that it could be slightly different given that vegetation has been lost so perhaps rates of accretion are lower? If the degraded area is experiencing any subsidence, or even just lower rates of accretion, then it would be experiencing a faster rate of relative sea level rise.

Lines 430-435: all of the processes described between Line 430 and 435 indicate that degraded marshes do experience sediment loss which would then make them vulnerable to higher rates of relative sea level rise.

Technical Corrections:

Section 4.1.1. Minor writing suggestion - three sentences in a row start with “This”, consider rephrasing to reduce redundancy. The section could likely also be condensed.

Line 355: remove “be expected to”

Line 368: remove one parenthesis after “Ganju et al., 2013”