An Estimate of the Global Marine Carbon Flux of Cold-Water Corals

Steinberg, Amelie T.; Frank, Norbert

doi:10.5194/egusphere-2026-5117

Preprints

https://doi.org/10.5194/egusphere-2026-5117

Preprints

26 Aug 2026

| 26 Aug 2026

Status: this preprint is open for discussion and under review for Biogeosciences (BG).

An Estimate of the Global Marine Carbon Flux of Cold-Water Corals

Amelie T. Steinberg and Norbert Frank

Abstract. Marine calcification constitutes an important component of the global carbon cycle by sequestering carbon. During skeletal growth corals accumulate carbon, with numerous studies have highlighted their potential contribution to long-term carbon storage. This study compiles data from 32 published studies attempting to provide first estimates of the annual global carbon accumulation through cold-water corals. Carbon accumulation rates from individual reef systems were averaged and calculated up to global scale. As the worldwide coral covered area remains incompletely constrained, two different approaches were used, based on (i) coral mound density and (ii) regional coral area percentages, to asses the global suitable area. In addition, the proportion of living coral cover and the percentage of exposed mounds were taken into account. The global carbon accumulation was estimated using Monte-Carlo uncertainty propagation. The resulting simulations yielded probability distributions for the global carbon accumulation, from which the mean, median and 5th and 95th percentiles were derived. The mound-density-based approach resulted in medians, 5th and 95th percentiles of 0.003 (8.7x10^-5–0.079) Gt C yr^-1, whereas the area-based approach yielded 0.0022 (2.0x10^-5–0.086) Gt C yr^-1. The calculated means are approximately one magnitude higher than the corresponding medians, reflecting the asymmetrical and strongly right-skewed distribution. Consequently, most Monte Carlo realizations produced relatively low carbon accumulations, whereas only a small number of parameter combinations resulted in substantially larger values. The uncertainty spans approximately four orders of magnitude, emphasizing strongly the need for additional observational data to better constrain the parameters. The estimated global carbon accumulation of cold-water corals is approximately one magnitude lower than published estimates for tropical corals and two orders of magnitudes lower than the pelagic carbon burial. Compared with other carbon reservoirs cold-water corals therefore likely play a subordinate role in the global carbon cycle, but may locally be of importance. The estimated accumulation corresponds to a burial of 0.01 % (mound-density-based) or 0.076 % (area-based) of the global annual oceanic net uptake of CO₂. As anthropogenic CO₂ emissions continue to rise, cold-water coral ecosystems are increasingly threatened by ocean acidification and warming. However, the implications of these environmental changes for the global carbon cycle are expected to be considerably greater for tropical coral reefs and pelagic carbonate-producing systems.

Received: 25 Aug 2026 – Discussion started: 26 Aug 2026

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Amelie T. Steinberg and Norbert Frank

Status: open (until 30 Nov 2026)

Post a comment Subscribe to comment alert

CC1: 'Comment on egusphere-2026-5117', Giacomo Medici, 03 Sep 2026 reply

General comments
Very good research on carbon fluxes in carbonate environments. Please, follow my feedback to improve the new version of the manuscript.

Specific comments
Line 31. Insert references for physicochemical process with dissolution of calcite (physicochemical dissolution on coral mounds) and production of dissolved CO₂.
- Medici, G., Marianelli, D., Cornacchia, I., Gori, F., Brandano, M. 2026. Multi-disciplinary approach to paleokarst occurrence in the Eocene–Oligocene succession of the Apulia Carbonate Platform (Salento, Italy). Facies, 72(2), p.17.
- Jiu, Bo, Wenhui Huang, Nana Mu, Ruilin Hao 2022. Petrology, mineralogy and geochemistry of Ordovician rocks in the southwest of Tarim Basin, implications for genetic mechanism and evolution model of the hydrothermal reformed-paleokarst carbonate reservoir." Marine and Petroleum Geology, 140, 105687.

Line 99. Be more explicit on the overall goal of your research at the end of your introduction.
Line 99. Describe the 3 to 4 specific objectives of your research by using numbers (e.g., i, ii, and iii).
Lines 100-199. The methodology description is too long. Do you need to divide in sub-paragraphs?
Line 110. Formula unclear with issues of rigorous formatting.
Line 111. Justify the choice for starting from Dorschel et al. (2007).

Figures and tables
The nine tables look good, some issues on the figures.
Figure 1. Do you need a spatial scale for the conceptual scheme of carbonate platform?
Figure 1. More information is needed in the caption.
Figure 2. You need explanation in the caption for upper limit to 137.
Figure 2. Also here, much more description is needed in the caption.
Figure 4. You are describing the graph in the caption. I would insert this part in the main body of the manuscript.

Reply

Citation: https://doi.org/10.5194/egusphere-2026-5117-CC1
RC1:
'Comment on egusphere-2026-5117', Anonymous Referee #1, 21 Oct 2026 reply
Referee Report Egusphere-2026-5117: “An estimate of the global marine carbon flux of cold-water corals”
In the present paper the authors provide a global estimate for cold-water coral carbon accumulation and compare it to tropical coral accumulation and pelagic burial. This estimation is based on the following equation:
Coral C accumulation rate x Habitable Area x Mound covered area x Live coral cover x Exposed mounds = Global C accumulation
Although the equation presented in figure 1 makes sense conceptually, except for the exposed mounds parameter (see below), the authors’ approach has several substantial weaknesses:
Incorrect usage of published data for habitat area, mound coverage area and coral cover, partly values are not consistent with their source

The consideration of buried mounds is not necessary

The intersection of the different datasets provides additional problems that cannot be approximated.

None of these issues are discussed adequately.

A detailed description of all these issues is provided below.
Even if the methodological approach was correct, the conclusions seem quite biased. The global C accumulation rate provided has an uncertainty spanning over 4 orders of magnitude, and hence it cannot be used as an argument that it is one order of magnitude lower than that of Tropical Corals.
Due to all the methodological issues and the added up biased discussion, I can only recommend to reject this manuscript. I did not proceed with further detailed comments as the raised major concerns are too substantial.

(I) Coral C accumulation rate
Considering this is the baseline data from which all the follow-up calculations are made, more care should have been taken to make sure data was correctly extracted from the original papers, and to check that the correct papers were being used. When looking at Table 1, for Greiffenhagen et al. (2026) all the background sediment values are wrong as they are only including inorganic carbon and not its organic fraction present in the sediment, which is outlined in the original paper. The same occurs with the data coming from Greiffenhagen et al. (2025). Additionally, some of the numbers are wrong when compared to the original paper. Almost all the aggradation rates are copied from the wrong column of the original paper and instead reflect the Total Carbonate Accumulation (Table 3 in Greiffenhagen et al., 2025). The data from Titschack et al. (2015), presents a small conversion mistake in the total accumulation of all cores.
A second aspect that should always be considered when approximating rates from sediment core data is the time averaging effect – rates decrease with decreasing resolution of the age model due to the interpolation over hiati. The values acquired from Titschack et al. (2009) are not usable here, as the aggradation rates have been derived from coral fragments with age uncertainties of 350 kyrs, most likely underestimating aggradation rates and overlooking the presence of mound development hiatuses. Data from Dorschel et al. (2007) cannot be used for the presented paper, because the core age models are only derived from 4 samples each, encompassing more than 100 kyrs and several mound formation hiatuses. The authors consider the entire cores as a unique mound formation period, completely underestimating the aggradation rates and hence, the carbon accumulation rates of this mound.
In Titschack et al. (2016), core GeoB11135-2 cannot be used. The gravity core was retrieved from a talus deposit at the foot of a submarine wall and not a coral mound, which becomes a problem when the two approaches used to get the global estimate are based on coral mounds densities and coverage. Using the data from Hebbeln et al. (2019) is also not an appropriate approach, as the authors did not measure such rates themselves, instead, they estimated them making many assumptions and likely underestimating mound aggradation rates. Both van Oevelen et al. (2009) and Cathalot et al. (2015) are dealing only with organic matter ingestion by the corals, not the carbon sequestered and stored in the sediments through calcification, which makes this data not usable to estimate global carbon accumulation. Lindberg and Mienert (2005) and Tamborrino et al. (2022) data have the same issues as Hebbeln et al. (2019).

(II) Habitable area
The authors use the species richness habitat suitability model from Fragkopoulou et al. (2025) to predict the distribution of corals mounds globally, assuming that among all suitable habitats one of the species predicted would be a framework-building cold-water coral. This is far from being the correct approach for many reasons:

It is definitely not correct to assume that one of the present species would be a mound-forming cold-water coral. Why should this be the case?

If the authors aim at estimating carbon sequestered by coral mounds, the habitable area cannot include areas outside the continental margins and/or deeper than the current known existence of coral mounds.

The model used includes more than 700 species many of which are octocorals, black corals and stylasterid corals. Scleractinian corals is just one of the groups included and mound-forming cold-water corals are a very small subgroup (max. 6 – 8 species).

Even if it was modelling scleractinian framework-building corals, this model is predicting presence of individual colonies, not mounds.

It is unclear which species richness value was considered as suitable for the presence of coral mounds.

(IV) Mound covered area
Although including this term is likely a good idea, there is a need for a better explanation on how the numbers were extracted. In any case, this term can become quite problematic. How is the total area explored decided? This is crucial as the total area should be exclusively determined by the outer mounds limiting a mound province extension. If the total area is only related to the available multibeam bathymetry, side-scan sonar etc., the background area used to make the calculations is completely arbitrary. Probably the approach followed by Tamborrino et al. (2022), where he takes a radius of 1 km from all mounds to define the mound province background area, is the most adequate one. However, this approach should be applied to all other datasets.

In addition, the data extraction from Tamborrino et al. (2022) is wrong. The values from Table 2 correspond to the Median footprint area in the original paper (Fig. 6A in Tamborrino et al., 2022), not the mean.
The paper from Di Giovanna et al. (2025) makes no sense in Table 3. This paper is showing a habitat suitability model for scleractinian CWC colonies in submarine canyons and based on ROV videos. It is not talking about mounds anywhere.

(IV) Live coral cover
According to the equation in figure 1, the live coral cover is needed to estimate the area where high carbon accumulation rates occur. In principle, there are two possibilities to obtain this data: (i) An ROV or AUV mapping of an entire mound with photography and a subsequent image classification (see Lim et al. 2017); or (ii) the usage of ROV transects and a subsequent sea floor prediction map as performed by e.g. Sowers et al. (2024), De Clippele et al. (2017) and Heindel et al. (2010; not cited in manuscript). However, the usage of many datasets provided in Table 4 are not adequate due to the following arguments:
Vad et al. (2017) provides a live coral percentage based on individual colonies, the thickness ratio of living to dead part of colonies, surface percentages of reef cover on a mound were not provided.
Sowers et al. (2024) provided an aerial estimate for various ‘mounds’ habitats (slope, ridge, peak). All three habitats for a ‘mound’ and should be considered in context (relative areal contribution to the mound) instead of being used separately.
Overlap of data between De Clippele et al. (2017; subarea of 2021) and (2021). Furthermore, taken the numbers from the first paragraph of the results, live and dead framework is not separated and estimated to cover 17.6 % and coral rubble 76.5%...
De Oliveira et al. (2022) cannot be used because this study does not cover an entire coral mound. De Oliveira et al. (2021), Lim et al. (2017) and Conti et al. (2019) studied the same mound!!! It is Lim et al. 2017, which should be taken not 2018. Lim et al. (2018) provided mound density data. The numbers provide by Conti et al. (2019) seem to be wrong. According to the paper the dead coral should be >~25 % and live coral between 3 and 4 %. Actually, only the data of Lim et al. (2017) should be taken as this is based on seafloor observation of an entire mound.
Cordes et al. (2023) estimated the coverage also from ROV and submersible surveys. It is unclear if all habitats were covered and if the percentages of the relevant habitats in the region were considered relatively.
Based on the comments above, it becomes clear that the sources to estimate live coral cover of coral mounds are extremely heterogenous and several studies are actually not usable for estimating the live coral cover of a coral mound.

(V) Coverage of exposed and buried mounds
The consideration of buried mounds does not make sense to me. Completely buried mounds will not be considered in morphological detections of coral mounds and mound density calculations (see methodological approach of e.g. Tamborrino et al. 2022), parasound or seismic data would be needed for their detection. For partially buried mounds only the part of the mound above seafloor would be detected and only this part might be partially covered by living corals. In terms of the equation provided in the paper and considering the explanations given above, the authors were never accounting for buried mounds to start with, as their coverage/density approaches are only accounting for exposed mounds. This parameter is thus, completely unnecessary. Finally, none-buried, as well as partially-buried mounds might not have any living corals on top in an entire province (Tamborrino et al. 2022) or only some of them might not have living corals as indicated by the data presented in Wienberg et al. (2018). Consequently, the consideration of buried mounds in the presented equation (Fig. 1) does not make sense. Even worse, the occurrence of coral mounds with and without living corals on top within the same coral mound province presents an unsolved problem in any upscaling approach of carbon(ate) accumulation rates.

(VI) Intersection of datasets in the global carbon accumulation equation (Fig. 1)
In the equation presented in figure 1, several datasets are used, each for itself presents important information for the evaluation of CWC reefs as carbon sinks but the intersection of these datasets is difficult. For example, the Habitat Area can be modelled and the Mound Covered Area can be calculated with bathymetric surveys but coral mound provinces are patchy and do not cover homogenously the Habitat Area. Currently, the coverage of coral mound provinces in the Habitat Area cannot be estimated. The same problem accounts to the Mound Covered Area and the live Coral Cover. Both datasets can be produced and datasets are available for both but there is no information on how many coral mounds have a living coral cover within a coral mound province and if this ratio varies from province to province (is to be expected) or is rather stable.

Carbon cycle discussion
Fossil fuel combustion transfers carbon from the lithosphere to the atmosphere. Main focus of carbon cycle research is the evaluation of atmospheric CO₂ sinks, hence the partitioning of carbon between the short-term reservoirs: atmosphere biosphere and hydrosphere (short-term carbon cycle; e.g., IPCC report; Friedlingstein et al. 2020), as this exchange processes are extremely relevant to reduce atmospheric CO₂ concentrations on short time scales (years to hundreds of years). On a long run, only the carbon transfer back into the lithosphere (Holmèn 2000) balances the anthropogenic fuel combustion. Therefore, an improved understanding and budgeting of carbon, organic and inorganic (carbonate) carbon, accumulation is essential. Consequently, the cold-water coral reef carbon accumulation rates should consider organic and inorganic carbon (see Greiffenhagen 2025, 2026) and should be compared rather to the other long-term carbon sinks (reefs, peat, etc.) that transfer carbon back into the ‘lithosphere’ reservoir than to the partitioning rates between the short-term reservoirs. Additionally, ignoring the hemipelagic background carbon(ate) because ‘it represents part of the general pelagic sedimentation associated with marine primary production rather than carbonate production by CWCs` (p. t, lines 129 – 130) is in my opinion wrong. The background sediment accumulation rates in coral mounds exceed often significantly the sediment accumulation rates of the adjacent seafloor (in Greiffenhagen et al., 2025; 2026 they can be 1 order of magnitude higher in coral mounds with respect to the surrounding seafloor). Consequently, also the background sediment carbon accumulation rate is enhanced and should be considered. Finally, as carbon is also transferred on the adjacent seafloor into the lithosphere, only the additional carbon accumulation should be considered.
Used references:
Holmén K (1992) The global carbon cycle. In: Butcher SS, Charlson RJ, Orians GH, Wolfe GV (eds) Global Biogeochemical Cycles. Academic Press, New York, pp239–262
Friedlingstein P, O'Sullivan M, Jones MW, Andrew RM, Hauck J, Olsen A, Peters GP, Peters W, Pongratz J, Sitch S, Le Quéré C, Canadell JG, Ciais P, Jackson RB, Alin S, Aragão LEOC, Arneth A, Arora V, Bates NR, Becker M, Benoit-Cattin A, Bittig HC, Bopp L, Bultan S, Chandra N, Chevallier F, Chini LP, Evans W, Florentie L, Forster PM, Gasser T, Gehlen M, Gilfillan D, Gkritzalis T, Gregor L, Gruber N, Harris I, Hartung K, Haverd V, Houghton RA, Ilyina T, Jain AK, Joetzjer E, Kadono K, Kato E, Kitidis V, Korsbakken JI, Landschützer P, Lefèvre N, Lenton A, Lienert S, Liu Z, Lombardozzi D, Marland G, Metzl N, Munro DR, Nabel JEMS, Nakaoka SI, Niwa Y, O'Brien K, Ono T, Palmer PI, Pierrot D, Poulter B, Resplandy L, Robertson E, Rödenbeck C, Schwinger J, Séférian R, Skjelvan I, Smith AJP, Sutton AJ, Tanhua T, Tans PP, Tian H, Tilbrook B, van der Werf G, Vuichard N, Walker AP, Wanninkhof R, Watson AJ, Willis D, Wiltshire AJ, Yuan W, Yue X, Zaehle S (2020) Global Carbon Budget 2020. Earth Syst Sci Data 12:3269–3340, 10.5194/essd-12-3269-2020

Reply
Citation: https://doi.org/10.5194/egusphere-2026-5117-RC1
- CC2: 'Reply on RC1', Norbert Frank, 28 Oct 2026 reply
  
  Please find enclosed our point by point response to RC1
  
  Reply
  
  Citation: https://doi.org/10.5194/egusphere-2026-5117-CC2

Amelie T. Steinberg and Norbert Frank

Viewed

Total article views: 523 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
284	170	69	523	73	64

HTML: 284
PDF: 170
XML: 69
Total: 523
BibTeX: 73
EndNote: 64

Views and downloads (calculated since 26 Aug 2026)

Month	HTML	PDF	XML	Total
Aug 2026	61	23	4	88
Sep 2026	118	60	50	228
Oct 2026	96	83	14	193
Nov 2026	9	4	1	14

Cumulative views and downloads (calculated since 26 Aug 2026)

Month	HTML	PDF	XML	Total
Aug 2026	61	23	4	88
Sep 2026	118	60	50	228
Oct 2026	96	83	14	193
Nov 2026	9	4	1	14

Viewed (geographical distribution)

Total article views: 515 (including HTML, PDF, and XML) Thereof 515 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 03 Nov 2026

Short summary

This manuscript presents a first global estimate of carbon burial through cold-water coral (CWC) ecosystems, addressing a critical gap in understanding marine carbon sequestration. The study compiles data from 32 published studies and uses Monte-Carlo simulations to estimate global carbon accumulation rates. The manuscript is well-structured and aligns with the scope of Biogeoscience, which focuses on biogeochemical processes in the Earth system.


Total:	0
HTML:	0
PDF:	0
XML:	0