Seven millennia of carbon accumulation in the Lower Danube Floodplain controlled by base-level change and anthropogenic forcing

Țuțuianu, Laurențiu; Zăinescu, Florin; Vespremeanu-Stroe, Alfred; Stan, Ionel Bogdan; Dobre, Mihaela; Luca-Ilie, Maria; Sava, Gabriela; Mănăilescu, Cristian; Preoteasa, Luminiţa

doi:10.5194/egusphere-2026-1693

Preprints

https://doi.org/10.5194/egusphere-2026-1693

Preprints

16 Apr 2026

| 16 Apr 2026

Seven millennia of carbon accumulation in the Lower Danube Floodplain controlled by base-level change and anthropogenic forcing

Laurențiu Țuțuianu, Florin Zăinescu, Alfred Vespremeanu-Stroe, Ionel Bogdan Stan, Mihaela Dobre, Maria Luca-Ilie, Gabriela Sava, Cristian Mănăilescu, and Luminiţa Preoteasa

Abstract. Floodplains are among the most important long-term terrestrial carbon sinks, yet the controls governing millennial-scale carbon accumulation remain poorly constrained, particularly in large fluvial systems. Here we reconstruct the history of organic carbon accumulation in the Lower Danube Floodplain (LDF), one of the largest floodplain systems in Europe, using stratigraphic and chronological data from eight sediment cores spanning the last 7000 years. Total organic carbon (TOC), grain-size distribution, dry bulk density, and sedimentation rates derived from Bayesian age–depth models were integrated to quantify temporal changes in carbon accumulation rates (CAR) and to evaluate the geomorphic and environmental controls governing long-term carbon burial.

Over the investigated period, large volumes of sediment accumulated across the LDF, forming a thick stratigraphic sequence that stores substantial amounts of organic carbon. However, long-term carbon burial did not occur at a constant rate but instead reflects three distinct phases controlled by changing boundary conditions. During the early part of the record, rapid floodplain aggradation associated with post-glacial base-level rise promoted efficient carbon burial through accelerated mineral sediment deposition despite relatively low organic carbon concentrations. Following base-level stabilization, sedimentation rates declined and the system shifted toward preservation-dominated sequestration, characterized by higher organic carbon contents but lower carbon accumulation efficiency under more stable hydrogeomorphic conditions. During the last two millennia, increasing human activity in the Danube basin enhanced sediment delivery, leading to a renewed increase in carbon burial despite declining organic carbon concentrations.

Superimposed on these millennial-scale trends, carbon burial also varies substantially among sedimentary facies, reflecting contrasting depositional environments and carbon sequestration mechanisms within the floodplain. Facies-based analyses reveal contrasting mechanisms of carbon storage across floodplain environments. Peat and organic-rich deposits exhibit the highest TOC values (average 17 %) but relatively moderate CAR due to low sediment accumulation rates, whereas paleochannel and overbank deposits achieve higher CAR through rapid burial of mineral sediments containing lower organic carbon concentrations. These results demonstrate that long-term floodplain carbon sequestration is governed by the interaction between accommodation space, sedimentation rate, and hydrological connectivity rather than organic carbon concentration alone.

These findings highlight the importance of maintaining hydrological connectivity and sediment delivery in order to sustain carbon burial in large floodplain systems, suggesting that restoration strategies focused on reconnecting floodplain surfaces to fluvial processes may enhance long-term carbon sequestration and associated ecosystem services.

Received: 04 Apr 2026 – Discussion started: 16 Apr 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

Laurențiu Țuțuianu, Florin Zăinescu, Alfred Vespremeanu-Stroe, Ionel Bogdan Stan, Mihaela Dobre, Maria Luca-Ilie, Gabriela Sava, Cristian Mănăilescu, and Luminiţa Preoteasa

Status: final response (author comments only)

RC1:
'Comment on egusphere-2026-1693', Anonymous Referee #1, 10 Jul 2026
The authors present a potentially valuable Holocene dataset of TOC and carbon accumulation rates from the Danube floodplain although I am not convinced that this merits publication in Biogeosciences as it does not present substantial new ideas or concepts. Below are my specific comments intended to improve the paper:

A map of the lower Danube floodplain is needed. The satellite view of the study region in Fig. 1 does not contain delineations of geomorphic zones and is hard to read for somebody who is not familiar with the region. It does appear that only the most distal sector of the floodplain has been considered.

Many local names discussed in text are not provided in figures so the reader can understand what is being discussed.

References are missing regarding previous work on stratigraphy of the region. Several old boreholes (1, 2, 3) that are presented graphically in Fig. 1 are not referenced and are not properly described in text.

In general Fig. 1 is too busy. Previous boreholes are not presented at their ground elevation as their newer equivalents. I suggest splitting the figure: map with locations spelled out, old cores, new cores.

The new cores in Fig. 1 should be simplified. Too many colors are used making them hard to read. I suggest adopting a black and white scheme for sedimentology/stratigraphy. Keep the facies interpretation with colors and extend to the bottom of each core. Why does facies identification end mid-core in most cases?

Fig. 2 is also very hard to read. I suggest using lines instead of filled pattern for TOC and sed rates as drivers and keep filled pattern only for CAR as a derived parameter. Text is unreadable due to font sizes – needs to be addressed. Take inset for the top 1 m out – see comment below.

The material dated by radiocarbon needs to be presented in Table 1 – what has been dated: bulk carbon, peat, plant parts???? This should be done for all dates ideally but certainly for the new ones and old ones that have not been documented in this manner in their original papers. Justifications are needed for each eliminated outlier rather than a blanket statement in methods.

Bayesian age–depth modeling does not fully account for the highly variable sedimentation on floodplains. This is major point that should be discussed: how representative are the TOC measurements for highly heterogenous stratigraphies? How is a simple averaging justified? A Monte Carlo approach is more warranted for such cases.

More documentation is needed (photos, x-radiography) as well as references on assigning depositional environment from facies. I checked some of previous publications on cores presented here and such documentation is also missing there. Assigned lithostrat units for the region are not clear in the heterogenous cores presented in this paper. Clearly delineate them in figures and support assertion by data.

The use of depth distribution in the upper 1 m to assess anthropogenic impacts is not warranted as sedimentation rates are presumably different in time for each core and certainly different between cores. A detailed age model is needed such as based on radiogenic isotopes with appropriately short half-lives. The discussion at the end of paper on the recent period is unwarranted without such data.

The OC budget for LDF as 1.7 GT is confusing. What area was used for computation? How is the average depth that was used supported? Is this the whole LDF as stated in paper or just a part of it?

What do the authors mean by “geomorphic stability” in Table 2? How was this assessed?

The fact that TOC is not correlated with depth is not a new finding in a heterogenous floodplain deposit. Any long-term record of OC in such environment shows that.

Authors say: “Areas subject to frequent or prolonged inundation – such as topographic depressions, abandoned channels, or infilled lake basins – consistently exhibit TOC values of 7–10 % higher than those of rarely flooded or well-drained environments.” Where do they present such data systematically? Peats and organic rich sediment are presented in Fig. 5. The CAR data in same figure is more interesting but needs to documented better in terms of facies detection – see above.

The sea level driven TOC, CAR scenario is the same as any similar environment – too many references to cite but the authors should integrate this finding in a larger knowledge context than the one or two cases they citedand discuss what makes the Danube different, if anything.

Bottom line: What is the added value of this study beyond the set of new data?
Citation: https://doi.org/10.5194/egusphere-2026-1693-RC1
RC2:
'Comment on egusphere-2026-1693', Thomas Hoffmann, 19 Aug 2026
Comments on „Seven millennia of carbon accumulation in the Lower Danube Floodplain controlled by base-level change and anthropogenic forcing „ submitted by Tutuianu et al to Biogeoscience.
The manuscript presents a very valuable and interceding dataset of long-term carbon accumulation in the floodplain of the Lower Danube, which is connected to sea level changes of the Black Sea. The study is based on a solid sampling and analysis design and provides an interesting story of carbon sequestration. Yet, the implications derived from the results should be strengthened. For instance, the authors should make statements about the relevance of their study on the importance of floodplain carbon sequestration in context of river management in the light of the global climate crisis. How should be floodplains managed and can floodplain management be optimized (if yes, in which way) to increase carbon storage in river systems. In general, the paper is well structured and written and of high quality and is a valuable contribution to the journal Biogeoscience. Given the general comments above and the more detailed line-by-line comments below, I suggest publication after minor revision.
Kind regards
Thomas Hoffmann

General comments:
The age models and derived sedimentation rates of the analysed cores are of fundamental importance for the calculate of the carbon accumulation rates (CAR). The used radio carbon ages are presented in Tab.1 of the manuscript. Given the importance of these data and the derived age models, the authors should give more information on this topic. As this is rather a fundamental basis, rather than the core of the topic, the details on the radiocarbon ages (outlier detection) should be documented in a supplementary file attached to the publication. I suggest that the authors provide plotted age models for each analysed core, including the outliers that were not used in this study and the derived sedimentation rates. This will be very valuable information to evaluate the robustness of the derived sedimentation rates. The authors should also give a table (within supplementary information) with the calculated sediment rates, TOC and CAR for each core.
Obviously, a strong focus is given on the stratigraphical analysis of the cores, however, I suggest to more strongly compare the upper and lower part of the study site, or longitudinal differences from up- to downstream. As the upper part of the study site is characterized by higher agricultural impacts and more pronounced anthropogenic stresses, compare to the lower part of the study site, with more extensive active flooding areas. This is especially valuable for the discussion of potential restoration measures.
The authors relate the observed phases of carbon accumulation rates to the geomorphic setting, the long-term environmental conditions and the human impacts during the Holocene. I have the feeling that the discussion is partially very general and could more strongly linked to the findings of the study. For instance, the authors state that during the past decades, after the implementation of major dams in the upstream catchment, resulted in reduced CAR. However, this is not strongly presented in the results. The authors argue that in the upper 50cm TOC is reduced 2.5-fould. This is in contrast to the general discussion that TOC is not controlled by depth. More in-depth presentation of this results is requested, e.g. by a more detailed comparison of the different cores in different environmental settings. As the potential for TOC storage is also linked to the areal changes of the floodplain, changes of the extend of these settings with time are relevant and could/should be discussed.

Detailed comments
Line 48: 0.5-1% to 0.5-8% is not really a convincing disproportionate contribution given the large variability.

Line 65: Use Hoffmann (2022) (https://linkinghub.elsevier.com/retrieve/pii/B9780128182345000699) instead of Hoffmann et al (2013) here.

Figure 1: I am wondering why the core on the right side are not vertically aligned to depth = 0m. Does this reflect the topography of the floodplain? If yes, please use scale masl, if not please aligned to depth = 0. Use same z-scale for each core. Axis labels and legend are very hard to read (too small)

Line 103: change to “… (2) an upper-stream, artificially modified section…”

Line 107: suspended loads should be preferentially given in Mt/a

Line 118: “previously” is misleading in this context, use “before anthropogenic transformation” instead.

Line 122: Is there any information available on ground water changes, as this is a crucial factor for carbon conservation/decomposition.

Fig. 2: This Figure is very complex. However, not much details can be derived due to the low quality/resolution of the graphic file. Furthermore, I wonder why the surface (y-axis = 0) of the cores are plotted at different heights? Does this reflect absolute topography? Is suggest to plot sedimentation rate in combination with TOC in one graph and CAR for all cores in a comparable way (i.e. time as independent variable) in an extra graph. See also my general comment above.

Line 146: Clay content is a very good TOC proxy. This should be considered in the analysis.

Line 156: bulk density from cored sediment is critical, as sediment in most cores is compacted during drilling, resulting in to high BDs. Mostly empty space in the upper part of the core is indicative for compaction. Was compaction of the cores considered to correct calculated BDs? Given this issue, a quality assessment / error analysis of BD from cores should be undertaken.

Line 161: Where the parameter tested here? What is the RSME of the regression analysis. I suggest to show the scatter plot and the fitted line.

Line 169: if TOC as a central parameter was calculated from two estimated values (TC and TIC), which are associated to measurement uncertainties, TOC will have a larger uncertainty as the uncertainty estimates from TC and TIC. Please provide an uncertainty analysis of the estimated TOC values.

Line 172: SR needs to be defined.

Line 177ff: This is indeed a very good correlation. However, I am interested to know the RMSE of predicted TOC values.

Line207: Can you precise what “large analytical uncertainties” are. Please give threshold. The sample LC8 with the 3^rd largest analytical uncertainty of 71 was not considered as an outlier, but many other samples with much smaller uncertainty. It would be good to plot the age-models for each core in the supplementary material, for better documentation of the data. Tab. 1 could then also be moved to the supplements.

Line 271: rephrase: “Linear regression analysis of TOC vs clay content showed a positive correlation (R2…”.

Line 275: I am wondering how the scatter plot of TOC vs depth looks, if you use TOC normalized to clay content instead of overall TOC? I assume that this might change the story. I agree that the pattern is controlled by the floodplain dynamics, but clay-normalized TOC might show some subtle differences between deep and shallow TOC.

Fig. 4: I suggest to set TOC on y-axis, as this is more likely the dependent variable, while Clay Content, depth are seen as the independent controls.

Line 284 + Fig 4c: Not sure how the redline was achieved in Fig. 4c. I have the feeling that all data points are located along one line and does not follow the broken lines.

Line 295ff: I suggest to plot cumulative carbon accumulation from bottom to top for all profiles. This could be done either using TOC_cum ~ depth or better TOC_cum~age. This should be done in a way that core can be compared between upstream and downstream.

Line 348 & 360: I wonder if decoupling is the correct wording here. Depth plays a crucial role in preserving the TOC (its not decoupled from depth), as the pattern of TOC-deposition with depth is preserved. The causes for the differences to the other studies should be discussed here in detail? AS groundwater plays a crucial role, this should be discussed. Are groundwater conditions in the upper and lower parts of the study site different?

Line 364ff: Please link this to your findings. How is this shown in your results, i.e. what are the spatial differences expressed by the differences in the cores.

Line 439: Please describe what you mean with hydrological connectivity.

Line 468: Limit to base-level impacted floodplains. This is not the case for upstream floodplains.

Line 498-506: How is this decline expressed in your data?
Citation: https://doi.org/10.5194/egusphere-2026-1693-RC2

Laurențiu Țuțuianu, Florin Zăinescu, Alfred Vespremeanu-Stroe, Ionel Bogdan Stan, Mihaela Dobre, Maria Luca-Ilie, Gabriela Sava, Cristian Mănăilescu, and Luminiţa Preoteasa

Viewed

Total article views: 394 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
288	79	27	394	32	23

HTML: 288
PDF: 79
XML: 27
Total: 394
BibTeX: 32
EndNote: 23

Views and downloads (calculated since 16 Apr 2026)

Month	HTML	PDF	XML	Total
Apr 2026	130	39	15	184
May 2026	84	6	2	92
Jun 2026	19	6	2	27
Jul 2026	41	19	5	65
Aug 2026	14	9	3	26
Sep 2026	0

Cumulative views and downloads (calculated since 16 Apr 2026)

Month	HTML	PDF	XML	Total
Apr 2026	130	39	15	184
May 2026	84	6	2	92
Jun 2026	19	6	2	27
Jul 2026	41	19	5	65
Aug 2026	14	9	3	26
Sep 2026	0

Viewed (geographical distribution)

Total article views: 354 (including HTML, PDF, and XML) Thereof 354 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 03 Sep 2026

Short summary

Floodplain sediments store large amounts of organic carbon, but the processes controlling how much is buried and preserved over millennia remain poorly constrained. Using eight sediment cores from the Lower Danube floodplain, we reconstructed seven thousand years of carbon accumulation and show that changes in sea level, sediment supply, and human land use influenced the efficiency of carbon storage, with implications for floodplain restoration and climate mitigation.


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