the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 10 Oct 2025
Fluvial biofilm phosphorus entrapment shifts from intracellular to extracellular dominance along a multifactorial longitudinal river gradient
Abstract. Phosphorus (P) entrapment by biofilms can be distinguished into the intracellular P entrapment (the P uptake by microbial cells) and the extracellular P entrapment via the extracellular polymeric substances. It is unknown how these two P entrapment pathways behave in natural ecosystems in which gradients of environmental drivers such as P, labile dissolved organic matter (DOM) and light occur. Another key aspect in P dynamics at the sediment-water interface is microbial activity. Microbially mediated mineralization of DOM and microbial activity in general is suspected to be a driver of internal P mobilization from the sediments. Here we brought such expectations into a real-world context by analysing the P entrapment patterns of benthic biofilms and the P release potential from the sediments along a longitudinal gradient in a third order river in front of the background of key microbial metabolic variables. The gradient consists of increasing P availability, DOM lability and light availability.
We found a gradual shift in the dominance of P entrapment from higher intracellular P entrapment in the upstream biofilms to higher extracellular P entrapment in the downstream biofilms. This shift towards dominance of extracellular P entrapment was accompanied by an increase in the P release potential from the sediment. The increasing P release potential was also connected to high extracellular enzyme activity of alkaline phosphatase, an enzyme involved in the mineralization of P from organic compounds. We further found that a balanced ratio between intracellular and extracellular P was connected to a higher C metabolic diversity.
All this evidence suggests an influence of the benthic biofilms on P dynamics at the sediment-water interface. This research advocates for a more integrated perspective that accounts for both intracellular and extracellular biofilm-mediated processes.
- Preprint
(2361 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-4844', Anonymous Referee #1, 02 Dec 2025
-
AC1: 'Reply on RC1', Simon Wentritt, 12 May 2026
Response by the authors of “Fluvial biofilm phosphorus entrapment shifts from intracellular to extracellular dominance along a multifactorial longitudinal river gradient” to the comments of the anonymous referee 1:
We would like to thank the associate editor Gabriel Singer for handling the manuscript and the anonymous referee 1 for their useful comments. These comments definitely helped us to further clarify and improve the manuscript. Responses to the referee’s comments are provided below. We tried to divide the comments into different sections where applicable (e.g. RC 1.1 = Referee 1 Comment 1) to give a well-structured response. The bold “headings” represent a one-line abstract of the comment as we understood it. The referee’s comments are in light grey italic and our own response in regular black font.
RC 1.1: Summary of the manuscript with regards to content and structure.
The authors present data on phosphorus and carbon turnover characterization in benthic biofilms of a river stretch. The data include results from a large number of laboratory classification approaches (including biofilm composition, enzyme activity and P release experiments) for 7 positions in a 30km stream stretch, sampled once in autumn of 2023 under baseflow. The results are imbedded in a research question that surrounds the role of the biofilms as P retainers of various form (cells and their biofilm) and how their relation is driven.
The manuscript is well structured, characterizing the field site and its boundary conditions and then successively introducing the dynamic biofilm characteristics. The discussion section describes the systems effectively without repetition from the results section. Methods and statistics seem rigorous.
Reply: We thank the referee for this recapitulation of our manuscript as it shows us that the methods and the frame our study is presented in are clearly understandable, especially in the context of the referee stating that they are not studying stream biofilms. We are very pleased by the assessment of the methods and statistics as rigorous and it reads like the general structure of the manuscript is working.
RC 1.2: Critique on novelty and seeming lack of benefit of the establishment of a connection between biofilm P retention and C metabolism due to lack of structure of the introduction.
Despite the technical benefits, I struggle with the overall novelty of the manuscript. I study aquatic biogeochemistry across various environments, not stream biofilms in particular, and it seems to me as if this manuscript suffers from an overly deep insight paired with too limited spatial and temporal coverage. With their dataset, the authors try to establish a novel connection between carbon metabolism and P retention mechanisms. It escaped me, unfortunately, what the larger benefit of the knowledge of such a tight-knit dependence would be. This may be due to my limited insight to biofilm mechanics, but also to the limited structure of the introduction section.
Reply: The critique of the referee regarding the novelty made us realize that the introduction to the topic of intracellular and extracellular P entrapment is not clear enough. To clarify, one novelty of the study is that we differentiated the total P entrapment of biofilms into the intracellular and extracellular P entrapment. To our knowledge, there are only very few studies that consider P in the EPS matrix of biofilms in non-engineered systems and even less in stream biofilms. There is at least as far as we know no study that specifically tried to estimate the P release potential from sediments and link that to the P in the EPS of biofilms. Further, linking C metabolism variables of the biofilm to the different modes of P entrapment is a more holistic approach, connecting biofilm level processes with environmental parameters on the ecosystem level (e.g. light availability, catchment land-use). As we showed, P entrapment in biofilms is not only mediated by intracellular P pools but also to a large degree by the retention of P in the EPS matrix. Intracellular uptake (e.g. cellular growth) and EPS production of biofilms both require C and/or energy and their acquisition can be linked to the C metabolism variables we measured (for heterotrophs at least). In our opinion there is definitely merit in the investigation of these dependencies. This knowledge increases our understanding of the underlying processes behind biofilm mediated P retention in streams and can help to develop strategies to reduce deleterious effects of P in downstream ecosystems. The ecosystem-level importance of these processes is certainly varying between systems. Based on this comment, we would like to revise and restructure the introduction to strengthen these points and address the specific limitations regarding the spatial and temporal coverage in the methods section.
RC 1.3: Insufficient interconnection of presented background, open questions and hypotheses.
Especially the “open questions” are mixed with background knowledge and presented independent from the (too many and unclear) hypotheses. I’d like to motivate the authors to review this section with great care.
Reply: This comment is connected to the RC 1.2 as it criticises a lack of connection within the introduction section. It does though specifically criticise the connection of the presented background to the hypotheses and the open questions. This is a very important comment to us, as it shows that our introduction does not entirely achieve to introduce the motivation behind our study that led to the presented hypotheses. To tackle this, we would like to restructure the introduction and to include the suggestions of the second referee (see RC 2.2), to focus the introduction on the core aspects. A more focused introduction section with clearer connections between the presented background, the arising open questions and the succeeding hypotheses should clarify the motivation of our study.
RC 1.4: Excessive re-characterization of the riverine gradient despite support from peer-reviewed literature.
This assessment it seems to me is mirrored in the discussion section. Next to the (solid, but to a large extent already published) presentation of the riverine gradient, there is a lot of text that revolves around the argumentation why the analyzed system is three-fold partitioned and why some of the other patterns observed in the previous studies trigger some of the observations made here (e.g. DOM quality, 490f).
Reply: This is a very valuable comment. As the referee points out, the riverine gradient was already established in peer-reviewed literature (Weitere et al. 2021, Kamjunke et al. 2019) and we refer to this literature. In light of this comment, we would like to revise this part by shortening the presentation and contextualisation of the riverine gradient (sections 3.1 and 4.1) and integrating the shortened part with the methods section (section 2.1) as this represents the background to our study and rather not a stand-alone result that adds to the manuscript by discussing it. The figure 1 would be moved to the Appendix A.
The three-fold portioning of the sampled river stretch is done on the basis of the similarity of the measured variables revolving around the C metabolism (BIOLOG Ecoplates). We would suggest to reduce the argument for the portioning in length and better fit it into the manuscript to not repeat this argument in the manuscript where it is mentioned right now.
RC 1.5: Overextension of the interpretation of the observed results and unclear section.
Section 4.5, which would be the heartpiece of the manuscript as currently presented, is not sufficiently clear. Overall my conclusion is that the authors go beyond what the limited observational density can provide.
Reply: As the referee states, the section 4.5 is really the core of the discussion. This section aims at considering the relationships between all biofilm related variables (P entrapment and sediment P release potential, structural variables, C metabolism and extracellular enzyme activities). The assessment of referee 1 that this section is rather unclear was also taken up by referee 2. To improve this section, we would follow the suggestions of referee 2 and revise this section by restructuring this paragraph, adding sentences that better contextualize the described relationship and eliminate repetition of parts already described in other subsections of the discussion. We would further divide the paragraph into sub-sub-sections (4.5.1 to 4.5.3) to make the paragraph more accessible. These sub-sub-sections would be based on the different river reaches we describe in the paragraph.
The assessment that our interpretation of the results overextends the observational density is an important point and we thank the referee for bringing it up. We realized that we need to better communicate the limits of our study, especially in light of the before mentioned lacking temporal and spatial scale. Further, we would emphasize the nature of our observations as correlations. We will clarify that the links between different variables we discuss are not mechanistic (if not already shown by peer-reviewed literature). We believe, that by implementing these clarifying changes, we can eliminate the overextension of the interpretation of our results.
RC 1.6: General recommendations to the authors for a revision of the manuscript.
For a revision, I think this manuscript needs a better focused introduction and a better imbedding of the results. I understand that the many analyses were time-consuming and could not be rolled out to the full network as in Weitere et al. network description. Despite that, I think the data has the potential to couple the observations from the 7 sites with the findings across the larger network and maybe even pair it with numerical data from other sites and studies.
Reply: We fully agree with the verdict that the introduction needs to be more focused. We understand the notion on better imbedding the results in the light of the discussion, especially what the referee commented in RC 1.5 and fully agree here as well. To tackle these points, we would implement the discussed changes (replies to RC 1.2 to RC 1.5). The suggestion of a deeper integration of our original data with data from other sites and studies is very interesting. We will look into the possibilities we have there. Our core variables, intracellular and extracellular P entrapment are pretty much singular in the context of stream biofilm research but explanatory variables such as light availability, SRP availability and possibly data on the DOM composition could be acquired for the Holtemme specifically. A problem with peer-reviewed literature from the Holtemme could be the temporal disconnect between these studies and our study.
RC 1.7: Revision of framing of the manuscript to achieve a better accessibility for a broad-scope biogeochemistry journal.
I don’t want this review to be read as a straight-out recommendation for rejection, as I see merit in the rigorous work. However, I feel for a strong broad-topic biogeo journal as BG, the study motivation, depths and conclusions should be more accessible to a broad community of researchers, which I think is not achieved in the current form, but authors and data have the potential for.
Reply: We thank the referee for this comment as it helps us understand how the manuscript can be perceived in the broader community. We agree with the assessment that the manuscript in its present form lacks a broad accessibility and would take on the task of increasing the accessibility of our manuscript by incorporating the changes as laid out in the replies to the referees’ comments. We would here especially highlight our replies to the comments that refer to the introduction (RC 1.2, RC 1.3 and RC 2.2) and the section 4.5 (RC 1.5, RC 2.12 and replies to specific comments of referee 2 on discussion section).
Referee 1: Specific comments
Some more specific comments
46-51 is it possible to add quantities to these mechanistic background. It becomes more tangible if the pool sizes of intra and extra-P are better known
Reply: We thank the referee for this suggestion. We agree that actual quantities or the relative pool sizes (ratio of intracellular to extracellular P) will make this part more practical. We will add the specific quantities in the revised version of the manuscript.
57 “increasing gradients” is maybe not a good terminology here. It is gradients of increasing XYZ
Reply: We will rephrase the sentence according to this comment. The referee was right to point this out.
64 not sure if Hupfer and Lewandowski, 2008 really had dissolved OM in mind, it is more about solid-phase OM
Reply: This has to be changed from DOM to OM. The referee is completely right here. In the cited article the authors refer to P release from organic matter at the benthic layer and not dissolved in the water column.
60-80 maybe it is possible to state the 2 open research question in a more systematic fashion (e.g. in two paragraphs, each focused on one of the topics: labile DOM->APA and the P entrapment gradient)
Reply: We thank the referee for this suggestion, as it would help to straighten out these two ideas and clarify the concrete background of the open questions we present.
100 write out upstream and downstream
Reply: We will implement this suggestion in the table.
149 do you put the sediments in the P buffer solutions?
Reply: Yes, exactly. We will clarify that the sediment was placed in the P solution and not only in a centrifugation tube.
250 I do not understand why the zero-inflation was necessary in the data structure
Reply: The gamma distribution does not contain 0. To make a glm with gamma distribution work with a data set containing zeros, we had to expand the model to handle 0 values. We would clarify this in the methods section. See also our reply to referee 2’s specific comment on the zero-inflation part of the model.
401 Increasing gradient.. again this seems like a wrong terminology. What is “light DOM availability”
Reply: We will change this. “Light DOM availability” is a typo which will be changed, thank you for pointing this out.
404 As an example here but also elsewhere, the discussion section could be clearer if it more directly links to the results figure in question.
Reply: Thank you for this important suggestion. We will include references to the specific figures, this could especially increase the connectedness in the last section of the discussion as both referees pointed out that this section needs to be restructured and clarified.
447 the biofilm as natural P buffer… this is a relatable concept that may be useful also elsewhere
Reply: This is a very valuable comment. We will use this pick up this concept again for the conclusion.
459 -465 Unclear what this statement is leading to. Discussion is not developing more than just the presentation of results.
Reply: The referee is right and we also realized that this statement is unclear. The concept we aimed to present here is that a mere increase in SRP concentration does not govern the characterisation of a sediment as a P sink or P source. By this we wanted to emphasize that even though the SRP concentration increased along the stream, this was not necessarily the reason for the classification of the sediment as a P source, even though it was found that EPC0 and SRP corelate across several orders of magnitude.
As the referee points out, these lines are not well developed. We will have to re-evaluate this part of the EPC0 discussion and possibly shorten the discussion here to concentrate on the essential interpretation of the EPC0 in the context of our study.
Citation: https://doi.org/10.5194/egusphere-2025-4844-AC1
-
AC1: 'Reply on RC1', Simon Wentritt, 12 May 2026
-
RC2: 'Comment on egusphere-2025-4844', Anonymous Referee #2, 09 Apr 2026
The manuscript presents the phosphorus and carbon contribution in biofilms along a longitudinal gradient across different reaches of a low-order stream. These patterns may be further influenced by climate change modifying nutrient biogeochemical processes. Results provide a comprehensive representation of both structural and functional characteristics of biofilms, particularly focusing the attention on intra and extracellular phosphorus acquisition, as well as the contribution of dissolved organic carbon. This study globally embedded the contribution of fluvial biofilm in this longitudinal gradient with extensive length covered.
The manuscript clearly introduces the key concepts related to biofilms and low-order streams, which are currently threatened by global change. However, despite the extensive research conducted by the authors, several issues and questions can arise. The introduction is relatively long, and particularly the first paragraph is too general. Focusing more directly on the key aspects would strengthen it. Additionally, some references are old (e.g., lines 39, 41, 43, 56, 74, among others). I recommend including more recent citations where possible.
The methods are well described. The field site is clearly characterised, and assessing half of the total stream length is an important aspect that should be explicitly highlighted. The authors also acknowledge challenges encountered during sampling, which is appreciated. The description of biofilm parameters evaluated is clear, and the statistical approaches are well explained with scientific rigour.
I suggest replacing Table 1 with a map showing the different stream reach categories to strengthen this section. Table 1 could be moved to the supplementary material. Additionally, incorporating land-use analysis of each reach’s surroundings with differentiating agricultural, forestry, or urban predominance could help interpret the characteristics of each reach and improve the discussion, particularly in subsection 4.1.
For clarification: how was phosphorus concentration in the biofilm extract analysed? If inductively coupled plasma mass spectrometry (ICP-MS) was used, I suggest briefly describing this method.
I think it is important to mention the six samples that were incompletely homogenised to obtain good results, but in this case, I suggest modifying the paragraph to start mentioning the relevance that a minimum of 3 observations for each sampling site were obtained, and due to this issue, these six samples were discarded.
The CLPP analysis from Perujo et al., 2020 is highly relevant and strengthens the interpretation of functional variables.
Some specific comments from the Methods section:
Lines 152-154: the explanation of the x-intercept determination is unclear.
Lines 166-167: Why was Optiprep density gradient medium used?
Line 171: why was dilution performed with Milli-Q water instead of Ringer solution (to avoid osmotic shock)?
Line 175: Chl-a was determined using ethanol 90% instead of acetone 90%, could you justify this choice?
Line 178: consider using “biofilm functional variables” instead of “metabolic variables”
Line 235: the absorbance range (250-400 nm) requires further clarification.
Lines 244-247: why was the Benjamini-Yekutieli multiple test selected over other non-parametric approaches?
Lines 250-251: why was Chl-a expanded for zero inflation? Is this comparable to a log(x+1) transformation?
Lines 254-255: why was one-way ANOVA used for microbial C metabolic profiles, substrate guilds, and EEAs? Would glm models be more appropriate?
Results were clearly representative to demonstrate the differences among sampling stream reaches and respective contribution of biofilm structural and functional variables. Most of the current studies in this field were considering longitudinal gradient patterns on nutrient release or uptake from urban wastewater treatment plants. This article goes even further with specific contributions on phosphorus and carbon dynamics to widely represent the biofilm capacities on their longitudinal shifts.
In several figure captions, the colour gradient indicating distance from the source is unclear and could be redundant, as distance is already shown on the x-axis. Additionally, given the limited number of replicates, presenting mean values with standard deviation for each distance site may be enough.
Specific comments from the Results section:
Lines 283-284: does an increasing slope ratio indicate decreasing DOM molecular weight (Fig.1C)? This is unclear or can be misunderstood. High variability may also affect the interpretation. Similar concerns apply to Fig.1D.
Line 286: I suggest defining BIX as “biological index” when first mentioned.
Lines 285-290: is statistical analysis missing here? Could it be an option to interpret statistically the variability of these variables against distance in section 3.1? Similarly, as you performed at section 3.2.
Line 309: in figure 3 caption, the P release potential should be added as EPC0 acronym.
Line 322: consider using “biofilm structural variables” only.
Lines 322-331: consider including standard deviation alongside mean values.
Lines 347-349: consider including key CLPP results in the main text and moving the diversity indices figure of Shannon and Evenness to supplementary material.
Line 354: define EEA abbreviations in figure 6 caption.
Lines 373-374: results from the log EEA ratios should be briefly described in the main text, only referring to the supplementary material for figure and table.
Line 387: consider replacing “uppermost sampling site” with “upstream”. And consistently use upstream/downstream throughout the manuscript.
Finally, regarding discussion and conclusions sections. I would appreciate the work done by the authors of this research, extensively discussing their results. Nevertheless, some specifications and recommendations are also needed to consider improving the readability and strength its interpretation. Subsections 4.1 and 4.2 are particularly interesting. However, incorporating recent citations could strengthen the discussion, especially with the connection between Jarvie et al., 2002 and Weitere et al., 2021 focusing on their temporal assessment, compared to your study which was focused on longitudinal and specific temporal period evaluated. Given your focus on longitudinal gradients, I expected to observe a stronger connection with WWTP effluents discharged into these low-order streams. For instance, SRP loads (lines 435-438) may be originated from urban WWTP discharges.
Only a specific mesocosm experiment was cited (lines 441-443). Nevertheless, I suggest adding other studies that can corroborate it.
In some subsections you referenced the hypothesis mentioned in the introduction. However, there are other subsections such as 4.3 and 4.4 that can be commented in the main manuscript for enhanced readability.
Additionally, there are several conditional sentences that some of them can be improved with specific citations. In this case, adding recent citations could strengthen the discussion.
Subsection 4.5 is highly relevant and well-written, although some parts are repetitive from the other subsections (e.g., lines 500-502). A few sentences are also unclear; for instance, between lines 539 and 542, you described a lack of downstream connection with APA, followed by correlations with Chl-a and EPC0. Clarifying these relationships would improve interpretation. Structuring sentences around observed correlations could improve readability and clarity.
Some specific comments from the Discussion section:
Line 453: replace with “upstream sampling sites” and homogenise terminology throughout the manuscript.
Line 464: could you mention additional underlying factors?
Line 470: specify land-use types? Agricultural or urban shift?
Line 480: avoid speculative sentences; support statements with evidence. For instance, in the sentence “likely caused by an increase in DOM diversity”. I suggest avoiding supposition and corroborating it, as you explained in the following sentences with the examples provided.
Lines 504-507: the reference of Zhou et al. (2017) may be out of context? They performed an experiment based on Synechocystis sp., but I am not sure if they considered different microbial communities such as occurred with biofilms.
Line 508: why did you consider the upstream reach as transient P storage due to EPS? A brief explanation about the reference you found could enhance this interpretation.
Overall, I recommend a minor revision, addressing the previous comments mentioned above.
Citation: https://doi.org/10.5194/egusphere-2025-4844-RC2 -
AC2: 'Reply on RC2', Simon Wentritt, 12 May 2026
Response by the authors of “Fluvial biofilm phosphorus entrapment shifts from intracellular to extracellular dominance along a multifactorial longitudinal river gradient” to the comments of the anonymous referee 2:
We would like to thank the associate editor Gabriel Singer for handling the manuscript and the anonymous referee 2 for their useful comments. These comments definitely helped us to further clarify and improve the manuscript. Responses to the referee’s comments are provided below. We tried to divide the comments into different sections where applicable (e.g. RC 2.1 = Referee 2 Comment 1) to give a well-structured response. The bold “headings” represent a one-line abstract of the comment as we understood it. The referee’s comments are in light grey italic and our own response in regular black font.
RC 2.1: Summary of the manuscript with regards to content
The manuscript presents the phosphorus and carbon contribution in biofilms along a longitudinal gradient across different reaches of a low-order stream. These patterns may be further influenced by climate change modifying nutrient biogeochemical processes. Results provide a comprehensive representation of both structural and functional characteristics of biofilms, particularly focusing the attention on intra and extracellular phosphorus acquisition, as well as the contribution of dissolved organic carbon. This study globally embedded the contribution of fluvial biofilm in this longitudinal gradient with extensive length covered.
Reply: We thank the referee for this summary as it picks up the core ideas of our study.
RC 2.2: Critique on the introduction and references used.
The manuscript clearly introduces the key concepts related to biofilms and low-order streams, which are currently threatened by global change. However, despite the extensive research conducted by the authors, several issues and questions can arise. The introduction is relatively long, and particularly the first paragraph is too general. Focusing more directly on the key aspects would strengthen it. Additionally, some references are old (e.g., lines 39, 41, 43, 56, 74, among others). I recommend including more recent citations where possible.
Reply: We thankfully take on the critique of the referee 2 on the structure and length of our introduction. We realize that the introduction lacks clarity and is partly too general. A lack of clarity was also criticised by referee 1 (i.e. RC 1.2 and RC 1.3). To tackle this, we would revise the introduction to emphasize the key aspects of our study as laid out in the reply to RC 1.2. We further would restructure the paragraph to follow a clear and logical order, also including a stronger connection between the presented background, the resulting open questions we investigated in this study and the hypotheses.
The suggestion to include newer references where possible is a very welcome one. In our effort to revise the manuscript, we would expand the citations as these research fields also progressed of course. However, we would keep most of the citations mentioned by the referee as these are the fundamental papers that coined specific concepts we want to introduce here.
RC 2.3: Summary of the methods.
The methods are well described. The field site is clearly characterised, and assessing half of the total stream length is an important aspect that should be explicitly highlighted. The authors also acknowledge challenges encountered during sampling, which is appreciated. The description of biofilm parameters evaluated is clear, and the statistical approaches are well explained with scientific rigour.
Reply: We thank the referee for this comment as it shows us that the framework of our study is laid out understandably in the manuscript as it is now. To clarify the aspect of only assessing a part of the whole stream, the study was designed to capture a continuous gradient of potential microbial energy availability and SRP availability.
RC 2.4: Suggestion for the improvement of the land-use characterization and visualization of the sampled reach.
I suggest replacing Table 1 with a map showing the different stream reach categories to strengthen this section. Table 1 could be moved to the supplementary material. Additionally, incorporating land-use analysis of each reach’s surroundings with differentiating agricultural, forestry, or urban predominance could help interpret the characteristics of each reach and improve the discussion, particularly in subsection 4.1.
Reply: We greatly appreciate the suggestion by the referee to visualize the stream, sampling points and the land-use characterization in form of a map. We agree that a map would increase the accessibility of the system we performed out study in. However, the map as it is described by the referee was published in Weitere et al. 2021 which we reference also for this reason. To make the connection clearer, we would refer specifically to the map presented in Weitere et al. 2021 in the methods section 2.1.
RC 2.5: Clarification of the ICP-MS measurements and better description of the method.
For clarification: how was phosphorus concentration in the biofilm extract analysed? If inductively coupled plasma mass spectrometry (ICP-MS) was used, I suggest briefly describing this method.
Reply: We thank the referee for this inquiry. We would better explain the digestion process of the solid sample, the initial calibration and the recalibration between measurements. Our proposed wording would be:
“The P concentration in the biofilm extract was then determined with inductively coupled plasma mass spectrometry (ICP-MS; 8800 Triple Quad, Agilent, USA) after complete microwave digestion of the sample (3 mL HNO3, 1 mL HCl, TMax=170°C). Calibration was carried out using a single element standard based on Na4(PO4)2 · 10 H2O (Merck, Germany) with the certified reference material CCS-5 (Inorganic Ventures, USA) as its verification. Additionally, potential drifts in instrument sensitivity were excluded by using an internal standard solution and frequently measuring a calibration standard every 12 samples. Results are given in μg-P cm-3.”
RC 2.6: Minimum observations for statistical inference reached for each sampling site.
I think it is important to mention the six samples that were incompletely homogenised to obtain good results, but in this case, I suggest modifying the paragraph to start mentioning the relevance that a minimum of 3 observations for each sampling site were obtained, and due to this issue, these six samples were discarded.
Reply: This is a very nice suggestion and we would revise this section exactly as it is laid out by the referee.
RC 2.7: Higher relevance of the CLPP analysis in the results.
The CLPP analysis from Perujo et al., 2020 is highly relevant and strengthens the interpretation of functional variables.
Reply: We fully agree with the assessment of the referee that the CLPP analysis is highly relevant to the interpretation of the functional variables derived from the BIOLOG EcoPlates incubations. We would move the figure C3 and the related description of these results to the section 3.4 (Biofilm metabolic profiles). As referee 1 suggested in RC 1.4 to reduce the excessive re-characterization of the riverine gradient and shortening the argumentation around the partitioning of the sampled stream stretch into three reaches, the number of figures would be the same. Further, the now figure C3 is also suitable to shortly affirm the partitioning of the sampled stream stretch into the three reaches.
Referee 2: Specific comments for Methods section:
Some specific comments from the Methods section:
Lines 152-154: the explanation of the x-intercept determination is unclear.
Reply: We would clarify the determination of the x-intercept in the revised manuscript. In short, we performed a linear regression with the P concentration of the concentration series in the beginning of the incubation against the concentration of P sorbed to the sediment after 24 h of incubation. A linear model was fitted through the data points and the x-intercept, representing the P concentration in the water where the concentration of P sorbed is 0 (no exchange of P between water column and sediment P) can be calculated. We used only the first three positive sorption points from incubations as it is described in the reference Simpson et al. 2019 (section 2.5 and electronic supplementary material 1), referenced in our section 2.4.
Lines 166-167: Why was Optiprep density gradient medium used?
Reply: The density gradient medium OptiPrep is used to purify the sample as it helps to reduce the amount of detritus particles based on the density of the particles (see Amalfitano and Fazi 2008). We will clarify this point in the manuscript. If your comment relates to why specifically OptiPrep was used and no other density gradient medium (e.g. Nycodenz), it was the density gradient medium that was available to us in Germany. OptiPrep density gradient medium is also used in other studies for the purification of sediment bacteria communities to prepare the samples for flow cytometry (e.g. Boadella et al. 2024).
Line 171: why was dilution performed with Milli-Q water instead of Ringer solution (to avoid osmotic shock)
Reply: The referee is right to point this out and we take this into account for future work. As we determined the total bacterial density, we hope that osmotic shock was not so critical as compared with a differentiation of live and dead cells. We acknowledge that the data might not be usable for a comparison outside this study, however for a comparison between the sampling sites within this study, we believe the data is safe to use. All samples were treated the same (same dilution factor and same incubation time after staining) which should retain relative comparability between the samples.
Line 175: Chl-a was determined using ethanol 90% instead of acetone 90%, could you justify this choice?
Reply: The extraction is based on a German industry norm (DIN-38409-60) for the spectrometric determination of chlorophyll-a in water and sludge. The use of ethanol as an extraction solvent is suggested under this norm and our spectrophotometer is calibrated on ethanol as an extraction solvent. Also, the standards we are using are based on ethanol. Moreover, our institute opted for the use of ethanol over acetone due to health concerns with acetone handling, as we try to minimize exposure to toxic chemicals in our laboratory analyses wherever possible.
Line 178: consider using “biofilm functional variables” instead of “metabolic variables”
Reply: We will change the wording as suggested by the referee. Going through the manuscript, we found another case where we will adapt the wording to align this throughout the manuscript (line 385).
Line 235: the absorbance range (250-400 nm) requires further clarification.
Reply: The absorbance range was used to calculate the absorbance based bulk DOM quality indices Slope Ratio and E2:E3. The referee is right to point out, that the sentence lacks the connection. We will clarify the connection of the chosen absorbance range. In short, the slopes of the log-transformed absorbance ranges from 275 nm to 295 nm and 350 nm to 400 nm were determined and the ratio of the two slopes was calculated to determine the Slope Ratio (see Helms et al. 2008). For the E2:E3, the ratio of the absorbance at 250 nm and 365 nm is calculated (see Minero et al. 2007). To streamline the measurements, we measured the complete absorbance range from the lowest needed value to the highest.
Lines 244-247: why was the Benjamini-Yekutieli multiple test selected over other non-parametric approaches?
Reply: We used the Benjamini-Yekutieli method to control for the false discovery rate due to multiple testing (i.e. false positive detection) since we performed multiple pairwise Spearman rank correlation tests. In contrast to other statistical procedures that control for the false discovery rate (e.g. Benjamini-Hochberg), this is a more conservative approach that does not assume independence of the data. For our data structure, we deemed Benjamini-Yekutieli the most suitable approach.
Lines 250-251: why was Chl-a expanded for zero inflation? Is this comparable to a log(x+1) transformation?
Reply: The gamma distribution does not contain 0. To make a glm with gamma distribution work with a data set containing zeros (i.e. chlorophyll-a measurements below the detection limit), we expanded the model to handle 0 values. This approach is not comparable with log(x+1) as it does not set 0 values to very small positive values. It rather expands our glm, where we model the chlorophyll-a concentration as a function of distance, assuming strictly positive values, by a model that models the probability of chlorophyll-a absence (==0) as function of distance. The glm part is modelling magnitude and the zero-inflation part is modelling absence/presence. We would clarify this in the revised version of the manuscript.
Lines 254-255: why was one-way ANOVA used for microbial C metabolic profiles, substrate guilds, and EEAs? Would glm models be more appropriate?
Reply: The choice of the glm vs ANOVA was done based on the question and the underlying assumptions towards the different variables. For the C metabolic profiles, substrate guilds and EEAs, we were not hypothesising a continuous spatial gradient but rather differences among the discrete sampling locations. So, local conditions were hypothesized to be dominating, and not the continuous change of the environmental conditions. The glms on the other hand really tackled the question if the variable is subject to a systematic change along the gradient. We choose these two different approaches based on a visual inspection of the data.
We realized that the choice of analysis was not sufficiently explained in the text and we will add the justification to the section “2.8 Data Analysis” in the manuscript. Further, we found that there was an error in the description of the ANOVA in line 255: We wrote that “Distance from the source was used as the explanatory variable” which implies a continuous variable but this is not true, we actually used the discrete variable “Site”. We will of course rectify this in the revised manuscript. The ANOVA results tables (e.g. Table C4) already show “Site” in the model as the independent variable, so this is just an error in the description, not in the model.
RC 2.8: Summary of the results
Results were clearly representative to demonstrate the differences among sampling stream reaches and respective contribution of biofilm structural and functional variables. Most of the current studies in this field were considering longitudinal gradient patterns on nutrient release or uptake from urban wastewater treatment plants. This article goes even further with specific contributions on phosphorus and carbon dynamics to widely represent the biofilm capacities on their longitudinal shifts.
Reply: We really appreciate this assessment by referee 2. It is reassuring to read that our approach is seen to go further than other articles in our field.
RC 2.9: Suggestion to increase the clarity of the figures.
In several figure captions, the colour gradient indicating distance from the source is unclear and could be redundant, as distance is already shown on the x-axis. Additionally, given the limited number of replicates, presenting mean values with standard deviation for each distance site may be enough.
Reply: The referee makes a very good point here. We internally discussed this topic of the colour gradient before submitting the manuscript to Biogeosciences. The gradient is strictly necessary for the multivariate analyses in our opinion and I (main author) decided to keep the gradient in all plots for consistency. However, the referee is right that this is redundant for most of the figures. For the sake of clarity, we would adapt the referee’s suggestion here regarding the use of the colour gradient and switch to uniformly black coloured mean values and the standard deviation to indicate the heterogeneity within one sampling site. We would change this in all figures where applicable, i.e. where the clarity is improved and removing the colour gradient is unambiguous.
Referee 2: Specific comments for Results section:
Specific comments from the Results section:
Lines 283-284: does an increasing slope ratio indicate decreasing DOM molecular weight (Fig.1C)? This is unclear or can be misunderstood. High variability may also affect the interpretation. Similar concerns apply to Fig.1D.
Reply: Yes, the increasing slope ratio indicates a decreasing bulk DOM molecular weight and an increased E2:E3 also indicates a lower bulk DOM aromaticity degree. As described in the reply to a comment by referee 1 (RC 1.4), we plan to integrate aspects of this section into the methods section (section 2.1) and to move figure 1 to the Appendix A. We will keep the unclear wording pointed out by referee 2 in mind when restructuring the characterization of the gradient and clarify these points.
Line 286: I suggest defining BIX as “biological index” when first mentioned.
Reply: Yes, the first mention is in the methods section, we will change this when we revise this section (see reply above).
Lines 285-290: is statistical analysis missing here? Could it be an option to interpret statistically the variability of these variables against distance in section 3.1? Similarly, as you performed at section 3.2.
Reply: We analysed the relationship between these variables and with the distance as described in section 2.8 (lines 239-247) with a Spearman Rank correlation. A figure, visualising the results is found in Appendix A (Fig. A1). These variables are the ones that really form the gradient and the mentioned correlation analysis is our justification to use Distance in the glms as the explanatory variable. Further analysing these variables is not necessary for this study as it was already pointed out by referee 1 that the characterization of the stream gradient is sufficient and not the main topic of this study.
Line 309: in figure 3 caption, the P release potential should be added as EPC0 acronym.
Reply: We will change this. By going through the manuscript, we found several more instances where we did not use the acronym (Lines: 70, 83, 88, 142, 144, 249, 313, 512). We have to determine the context of the use of “P release potential” individually in these instances to determine how to change the wording.
Line 322: consider using “biofilm structural variables” only.
Reply: We will change this of course, the wording “biofilm structural variables” is more precise than “biofilm structural and microbial variables”.
Lines 322-331: consider including standard deviation alongside mean values.
Reply: Thank you for this comment, we will include the standard deviation in the revised manuscript together with the mean values in the running text. Also, considering our reply to RC 2.9, this would streamline the presentation of the results. This will also be changed for the sections 3.2, 3.3 and 3.4.
Lines 347-349: consider including key CLPP results in the main text and moving the diversity indices figure of Shannon and Evenness to supplementary material.
Reply: We thank the referee for this good suggestion. We would here refer to our reply for RC 2.7. Our suggestion would be to keep the diversity indices, add the figure C3 (CLPPs) to the section 3.4 and move figure 1 to Appendix A. The diversity indices are key results that show strong metabolic differences between the biofilms at the different sampling sites.
Line 354: define EEA abbreviations in figure 6 caption.
Reply: Yes, we would of course change that in the revised version of the manuscript. The abbreviations for the EEAs were already coined in section “2.6 Biofilm metabolic variables”.
Lines 373-374: results from the log EEA ratios should be briefly described in the main text, only referring to the supplementary material for figure and table.
Reply: We agree with the referee and would shortly describe the results for the EEA ratios here in the revised manuscript. This could help the reader later in the section 3.5 and the discussion (section 4.5) to better contextualise the discussed results.
Line 387: consider replacing “uppermost sampling site” with “upstream”. And consistently use upstream/downstream throughout the manuscript.
Reply: We agree that the consistent labelling is important for the manuscript. We would adopt the suggestion of the referee and change this in the revised manuscript.
RC 2.10: Increase the strength of the interpretation in the discussion section with more recent and better fitting citations.
Finally, regarding discussion and conclusions sections. I would appreciate the work done by the authors of this research, extensively discussing their results. Nevertheless, some specifications and recommendations are also needed to consider improving the readability and strength its interpretation. Subsections 4.1 and 4.2 are particularly interesting. However, incorporating recent citations could strengthen the discussion, especially with the connection between Jarvie et al., 2002 and Weitere et al., 2021 focusing on their temporal assessment, compared to your study which was focused on longitudinal and specific temporal period evaluated. Given your focus on longitudinal gradients, I expected to observe a stronger connection with WWTP effluents discharged into these low-order streams. For instance, SRP loads (lines 435-438) may be originated from urban WWTP discharges.
Only a specific mesocosm experiment was cited (lines 441-443). Nevertheless, I suggest adding other studies that can corroborate it.
Reply: We thank the referee for this assessment of the sections 4.1 and 4.2 of our discussion. The referee is right in pointing out that referencing temporal assessments is not ideal in regards to our longitudinal study. The following should not be read as a justification but rather as a contextualization: There are only very few studies that differentiate the total P in the biofilm into the intracellular and extracellular P pools in natural stream biofilms. Jarvie et al., 2002 is an example where the authors tried this, albeit in a temporal context. We always give the context of the referenced studies so that the reader can contextualize the discussion of our results. More studies that consider the extracellular and intracellular P pools in biofilms were done in experimental systems (e.g. the referenced microcosm experiment of Xing et al., 2021 or Zhou et al., 2017 and Li et al., 2015 which are both referenced in section 4.5). We agree that an explicit mention of the urban WWTP discharge as the point source for SRP in our study should be added in this section (lines 434 -437) and we would revise this.
In the revised manuscript, we would again look for more fitting literature where we can find it and better contextualise the references we already used, so to make our interpretation stronger. This comment also seems to tie into RC 1.5 and a possible overextension of the interpretation of our results. We would therefore refer here to our reply to RC 1.5.
RC 2.11: Missing references to the hypotheses in the discussion and missing citations
In some subsections you referenced the hypothesis mentioned in the introduction. However, there are other subsections such as 4.3 and 4.4 that can be commented in the main manuscript for enhanced readability.
Additionally, there are several conditional sentences that some of them can be improved with specific citations. In this case, adding recent citations could strengthen the discussion.
Reply: We agree with the referee, that referencing the hypotheses from the introduction in the subsections of the discussion increases the readability and better connects the whole manuscript. We would revise this and add the corresponding hypotheses to the sections where this connection is missing.
We also agree with the referee that we should add recent references to conditional sentences to give them explanatory power. This will be revised in the manuscript. A specific example at the end of section 4.3 (lines 463 – 464) will be rephrased completely in response to referee 1’s comment on this section. For the case of section 4.4 (lines 483 - 495) we would add Säwström et al 2016 who showed that the differences in organic matter composition in estuarine and coastal sediments influenced microbial communities CLPPs. We would also embed the two references Oest et al. 2018 and Freixa and Romaní 2014 better in the running text as their connection to our own observations are not entirely clear.
RC 2.12: Restructure the section for higher clarity and readability.
Subsection 4.5 is highly relevant and well-written, although some parts are repetitive from the other subsections (e.g., lines 500-502). A few sentences are also unclear; for instance, between lines 539 and 542, you described a lack of downstream connection with APA, followed by correlations with Chl-a and EPC0. Clarifying these relationships would improve interpretation. Structuring sentences around observed correlations could improve readability and clarity.
Reply: We thank the referee for pointing out that the section 4.5 is well written but suffers from repetitions and unclarities. It is very important to us to get this outside perspective. We like to refer to our reply to RC 1.5 as both referees have similar conclusions. We would like to revise this section as follows: We would restructure this paragraph by dividing it into sub-sub-sections (4.5.1 to 4.5.3) treating each river reach. These will be connected by connecting sentences and to eliminate unclarities, we would follow the suggestions by referee 2 and add clarifying statements that better describe the observed correlations and contextualise what we discuss here. Lastly, we would eliminate the repetitions and streamline the discussion across the sections.
Referee 2: Specific comments for Discussion section:
Some specific comments from the Discussion section:
Line 453: replace with “upstream sampling sites” and homogenise terminology throughout the manuscript.
Reply: We will adopt this suggestion as it was also brought up in the specific comments of referee 2 to the results section.
Line 464: could you mention additional underlying factors?
Reply: Yes, this statement as it is does not help the reader. We will rephrase this part of section 4.3 in response to referee 1’s specific comment on this. See also our reply to RC 2.11.
Line 470: specify land-use types? Agricultural or urban shift?
Reply: We reference Weitere et al., 2021 here as there, the major land-use shifts in the catchment are laid out. We would clearly state which are the land-use shifts here in the revised manuscript to make the connection easier for the reader.
Line 480: avoid speculative sentences; support statements with evidence. For instance, in the sentence “likely caused by an increase in DOM diversity”. I suggest avoiding supposition and corroborating it, as you explained in the following sentences with the examples provided.
Reply: We thank the referee to point this out, as we realize that this reads like an interpretation of Freixa et al., 2016 results. This sentence was meant as an abstract of the reasons laid out by Freixa et al., 2016 in their discussion of the observed EEA patterns. We would revise this and the following sentence and make clear that this is an observation by Freixa et al., 2016 and we found similar underlying conditions in our system (e.g. mixing of anthropogenic and natural discharge due to land-use shifts).
Lines 504-507: the reference of Zhou et al. (2017) may be out of context? They performed an experiment based on Synechocystis sp., but I am not sure if they considered different microbial communities such as occurred with biofilms.
Reply: We agree that the context of this reference is not the same as in our study due to the nature of a single-organism microcosm versus our mixed microbial communities. This is why we specifically give the context of this reference (see reply to RC 2.10). However, here we discuss the extracellular P entrapment in the EPS and Zhou et al., 2017 showed two very important things for the context of extracellular P entrapment: First, that the EPS matrix is a transient P storage, meaning that P can be first adsorbed and then taken up into the cells and second, the strong correlation between the protein content in the EPS and extracellular P. Both of these findings from Zhou et al., 2017 are probably universal in the context of the EPS matrix.
We realize that the part from line 504 to 510 can be confusing and we would increase the clarity in the revision of this section.
Line 508: why did you consider the upstream reach as transient P storage due to EPS? A brief explanation about the reference you found could enhance this interpretation.
Reply: This comment ties into the previous one. We agree that this need further explanation and lacks context. In our effort to revise section 4.5, we would address this point and clarify the statement of transient P storage.
RC 2.13: Final recommendation for the editor.
Overall, I recommend a minor revision, addressing the previous comments mentioned above.
Citation: https://doi.org/10.5194/egusphere-2025-4844-AC2
-
AC2: 'Reply on RC2', Simon Wentritt, 12 May 2026
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 1,611 | 807 | 167 | 2,585 | 160 | 203 |
- HTML: 1,611
- PDF: 807
- XML: 167
- Total: 2,585
- BibTeX: 160
- EndNote: 203
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The authors present data on phosphorus and carbon turnover characterization in benthic biofilms of a river stretch. The data include results from a large number of laboratory classification approaches (including biofilm composition, enzyme activity and P release experiments) for 7 positions in a 30km stream stretch, sampled once in autumn of 2023 under baseflow. The results are imbedded in a research question that surrounds the role of the biofilms as P retainers of various form (cells and their biofilm) and how their relation is driven.
The manuscript is well structured, characterizing the field site and its boundary conditions and then successively introducing the dynamic biofilm characteristics. The discussion section describes the systems effectively without repetition from the results section. Methods and statistics seem rigorous.
Despite the technical benefits, I struggle with the overall novelty of the manuscript. I study aquatic biogeochemistry across various environments, not stream biofilms in particular, and it seems to me as if this manuscript suffers from an overly deep insight paired with too limited spatial and temporal coverage. With their dataset, the authors try to establish a novel connection between carbon metabolism and P retention mechanisms. It escaped me, unfortunately, what the larger benefit of the knowledge of such a tight-knit dependence would be. This may be due to my limited insight to biofilm mechanics, but also to the limited structure of the introduction section. Especially the “open questions” are mixed with background knowledge and presented independent from the (too many and unclear) hypotheses. I’d like to motivate the authors to review this section with great care. This assessment it seems to me is mirrored in the discussion section. Next to the (solid, but to a large extent already published) presentation of the riverine gradient, there is a lot of text that revolves around the argumentation why the analyzed system is three-fold partitioned and why some of the other patterns observed in the previous studies trigger some of the observations made here (e.g. DOM quality, 490f). Section 4.5, which would be the heartpiece of the manuscript as currently presented, is not sufficiently clear. Overall my conclusion is that the authors go beyond what the limited observational density can provide.
For a revision, I think this manuscript needs a better focused introduction and a better imbedding of the results. I understand that the many analyses were time-consuming and could not be rolled out to the full network as in Weitere et al. network description. Despite that, I think the data has the potential to couple the observations from the 7 sites with the findings across the larger network and maybe even pair it with numerical data from other sites and studies. I don’t want this review to be read as a straight-out recommendation for rejection, as I see merit in the rigorous work. However, I feel for a strong broad-topic biogeo journal as BG, the study motivation, depths and conclusions should be more accessible to a broad community of researchers, which I think is not achieved in the current form, but authors and data have the potential for.
Some more specific comments
46-51 is it possible to add quantities to these mechanistic background. It becomes more tangible if the pool sizes of intra and extra-P are better known
57 “increasing gradients” is maybe not a good terminology here. It is gradients of increasing XYZ
64 not sure if Hupfer and Lewandowski, 2008 really had dissolved OM in mind, it is more about solid-phase OM
60-80 maybe it is possible to state the 2 open research question in a more systematic fashion (e.g. in two paragraphs, each focused on one of the topics: labile DOM->APA and the P entrapment gradient)
100 write out upstream and downstream
149 do you put the sediments in the P buffer solutions?
250 I do not understand why the zero-inflation was necessary in the data structure
401 Increasing gradient.. again this seems like a wrong terminology. What is “light DOM availability”
404 As an example here but also elsewhere, the discussion section could be clearer if it more directly links to the results figure in question.
447 the biofilm as natural P buffer… this is a relatable concept that may be useful also elsewhere
459 -465 Unclear what this statement is leading to. Discussion is not developing more than just the presentation of results.