Review Biogeosciences MS No.: bg-2021-69

The manuscript “Partitioning carbon sources in a tropical watershed (Nyong River, Cameroon) between wetlands and terrestrial ecosystems – Do CO2 emissions from tropical rivers offset the terrestrial carbon sink?” by Moussa Moustapha et al. address an important question of the sources of organic materials fueling the high CO2 emissions from tropical rivers and streams, which is within the scope of BG. The manuscript contributes especially with relevant information increasing the understanding of the hydrologic influence on the C dynamics in African aquatic environments. The discussions that drive the conclusions still need more support from the literature. The overall methodology is robust but needs some clarifications. The results would benefit from some extra description especially regarding the relationships between discharge and the main variables considering the entire period instead of the binned evaluation for only 3 seasons that do not show relationship. These extra results would strengthen the conclusions. Several key results are not shown (“data not shown”). The title is not very clear and the question in the title is not really answered. The abstract could have a short justification highlighting the importance of such study, and the sentences describing the results need to be rewritten to improve the flow. Data could be further explored, especially regarding the temporal variability, and extra figures could be added as supplementary information. In conclusion, the manuscript presents a valuable database that will contribute to the understanding of the carbon cycle and CO2 emissions from streams in the understudied Africa region. However, several points need to be considered before the manuscript is suitable for publication.

General comments:

The as many parts of the text that is hard to follow and not so precise. Careful English revision is needed.

The temporal evaluating would benefit from monthly estimates of C degassing, lateral input, metabolism, and export to the ocean. The annual budget should then be calculated considering the temporal variability instead of the annual average. It would be interesting to compare the final result of both estimates.

Since the seasonal variability is regulated by hydrology, it would be good if the authors could explore correlations between discharge and the different C compartments observed in the rivers.

Section 2.4.2 is hard to follow, see specific comments below. Please reformulate it. The authors could also make a schematic figure to help explain the lateral input of carbon from these two sources. Please, also include information about the direct input of POC from the forest to the streams.

It is unclear if the site Mengong outlet is a wetland or if it is a 1 order stream draining a wetland. Also how representative is this environment as a 1^st order stream in the entire Nyong River basin?

It seems that the authors assume that the input of C via forest groundwater observed in the spring is the same for all streams regardless of order. Wouldn’t the input of groundwater increase with stream order, since the catchment area is much larger. Consequently, wouldn’t the potential input of pCO2 from groundwater directly to the streams increase with stream order?

Plotting and testing the relationship between the main variables and discharge using sampling events instead of binned into only 3 seasons would strengthen many points of your discussion.

The authors could include more information to cover what is the proportion of carbon derived from wetlands in high-order streams? An estimate of the input of POC from the forest canopy would also be interesting.

The discussion in many cases is speculative and lacks support from the literature. I acknowledge the lack of data for African rivers, but you make comparisons with a few temperate systems without mentioning the many studies carried out in the Amazon basin containing useful information that should be included in your discussion. Check for example Johnson et al 2008, Rasera et al 2008, Amaral et al 2019, Salimon et al 2013, Neu et al 2011, Scofield et al 2016, Ellis et al 2012, among others. The discussion would also benefit from information about other potential fates of C.

Borges et al 2015 discuss the different wetland-river connectivity between rivers in the Congo and Amazon basin. The connectivity in the studied catchment is similar to which one? How would you describe how the patterns you found in this study for both conditions and larger rivers?

The main conclusion of the larger contribution of wetlands is speculative because groundwater input of C directly to high-order streams was not measured. So, the limitations for this conclusion need to be mentioned.

The terrestrial sink is barely mentioned in the discussion, and you don’t really answer the question in the title.

Figures

You could make separate figures showing the annual variability and the statistical difference among sites. This would give a better overview of the variability. Then you could replace this new figure as Figure 3 and use Fig 3 as supplementary material to describe the season variability of physicochemical parameters. TA could also be moved to the supplementary materials. Please also present figures or tables presenting the relationships stated in the results where you say “data not shown”.

Specific comments:

Abstract

L20-21. Explain the gradient groundwater to the main stream to mention that groundwater was measured in the forest (non-flooded?) and wetlands.

L23. Please report the pCO2 value.

The abstract is hard to read. Maybe start with a short contextualization and focus on the overall results instead of describing every single result.

L30-31. Please mention the heterotrophic respiration and CO2 emissions from the river above when you describe what you measured.

L34-36. Please clarify what you mean by “unique terrestrial source”, and specify what you mean with the “whole amount of carbon”. I read this sentence many times and I’m still not sure I got the point. Are you highlighting that wetlands are the most important source of carbon to rivers? Please rewrite this sentence to make your message clearer.

Introduction

L38-41. Please rewrite in fewer sentences, merging some of this information to give better flow.

L69. I noticed you cite three different studies by Borges from 2015 in the introduction (Borges et al 2015, Borges et al 2015a, Borges et al 2015b) but you only have one reference in the Reference list. Please correct this inconsistency.

L77-79. You could mention how this study benefits from this M-TROPICS effort. Otherwise, this piece of information is loose here and could be removed.

L79-82. Rewrite the study objectives more directly and clearly. You are evaluating the changes in C concentration across groundwater to different stream order over the seasons.

L86-87. You did not assess the complete terrestrial C export. Please specify only the component analyzed (groundwater). Change “terrestrial ecosystem” to non-flooded forest groundwater or something similar.

L89-90. You need to give more context to clarify the study hypothesis. What is the link between the net C sink and the riverine C budget?

Methods

Describe the wetlands in more detail. Are they flooded forest or open areas with grass? Do they differ between the headwaters and high-order streams?

Please number the equations.

L97. Remove the “the” before swamps.

L99- Change “experiences” to has.

L106. Mengong source and outlet are not displayed in Fig 1.

L107-109.

L110. Export is misspelled “epxorted”.

L113. Start saying that this is the first order basing.

L115-117. Needs a reference or should be rephrased using “likely” instead of “eventually”.

L113-126. Please rewrite more concisely and clearly. Try going straight to the point of why this information is useful.

L.129-130. It is not clear if the Mengong source is only groundwater.

L130. Explain how samples were taken from the stream. Using the Niskin bottle?

L132. State what TA stands for.

L134-142. This part needs to be rewritten. The physicochemical data explanation is very fragmented.  You could just say that from Jan to March you used one probe and after that another. After that the details about calibrations.

L143-161. State the number of replicates you have for each analysis.

L217. What do you mean by terrestrial groundwater?

L220. What the “r” in “[C]GWr” stands for? Define what is FexGW.

L222-223.  Please reformulate this sentence. The flow of what? Please specify. Why the flow rate unit is in metric not volumetric? Clarify what surface area the swamp drain? Is it the sub-basin area of the non-flooded forest? What is the total area of wetland in the catchment?

L223. Please be consistent using wetland or swamp throughout the text.

L228-229. Explain how did you estimate DOC and DIC from this study.

L230. Need citation for the negligible surface runoff.

L239. Include the site depth in table 1 and mention in the methods that it was measured.

L279-282. Merge these two sentences.

L280. Substitute “peaked” by significantly higher. Peaked works for a temporal description of a site, but to compare different sites is better using higher or lower than.

Results

Section 3.4. Please also describe the results after accounting for the respective areas of the streams in t Cyr-1.

Discussion

L349. If you are not testing this hypothesis, rephrase the sentence using “suggesting that…”. Also here, mention how deep is the groundwater level and how far the groundwater is from the organic layer of soil. Information regarding the C distribution in the soils would be helpful.

L358-359. Wouldn’t this then be groundwater respiration?

L362-363. pCO2 does not change significantly between seasons. How can this explain the high O2? The O2 is higher than I would expect for groundwater. Could this be a sampling artifact? Since this was a seep and you have the water flowing through a pipe, the water may have been oxygenated, and a lower volume of water flowing during the base flow period would get oxygenated faster. If that is the case and considering that most of the DIC in the groundwater was as free CO2, is it possible that your pCO2 may be underestimated?

L372-374. What type of vegetation predominates in the wetlands, C4?

L377-379. Couldn’t this also be attributed to the deposition of inorganic materials due to the reduced water flow in wetlands in comparison with streams?

L379-382. A similar process wouldn’t also happen in forest soils with trees and their roots reaching the groundwater and supplying labile OM below the lateritic layer?

L403-404. In Figure 5 you don’t show a relationship, but only a difference between low and high water. Make a correlation test.

L419. Change “confirms” to suggests.

L421-423. Explain why substantial input of pCO2 via groundwater is not expected to sustain high pCO2 in high-order streams.

L443. What do you mean by “invested”?

L461. Add inside parenthesis what are the components of this ratio, and why you show three components and only two for Ciais et al 2013?

Tables and Figures

Table 1. Is the slope unit correct? What slope is this? The average basins slope?

Tables 2 and 3. Explain if the NA is because samples were not collected or because it was below the detection limit, or another reason.

Figure 1. Add River order in the legend. It could be good to display river order in a shade of blues to differentiate more from wetlands. What is the unit of the coordinates? Why the site Yaounde is shown with a large red dot?

Figure 2. Do you mean river discharge? Please show the historical mean and variability of monthly discharge.

Figure 6. Present the results after multiplying by the respective areas of the entire basin and river's surface. Also, add the unit in the figure caption.

Review for: Partitioning carbon sources in a tropical watershed (Nyong River, Cameroon) between wetlands and terrestrial ecosystems: Do CO2 emissions from tropical rivers offset the terrestrial C sink?

Overview

            This manuscript by Moustapha et al. present a substantial collection of physicochemical and carbon data across stream orders in the Nyong basin in Cameroon to partition fluxes and attempt to close the C budget in this basin. The contribution of C flux data from tropical streams and rivers, groundwater, and from Africa is exciting to see, though the manuscript has several points that need attention before publication. There are general editing issues (typos, missing words) that will help focus the paper and a general polishing of the writing will help. In the results and discussion, words like ‘obviously’ and ‘probably’ should be removed following interpretation of the results the statistics. I believe a hypothesis driven approach will help the authors examine their data at a finer temporal scale and focus the broad application of statistics at a finer level to account for more of the variability in the dataset.

General comments

            The data collection spans one year from 6 sites in the Nyong basin and attempts to separate inputs (terrestrial vs wetland groundwater) and exports (evasion and export). However, there should be greater focus towards a higher temporal resolution of the fortnightly measured variables and the hydrology. Separating the hydrograph and seasonality into 3 categorical sections is too coarse of an approach when a higher resolution is capable and likely overstates the continuous nature of seasonality. Further to this point, I don’t recall much discussion of 2016 compared to ‘the average’ year. Particularly for evasion, more data spanning hydrologic variability is needed from across the globe and is in the dataset for the manuscript, but not presented.

            The evaluation of C inputs and exclusion of respiration needs further discussion. The methods to measure pelagic respiration are stated, presented, and discussed, but not included into the budget. I fully agree that including this small amount of CO2 from in-stream processes is minimal compared to groundwater and wetland contributions but excluding it does not make sense to me. I see two options, though there may be others: 1) include the in-stream component respiration into the larger budget and empirically show this flux is much smaller the other input fluxes or 2) remove the respiration component entirely and refer to these data in supplementary material or as unpublished data that are not on the same order of magnitude as the other fluxes.

            The chamber method used leads me to think option 2. While the dark chamber or respiration chamber method is fine for large rivers and lakes (e.g. Borges et al. 2019), this approach focusing solely on pelagic processes in low order streams and rivers are not sufficient and understate the influence of the benthos in the transition from benthic to pelagic processes that occur in mid-order rivers (Reisinger et al. 2021). The authors acknowledge some of the issues with respiration in the discussion section, but they fail to include the data even though it is available.

            There are broad issues with units throughout the paper, and I recognize conversion between the units varies between scientific communities or journals. Presenting concentrations and fluxes as both moles and grams is a little confusing and the units need specification of what is being presented (mmol CO2-C or mmol CO2). Basin scale fluxes are presented as both Gg and tons of C. I would stick to the metric unit (Gg) or convert to Pg, which are used in other C flux studies and the readership for this paper will be more familiar with. The presentation of units between mol and g is something I deal with in my own work, so I empathize with the authors.

            There is a structural issue regarding the statistics that I think can be resolved with presenting hypotheses. At the end of the introduction, only one hypothesis is stated and is unclear to what extent this is revisited later. While this paper is a C budget and perhaps not best suited to hypotheses, I suggest adding several hypotheses to guide the presentation of the data and focus the statistical approach. There are interesting questions about temporal and spatial hydrologic variability, stream order position, rainfall, etc. that can be used to ask questions and lead to testable hypotheses within the dataset. These hypotheses can help clear up the statistical approach, which appears to have been a broad application of ANOVA to all the data (see specific comment below). I think a list of focused hypotheses will lead to a cleaner presentation of the statistics and results section of the paper, while also allowing the main question in the title of the paper to be answered explicitly.

Specific comments

• L30-31: what are the units for respiration here? As mmol C, mmol CO2-C, mmol O2? Be specific. Also, in L29 can the units here be in metric (e.g. Pg C)
• L40: I think the word ‘evasion’ is missing before the Raymond 2013 citation.
• L41: ‘compare’ change to ‘compared’
• L44: See Drake et al. 2018, Tank et al. 2018, or Gómez-Gener et al. 2021 for updated values of global CO2 emissions from inland waters.
• L82: I appreciate this explicit designation of the fluxes measured in this study. However, in the abstract, estimates of heterotrophic respiration were mentioned, but not here even though this production of CO2 through in-stream metabolism can be a small but non-trivial source of CO2 (Rocher-Ros et al. 2019).
• L83: Only one hypothesis?
• L99: Scientific names for these plants might be more useful to a broader audience
• L102: Is the Mengong catchment within the Nyong (I see this is answered in L113)? Is the rainfall measured here characteristic of the wider basin? Help the reader by giving context to your study area
• L106: I would re-cast ‘stream orders’; groundwater is not a stream order. Something like: ‘We sampled groundwater and surface waters, including streams across Strahler orders 1-6’ (if that is indeed the case).
• L107: ‘gauging gauges’. Change to ‘gauging stations’. The table has ‘stations’, I would follow that.
• L110: Is 200 m3/s the annual mean? What is the temporal variation, as you’ve indicated there is seasonality in flow? Also, typo ‘or’ is meant to be ‘of’. ‘Epxorted’ typo as well
• L113-126: This section should be shortened and edited
• L128: Personal preference for the Oxford comma
• L163: I have to assume the cool box is also a dark box that prevents light. I’m not sure the pelagic approach to respiration is the most representative approach to study in-stream CO2 production especially in streams and small rivers, as much of the biological activity is occurring in the benthos. You may be underestimating the in-stream contribution to CO2
• L203: We are in Section 2.4, I think you mean Section 2.3
• L205: be specific with units: mmol CO2-C or mmol CO2? You then convert to Gg in the next sentence. Pick one of grams or mols and stick to it through the whole paper. Again, in L209, why convert into t C? Most C flux units are as Pg or Gg. Make it easy for your readers by not over-converting between units
• L236: what are the units an^-1? Is this an annual basis (i.e. year^-1)? Be consistent
• L238: What are the explicit units here (CO2-C or CO2)?
• L247: Unit issues again
• L264: ‘a given parameter’- be more explicit. You have measured a tremendous number of parameters, as fluxes, concentrations, etc. How is the reader to know if you 1) ran a correlation for everything measured or 2) focused on specific fluxes? I think there is an opportunity to be specific here in the statistical approach that would be aided by defining hypotheses or explicit relationships in the introduction that are missing in the introduction. I appreciate that the C accounting is not as a hypothesis driven approach, but you are also examining seasonality, stream order, and Q-C plots that could benefit from generating testable hypotheses in the data.
• L272: what are the O2 units? Be specific and say percent saturation.
• L280: ‘peaked significantly’; peaked suggests change over time, but this comparison is between sites. Re-cast as ‘DO was highest in the So’o’. The wording of the statistical inference in L281-2 needs cleaning up.
• L282: Here are the data to answer a hypothesis related to temporal variation of these variables
• L324: 16% seems higher than ‘fairly balanced’. In the results section, I would simply state the ‘difference was 16%’ rather than qualifying as ‘fairly balanced’, which is a judgement that merits discussion later in the paper.
• L341: ‘soil OM respiration’ reads as if the soil OM is doing the respiration. Re-cast as ‘respiration of soil OM in the unsaturated zone’
• L343: ‘probably’- do the papers cited at the end of this sentence give any clarity or more definitive data to guide this statement?
• L352: ’50 times higher’; be explicit, what is the concentration or ppmv?
• L355-6: ‘During base flow, precipitation was low…’ I hope so! Switch the order of this statement ‘Low rainfall resulted in lower flows than the other seasons…’ or similar. Same language issues in L357.
• L443- ‘invested’; not sure that is the word to use in this case
• L449: Based on your budget, but you acknowledge you didn’t include respiration, which is a flux you measured but chose not to include! I agree that groundwater and wetlands are likely large contributors to stream C but you have the data to make the comparison to in-stream processes. You make this comparison in L454, but I don’t see why not include in the budget, even it its less than the error of the other input fluxes
• L453: typo ‘trough’; delete everything after ‘atmosphere’
• L 474: there is no discussion of the 16% difference mentioned in the results, that seems important to bring up again

Tables and figures

Table 1- how representative are each of these streams of the broader orders they represent across the basin? ‘Averaged annual’ change to ‘Mean annual…’ and use yr^-1 in the units. Can you provide a brief overview of the gauging stations as a footnote or in a supplementary file?

Table 4- Is the first column the different stream orders? Why was respiration only measured in 2 sites, the text says in all sites? The units in the table are an issue: umol, mmol, and Gg. The gas exchange rates seem low; was there any attempt to evaluate change in k600 over time and changes due to changes in discharge?

Figure 2- are ‘Days’ day of the year? Day since start of the project? Days in the water year? Please change to a date to help your readers. Also, why not show the data from all the streams with a gauging station?

Figure 3- If Tukey’s post-hoc test compared the seasons, why not use the groups from that test above or below each boxplot to designate the significant groupings? The horizontal bars and asterisks are distracting. The axis text and titles could be bigger. Also, is this figure and Table 2 showing the same information? I think the figure is more valuable than the table.

Figure 4, 5- same comment about Tukey letter groupings as Fig 3

References

Drake, T. W., P. A. Raymond, and R. G. M. Spencer. 2018. Terrestrial carbon inputs to inland waters: A current synthesis of estimates and uncertainty. Limnol. Oceanogr. Lett. 3: 132–142. doi:10.1002/lol2.10055

Gómez-Gener, L., G. Rocher‐Ros, T. Battin, and others. 2021. Global carbon dioxide efflux from rivers enhanced by high nocturnal emissions. Nat. Geosci. 1–6.

Reisinger, A. J., J. L. Tank, R. O. Hall, E. J. Rosi, M. A. Baker, and L. Genzoli. 2021. Water column contributions to the metabolism and nutrient dynamics of mid-sized rivers. Biogeochemistry. doi:10.1007/s10533-021-00768-w

Rocher-Ros, G., R. A. Sponseller, A. Bergstr, M. Myrstener, and R. Giesler. 2019. Stream metabolism controls diel patterns and evasion of CO2 in Arctic streams. Glob. Chang. Biol. 0–3. doi:10.1111/gcb.14895

Tank, S. E., J. B. Fellman, E. Hood, and E. S. Kritzberg. 2018. Beyond respiration: Controls on lateral carbon fluxes across the terrestrial-aquatic interface. Limnol. Oceanogr. Lett. 3: 76–88. doi:10.1002/lol2.10065

The authors have been compiling a large carbon data set from a river in Cameroon (monitored since 1993!) and use it to weigh in the hot topic of river metabolism and C budgeting. Their main goal is to assess the relative contributions of wetlands and uplands to river carbon, finding that wetlands are somewhat more important with a 60/40 split. The authors also conclude that upland forests, in addition to being less important than the wetlands for C export, apparently only export some 4% of their net C uptake, suggesting that they are indeed potent carbon sinks and not simply “leaking” carbon away to be re-emitted by streams.

Beyond the trendy C attribution angle, the authors also have a loosely-defined goal of describing spatial and temporal variations in their river C data set as part of assembling an overall budget for the catchment and they devote much of the results section to describing different trends and patterns. These deep technical dives are not so well-conceived, lack a conceptual framework and are largely divorced from the broader narrative arc of the paper. There is no doubt important information here, but the reader has little help gleaning it from paragraphs without topic sentences.



Overall, this dataset is definitely worthy of publication and makes an important contribution to understanding riverine C dynamics in a poorly understood region (tropical Africa). Below I point out a few key issues to address and make some recommendations for improving the narrative structure.

Major comments

Issues of narrative

The idea of assessing relative contributions of wetlands/uplands to the river C is well-grounded in active literature discussions and effectively pitched as a topic of interest (and is rightfully highlighted in the title).

In contrast, the goal of: [describing spatial and temporal variations in river C], absent any problem statement or hypothesis is not well-conceived and sets up a results section that is largely strings of facts. There are plenty of significant patterns to be found in such a large data set, but what do they mean? How are they useful? What questions do they answer? Without any of this framing the reader is left wondering: what is the point of all this work and analysis? After re-reading these sections several times I had an idea of why this information was useful, but this was not so easy. There is no problem with the mechanics writing (few typos aside), but rather an issue of conceptualization and paragraph design.

The discussion subsection topics similarly lack a thread holding them together. They give detailed glimpses into specific parts of the system: forest groundwater, headwater stream, catchment scale patterns… They could really us a narrative arc that ties them together—a clear objective or a question. The descriptions are great, well-supported with data and references, but lacking an overarching motivation or glue.

I would suggest laying out a framework in the introduction that explains why looking in detail at these different components of the catchment C system are important for understanding the functioning of the whole and then referring back to the framework with topic sentences. Give the reader some guidance as to why this all matters.

Terminology for wetlands vs. non-wetlands

In the introduction and throughout the manuscript, take care with terminology. The authors might try using “upland” to refer to well-drained ecosystems/forests. They occasionally use “Terrestrial” to refer to non-wetlands, but this would likely would include within it many types of wetlands (room for interpretation/confusion in the mind of the reader) and even “forests” will likely also include many wetlands since they are treed with palms. This is quite important to define clearly since distinguishing between the role of wetlands and non-wetlands within the catchment is main goal of the study.

How solid is the estimation of catchment wetland area?

Is there no more localized data source for the wetland area? Gumbricht et al. 2017 is a global-scale wetland map and probably is not very accurate at such a small spatial scale. Check if there are better wetland maps for the region, if not then make some sort of a statement that local/regional maps are unavailable and that using a global map is the only option. How sensitive are the results to accurate delineation and counts of areal coverage of wetlands vs. uplands?

Logic of transport pathways (Fig. 6)

The authors model two C transport process between terrestrial ecosystems independently: Forest -> Stream; and Swamp -> Stream. But in a high percentage of cases will these not be linked? i.e. Forest -> Swamp -> Stream ?  It seems that this logic opens the possibility that some important portion of wetland C export might have been “inherited” for surrounding hillside forests. This might mean that the importance of wetland C export relative to forest C export might be exaggerated by this analysis.

Framing of conclusions

“The current paradigm…”

I had to refer back to the introduction, to try to figure out where this paradigm comes from and can find no sign of it. This appears to be a bit of a “straw man” argument? Yes, the paradigm is that headwater streams are heterotrophic, but I don’t see any sources stating that the DIC/CO2 is produced exclusively or dominantly through heterotrophy, rather than inherited in inorganic form from groundwater…

Minor comments

Line 58: “anthropogenic budget”? So far the processes being described do not involve human activities…

Methods: Describe handling times of samples, preservation methods (kept cool?) and analysis location (Europe somewhere?)

Statements of precision/repeatability

“The repeatability was better than 0.1 mg/L”

“Precision was +/- 0.1 mg/L”

It isn’t totally clear what parameters these statements refer to or what exactly they mean. Is it analytical precision (ie the balance measures to the nearest 0.1 mg?) Or did the authors take replicates and calculate standard errors?

Lateral inputs.

I struggled to follow the logic in this paragraph regarding a 0.48 km2 “hillside.“ Unclear how this relates spatially to or represents behavior of a catchment of 27,800 km2. How does a hillside have “base flow” ?

Line 310 (data not shown)

Why not add to supplement? If important enough to mention in the text, the authors should somehow report the data.

Line 323 why reverse order here? DIC, DOC, POC vs POC, DOC, DIC… best to be consistent…

Section 4.1



The key information is buried way down on line 350. The authors are interested in understanding whether groundwater CO2 is coming from rock-water interactions, deep soil respiration or surface soil respiration. Based on their evidence, they posit the latter. Make the goal of this section clear from the start. The reader is left wondering for 22 lines of evidence what the point is…

Section 4.2

Lots of descriptions here, but not connected to any explicit goals. What’s the point of all of this?

Section 4.3

Same as above. The authors reach several conclusions in this section… “It confirms…” “This confirms…” But there was never any explanation of what tests were being done or any specific hypotheses. This is both a challenge for readability (lots of evidence is presented before the research question has been explained, making the logic difficult to follow); and also scientifically: we should be laying out tests, with clear possibilities that the data and other evidence from literature can support or refute. Just describing the data and then saying it confirms something is not exactly the scientific method.