The focus of this manuscript is to explore the effectiveness of working with citizen scientists to collect long-term environmental data with support from a project team, specifically focused on the impact of a small dam removal.  In its current form, the manuscript does not clearly align with the stated aims and scope of the journal, it would be more appropriate elsewhere, or requires substantial reworking with major revisions at least.

1. The manuscript is presented as a citizen science focused project, yet there is actually very little detail on the citizen scientists themselves, including how they were recruited ethically, when they engaged with the project, how their engagement was distributed throughout time and perhaps most importantly whether they themselves benefited from the project.
2. Building on the first point, the manuscript can be read as implicitly attributing data loss or quality issues to citizen scientists, which may be unintended. Careful use of language is required, especially when the authors are claiming that citizen scientists can collect useful data (a message which I agree with).
3. Much more detail is needed around the results - the methods section 3.1. is fantastic, lots of great detail but then only d50 and WAV are presented, and it is not clear why. Linked to this, the data analysis and result interpretation requires further detail, for example how were the error bars determined? Did you perform any statistical analysis regarding the difference between d50 at different locations?
4. The overall structure of the manuscript needs reworking, particularly ensuring that the correct text is placed in the correct location. Introductory text are presented in the discussion (e.g., L329-339),  results are shown in L306-316 and L341-350 and so on. 
5. The supplementary data are lacking, with no clear readme file nor information regarding what the data relate to.

Below are line-by-line comments with further points that I hope that the authors find constructive.

L27    "Both species" - what species are you referring to?

L36    There are nine dams in the abstract, why are there now seven?

L53    "Local streams" - what are the size of these and location relevant to where the citizens are coming from?

L60    By "above-dam levels" do you mean upstream of the dam?

L66    What about research outside of the US? Such as in central Europe on check dam removal?

L78    9177 m is higher than Everest! Assume there is a typo here.

L88    Consider adding this information regarding the stream gauge to the Figure 1, alongside the actual data for the monitoring period as this will provide useful context for the rest of the manuscript.

L94    Seven dams again - please clarify whether this is correct or not and update the abstract accordingly.

L105-114    "This paragraph is a little confusing and needs restructuring. At the moment, it is presented as: "citizen science is good for spatiotemporal scales, yet data quality is potentially not robust. Research has found it to be non robust enough, but this is like comparing apples to oranges, so citizen science data is robust."

I think this needs unpicking more as you can compare citizen science data to that collected as standard, by using the standard data as a reference point. While citizen science has demonstrated value in many contexts, it is important to clearly articulate where it is appropriate and where limitations exist.

L125    How were these hours distributed across the study period? Were they equal, or did more monitoring occur in 2018, for example. How did this effect your results? How many people were engaged and how many were the same over the monitoring period?

L126-144    This information should all come in further up in my opinion when you are introducing the study site. This would then leave a dedicated section to discuss the citizen science objectives/protocol.

L157    This sentence doesn't make sense, please clarify/revise.

L162-163    How were these sites identified? Why only two upstream but eight downstream?

L171    More information is required regarding the width of the streams at the 10 monitoring sites, and why 20-50 intervals were chosen. Did you normalise the stream widths?

L230    "Winter season WEN" - this suggests that volunteers had different roles, please expand on this.

L230    Can you share the Excel sheet so that others with similar objectives can learn from your design?

L269-271    Why use the upstream sites as comparisons when you have three years of data before dam removal? This is the point of BACI. Furthermore, there might be changes upstream due to flow being no longer restricted.

L273-274    What about the data before/after then as this would give greater detail into how the river responded to the dam removal?

L275    How many observations? More detail required

L279    Why did you only calculate d50? If there are 100 measurements per sample site why not include the entire grain size distribution, including d16, d50 and d84 as stated in the methods?

L285    Figure 2 - I think that these figures would be better combined, showing the change throughout time based on the year. It is unclear where the error bars have originated from, why not use a percentile metric to calculate the 95% confidence intervals? See Eaton et al., 2019, ESurf, https://doi.org/10.5194/esurf-7-789-2019 for example.

L289    Figure 3 - similar comment to Figure 2, I don't think separating the data in this way is additive, therefore please think about combining the plots. Also it's unclear how the error bars were calculated.

L293-299    This paragraph would benefit from reframing, as it reads as if the citizen science data were effective in their collection, however a third of the data collected was not usable. Please consider how this is framed.

L301-305    This is not discussion as it doesn't link to the presented results as there was no significant change.

L307    "sediment flow" - you have not reported measuring sediment flow, rather grain size distributions and WAV. Please reword.

L308-312    This is results rather than discussion, why isn't this presented earlier in the results section? It contains important information that has not been presented.

L312    What irregular variation? More information is required.

L319-324    This is introduction to the area rather than discussion. Equally for the sluice gate statement, this should come earlier in the text.

L329-334    This is not discussion, it is introduction. Secondly, the results presented don't back up these claims as there's no evidence to support them. It's also not what the focus of the manuscript currently is. If you have anecdotal stories/evidence that would be more effective.

L336-339    Again this is introduction, not discussion.

L341-350    This raises a quality control issue that has not been addressed - why was all of this potentially fantastic data not quality controlled or checked regularly?

This paper is a useful exposition of the potential and challenges of using crowdsourcing techniques to gather a timeseries of scientific data over several years with changing teams.  The information presented is valuable, but requires more context, organization, analysis, visualization, and reflection to be useful to other readers of the journal.

CONTEXTUAL:

Dam removal is gaining traction primarily in the US and Europe in recent years.  The authors note correctly that research on this is primarily on larger dams although smaller dams and barrages may be more common.  Even in the larger dam cases, often little is known about the before and after timeseries of critical stream physical and ecosystem changes because of the dam removal.  The case that is made here that in the smaller dams/barrages, it may be useful to have more citizen science use as they have shown in this case to do such analysis.

The lack of any discernable changes in stream due to this small dam removal could perhaps be a function of not only the size of the dam, but also the hydrology (including intra and inter-annual variability), river and floodplain cross section, sediment load and characteristics, flora and fauna, history, and use trends in the area.  The relative amount of live and dead storage in the dam that was removed is unclear and could have played a significant part in the results – since the live storage was already removed for this dam since 2012 when the gates were removed – so the changes in 2020 would just be the removal of the dead storage.

The relative sizes of the 7 dams (major?) of which one was recently removed would be useful to indicate along with their locations.  Even for the site of interest, some characteristics of the dam that was studied (even basics like a name/codename if any, height of sluices, live/dead storage, catchment area, inundation area on the map, rainfall, flows, original operations, etc.) and the characteristics of the u/s and d/s monitored areas (incl. some visualization of the cross sections and other information) would be useful to include.  A timeline of the events on the dam would be a great graphic to include.

Such approaches could actually also be useful to understand changes when dams are first constructed if there is a good baseline for a few years before the dam is constructed. 

The description mentioned that one concern that may have led to dam removal was the spring fish migration problems during high flow – did this change after the full dam removal?

METHODS:

The primary thrust of the paper is to showcase how crowdsourcing methods could be standardized and used to collect useful information using simple tools and techniques.  It does seem like these can not only be a great way to teach students and get them more connected to the nature around them, but also provide some data for insights even with the quality concerns.

A challenge indeed is that 3 years data on either side of a dam removal may not give enough time for the geomorphological and ecological processes to change and may just reflect the variability in the streamflows and sampling.  The use of basic data quality management is also highlighted (e.g. missing data and not knowing which cross section measurement was from which site).  The websites for sharing the data and code should be checked as they do not exist as included.

The use of simple tools and simplified standard operating procedures and their description is useful.  Some of these techniques may also be modernized over time – e.g. writing on paper pads in such an environment and then digitizing into spreadsheets for analysis can perhaps be replaced by using smartphone tools (online spreadsheets or even simple apps that can have built-in error checking upon entry and also record locations, etc.).  Simple probes/sensors that are getting increasingly cheaper could be utilized. Some of the lidar scanners built into some smartphones and tablets could also be great for cross sections that seems to have been a weak spot in this analysis due to unusable data from traditional methods.

Overall, it looks like this study can provide pointers for both what to do and certainly on what not to do in the case of crowdsourcing such data.  It also points to perhaps moving towards a more standardized approach ensuring better data quality that can provide crowdsourced harmonized data for every stream based on the work of nearby educational institutions that facilitate crowdsourcing.  This could provide a great way to just look at stream physical and ecological changes over time whether dams are involved or not.  It may be useful to also explore strengthening some national knowledge network for crowdsourced stream monitoring that can also develop simple online ways to organizing, analyzing, visualizing, and disseminating these kinds of open-source data for open science.  This may be a great approach to be further adapted to encourage open science especially for the developing world where even less is often known about local streams or rivers.