This study addresses the impact of hydrometeoerological extremes on various landscapes – particularly with respect to TSS and DOC. Uses North Saskatchewan River between 2019-2022 as case study. Study finds that warming and wetter periods resulted in greater TSS and DOC export, with DOC behaviour being more predictable than TSS during these periods. 

I could easily understand the paper and it was the methods seem sound and appropriate for answering the key research questions of the impact of hydrometerological extremes on TSS and DOC export across different landscapes. 

I only have a few suggestions that will hopefully make it easier for the reader to fully appreciate and understand the study and its findings. 

Line 41-42: “During higher water” -> consider changing to “During wetter years”

Line 53: “Recently” -> consider specifying when

Line 103: “Ongoing changes” -> for international audiences, please state what changes are being experienced

Line 105: “recently endured” -> for international audiences, please specify when these extremes have been occurring (how recent have they been?)

Line 105: “High runoff years” -> please specify how high the runoff is

Line 245: “yet” -> implies that one day these data will be available. 

Line 252: “Phosphorus” -> in this sentence, unclear if it is total and/or dissolved phosphorus here

Line 285-287: “The three fixed factors…autoregressive approach” -> Having the equation structure shown here might help clarify how the model was set up for the reader.

Line 288: “normality standards” -> would like some indication of whether they actually ended up meeting normality standards, or whether they just got a little bit closer to normality after transformation

Line 289: “Sidak multiple comparisons” -> please provide a citation for this as I (and other readers) might not be familiar with this approach.

Line 295: This is a great way to get around the fact that many catchment characteristics are cross-correlated with each other 

Line 337: “NSR-Edmonton Station” -> is this the downstream most station? Would be good to specify where this is in the context of the catchment for readers unfamiliar with the area. 

Line 352: “substantially” -> would be good to quantify this here

Line 383: “statistically” -> I’m interpreting this as the slope having p<0.05

Line 387: “statistically stronger” -> I’m not actually sure what this means – does this mean that the slopes are steeper?

Figure 4: I’m struggling to understand what the letters mean in the mixed modelling results here. If you can provide an explanation in the caption, it would be helpful. 

Line 524: “C-Q” -> these relationships feature a lot throughout especially the discussion. I wonder whether it is worth having a figure/table of these relationships in the main text instead of just in the supplementary materials. Or at least putting some examples of the relationships in the text.

Line 545: “we observed…changes in flow” -> at this point would like to see which specific result this finding is being drawn from. I think this is pretty important, but am unclear as to how the results of the analysis have led to this conclusion.

Line 552: “model results” -> I assume this refers to the linear mixed models, but would be good to specify

Line 553: “high end” -> I assume this means high C and Q?

Line 583: “weaker” -> again here - providing some numbers from comparison (as evidence) would be really useful to help the readers understan d the conclusions being made.

Line 592: “stronger C-Q associations” -> I assume this means a steeper C-Q slope?

Line 621-623: “little evidence in…DOC mobility” -> would be great to refer specifically back to what findings/results points to this not happening

Table S4-S5: would be really good to show error bars in load estimations too (max-min)

The authors assess how hydrometeorological conditions and catchment structure influence total concentration and fluxes of total suspended sediment (TSS) and dissolved organic carbon (DOC) in a very large glacial catchment (upper North Saskatchewan River). I enjoyed reading the manuscript and I think it falls within the scope of HESS. It makes an important data contribution, and the authors made a big effort to summarise and nicely show a large amount of data and results. The paper is well written. However, I have a few concerns that I would have liked the authors to better addressed. The most relevant one related to the method and way they chose to predict and report annual fluxes.

• I find the title misleading, “organic matter mobility” might also include particulate forms, whereas the article only looks at dissolved organic carbon. There are also other forms than carbon contributing to dissolved organic matter. Consider replacing by “dissolved organic carbon” to be precise. This should also be revised in the manuscript.
• The paper is missing a clear research question and/or hypothesis. The authors “assess how fluctuations in hydrometeorological conditions (…) impact the downstream delivery” of DOC and sediments. However, I think there is not enough information in the introduction explaining what is already known about the processes and leading mechanisms, and which knowledge gaps the research intends to fill. The authors expect “differences between TSS and DOC mobilized through the river network (…)” what remains very vague and obvious. This could be rephrased into a more precise hypothesis, and I encourage the authors to formulate a research question.
• Until I got to 2.3.1 I thought the definition of CSUs was “to quantify the influence of major landscape types on modulating hydrometeorological conditions” (lines 142-143). However, it is then presented in section 2.3.1 that the method is used to classify (with replicates) the catchments in different types (“the number of clusters selected for further assessment was chosen manually to balance replication of catchments within a reduced number of clusters”), is this right? I think the aim of the analysis should be better presented earlier (in section 2.2.1). I also find a bit confusing that sections 2.2.1 (data) and 2.3.1 (numerical assessments) are separated. I also was confused in Figure 1. Maybe a cross-reference to 2.3.1 could be added in the figure caption.
• Figure S1 shows the sampled flow percentiles. Which flow data did you use to make this figure (15-min, hourly, daily, monthly?). Without this information it is very difficult to interpret the significance of the figure. It seems impressive to be able to sample 90% of historic flow values in most stations with 4-years monthly samples. Maybe this is achieved with the “flow-targeted sampling”. I think more information is needed.
• I think the methods used to predict annual SSC and DOC loads should be better explained. So far there is only reference to the R package loadflex (Appling et al 2015 – a reference that does not appear in the reference list). Results are expressed as “Export ± standard error” (Tables S9 and S10), the reader should know how this is computed. With the given information it is very difficult to address the difference between methods and their accuracy and precision.
• Loadest and log-log regression export model concentration-flow relationships fits for the 2019-2022 periods are reported in Table S3, but what about Loadset composit? Also, there is a few manuscripts investigating the uncertainties associated with these methods, which are rather high, so this should be better addressed. Annual fluxes (and storage) are calculated/reported using the median export of the four models (supplementary tables S4 and S5). I think this adds an extra layer of bias in the results that could be avoided by finding out which method works best (an analysis that is completely missing). I do not think using the median of the four methods included in the loadflex R package is the most accurate way to report annual fluxes of DOC and TSS. This is especially true because r2 values for some sites are highly variable when using different methods, e.g. 0.17 - 0.99 (DOC Strawberry Ck.). The authors also provide the values reported with the four methods, but it is very difficult to assess how good the methos are when only looking at the results. The same happens when looking at the supplementary material, where only r2 values and total fluxes (also per year) are reported. Maybe figures would help here.
• Figure would also help when the authors discuss how extreme high events perturb model fits at the high end of C-Q associations (lines 551-553). This is very difficult to assess as we do not see the data at all. Maybe an
• hydrograph with the sampling points would also help to evaluate how representative are the sampling times.
• Is there any other evidence of in-channel/bed or bank erosion or deposition that could be used to cross-check the results? Sediment storage in the upper two station reaches seem to be happening independently of runoff (lines 484-487), is this related to the dam?
• The authors discuss the projection and consequences of their results in the future (lines 653-705). I find some of the discussion going a bit too far. For instance, the authors discus in lines 663-674 how changes in DOC and SSC will impact water treatment cost in the future or ecosystem health. I can see the importance of this, but I think it goes beyond the scope of the paper to describe future needed adjustments in water treatment facilities.
• I also miss a clearer evaluation of the limitations of the used approach and acknowledge of the uncertainties involved. Especially to address the uncertainties associated with the sampling and the estimation of annual loads. Between 37 and 55 samples (four years) are used to estimate annual loads (between 9 and 14 samples per year). The authors should be clearer about the high uncertainties associated with the results.   
• I would also advice to revise the conclusions to mainly address the most relevant findings. I do not think that some of the facts described (e.g. that a broader application of monitoring approaches that capture landscape heterogeneity (…) will lead to better decisions) are direct conclusions of the results shown.

MINOR COMMENTS

• Line 23: In the introduction (lies 83-84) the authors stated the opposite, that there are many factors that challenge the use of rivers as indicators of change.
• Line 24: “river material” is too vague.
• Lin 62, 64: I would also recommend using another word as “material” is vague.
• Line 64: I wonder if the sentence is 100% true, as if we quantify the “export of material by rivers” it does not tell us anything about response to change. One must look at how the export varies with time.
• Lines 74-76: This reference might be misleading as it does not apply to all types of catchments. Does this only apply to Canada? Is this also true when 90% of the area is covered by forest? Too vague.
• Line 78-80: pristine instead of intact? The sentence does not seem grammatically correct: runoff can modify surface water quality by storing water?
• Line 87-89: Not clear to me how a “river network monitoring” allows distinguishing whether sediments come from tributaries or from in-river sources (banks, channels).
• Line 132-134: where does this data come from?
• Line 138-139: how big are these reservoirs? Difficult to figure out their impact. It would be nice if they appeared in Fig 1.
• Figure 1. I can’t see the greyed catchment names. It would also nice to identify the 13 sites that are included in the analysis (2.3.3).
• Line 144: I do not recall Nippgen et al., 2011 use the catchment structural unit approach. I might be wrong.
• Lines 161-162: What’s the origin of these data?
• Line 164: There are only 17 sites lister for water quality in Table 1, but more than 20 points in the map.
• Line 165-166. I guess a rating curve I needed. A bit more info would be appreciated here.
• Line 166. How are the predefined flow levels defined for the flow-targeted water quality sampling defined? These should be better explained.
• Lines 203-205. I would appreciate more details here. I would be unable to reproduce this. Has any king of delay has been taken into consideration when flows not expected to peak at the same time?
• Line 215: this report is difficult to access.
• Line 215: At this point we have not seen DOC values, but as the paper only deals with 2 parameters (DOC and SSC) it would be interesting to provide more information about the data quality. I think this is necessary to interpret the results. I looked at the Laceby et al 2022 report, and it seems that 13% of the field blanc samples present DOC (environmental contamination). I would appreciate to know a bit more about this and the contamination sources.
• Line 224: basins?
• Line 322. Strange to add references in the Results sections – where authors should only report their findings. Maybe this should be moved to the discussion section.
• Line 359: how is the standard error of the median computed, bootstrapping? Would it be more appropriate to report the standard error of the mean (if values normally distributed)? It is normally easier to calculate (st dev/sqrt(n)).
• Figure 3. Are dots outliers? Information missing in the caption.
• Figure 4. why is TSS yield data missing for 2021 in Redwater and Tomahawk?
• Table S5. Correct uppercases in the table headers.
• Figure 5. Mention in the caption which data is reported (i.e. the model used) to compute TSSs.
• Line 545. Is “proper” a common term?