The paper by Croghan and co-authors entitled "Seasonal and interannual Dissolved Organic Carbon transport process dynamics in a subarctic headwater catchment revealed by high-resolution measurements" presents a comprehensive data set of high-frequency flow and DOC data (and associated hydroclimatic variables) over four years to assess the DOC dynamics in an arctic catchment (without permafrost). The authors utilize this data to explore high-frequency metrics and other indices. They use random forest models to assess the drivers of these metrics, and find some interesting and some confounding patterns.

This paper is well structured and clear - with a thorough  analysis that doesn't stretch too far beyond the data in hand and seeks to advance our understanding of the system. The data set is valuable and unique, and the process insights strong. There are a number points I would ask the authors to consider, and also explore a newly published paper that has a similar analysis in a subarctic environment with at times similar and at times different results that likely was not spotted before submission. Some of the authors interpretations may (or may not) be informed by this paper:

Shatilla, N. J., Tang, W., & Carey, S. K. (2023). Multi‐year high‐frequency sampling provides new runoff and biogeochemical insights in a discontinuous permafrost watershed. Hydrological Processes, 37(5), e14898.

Other comments:

~Equation 1. Is there a reference for this equation? 

~Line 237. Only single peaks were used for analysis. How many events were single vs multiple peaks in the entire data set? Did the authors examine multiple peaks or was this simply 'too messy'?

~I would like the authors to consider including summary hydroclimatic information that is easier to 'digest' than the time series. This could be in the SI, but seasonal and yearly temperatures and precipitation (as well as rain v snow) may be informative to the reader as they interpret the data. At several times during the discussion I was trying to evaluate warm vs cold seasons or wet vs dry and it was difficult from the time series alone. 

~If I am interpreting this right, DOC is strongly influenced by water temperature (and all the other factors as well) in the winter (Table 2). I'm struggling to interpret the very large values of node purity (compared with the summer months) and the % variance explained (which is less in the winter). This likely is due to my lack of knowledge with the random forest analysis although it is quite commonly applied. I'm more used to seeing the %Var attributed to each predictor variable.

~Perhaps a small thing, but air and water temperature are highly related, does this affect the random forest models at all (co-linearity issues) 

~Line 426. How as the event water determined? Via isotopes? Please simply state the method.

~For Figure 6, how much of the inter-annual differences is driven by the differences in range of event water yield? Some years have very high flows, whereas others are much more modest. I'm wondering if the 'differences' are simply driven by the end-members. Can a test/consideration be done for values within the same range (say < 50 mm). Largely I'm wondering if doing this constrains the process drivers to more common events, etc., and whether the inter-annual variability in load vs event water yield holds. I'm struggling a bit here and in the interpretation to figure out 'why' this would be. Obviously different climate conditions would be the first place to look, but it remains somewhat unresolved.

~Line 469-470. While the forest is mostly conifer, is there much of an understory or mixed forest input of leaf litter? Do you think this affects your interpretation of the DOC sources?

~At the end, you mention stable isotopes of water as an area of future focus (line 537). Do you think that looking at DOM quality would be helpful in advancing our interpretation of DOC data? 

This paper used high frequency in situ measurements of DOC and discharge from 2018 to 2022 to examine the seasonal variation in concentration-discharge (C-Q) relationship. Overall, this paper presented a novel data set with comprehensive analyses of the C-Q relationship, including the C-Q slopes, the hysteresis patterns, and DOC load yield. The paper is mostly well written and clearly presented although some clarification of concepts and  interpretation of results may need improvements . I hope the comments below help the authors improve the manuscript.

1. In the introduction, the authors used the term “transport process” as a general term for what their research focuses on. This term, however, has a very broad meaning and thus it may not be clear to readers what exactly the authors will investigate. I thus suggest the authors clarify what they mean by transport process. Although the scope and meaning of this term becomes clear in the methods, it would help readers understand the paper better if the term can be explicitly explained in the introduction.
2. The introduction section does not have a clear statement on the current knowledge gap and how this paper will address that gap. It is not clear what scientific question this paper tries to address. The analysis essentially computed almost all metrics one can do to a concentration-discharge relationship. I think a more clearly defined research question/hypothesis would make the paper’s motivation more clear.
3. The authors suggest that the variation in C-Q slope between years were generally smaller than seasonal variation (e.g., line 321). Is this really the case? If we look at figure 3(a), it appears that, at least in November, the among-year variation in C-Q slope is larger than month-to month variation within the same year. I thus think the statement that month-to-month variation is larger is not fully supported by the data.
4. The authors only used single peak data for analysis of event C-Q relationship. How many events have multiple peaks? If multi-peak events were not excluded, would the results remain the same? Some additional analysis including the multi-peak event would be helpful.

Below, I listed detailed line-to-line comments.

Line 187: please provide citation information for equation 1.

Line 209-214: Could you please provide numeric criteria you used to define snow cover, snow melt and snow free season. From figure 2, snow depth fluctuates even in what you classified as snow cover season. Wouldn’t that cause some period of time to be classified as snow melt season within the currently defined snow cover season?

Line 225: is it each season or each month? From the figure, it seems C-Q analysis was done to data in each month.

Line 237: A brief explanation on how HI index was calculated could be helpful.

Line 248: the term “dynamics” is a bit vague here. Could you please be more specific and explicit about the meaning of this term here?

Table 1: Is the unit of flow wrong? Shouldn’t it be L s-1 km-2?

Figure 3: how was coefficient of variation calculated? What is the standard deviation used in the calculation of CV? If I understand correctly, the C-Q slope here is derived by linear regression using data in each month. Is the standard deviation used in CV calculation obtained from the standard error of slope from the regression? If so, the CV calculated here is not meaningful because the SE of slope from regression shows the uncertainty of estimation, not true variation in slope over time within a month.

Figure 6: similar to my comment to table 1, shouldn’t the unit of DOC load yield be kg DOC km-2?

Line 364: please give exact p-value, not just a range.

Line 375: please give the degrees of freedom of the F test statistic.

Line 382: what’s tested here is that the slope is statistically different from zero assuming there is a linear relationship. Whether the relationship is linear or not is not tested. Thus, the term “remained strongly linear” is a bit misleading.

Line 420-432: the explanation here may need further consideration. The trend in C-Q slope seen using data within each month (figure 3) is not evident when analyzing event C-Q behavior. Thus to say that the patterns seen in figure 3 suggests limited source does not reconcile with what is shown in figure 4(a), particular considering that ~60% of flow occurs in events.