Review of hess-2021-358: “A parsimonious model of longitudinal stream DOC patterns based on groundwater inputs and in-stream uptake”

Content This manuscript describes a parsimonious model application for instream DOC concentrations along a stream reach, basically consisting of a mixing model and an instream uptake model. While the DOC concentrations were directly taken from groundwater sampling at each of the 25 stream segments, the amount of GW inflow was calculated from the discharge discrepancy between inflow and outflow of the reach. The distribution of water inflow per segment was either equally distributed (diffuse inflow scenario) or according to the area of specific sub-catchments contributing to groundwater inflow at a segment (UCA, upslope contributing area). The model is used to test the hypothesis, that previously defined ‘discrete riparian inï¬ow points’ (DRIPs, Ploum et al. 2020) have a major influence on DOC stream concentrations and that this is connected with spatially variable groundwater inputs.

explaining the strategy to account for uncertainty in the simulations (L.232-244) is on the one hand hard to follow and some decisions seem arbitrary. On the other hand, simulation uncertainties are not discussed at all in the manuscript, even though they are shown in the main results figure 3.
The model performance is rated by using four different likelihood measures and predefined thresholds for behavioural models. When the model performance is discussed, it seems arbitrary which likelihood measure is highlighted. The discussion would be stronger if the model performance evaluation would be more systematic. Further, the influence of DRIPs on DOC concentrations is only visible (in the sampling data) if the gain in streamflow is > 40% ( Figure S1). Otherwise, the lake outflow dominates, which was also stated in the discussion. Thus, events A, B, F, G and H are the most important for the testing of the hypothesis. These events should be discussed in more detail. For example, 'diff_nobio', which does not consider UCAs, often performs better or similar good compared to the UCA-version. Is the UCA-concept really helpful or are the stream concentrations rather driven by the groundwater concentrations and not the inflow volumes? If the UCA-concept is helpful, in which hydrological situations? After a structured analysis and discussion, more specific conclusions should be drawn.

Specific comments
L.1. Title: In my opinion, the title does not correctly represent the content. In the manuscript, the model itself is rather used as a tool to analyse processes. The title, however, suggests that the main focus is on model structure and testing. L.233-234. '…we compared the range of uncertainty in our simulations to the range of uncertainty in the observations.' Where did you do that? I can't find it in the manuscript.
L.234 The number of model runs (100 for each scenario) appears really low, especially for such a simple and fast model (in terms of computational time). Usually, several thousands of runs are performed for uncertainty estimation. How did you check that 100 runs are enough for stable uncertainty bounds? L.236. Why did you choose to use 66% uncertainty bounds? More common are 80% (10 th -90 th quantile) or 90% (5 th -95 th quantile).
L.246ff. I understood that 100 model runs were performed for each scenario. Which one was evaluated with likelihood measures? All? Only the best? Which values are shown in table 2?
L.252-254. The description of R² is -at least -unusual. I would recommend using a more straight-forward description. In this context, the most important thing might be, that R² shows the ability of the model to simulate the dynamics of the measured time series, but not the absolute value. L.364. Change 'that' to 'than' L.364-365. Change '… spatial variability in groundwater …' to ''… spatial variability in groundwater concentrations…' L.402. '…demonstrated that UCA can be useful to identify "reactive" reaches…' Actually, the model results don't show a clear signal towards UCA explaining the variability better. Often, the Diff-version is similar good or even better.
L.419ff. In the conclusions, UCAs are not mentioned at all. Since they play an important role in the whole manuscript, please add your conclusions about this concept. L.425. 'occasionally' is imprecise language. Please, be more specific.
L. 429-431. 'This study …. (Mineau et al., 2016)'. This is not a conclusion of this study. Delete or shift the sentence. Figure S1. I would recommend adding figure S1 to the main text. I think the different hydrological conditions, even though they are provided in the text, are much easier to understand with Figure S1.