The authors have run some clever experiments and have collected a wealth of interesting data over several field campaigns. The math involved is above my paygrade so, hopefully, the other reviewer will be able to look at it more critically. I have uploaded a pdf of the manuscript with detailed comments. Here are my more general comments.

1) The arguments about how transport regime changes over time after the fire seem to hinge primarily on the condition and density of the vegetation; however, there’s no actual data presented on this, just qualitative observations. There are established techniques for making these types of measurements at the ground level and not employing them in this study was a missed opportunity considering how important the condition of the ground is with respect to frictional resistance. Perhaps the authors took a bunch of pictures and could use them to provide quantitative information? Moreover, the descriptions that are given regarding the vegetation aren’t very systematic; it would be helpful to provide a table describing the condition of the vegetation and the surface during each field campaign, segregated by slope aspect.

2) Also, considering the importance that vegetation and aspect presumably have in modulating the post-fire dry ravel response and the recovery to ‘background’ conditions, there was little explanation of why or how. Indeed, the last sentence in the paper emphasizes the importance of vegetation and aspect, but this idea isn’t explored in the manuscript. For the revision, I would recommend diving into this concept a bit more, otherwise, it’s not clear why it matters that you’re comparing north- and south-facing slopes.

3) The study bills itself as an examination of ‘post-fire’ variability in dry ravel, which implies that what is being measured is unique to recently burned hillslopes. However, the only way to know whether that’s true or not would be if you compared your results to north- and south-facing unburned hillslopes. In other words, including control sites would have been important for determining how much of the effect that you saw was due to the fire. Although it’s obviously too late to implement this suggestion, consider including control sites in the future, especially if you’re documenting changes over time. I would recommend addressing the absence of control sites, and its implication, somewhere in the Discussion. At a minimum, it might spur future studies to consider this.

Manny

This work represents an exciting test of previously explored frameworks for quantifying the complex nature of nonlocal sediment transport as presented in Roth et al 2022 and the pair of Furbish papers. The authors take advantage of a disturbance and set up a natural experiment, and they make good use of hard-won field and experimental data. However, I agree with Manny in that there is a disconnect between the focus of the motivation for the study, the data collected, and the discussion/conclusions drawn by the study. This confusion makes it difficult to interpret and contextualize the results and determine if the hypothesized relationships have been tested.

My comments and suggestions, from “big picture” to technical:

• As Manny wrote I find it hard to reconcile the data as presented with the results and then discussion. I found that the introduction, results, and discussion did not have similar focuses and I recommend realigning these sections so that the nature of the data collected and analyzed matches the scope and validity of the discussion content. One recommendation I have for this is a reimagining of the discussion section, which is right now two large paragraphs. Instead, I recommend using the discussion to systematically step through the results, challenges, and implications of each variable tested here – I get (1) particle size, (2) time since disturbance, (3) aspect, and (4) hillslope position/slope (e.g. Figure 5 says to me hillslope position isn’t an important control – is that also your interpretation?). Right now I only see size and time reported but not systematically discussed. I also recommend using the second paragraph of the conclusion section to lead the discussion (and then keeping the conclusion focused on implications for the evolution of the distribution shapes).
• Again to echo Manny it seems as though vegetation plays a key role in these results despite not being presented thoroughly with the data – my understanding is that aspect strongly modulates vegetation, to the point where the text focuses on north- versus south-facing slopes when for the reader it might be clearer to discuss “grassy” versus “wooded” slopes (because insolation differences drive vegetation community?)? In addition to adding any and all observations of vegetation as Manny recommended, I recommend revising the abstract, introduction, and discussion to make the distinction of vegetation rather than simply aspect between the two slope directions.
• Although literature on dry ravel is cited in the introduction, I noted a lack of engagement with the results from previous work in the discussion section beyond a broad allusion to previous work thinking about vegetation– has this study changed or reinforced how we think about these processes? Please add these outward-looking details in the discussion as well.
• Alas I am also no math wiz, but I did my due diligence by (1) skimming Roth et al 2020 and the Furbish papers and (2) downloading and executing the Matlab code. These experiences indicate to me that the Lomax distribution reasonably captures the physical characteristics of the system, but that you’re “A” value is only as good as the number and nature of your input data. Sections 2.2 through 2.4 are rather dense (although I really like Figure 3), and it’s not clear how the method in this work differs from Roth et al. 2020. I recommend citing out what you can and then explaining, with less technical terms, how your real life messy data were fit into the Lomax mold, especially focusing on your method for truncation and censorship (which I would have liked to have been defined) and your sample size, because these seem at least to me where the most slop in your A value will be.

I was interested in running the code because I wanted to experiment with the sensitivity of A to various data shapes and sizes because the authors alluded to the downsides of having (relatively) small sample sizes. Alas, after I installed Matlab and all the required toolboxes I determined that the Lomax optimization step is computationally intensive and would take too long (gentle suggestion to explore parallelization schemes? If each simulation is independent from other simulations perhaps you can run many at once and then compute?). So I am left without an answer to the question: “How many data points do we need for the calculated A to not be useless?,” which I think is important for the authors to answer at least numerically (as I was hoping to do, as I didn’t quite find the figure I wanted from the Roth or Furbish papers). Can you show the readers with a figure (perhaps in the supplement) how the sample size affects A and how uncertain we are about A if we’re only getting 100 samples? My idea was to generate a synthetic distribution with 1000 grains and then simulate randomly sampling 50, 100, 200, 500 of them and then quantifying the spread of subsequent A values would be.

• In general I recommend revising each paragraph to have a strong topic sentence to focus the remaining sentences, and having those topic sentences follow in a coherent order for each section. I also noted many complex phrases and noun chains (e.g. “dry ravel travel distance exceedance probabilities”) that hindered my ability to follow some concepts, so I recommend simplifying some of these.

Some line-by-line comments:

Lines 19-21: These two ideas (“tracked evolution” and “fitting a distribution”) are not connected in a clear way.

Line 26: Abstract should wrap up with a “why we care” sentence or two.

Line 59: The “lower gradient on smoother hillslopes” phrase appears to be a red herring, as is the later reference to the Central California landscape – why is this important? Does Central California have these low gradient slopes to test a hypothesis you seem to be alluding to in the topic sentence of this paragraph?

Lines 85-92: This is a great paragraph and sets the paper up well.

Figure 1: Maybe pop out Row C photos so they are big enough to see the contrasts. Also feels like Panel A should have a topographic cross-section as an inset.

Line 129: Does the Donaldson citation here mean that your observations are consistent with this other work?

Lines 140-51: This section is very connected to the conclusions but is not addressed in results and discussion – does this need to move sections and be placed in context of results? Lines 147-149 in particular read like results.

Lines 330-334: In the Results tell us the number or percentage of 0<A<0.1

Figure 4: Add the epoch/time to the figure titles/labels

Figure 5: Use smaller gap size for dashed lines

Figure 7: Use point symbols at observation locations. Also (and this is a bigger question) can you show the reader a +/- uncertainty on A (for example, some of the relatively small changes in A with days since fire may be within uncertainty? Table A4 implies there’s some spread)