Reply on RC2

This is an extremely well-written paper which tackles the interesting problem of windgap migration with a rigorous methodology. Given the interest and recent attention to drainage divide migration in the Earth surface processes community, I think this paper will be of great interest and lead to more studies exploring divide migration in valleys. I have some relatively minor comments, but after these are addressed, I think the paper is very suitable for publication in ESurf. I look forward to seeing the final version published!

This is an extremely well-written paper which tackles the interesting problem of windgap migration with a rigorous methodology. Given the interest and recent attention to drainage divide migration in the Earth surface processes community, I think this paper will be of great interest and lead to more studies exploring divide migration in valleys. I have some relatively minor comments, but after these are addressed, I think the paper is very suitable for publication in ESurf. I look forward to seeing the final version published! Thank you. Fig 4 all seem to be from arid regions with large alluvial fans where avulsions can happen frequently. What would happen in vegetated regions where the tributary channels may be more fixed in their original course? Would you end up with a windgap in a stable position relatively close the original capture point, as shown in the simulations with no avulsions? While I think simulating this is beyond the scope of the paper, it would be good to see some discussion of the types of landscapes where the fixed confluences vs. avulsions scenarios might be applicable.

The avulsion of tributaries at windgaps is a really interesting concept for windgap migration. I was wondering at the ability of this process to occur in vegetated landscapes: the examples shown in
We appreciate this point and its broader implication that the climatic conditions could influence the stochasticity of landscape change and, as a consequence, landscape accessibility to energetically more favorable topologies. To acknowledge the potential vegetation/climate effect, we added a section to the discussion: (L286-289) "We also did not attempt to explore the influence of vegetation (and by extension climate), which can have competing effects of stabilizing channel banks and reducing the frequency of avulsions (Tal and Paola, 2010), on the one hand, but obstructing flow, and causing aggradation and avulsions (McCarthyet al., 1992;Jones and Schumm, 1999), on the other hand." We think that a broader discussion should be left outside of the current manuscript, which we try to keep relatively focused.
Following on from this, in agreement with reviewer comment 1, I also think it would help the manuscript to acknowledge in which scenarios wind gaps are likely to migrate and where they may be stationary (either in the introduction or discussion).
Thank you, a criterion for stability is presented in the discussion (L194-207) section that addresses static settings. We now added the following text to acknowledge additional cases that can perturb stable windgap positions: (L289-290) "We also note that processes such as valley damming by landslides or glaciers can cause overflows across windgaps and perturb stable windgap positions".
For the landscape evolution modelling, I think you ran all your scenarios with n = 1? I suggest running a sensitivity analysis to test the variation where n is not equal to 1 and there is therefore a non-linear relationship between erosion rates and slope, similar to your tests on the scaling between erosion rate and drainage area. This might have an impact on windgap migration rates if some tributaries are steeper or if there is migration through a shallower part of the main valley.
Thank you for pointing at this gap. We now include the results of such simulations in an SI. Overall this changes some of the details but the overall pattern remains.
Would the junction angles at tributary junctions influence the rate of windgap migration? It would be interesting to explore whether, if the junction angles in the victim catchment are larger (more perpendicular to the trunk channel), there is less variation in migration rate across a tributary junction. Perhaps for a future paper! Thank you for suggesting this. In the simple 1D perspective presented here the junction angle cannot be explicitly represented. In a more realistic 2D setting, we think that this angle could influence the likelihood of avulsion in the downstream vs upstream valley direction and thus the rate of windgap migration. In natural settings, the junction angle could depend on the relative slope, and potentially, order, of the tributary with respect to the main valley, as well as surface and subsurface hydrology, giving rise to the possibility that additional aspects of network topology control the style of widngap migration. Exploring these issues in detail is beyond the scope of the current manuscript but we now acknowledge its potential role by addition the following sentence (L292-293) "Twodimensional simulations might therefore reveal a more detailed response, which could depend on the 2D valley and confluence geometry."

The results of the modelling seem to show that windgaps like to form a stable position at tributary junctions. Is this borne out by results from real landscapes? The earlier figures in the paper show a lot of detected windgaps across the Himalayas and Appalachians. It seems like it would be relatively simple to detect if these are located at major confluences, which would help to strengthen the conclusions of the modelling by showing a nice correlation with observations.
Thank you for pointing at that. We think that this may be the case if avulsions, or similar processes that perturb stable windgap configurations (e.g., landslide and glacial valley damming) were not having a meaningful effect on windgap migration. Perhaps future work can explore this in conjunction with the influence of vegetation (in light of the prior comment), by comparing windgap locations (i.e., relative to junctions) between vegetated and non-vegetated settings (i.e., a difference may arise if avulsions are meaningfully less frequent in vegetated areas).

I find it odd that the simulations with avulsions shows a steadier migration rate of the divide compared to the ones with no avulsions (e.g. Figure 5c). I would expect that, if you have a sudden increase in the discharge to the aggressor basin and a corresponding increase in erosion rate, you should have an increase in migration rate with each avulsion? Is there an explanation for this steady rate of migration in the simulations with avulsions?
Thank you for pointing at this lack of clarity. We now address this in the discussion (L245-249)." Note that avulsions can effectively reduce or prevent windgap slowdown before large tributaries, reducing the temporal variability in migration velocity. The expression of avulsions in the time-location space of Figure 5c, therefore depends on the frequency of avulsion and the spatial resolution of the simulation." Thank you, to address this comment, as well as a similar comment by reviewer 1, we made substantial changes to the captions of most figures, moved methodological descriptions to the SI, and describe the values of model parameters in a separate table.
Line 37: small typo, should be "loses" rather than "losses" Thank you for this suggestion, we changed the line thickness between channels of different order (to avoid color-related confusion). Thank you for pointing at that. To better show the association of windgaps with preexisting structures and antecedent topography we changed the color-scheme of figure 2a, added zoomed in panels to figure 2 (now 2b, 2d) and changed the point size in all panels. Thank you for pointing at that. We replaced the thin line with a bold, yellow line