Dear Authors,

I have really enjoyed reading your ms, which I think represents a very valuable piece of work for improving the quantitative estimation - and most importantly the knowledege about related uncertanties - of wood fluxes during flood events. In my opinion the text is generally clear enough and figures/Tables quite informative. I just have a few suggestions to improve the work, as listed below:

- in the description of the FGA, I think it would be useful for reader to have the relationship between the max distance and slope steepness for the landslide "connectivity" assessment. At the moment it says just "further away if slope >40% (line 318). But how far upslope? this is an important parameter, I believe.

- for both EGA and FGA please make explicit whether for bank erosion the erodibility of the lateral channel boundaries are considered or not. In other terms, can the models exclude (or greatly limit) bank erosion inputs in case of bedrock banks? If I am not wrong this is not the case, and thus I suspect bank erosion contribution may be overestimated, especially in steep reaches by FGA as the use of the width ratio is related to channel steepness, but  steep channels may be extensively bordered by stable bedrock areas. I think the authors may wish to discuss how the introduction of erodibile vs non-erodible areas for both models may improve the performance of models. Such an approach was introducted by Franceschi et al. (2019)  in combination with the use of width ratios (whose values can be set by the users)

- Unfortunately Franceschi et al (2019) is not published yet, as you know. However, please not that the preprint has a longer list of authors that those reported in the references. Please check it here "https://www.researchgate.net/publication/330397920_GIS-based_approach_to_assess_large_wood_transport_in_mountain_rivers_during_floods". Also note that this model accounts for both landslides and bank erosion inputs, differently from what reported at lines 625. In addition, Franceschi et al model is capable to work with single trees and thus provide statistics about number of elements and their size (For table 1). 

- The authors should comment  about the need/benefit for hazard prediction to modiel - in addition to the wood supply presented here - the propagation of LW in channels during floods (1D or 2D,  including critical sections as bridges where most LW can get trapped), I think that especially for the larger basins the lack of propagation (and thus the longitudiinal disconnectivity) may be partially responsible for the important overestimations showed in this study. 

I hope you will find my comments of use. Best wishes!

Francesco Comiti