This article provides a method for estimating the maximum volume that can be released during a GLOF. 

 I agree that lakes typically do not completely drain, so this method would be helpful to estimate a more realistic flood volume.

-Why do you assume that the crest is at the center of the moraine? 

-How do you justify using a specific slope for calculating the PFV? For example, for Galong Go, considering an angle of 0 degrees, we will have a breach that goes from the crest to the toe, so that we would have a breach of 50 meters (5073-5023m). This would give us, assuming the lake area is constant, a volume of 50x5630000 = 28.150.000 m3. So it seems that a significant factor is the width of the moraine. If we have a narrow moraine, a breach at 3 degrees will be close to 0 degrees, but if you have a wide moraine, as in this example, the breach height decreases. The same holds for all the examples in Table III. So, how does the width of the moraine correlate with the angles in the table from Supplement I: The list of analysed events, measured and calculated characteristics?

- How certain or defensible is conclusion two from this work, considering that your GLOF sample used is small? It seems like a reasonable guess more than a conclusion derived from the results. It is important to note that the range of angles in the supplementary table is considerable.

- It would be good to mention how the volumes were calculated in the previous calculations. For example, Gepan Gath and Luding have similar configurations and PFV. However, Fujita et al.'s (2013) estimate was much closer (85.6%) than Worni et al.'s (2013) (37.8%) to the revised PFV. My guess is that the difference between Fujita et al. (2013) and this method is within the error of the methods.

This communication addresses a critical inefficiency in current GLOF modeling: the tendency to assume a "100% drainage" scenario as the default Worst-Case Scenario (WCS). From a hydrological modeling perspective, I appreciate the authors' attempt to constrain boundary conditions with empirical reality rather than theoretical absolutes. The assumption that deep, overdeepened basins will empty completely is physically counter-intuitive and leads to "inflationary" hazard assessments that may misallocate risk management resources.  This paper serves as a valuable "correction mechanism" for the GLOF hazard community, moving us away from unrealistic catastrophe modeling. It provides a pragmatic, defensible way to reduce the dimensionality of the WCS problem. However, I would recommend the author to consider the following thinking.

One problem with this research is that it lacks the process-based rigor required for high-fidelity hydrodynamic modeling. The reliance on a single scalar parameter (α) derived from a small dataset (N=24) simplifies complex geomorphic feedbacks into a geometric rule of thumb. The authors propose using α to limit breach depth. They report a median slope of 5.0° and a mean of 5.5°. However, for a true WCS, we should not be looking at central tendencies (means/medians), but rather at the tail of the distribution—specifically, the flattest possible channel slope which maximizes breach depth. The paper notes a minimum observed slope of 2.3°. If the goal is to define a safety margin for WCS, the methodology should explicitly standardize the use of this lower bound (or a specific percentile, e.g., 5th percentile), rather than discussing mean values which might lead practitioners to underestimate the potential discharge in outlier events.

A related problem is the "black box" nature of α unsatisfying. The paper argues that incision stops when the channel flattens to this angle, but it does not explain why. Is this angle a function of the internal friction angle of the moraine material? Is it determined by the armoring of the channel bed by coarse lag deposits? Is it a hydraulic control issue where shear stress drops below the critical threshold for sediment entrainment? By ignoring the physical drivers (sediment transport, dam heterogeneity, buried ice), the method assumes stationarity—that future breaches will behave exactly like this specific set of 24 past events. In a changing climate or in different lithological settings, this empirical relationship might break down.

The dataset comprises only 24 events. While I understand the difficulty in obtaining high-quality post-GLOF DEMs, relying on such a small sample size to derive a global parameter is statistically risky. Furthermore, there is a potential survivorship bias. Are we only analyzing dams that partially breached because they were the ones that left distinct channel geometries we can measure? If a dam were to fail catastrophically and wash out completely (eroding below the toe), would it fit this geometric model? The authors note the South Lhonak exception where erosion continued below the toe, which suggests the "toe limit" rule is not absolute.

Another problem is about predictive modeling associate with breach depth. The authors correctly identify that breach depth controls the hydrograph shape. In my experience with hydraulic analysis of megafloods, the sensitivity of downstream routing to the input hydrograph is massive.

While this method provides a better "upper bound" for total volume than the "full drainage" assumption, it does not necessarily help with the rate of release. A 50% volume reduction is significant, but if that volume is released in half the time due to a rapid breach, the peak discharge could still be catastrophic. The paper acknowledges this limitation but does not resolve it.