This manuscript reports some estimates of throughfall kinetic energy at the soil surface. The motivation for the report is that very high kinetic energies were measured at a selected location where there was concentrated drip from a tree.

The experiment tests a trivial hypothesis L85 L162. We already know that throughfall energy is greater when and where there is more throughfall and when and where drops are larger. The effect of leaf status is more interesting.

I do not think this sampling effort is sufficient for the data to be reliable. L113 one drip point and 6 non-drip points, measured in 10 rainfall events is a very small sample size, given that spatial variability of throughfall is usually very high. Also, measurements at the single concentrated throughfall point failed in all but three events. It appears (L144) that there was even some pseudoreplication in the ANCOVAs because regression was used to obtain throughfall estimates in four events (L122). Finally, the gap-filled data were all in the leafed period, and the conclusion is that the unit kinetic energy during the gap-filled events was about half or less than the one measured event (Table 1)—this is an example of how poorly constrained the estimates are.

All estimates of the magnitude of energy concentrated at drip points depend on overinterpretation of data. With one, non-randomly-placed, drip-point sampler, we do not know how widespread these points are nor how variable they are, so there is no way to make any estimate of the importance of drip points at any scale.

Minor comments

Fig 2 I think the right-hand panel is a blowup of the left but there are no labels to support this guess. It would be much easier to read this figure if there were labels instead of text to describe the symbols.

L153-154 but the branch height for the concentrated drip point was the same as the others and leaves were not measured by location, so the experiment did not address these questions.

L164-171 the points about terminal velocity do not lead to the conclusion L169.

L181 what is risk exactly, and how can it be lower at the drip point than elsewhere?

Table 1 column headers say “Impact locations” but there was only one.

L207 there are no drop-size distribution data presented. I think the inference is correct but the wording must careful not to imply this research supports the statement directly.

Th English could be improved substantially. Some problems: L46 punctual is not the right word; L50 grazing (although grazing is probably too specific of a word and “feeding” would probably be better); L71 differs; L73 “is different”; L83 occurred with splash cups? L103 sentence makes no sense; L104 weighed dry; L118 “failed” instead of “missed”; “was,” not “were”; L140 ANCOVA does not only examine significant differences; L149 considerably; L159 “The It”; L167 do not reach terminal; L178 than that at; L183 widely?

This study measured the throughfall kinetic energy (TKE) and throughfall precipitation at concentrated and general locations beneath canopies of the Shiiba research forest in Japan, which were compared with freefall kinetic energy (FKE) and freefall precipitation at both leafed and leafless stages, respectively. Thus, the splash erosion caused by droplet impacts could be investigated, accordingly. The authors found that TKE at the concentrated impact location was 15.2 and 49.7 times higher than that at general locations under beech canopies and FKE, respectively. This study confirmed that canopy drip from woody surfaces can be a hotspot of soil erosion in temperate forest ecosystems. The potentially high rates of sediment detachment could be induced by not only throughfall precipitation but also larger throughfall drop size distributions at concentrated impact locations. This topic is of scientific significance, and falls in the research scope of Earth Surface Dynamics. I recommend accepting this study after the below-mentioned revisions have been addressed.

1. Recommend to add a figure in Section 2 to show the concentrated and general locations for measuring throughfall, and the location where the freefall was measured. It benefits a clear introduction of experimental design in this study.
2. Detailed descriptions of the splash cups, such as their diameter, height, etc., are needed, because these cup characteristics affect the quantitative measurements of loss of soil (LOS) and consequent TKE via linear regression.
3. Lines 121–124: There were no introductions on how to get these quantitative relations of freefall precipitation with TKE and throughfall precipitation. If doing regressions based on the measurements in this study, please add the data and analysis. If citing other research, add the references, please.
4. The authors installed seven splash cups to measure TKE, with six cups at general locations and one cup at possible concentrated location. However, throughfall measurements were not clearly described in this study. Is it that throughfall precipitation and TKE were measured at the same location? If so, how to precisely measure TKE by using the splash cup and avoid the disturbance of throughfall precipitation measurements at the same time and locations via installing rain gauges?
5. The authors measured tree traits, such as diameter at breast height, tree height, LAI, leaf area, leaf mass per area, etc. They particularly addressed the effects of structurally designed high energy points on TKE in Section 3.1. However, there were no quantitative descriptions to introduce what is the structurally designed high energy points like, and no quantitative analysis to defend the claim of its effects on TKE.
6. The authors discussed the effects of leaf status (i.e., leafed and leafless) on TKE and consequent splash erosion risks. They conducted these measurements in spring and summer from March 3^rd to April 5^th, and in autumn and winter from August 19^th to October 11^th, respectively. However, in addition to the influence of different leaf statuses, the distinct meteorological conditions also significantly affected throughfall precipitation and TKE. Therefore, the authors might need more evidence to support their claim that leaf status, not the meteorological conditions, dominated the influence on splash erosion risks.

Minor suggestions:

1. Line 85: No need to start a new paragraph to state the hypotheses.
2. Line 159: Delete “The” before “It”.