This is a well-conceptualised and thoroughly documented study that adds to our knowledge of the occurrence of loess funnels on the Chinese Loess Plateau. The authors conducted field research using UAVs (photogrammetry and aerial laser scanning) and handheld laser scanning to inventory and quantitatively estimate the occurrence of loess sinkholes in a typical loess catchment. The authors proposed indicators and methods for the morphological quantification of loess sinkholes and compiled a dataset containing 1,194 records at the catchment scale. They analyzed the spatial distribution of loess sinkholes in the catchment area and identified spatial patterns, performed a morphological analysis, and analyzed the fit of morphometric parameters. They also estimated subsurface soil erosion and quantified the impact of various factors on the formation of loess sinkholes. The article thus provides a better understanding of the causes and mechanisms of loess sinkhole formation. The structure of the manuscript is correct, the concept and purpose of the study are clearly defined, and the methodology is correctly selected and well described. The results and discussion introduce new knowledge, and the conclusions accurately reflect the achievements of this study.

# General comments

The paper describes a comprehensive survey of sinkholes in a catchment of the Chinese Loess Plateau using UAV- and handheld LiDAR-based remote sensing. The authors report statistics on sinkhole characteristics and examine spatial statistical relationships between topographic and geomorphological factors on the one hand, and the occurrence of sinkholes on the other. The three-dimensional measurement of sinkhole volumes allows for much more accurate estimates of soil loss than simplified geometric methods. The research objective is clearly defined, the methodology is appropriate, and the results are reproducible. The figures are generally informative and well-designed, aside from minor errors (see below). Results and discussion expand the knowledge of sinkholes in loess deposits, and open sharing of the collected field data allows for scientific reuse. The submitted data is complete and supplied in accessible formats.

# Specific comments

The percentages reported in the "Results" section do not seem to add up. Most are correctly calculated based on a denominator of 1,194, while others appear to refer to a denominator of approximately 809. I have listed a few examples below, but I recommend that all values be carefully checked again.

l 520f: The authors state that the piping developed at a former footslope position, which implies it formed on the elevation of the then-active main channel. Subsequently, the channel incised as a result of uplift, such that the sinkhole is now situated at an elevation of approximately 8 m above the present channel level. I am surprised, however, that the massive landslide on the opposite slope is not considered in this interpretation. As illustrated in Fig. 14a, the toe of the landslide is being undercut, indicating that the river continues to erode its deposits. Progressive reworking of the landslide-derived valley fill, and the associated temporary damming of the channel, could also account for the backflood deposits within the pipe (l. 535) as well as the stratified flood deposits observed on the fluvial terrace (l. 529). This implies a much younger age of the piping and sinkhole formation than the one suggested by valley incission.

l 589–622: The authors describe covincingly the influence of gullies and landslides on sink holes and piping. I suggest adding, that this relationship is not unidirectional, but the results of interconnected processes and overlapping external influence factors (e.g. loess thickness, landscape position, surface and subsurface flow, impermeable base layers, etc.).

l 774: The conclusion section refers to "hazard curves incorporating the time dimension (ie., timing of sinkhole occurrence)". However, this doesn't seem to be subject of the manuscript presented.

# Technical corrections

l 153: Clear object missing in "will make available"

l 193: Figure 2b caption: The 3D view behind the QR Code is very helpful. Consider adding the URL directly to the figure captions. Also consider english translations.

l 193: Figure 2c legend: "Quingshuiying group (Paleogene/EOGENE)" seems to be a typo, please use ICS terminology. Do you mean the 'N'eogene (system) or Eo'c'ene (series)?

l 193: Figure 2d: Lithology "Neocene" -> Neogene

l 213f: Add software version numbers

l 222: Figure 4 says "Artificial extraction of sinkhole polygons". Did you mean "Manual", as stated in l 215?

l 229: Which lidar detection mode (linear, geiger, etc.) does this scanner provide? URL is inconclusive.

l 237: Add version number

l 262: Add version number

l 267: "Slop" -> slope

l 293: delete "some"

l 359ff: The absolute and relative numbers add up to 545+216+28=809 instead of 1194 sinkholes in total. Where's the rest?

l 338: "un" -> in

l 402: cartographic -> spatial

l 410f: The relative numbers don't add up to 1194. 382/1194 is 32% (not 47%); 58/1194 is 5% (not 7%).

l 415: 428/1194 is 36% (not 53%)

l 424: 569/1194 is 48% (not 70%)

l 426: 240/1194 is 20% (not 30%)

l 541: "root systems, AND human activity"

l 729: delete duplicate hyphen

l 825: "Brlmo" -> Bruno