Review comments on “Seasonal dynamics and spatial patterns of soil moisture in a loess catchment” by Liu et al.

General comments:

In this study, intensive measurements of soil water content from 5 m profile in a watershed were made during 5 years. Spatial pattern and temporal dynamics of soil water content have been explored. The authors found some interesting results and further highlighted that evapotranspiration is the dominant mechanism of water flow under both wet and dry conditions on the Chinese Loess Plateau. The paper is well written and easy to understand. I think it worth publication in HESS. However, I think authors can further improve the expression of some equations and some other minor issues need to be addressed before it can be accepted.

Specific comments:

Lines 17-18: this is too general and may not be true. Please be more specific regarding the knowledge gap please.

Line 45-49: This was also observed in the study area before at the transect scale for different land uses. In addition, the associated patterns will be depending on what indicator (SD or CV) is used for characterizing spatial variability (https://doi.org/10.1016/j.geoderma.2011.02.008).

Line 56-57: spatial pattern of soil moisture was found to be dominated by topography in a watershed on the Chinese Loess Plateau (https://doi.org/10.1016/j.jhydrol.2013.10.002). I would encourage authors to discuss dynamics and mechanisms of soil moisture by drawing a bit more literature from the study area.

Line 76 and Line 95: please see my comments above. The most relevant literature on the Chinese Loess Plateau should be discussed for yielding knowledge gaps.

Line 134: Authors please comment why 1 mm rather than 2 mm was used for soil texture analysis.

Line 155: too many abbreviations to remember, how about using s, m, d and y, respectively, to represent site, month, depth and year?

 Line 161-169: I would like to consider to embody soil moisture of different sub-regions or the whole area of the watershed in equation (2) by introducing another variable. The same may apply to eq. (3). As it is, the equations are not mathematically robust enough although they are understandable. You need to explain why you are specifically interested in 0-100 cm, does it have anything to do with your finding that soil moisture at 0-100 cm is more temporally dynamic?

Line 179: it is a bit confusing as k refers to soil depth above (e.g., 20 cm, 40 cm, …), but here it represents the rank of the layer from top (e.g., 1, 2, …). Please be sure they are referring to the exact same thing.  Can you use eq. (4) to calculate? In my mind you just need to make k more flexible, it can be a certain depth or a certain layer. The same to Eq (6), can you please use one equation to explain it clearly? To me, the main difference will be whether the whole watershed or just part of the watershed are involved in calculation. So please try to use as small number of equations as possible.

Line186-192: why these needed to be calculated, how does this relate to the main objective you want to target? They need to be better clarified.

Line 198 and Line 203: you need try to find a way to embody slope in the equation as I commented above. Why 𝛿′rather than 𝛿 is used here?

Line 224: why not other topographic properties such as slope?

Line 232: why not try cos(aspect) as it looks like a good indicator for soil moisture (https://doi.org/10.1016/j.jhydrol.2017.05.054)

Figure 6c: title from y axis is missing. Can you please add measurement error for each graph？

Line340-342： this is not that obvious visually. I would suggest authors to improve figure 7 by also making sure they are readable in white and black.

Line 350-356: I don’t think the convex-upward model is necessarily applicable to Chinese Loess Plateau. Even mean soil water content higher than 20% was measured, this pattern was not observed in a previous study (https://doi.org/10.1016/j.geoderma.2011.02.008). Authors may want to discuss a bit more here.

Line 373-374: again, this is not that obvious though this may be true. Please consider to improve the figure 7. Probably rather than showing point value, it would be good idea to mapping the whole area?

Line 405-406: the statistical analysis here can be misleading as less samples from gully than slope.

Line 411-413: I would be cautious to draw this conclusion because gully was much wetter in this study.

Line 426: the role of aspect in driving water variation was also documented in previous studies in the same areas. These papers need to be included in discussion.

General Comments:

This study synthesizes information collected from a dense soil moisture monitoring network in a watershed with deep loess soils. They have highlighted the limitations of using the topographic wetness index for understanding the spatial variability of soil moisture in places that do not have a shallow confining layer, and that the spatial variability in these hillslopes is largely driven by evapotranspiration. The introduction is well written, while the discussion could use some editing to highlight the main findings and the more clearly share the associated story. Below in my specific comments I give suggestions on how to edit and trim the discussion to help focus on the most interesting findings. 

Specific comments:

L23 – confusing since only part of the year is captured in these distinctions

L114 – Gully land consolidation --- I’m not sure if this is necessary here, but I am curious why this is being done and if it is occurring in other watersheds in the area.

L187 – why the 20% trimmed standard deviation?

L238 & 305 – suggest to edit sentences as to not start with “Fig. 2/6”

L290 – figure 4 caption is a little unclear, it seems as though the figure labels should be before the description, e.g. a) depth of max trimmed stdv, b) depth of collapse. How are 90% of sites “collapsing” between 20-100cm and 82% of sites collapsing between 160 and 260?

L282-286 is a longer description of the content in L 316-321 which is easier to read and is associated with a mechanism, consider removing the former.

L298 – What is the color scheme in Fig 5 for?

Fig 6 could be easier to read as a vertically stacked figure so that months are aligned

L341 – figure color descriptions should be in a legend or in caption, not in the text

L 349 – this is slightly misleading as written given that the max soil moisture is below 20% in the hillslopes. You get there in the paragraph, but the set up suggests that there is something different happening here than in the literature.

Figure 7 –an aspect layer as a basemap to help support the findings re: aspect and moisture patterns would be helpful; label the legend and edit the caption to start with "Deviation in VWC from ...

L374 – is this statement supported given the high amount of overlap between boxplots in Fig9? Is it statistically significant? There can be a pattern without being statistically significant, but it should be clarified.

L410-412 – this is the thing that I find most interesting and broadly relevant, but clarify this is referring to spatial variability in the hillslopes (not the hillslope vs gully)

Section 4.3 –consider editing these first two paragraphs to be more integrated, e.g. L422-425 reads like results while L435-437 gets into the context that makes them interesting.

L450 – confusing as written – most of the precipitation fell between driest and wettest months?

L449 – this paragraph isn’t particularly compelling as written, it walks through calculations and results presented in the table. Move the table to results and retain the last sentence in the discussion.