This paper explores a little further data from the same field observations as were discussed previously by some of the same authors in their 2019 HESS paper Temporally dependent effects of rainfall characteristics on inter- and intra-event branch-scale stemflow variability in two xerophytic shrubs (Yuan et al. 2019).

[Yuan, C., Gao, G., Fu, B., He, D., Duan, X., & Wei, X. (2019). Temporally dependent effects of rainfall characteristics on inter- and intra-event branch-scale stemflow variability in two xerophytic shrubs. Hydrol. Earth Syst. Sci., 23(10), 4077-4095. doi:10.5194/hess-23-4077-2019].

            The same two species of shrubs are studied, and the field observations analysed here come from the same 2014-2015 data collection as were analysed by Yuan et al. (2019). The field data collection appears to have been one and the same for both papers.

            Both Yuan et al. (2019) and the present ms. (An et al.) seek to explore the role of rainfall variability and plant architecture on stemflow, throughfall, and interception, paying attention to how these work at intra-event timescales. Their ability to do this is however hampered by their reliance on rainfall observations that were aggregated and logged only every 10 minutes. This is hardly sufficient temporal resolution to permit analysis of time lags before the commencement of stemflow, and various other analyses that the authors seek to make.

            Given that the two papers explore the same shrub taxa in the same field area during the same two years (2014-2015), and that both explore the intra-event workings of stemflow and throughfall, I think that a key requirement is for the Introduction to make it clear and explicit how the present paper differs in scope and results from the earlier paper of Yuan et al. (2019). The earlier paper appears to have focussed more strongly on branch-scale mechanisms, but a clear distinction requires very careful reading and differences in data processing make it very difficult indeed to see what is new in the current ms.

            In particular, in their Introduction (and perhaps also in a covering letter to accompany their submission) the authors should highlight what can be learned about stemflow and throughfall in the two shrub taxa that was not already demonstrated by Yuan et al. (2019). I think that it would be helpful for the authors to compare and contrast what was learned by Yuan et al. (2019) and what similarities or differences emerge in the present study (An et al.).

            In other respects the paper seems entirely routine, containing nothing new in method, theory, or argument. I do think that the authors should evaluate the adequacy f their data and field sampling, however. Do the 38 rainfall events in two years form a sufficiently large sample of events? Does sampling of just four branches (line 202) represent a sufficiently large sample? What is the evidence for this being the case? And only a single individual plant of each species was used to measure throughfall and stemflow dynamics (line 198). How was the single individual selected? Is a sample of one really sufficient to draw meaningful conclusions concerning the entire species, as the authors do? These matters and similar considerations should be discussed and critically evaluated. As the paper stands, the authors merely report that they studied only a single shrub of each species but provide no evaluation of whether this is a sufficient and representative sample. (In the same vein, the use of aggregated, 10-minute rainfall amounts warrants critical comment by the authors. The authors should also report fully and properly the characteristics of the rainfall events that they monitored. To judge from the data in Yuan et al. (2019) these were mostly rather brief - from one to a few hours. But the present ms. (An et al.) does not even mention the event duration (nor, for instance, whether the rain was during daylight hours or at night - which is surely relevant to evaporative losses and hence to interception amounts). All of this must be corrected in a revision to the current ms. Some evidence of the nature of just four rainfall events can be found in Figure 5 but this is hardly sufficient.

David Dunkerley

Faculty of Science

Monash University

Reviewer comments to hess-2022-4

Rainfall partitioning (interception loss, throughfall, and stemflow) is an old theme in the field of forest hydrology, and numerous studies have been done on quantified rainfall partitioning and their influencing factors at event scale. Nevertheless, the relevant studies are lacking for shrubs of drylands, and the intra-event dynamics have been less explored. An et al. characterized and quantified the rainfall partitioning of two xerophytic shrubs at both inter-event and intra-event scale in the Loess Plateau of China. What's really interesting to me is their concurrent finer investigation (10 min) on the intertwined rainfall partitioning processes. It seems that another paper by some of the same authors has been published in HESS (Yuan et al., 2019, 23(10): 4077-4095) in digging into the branch-scale dynamics of stemflow (only one of the rainfall partitioning processes). In my view, this study steps further and provides a full view of the reciprocal dynamics among interception loss, throughfall, and stemflow at the shrub-scale and thereby discussed the underlying mechanisms. In this sense, this study adds some new insights into rainfall partitioning and has the potential for a better understanding of the shrub-dominated eco-hydrological processes in drylands. Of course, the authors should explicitly explain the difference between two papers. Moreover, this paper is in general well-written; the experimental design and data analysis are normal and acceptable; results and discussion are informative. I have some moderate/minor comments that are required before considering the manuscript for publication.

1. L49: Water loss due to interception evaporates but not transpires back to atmosphere. “transpiration” here is NOT a correct term for interception loss.
2. L133: A citation is missed for Flora of China.
3. L271: What does values such as “11.1±8 mm”? Mean±SD or Mean±SE? Better explain for the first time as they appear.
4. L305-311: Are the thresholds for the generation of TF and SF derived from regression equations comparable to or in the range of that measured using tipping bucket rain gage? A brief discussion somewhere in the Discussion section is desirable.
5. L370: significantly → significant
6. L404-414: The authors contributed the significant higher IC % in C. korshinskii than in S. psammophila to the higher water storage of S. psammophila. However, they used the absolute values (4.9 L versus 6.0 L) but not the normalized ones. Surely, a large canopy tends to absorb more rainwater than a small canopy, but that does necessarily mean that the large canopy has a higher capacity in retaining rainwater. Actually, the intercepts (0.92 for C. korshinskii and 1.15 for S. psammophila) in the fitted formulas between interception loss (mm) and rainfall amount (mm) in Figure 4e are indicative that C. korshinskii has a lower canopy water storage, hence a potential lower interception loss.
7. L443-447: I would like to argue with the authors that the dynamic characteristics for TF such as TFI, TFI10, LETF, TFD, LGTF and LMTF are just different variables indicating the behaviors of TF but not the reasons for the generation of TF. Those variables are results but not the reasons. That means, according to those variables, it is reasonable to say which variables are indicative of an earlier or later generation of TF, but they are not the reasons that a shrub species is “beneficial to the generation of TF”. This is also the case for SF.
8. The authors made a detailed description of intra-event rainfall partitioning dynamics. I suggest that they elaborate more on the potential ecological implications.

The present work collected very detailed data to conduct a concurrent and in-depth investigation of throughfall (TF), stemflow (SF), and interception loss (IC) at both inter- and intra-event scales for two typical xerophytic shrubs on the dry region in the Chinese Loess Plateau, and the effects of bio-/abiotic factors were investigated. Previous publications from some of the same authors (Yuan et al., 2019, HESS) and the other researchers (Zhang et al., 2018, Science of The Total Environment; Yang et al., 2019, Journal of Hydrology etc.) only focused on TF or SF in shrubs, and the most of rainfall partitioning investigations are limited at inter-event scales. The intra-event rainfall partitioning dynamics which could help have a better understanding of soil water replenishment and its distribution in soil and the key ecohydrological cycle in arid regions have been rarely explored. As far as I know, this study is the first time to investigate the intra-event variations of all the rainfall partitioning components (i.e. TF, SF and IC) for shrubs. This is the main novelty and a step forward compared with the previous related studies (Yuan et al., 2019, HESS; Yuan., 2017, HESS; Yang et al., 2019, JH). This study obtained new insights to understand the fine characterization of shrub-dominated eco-hydrological processes, and improve the accuracy of water balance estimation in dryland ecosystem. The paper is well written and interesting to the general readers of HESS, and I think it can be published in HESS. I have the following comments to further improve it.

1. The authors should explain explicitly the novelty of this study, especially how it advances from Yuan et al. (2019) and Zhang et al. (2018).
2. Compare your stemflow data with that reported by Yang et al. (2019) in Journal of Hydrology for the same shrub species.
3. The authors selected three representative shrub plants to investigate inter-event rainfall partitioning. Eight TF manual gauges were placed under each korshinskii plant, and for S. psammophila, twenty TF gauges were placed under each plant. For SF yield, a total of 53 branches of C. korshinskii and 98 branches of S. psammophila were used. Compared to the thorough measurements at inter-event scale, the measurements at intra-event scale were somewhat limited (four TBRGs (tipping bucket rain gauges) for intra-event TF, and four TBRGs for intra-event SF). I know it is mainly due to high cost of equipment, and it is difficult to place a lot of TBRGs to measure intra-event rainfall partitioning. The authors should discuss this issue.
4. Some newest references are lost from this paper, such as "Yue et al., 2021, Global patterns and drivers of rainfall partitioning by trees and shrubs, Global Change Biology". The authors should check it.
5. Whether the rain ended at daylight hours or at night? How long after the end of rainfall to collect throughfall in TF manual gauges? The effects of relevant evaporative losses in TF manual gauges should be discussed, as they are open to atmosphere as shown in Fig. 1b.
6. Line 49, transpiration should be evaporation.
7. While describing the intra-event rainfall partitioning dynamics, the authors should elaborate more on its potential ecological significance.
8. The authors did not express clearly whether the 38 rainfall events were all rainfall events in 2014-2015 rainy seasons or those producing throughfall and stemflow.
9. The authors should describe the relationships between intra-event rainfall partitioning variables and meteorological factors such as wind speed and wind direction, even there were no significant relevance.
10. The possible limitations of your study and the future research focus are suggested to be included in the final section the Discussion part.
11. In some references, the authors' first and last name is incomplete. Please revise.