This manuscript presents the HDM-Plot dataset, a plot-based vegetation dataset across the Hengduan Mountains region, comprising 314 plots sampled between 2022–2024 and covering a wide elevational gradient (754–4,932 m), with detailed species, structural, and environmental information. Overall, the dataset is valuable, timely, and potentially highly useful for studies of mountain biodiversity, vegetation classification, and macroecological analyses. 

However, as a data paper, the manuscript still requires substantial revision to improve transparency in sampling design, clarify dataset scope, and strengthen technical validation and usability.

Major comments:

(1) One critical issue concerns the geographic extent of the dataset.

Based on Figure 1 and other maps throughout the manuscript, the spatial coverage—particularly in the southeastern part—clearly includes a substantial portion of the Yunnan–Guizhou (Yun–Gui) Plateau, rather than being restricted to the Hengduan Mountains sensu stricto. This has important implications for both the description and interpretation of the dataset.

Relatedly, the manuscript describes HDM vegetation as being dominated by subtropical evergreen broadleaf forests (Lines 73–74; light blue region in Fig. 1) and reports a relatively high proportion of tropical areal-types (35.4%; Line 24). These statements appear inconsistent with the widely accepted view that the Hengduan Mountains flora is predominantly temperate in character, within the Sino-Himalayan floristic region. While the HDM is indeed a transitional zone between tropical Southeast Asia and temperate East Asia, its floristic identity is generally shaped by temperate elements at mid–high elevations, with tropical components being comparatively limited and largely confined to lower elevations.

This discrepancy may arise from the inclusion of non-HDM areas (e.g., the Yun–Gui Plateau) in the dataset. I recommend that the authors explicitly clarify the sampling extent and adopt one of the following approaches: (1) restrict the dataset strictly to the Hengduan Mountains by removing sites appearently located on Yun-Gui Plateu & revise all corresponding maps and summaries, or (2) acknowledge that the dataset covers adjacent Plateau, revise the manuscript accordingly, and avoid presenting results as representative of the Hengduan Mountains alone.

In addition, Figure 2 appears to include several plots in the northwestern corner that fall outside the defined study region. The authors seems to include these plots in downstream analyses (and maps), and if so, apply a consistent treatment to all peripheral plots (including those in the southeastern plateau).

(2) The sampling design and representativeness

The manuscript states that plots were placed across representative mountains and valleys, while also considering logistics and accessibility. However, the sampling design remains insufficiently documented for a data paper -- there is no clear stratification scheme (by elevation / vegetation / region), potential accessibility bias is not quantified, and the overall sampling rationale remains unclear (representativeness, coverage of extremes, opportunistic survey). For transparency and future data reuse, the authors should clarify the sampling design rationale and goals. Additionally, the authors could provide a table or supplementary figures to show: (a) numbers of plots per vegetatin types (ideally alongside the relative area of each vegetation type in the study region); (b) numbers of plots per elevational band (e.g., 500-m intervals). These would be helpful information to interpret the data.

(3) Interpretation of ecological patterns

The manuscript reports patterns such as: unimodal richness along elevation and growth-form shifts across gradients. Given the non-random sampling design and uneven spatial distribution of plots, these results should be presented more cautiously. They are best interpreted as descriptive summaries of the dataset, rather than as general ecological conclusions about the Hengduan Mountains region. I recommend explicitly reframing these results to avoid overinterpretation.

(4) Figure elements

Several figures could be substantially improved in terms of clarity and presentation.

For example, multi-panel figures should include panel labels (e.g., a, b, c, d) to clearly separate different components. In Figure 1, the left panel could be (a) horizontal vegetation type distributions, and the right panel could be (b) elevational distribution. For Figure 3/4/5/6, there could be four pannels, (a) for spatial patterns of Families, (b) for genera, (c) for elevational patterns of Families, and (d) for genera. Similar suggestion for Figure 11. And figure captions need to be revised correspondingly.

For some figures, units are missing. Elevation units (“m”) should be explicitly indicated in Figure 1/2/3/4/5/6. Also, units for MAT (°C) and MAP (mm) are missing in Figure 2 and should be added

For Scatterplots by groups, such as elevational patterns of plots in Figure 3/4/5/6/11, though jittering helps to separate points, but many are overlayed due to the large amount. I recommend that these figures could be modified to volin plots to show the density of plots along elevation, which would better represent distribution patterns.

For Figure 7, since piecharts are not ideal (and largely redundant with numeric labels). Stacked barplots would provide clearer and more comparable information

Minor comments:

- Line 9: "investigations": investigation.

- Line 12: "in altitudes of 754–4,932 m": of/spanning altitudes of 754–4,932 m.

- Line 126: "The HDM cover...": The HDM covers ..., or The Hengduan Mountains cover...

- Line 254: “Along the ranges of elevation”: Across the elevational gradient, or Along the elevational range surveyed...

- Line 182: what is community height? How community height and total coverage are visually estimated in the field? Were standardized protocols used across observers to ensure consistency?

- “Abundance” is recorded (Line13), but it is unclear whether this refers to: counts, coverage, or others, across forest, shrubland, and grassland plots? The authors need to clarify this in the main text to explain how data collected behind Table 4.

- Table 3 is very long (nearly four pages) and would be better placed in the Supplementary Material or provided as part of the dataset rather than in the main text.

The manuscript introduces the HDM-Plot dataset, compiled from four field campaigns conducted in 2022–2024 in the Hengduan Mountains of southwestern China, which is a global biodiversity hotspot. The geographic and elevational coverage is impressive, and the data structure appears rich enough to support vegetation classification, conservation planning, and ecological analyses. The inclusion of raw plot data, species lists, importance values, and classification tables is a strong asset. I acknowledge the difficulty to conduct field work in this region and appreciate the authors’ effort in making such dataset publicly available.

However, I still have a few concerns and think that the preprint and the usability of the dataset would be improved by the following comments. 

1.Plot-size heterogeneity needs more clarification.

The authors note that “plot size was determined following community physiognomy and stand heterogeneity,” with forest plots typically 10 m × 10 m to 20 m × 20 m, shrubland plots 2 m × 2 m to 10 m × 10 m, and grassland plots 1 m × 1 m to 2 m × 2 m. That is a very large difference in sampling area among vegetation types. It would therefore be helpful to explain more explicitly how this was considered in the analyses. Please clarify whether any standardization or rarefaction was used for species richness comparisons, and whether plot-size distributions in space, along elevation, or for different vegetation formations could be summarized in the main text or supplementary materials to give a clear idea of the sampling patterns.

2. Strengthen the dataset description by summarizing more of the measured variables beyond species composition.

The assessment of the dataset largely constrained to species diversity in its current form. Since the dataset includes structural and abundance-related information, a few additional summaries would help readers better appreciate its value and scope. For example, patterns of DBH, community height, coverage, or abundance would be informative, especially for forest and grassland plots. It may also be useful to show how species richness varied across disturbance levels so the readers would have a better understanding of the potential limitation of the dataset, given that disturbance intensity was recorded for each plot and many of the plots were sampled along the road.

3. The vegetation classification may benefit from a quantitative classification approach.

In the vegetation classification section, 314 plots are divided into 142 alliance groups, and many groups contain only one plot based on Table 3. It raises the question of whether this level of detail is appropriate and whether all units need to be listed in the main table. The manuscript could also benefit from commenting on whether a quantitative classification approach, such as clustering based on species composition, might serve as a useful complement to the current descriptive framework.

In addition, figures could be improved in regards of clarity.

Fig. 2: Better to only highlight the few mountains that mentioned in the following analyses than numbering all the mountains in this region. Readers may feel it difficult to find where the five mountains used in Fig.8 for showing elevational gradient are. It would also be helpful to supplement the figure with plot density distribution against climatic space, as the deep valleys and mountains prevent the visualization of climatic condition change on the map and the dots representing single plot stacking on each other.

Fig. 3-6: Consider adding density plots or frequency maps to show the distribution patterns more clearly. This would make it easier for readers to assess how frequently each family or genus occurs across the surveyed plots

This manuscript presents a vegetation plot dataset (HDM-Plot) for the Hengduan Mountains, which is potentially valuable given the region’s ecological importance and data scarcity. The dataset is relatively rich in taxonomic and structural attributes and could serve as a useful baseline for vegetation classification and biodiversity studies. However, in its current form, the dataset coverage is too limited for widespread community adoption. And the manuscript suffers from insufficient methodological rigor, unclear sampling representativeness, and overinterpretation of descriptive patterns. I have several comments:

1. A major concern of this manuscript is the relatively limited spatial coverage of the dataset, which substantially restricts its broader applicability. While the dataset may be valuable within the specific study region, its geographic scope is too narrow to support generalizable insights or to serve as a widely applicable resource for the research community. As currently presented, the dataset’s utility appears largely confined to localized or case-specific analyses, limiting its relevance for comparative studies, large-scale synthesis, or integration into existing global or continental databases. This significantly reduces its potential impact as a community resource, which is a key expectation for data papers.

2. The manuscript does not provide sufficient justification or evaluation of the sampling design and its representativeness across the Hengduan Mountains region. Plot locations were selected partly based on logistical feasibility and road accessibility, which raises concerns about potential spatial bias. However, no quantitative assessment is provided regarding how well the sampled plots cover the environmental gradients (e.g., elevation, climate, vegetation types) of the study region. Without an explicit analysis of sampling coverage and potential spatial bias, it is difficult to assess whether the dataset can reliably represent regional vegetation patterns.

3. The dataset includes plots with highly variable sizes across vegetation types (e.g., forests: 10–20 m; shrublands: 2–10 m; grasslands: 1–2 m). However, the manuscript does not address how this heterogeneity affects the comparability of ecological metrics derived from these plots. Key variables such as species richness, importance values (IV), and structural attributes are strongly dependent on plot area. Without appropriate standardization (e.g., rarefaction or size-controlled analyses), comparisons across vegetation types and along environmental gradients may be biased by plot size rather than ecological differences.

4. The vegetation classification framework relies on a set of predefined thresholds (e.g., importance value criteria for dominant species) and expert-based decisions. However, the manuscript does not provide sufficient justification for these thresholds, nor does it evaluate the sensitivity of classification outcomes to these choices. Furthermore, the authors introduce modifications to existing classification schemes (e.g., new shrubland formation types), which may reduce comparability with other datasets and classification systems. The absence of data-driven validation (e.g., clustering or ordination analyses) further limits the reproducibility and robustness of the classification.

5. The manuscript presents several ecological patterns (e.g., unimodal elevational richness patterns, shifts in growth forms) as generalizable findings. However, these interpretations are not sufficiently supported by analyses that control known sources of bias, including uneven sampling effort, spatial clustering, and plot size heterogeneity. Without accounting for these factors, the observed patterns may reflect sampling artifacts rather than true ecological gradients.