the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 11 Aug 2025
Carbon and nitrogen allocation in montane vegetation: Understanding the impact of environmental change on ecosystem processes
Abstract. Mountain forests, comprising nearly 20 % of the world’s forest cover, are among the most ecologically fragile ecosystems due to their strong topographic and climate sensitivity. Understanding how climate influences these ecosystems requires examining the allocation of key nutrients like carbon (C) and nitrogen (N). We investigated foliar C and N allocation in the Himalayan vegetation by sampling 141 leaf samples from 14 species across three climatically different transects spanning 1,900–5,200 m. The mean total nitrogen (TN) was 6.6 ± 4.5 %, with Juniperus and Abies exhibiting the highest TN (8.3 %) and grasses the lowest (1.6 %). Foliar δ¹⁵N ranged from +2.1 ‰ in Juniperus to +8.9 ‰ in grasses, varying inversely with TN. Total organic carbon (TOC) averaged 37.5 ± 6.2 %, peaking in Juniperus (40.6 %) and lowest in Abies (34.2 %). Foliar δ¹³C in C₃ species clustered near –27 ‰, while C₄ grasses reached –13.4 ‰. In the Central Himalaya, Juniperus maintained high TN and TOC at altitude, whereas in the West, Rhododendron had elevated TN and TOC relative to other genera. Statistical analyses showed that warmer growing seasons strongly reduced TN in Abies and Rhododendron, and wetter conditions increased δ¹³C in Abies and TN in Pinus. Clustering of δ¹³C and TN effectively distinguished mycorrhizal types, with arbuscular‐associated grasses displaying higher δ¹³C and lower TN than ectomycorrhizal and ericoid‐associated taxa. Geostatistical modelling produced a dual‐isotope map revealing “arid hotspots” (high δ¹⁵N and δ¹³C, low TN) in dry western valleys and “humid cold spots” (low δ¹⁵N and δ¹³C, high TN and TOC) in moist eastern slopes. These findings demonstrate that steep precipitation and temperature gradients drive N much more than C in Himalayan plants, underscore the tight coupling of C and N, and identify areas most sensitive to climate‐induced shifts in nutrient allocation.
- Preprint
(1368 KB) - Metadata XML
-
Supplement
(1028 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-3687', Anonymous Referee #1, 25 Sep 2025
-
AC1: 'Reply on RC1', Bibhasvata Dasgupta, 02 Dec 2025
RC1: 'Comment on egusphere-2025-3687', Anonymous Referee #1, 25 Sep 2025
Sampling and spatial extent
Reviewer’s comment:
In this manuscript, total C and N concentrations and C and N stable isotope ratios of leaves from 144 plant individuals across three climatically different elevational transects in the Himalayas are reported and discussed with respect to potential plant responses to changing climate. Moreover, the data was extrapolated to produce countrywide maps of C and N isotope ratios for Nepal with a geostatistical approach.
While I think that the presented results are novel and could finally be a valuable contribution to our knowledge, I have two major concerns regarding the current presentation and discussion of the results.
The authors sampled 14 plant species which are not all listed at the species level in the supplemental information and several names are cut and cannot be fully read. In the main text, they grouped these 14 plant species into the five genera to which they belong and only discussed results at the genus level. This is problematic because there can be a considerable variation in genetic properties among species belonging to the same genus resulting in varying C and N concentrations and isotope ratios. It would be better to differentiate the presentation of the results and the discussion to the species level.
Response: We thank Reviewer #1 for their constructive feedback and thoughtful suggestions. In the revised manuscript, we have clarified and justified our choice to present and discuss the results primarily at the genus level, noting that this approach allows us to capture broader climatic and ecological patterns given the distribution of species along the Himalayan transects. All sampled species and corresponding taxonomy, environmental data, etc, are fully listed in Supplementary File I, which will be published alongside the manuscript and can be investigated further by the community at a species level.
Our study reports patterns at the genus level for two pragmatic and ecological reasons. First, the sampling distribution is unbalanced at the species level: the 141 leaf samples derive from 14 named species that aggregate into five genus/functional groups (Rhododendron, Juniperus, Pinus, Abies, Grass), with genus sample sizes of 46, 39, 30, 10 and 10, respectively. Many individual species therefore have small sample sizes (often <10), which limits statistical power and the ability to fit multivariate climate–trait relationships robustly at the species level. Second, genus aggregation corresponds to clear functional and mycorrhizal groupings that are ecologically meaningful and capture much of the systematic variability in our traits of interest: in variance partitioning, genus membership explains ~69% of the total variance in foliar δ¹³C, and explains a smaller but still informative fraction of variance in TN, TOC and δ¹⁵N (≈15–22%). Aggregating to genus therefore increases analytical robustness while preserving biologically relevant contrasts (functional type, mycorrhizal guild, leaf economics). We emphasise that the full species-level dataset (Supplementary II) is provided and that species/subspecies level analyses are feasible; we propose these as a follow-up that will be able to exploit the published data together with additional samples to achieve adequate within-species replication.
To produce Nepal-wide stable isotope ratio maps of C and N, the authors use multiple regression models of the observed 144 results on growing season precipitation and temperature, soil microbial N concentrations and soil moisture contents (for d15N) or elevation and C3/C4 plant ratio (for d13C). In my view, 144 data points are not enough for geostatistical modelling, particularly because several of these samples likely originate from (almost) the same site. It is unclear from where the data of the explanatory variables come and particularly, at what time the microbial N concentrations and soil water contents, which vary strongly in time, were measured. Are these data really available in a representative form for all grid areas shown in Fig. 3g? Moreover, the model validation is based on the data shown in Figs. 3a and b where I see two dislocated data clouds without internal structure that a linearly linked which appears like a regression through two points. To me, this looks as if the model does not fit the data well. Finally, I am missing an external validation of the model by comparing predicted with measured independent data.
Response: Thank you for this important point. We agree that spatial clustering and sample size are key considerations for geostatistical mapping and model interpretation. To clarify with our dataset: the 141 foliar samples derive from 72 unique locations (mean ≈ 2 samples per location; max = 5; each sample comprising of mature leaves from 3 plants), and sampling is uneven across transects (Central 33, East 44, West 64). Because multiple species were often sampled at the same coordinates (different individuals from the same site), the effective number of independent sampling locations is smaller than the nominal sample count. We therefore treated the modelling and interpretation conservatively, used an 80:20 train–test split (≈113 train / 28 test) to estimate model performance (reported as explained variance / R²), and explicitly caveat limited independence and spatial representativeness in the manuscript.
Regarding explanatory variables:
Growing-season precipitation (GSP) and growing-season temperature (GST) were extracted for each sampling location from high-resolution gridded climate products representing multi-year growing-season means. GSP was obtained from the GPM IMERG Late Precipitation L3 Half-Hourly 0.1° × 0.1° V07 dataset (Huffman et al., 2024; accessed 21.02.2023), and GST from the FLDAS2 Noah-MP GDAS Land Surface Model L4 Central Asia Daily 0.01° × 0.01° dataset (Slinski & Sarmiento, 2023; accessed 21.02.2023). Soil microbial N (SMN) was derived from the Global Soil Microbial Biomass C, N, P dataset (Xu et al., 2014; accessed 27.11.2022), while soil moisture (SM) was obtained from the Smerge-Noah-CCI Root-Zone Soil Moisture 0–40 cm L4 Daily 0.125° × 0.125° V2.0 dataset (Crow & Tobin, 2018; accessed 21.02.2023).
Given that the botanical control (C3 and C4) over the δ13C values could be extrapolated spatially by determining the relative abundance of C3 and C4 plants in the study area, we employed the ISLSCP II C4 Vegetation database hosted by NASA Earthdata (source resolution 1*1 degrees; Still et al., 2009) and downscaled it to the Himalayan region using precipitation and temperature ensemble means (model resolution 0.01*0.01 degrees).
We also recognise the reviewer’s concern about model validation.
Our regression models were trained with an 80:20 random train–test split (≈113 training samples, 28 test samples). Reported model ‘accuracy’ corresponds to explained variance (R²) on the held-out test set. Because our 141 foliar samples originate from 72 unique sampling locations (mean ~2 samples/site) and transect coverage is uneven, the number of independent spatial observations is modest; this reduces effective degrees of freedom and increases uncertainty in spatial extrapolation. We therefore treat the resulting isoscapes as first-order, spatially explicit hypotheses of foliar δ¹³C and δ¹⁵N distribution that require independent validation. Residual diagnostics and test-set performance are provided in Fig. 3 and Supplementary Table I.
I would skip “most ecologically fragile ecosystems”, because this applies to many more near-natural ecosystems. It is sufficient to point at the strong topographic and climate sensitivity.
Response: We revise the sentence as ‘…some of the most ecologically fragile ecosystems..’
15: …the mean total N concentration…
Response: Complied
16: “%” is not an SI unit. Change to g/kg.
Response: TOC and TN are commonly written as both % (Bisutti et al., 2004; Samantaray and Sanyal, 2023) and g/kg (Schumacher, 2002; Walthert et al., 2010)
16: Genus (and species) names are usually shown in italics.
Response: Complied
17: The foliar d15N values…
Response: Agreed
28-29: This is contradictory. Why should anything drive N much more than C if both are tightly coupled?
Response: We have rewritten this as:
These results show that while C and N cycles are tightly linked physiologically, nitrogen allocation responds more strongly and spatially predictably to climatic gradients, making N a sensitive indicator of climate impacts on Himalayan montane vegetation.
33: “sequestration” in the vegetation, the soils, the ecosystems? Specify.
Response: We have rewritten this as:
‘…and C sequestration across vegetation and soil…’
33: “Nitrogen is likewise essential” sounds strange. Nitrogen is frequently the most growth-limiting nutrient. It would be good to know if this is also case in your study area. Moreover, C is not a nutrient.
Response: Thank you — we agree that calling carbon a “nutrient” was misleading. We have revised accordingly.
40: I don’t understand “while respiration drives root allocation”? Plants respire all the time.
Response: Rewritten for clarity
Physiological processes such as photosynthesis and protein synthesis concentrate N in leaves, whereas allocation to roots increases when plants prioritise nutrient uptake, storage or belowground maintenance (Evans & Clarke, 2019).
47: It is unclear what you mean by “nutrient-poor” soils. Soils of mountain ecosystems frequently show high nutrient stocks in partly thick organic layers. Probably, you refer to nutrient availability? Please add a reference.
Response: Rewritten for clarity
Plants growing at very high elevations tend to be constrained by a combination of poorer nutrient uptake, partly caused by poorly developed soils, and lack of water.
49: I don’t think that it is possible to “disrupt” C-N linkages, which would be the end of life.
Response: replaced with affect
53: What do you mean by “enhance or suppress TN”? Please try to use more precise wording.
Response: replaced with ‘influence’.
65: There are no specific objectives and/or testable hypotheses. It would increase the value of your manuscript and also improve the structure if you included objectives/hypotheses.
Response: We appreciate the reviewer’s suggestion and agree that the objectives and hypotheses should be stated more explicitly. Our study does in fact operate with a clearly defined objective—to evaluate how foliar C and N allocation respond to strong climatic gradients across the Himalayas. The underlying hypothesis is that these allocation patterns will not follow a single linear or generalisable trend, but instead exhibit region- and genus-specific variability due to contrasting temperature, precipitation, and mycorrhizal environments. In the revised manuscript, we now articulate these objectives and hypotheses explicitly in the Introduction to improve clarity and structure.
75: Are all sampled plants native to the study region? Please add a list of the species names to the supplemental information.
Response: Historically, yes, as per botanical archives from the study areas and the review from Singh and Singh (2008). The dataset in the supplementary material (S1) has comprehensive genetic information of the samples upto the species level.
80: “in the … laboratory”
Response: Complied
87: And which standards were used for C?
Response: As mentioned in Lines 88-89, several known weights of acetanilide (C = 0.71%) and sulphanilamide (N = 0.039%) were analysed and calibrated for the concentrations of total organic carbon (TOC) and total nitrogen (TN) in unknown samples.
88: Where the calibration lines really (nonlinear) “curves”? The later listed equations are linear functions. Perhaps, “calibration functions”?
Response: Replaced with functions
95: There can be a considerable differentiation in N demands and d13C and d15N values of leaves/needles of different species of the same genus, which needs to be discussed. Furthermore, you obviously didn’t sample Pinus trees in the Eastern Himalayas, which should be mentioned. What was the reason for this?
Response: Thank you — we agree that species-level differences within genera can be important and have now (1) added text to the Discussion that explicitly acknowledges intra-genus variation in N demand and isotopic signatures, and (2) clarified why we report genus-level patterns in this manuscript (limited per-species replication; variance partitioning results that show genus captures the main functional signal; species-level analyses deferred to follow-up work using the published dataset). Regarding Pinus, we confirm that no Pinus samples were collected in the Eastern transect. We have added an explicit note in the Methods and Results to explain this absence and to point readers to the Supplementary Table that lists species presence by transect.
103: Explain abbreviations, when you first use them.
Response: Thank you — we have reviewed the manuscript and expanded the first occurrences of all abbreviations. Examples include TN (total nitrogen), TOC (total organic carbon), GSP (growing-season precipitation), GST (growing-season temperature), SMN (soil microbial N), SM (soil moisture), LAI (leaf area index), and GPP (gross primary productivity). All abbreviations are now defined at first use in the main text.
117-119: Why did you use different correlation approaches (Kendall vs. Pearson)?
Response: Both tests were performed and reported in the SI to account for the parametric and non-parametric distributions, to avoid sampling and/or transect biases. In this section, we describe Kendall tau only when the distribution is non-parametric, i.e., non-Gaussian.
131-135 belong to the Methods section.
Response: Agreed, but since this only applies to the clustering method and goes hand-in-hand with the results, we chose to keep it in section 3.3.
138: Does the “I” in “Supplementary I” stand for Information? I would spell this word out.
Response: That’s roman numeral 1, i.e. Supplementary 1, as there are 2 sections to the supplementary file.
140-142: This sentence is unclear. I don’t understand the part concerning C and doubt that the part concerning N is true. The d15N value of plant tissue is the result of complex processes, not only N2 fixation, which results in a d15N value of 0‰. That TN concentrations reflect the overall N content is trivial. Moreover, I would avoid using the ill-defined “content” (sometimes meaning stock and sometimes concentrations) and replace it by “concentration” throughout the manuscript.
Response: We removed this sentence as it is incomplete and replaced it with:
The δ¹³C–TN pair produced the highest discriminant score among tested pairs, but clustering performance was modest overall (Silhouette ≈ 0.53; ARI ≈ 0.13), indicating only partial separation of fungal groups. Mechanistically, this limited discrimination is plausible: foliar δ¹³C primarily integrates photosynthetic pathway (C₃ vs C₄) and stomatal/biochemical discrimination (i.e. season-integrated water-use efficiency; Farquhar et al. 1989; Dawson et al. 2002), whereas foliar TN is a direct measure of leaf nitrogen concentration (an index of leaf N status). By contrast, foliar δ¹⁵N is an integrative signal affected by multiple sources and processes — the isotopic composition of N sources (atmospheric deposition, soil NH₄⁺ and NO₃⁻, biological N₂ fixation), microbial transformations (mineralization, nitrification, denitrification, volatilization), depth of soil N uptake and mycorrhizal mediation — and thus does not uniquely or solely reflect biological N₂ fixation (Robinson 2001; Craine et al. 2009). For these reasons, the δ¹³C–TN combination can partly separate plant groups that differ in photosynthetic strategy and leaf N status, but δ¹⁵N requires careful, process-level interpretation and should not be used alone as an indicator of specific N acquisition pathways.
8: I suggest to use different symbol types instead or in addition to the different colors. Moreover, you just infer the mycorrhizal association from a statistical approach but did not determine them. How sure can you be that your inference is true?
Response: Symbols and Colours added in Fig. 2. We agree that clustering of foliar chemistry cannot substitute for direct identification of mycorrhizal fungi. Our manuscript already uses literature-based genus (Singh & Singh 2008) as labels for evaluation. We now clarify the limitation in the discussion text (4.3.2) and present the clustering result as exploratory: the δ¹³C–TN pair produces partial separation (Silhouette ≈ 0.53; ARI ≈ 0.13) consistent with guild differences, but the low ARI prevents firm assignment of fungal identity. We propose this as a hypothesis-generating analysis and note that direct fungal identification (e.g., root ITS sequencing or PLFA/NLFA assays) would be required to validate the chemical signatures; we have added this caveat to the Discussion and suggest targeted mycorrhizal assays in future work.
145: “modelling of”
Response: Complied
170: What is the number of data to which you fit the regression model? 144? Are all of these data really statistically independent of each other? Don’t the four and for C even five explanatory variables plus constants result in an overfit? What do you mean with “accuracy”? Is this just the R² value of your fit? This would rather be the explained variance of your data. You should try to predict the isotope ratios of independently measured values at other sites with your model and then report the quality of the prediction.
Response: Thank you — we have revised this clarity. The δ¹⁵N regression was fitted to n = 141 leaf samples using an 80:20 train/test split (ntrain = 113, ntest = 28; Test R2=76%). We also tested the null hypothesis that the observed Adjusted R² could arise by chance using a permutation test (n = 1000 shuffles), yielding p < 0.001. We emphasise two important points in the revised text. First, foliar δ¹⁵N shows strong spatial autocorrelation in the region (Moran’s I ≈ 0.93), so the independence of individual leaf samples is limited. Second, even with an apparently high predictive R² on a random hold-out set, OLS models can overestimate predictive skill when spatial structure is present. To address these issues, we (i) report held-out test metrics above, (ii) tested model predictive performance using spatially-blocked cross-validation (leave-one-transect-out; see Supplementary), and (iii) recommend, for future work, explicit spatial models (e.g., spatial autoregressive models or regression-kriging) and independent validation at additional sites.
174: What does “highest data density” mean? Just the highest number of plant samples of your three transects or of the explanatory variables?
Response: The δ¹⁵N isoscape shows the smallest residuals in the Central Himalaya, where our sampling density (number of leaf samples per unit area) and coverage of key predictors were highest; the Western and Eastern transects exhibit larger residuals, consistent with sparser sampling and/or greater environmental heterogeneity.
199-206: I would omit this paragraph which includes textbook stuff and an announcement.
Response: Removed
209-218: This is pure repetition of results, no discussion although it would be interesting to read your interpretation of these results.
Response: We have revised the text to avoid repeating results and discuss them further.
215: You don’t present analyses on the species level.
Response: Agreed and removed
233-234: There seems to be a general misunderstanding. The C concentration cannot vary freely, because it is the main component of the organic matter that builds up the plant tissue. The C concentration can only decrease (to a small extent) if the concentrations of mineral components including e.g., nutrients, Si or not needed metals increase. For grasses the Si uptake plays a particular role.
Response: We agree that TOC is constrained by leaf mass composition and cannot vary independently in the way a truly unconstrained variable would. To clarify: our TOC values are measured as organic carbon per unit dry mass (i.e. TOC of oven-dry leaf). Variation in TOC therefore reflects changes in the relative proportions of (a) structural organics (lignin, cellulose, hemicellulose, cutin), (b) non-structural organics (soluble sugars, starch, phenolics, resins), and (c) inorganic/ash constituents (minerals, Si, salts). In our Himalayan samples, these pools differ systematically with the environment. For example, grasses at higher elevation and in drier sites often accumulate greater silica and ash (lowering TOC), while many gymnosperms maintain high proportions of lignin and resin (keeping TOC high despite declining leaf N). Likewise, leaf morphological changes (smaller, thicker leaves with higher LMA) shift the balance toward structural polymers and can raise or buffer TOC relative to TN. Because we measured TOC on homogenised dry tissue, the reported variation is consistent with these compositional shifts rather than implying unconstrained carbon variation. We have added text to the manuscript, making this explicit, and suggest direct ash/Si and cell-wall compositional analyses as a useful next step to partition the mechanisms driving TOC variation.
234: I don’t think that you can infer the carbon allocation to specific plant metabolites just from an increasing C concentration. The C/N ratio reflects the nutrient-use efficiency, which is generally higher in conifers than deciduous trees. With increasing elevation, trees are adapted to decreasing nutrient availability and become more efficient (see e.g., Vitousek 1982, Am. Naturalist 119, 553-572).
Response: Text revised
However, increasing foliar C concentration alone cannot unambiguously be attributed to shifts toward specific biochemical compounds such as lignin or resins. Rather, it more broadly reflects changes in C/N ratios and nutrient-use efficiency, which typically increase with elevation as trees adapt to lower nutrient availability (Vitousek, 1982). This pattern is consistent with the generally higher nutrient-use efficiency observed in conifers relative to deciduous species.
238: A low C concentration of 34% can in my view only be explained by a high Si concentration in grasses. Thus, the increasing contribution of grasses with increasing elevation seems to explain the change in C concentrations.
Response: Text revised
This low TOC value at high elevation likely reflects the increasing dominance of grasses with high silica content, which dilutes C concentration in leaf tissue. The shift in functional group composition toward silica‐rich grasses with elevation therefore provides a parsimonious explanation for the observed decrease in mean foliar C concentration.
241: The reason for the decreasing N availability with increasing elevation is mainly increasing waterlogging of the soils which hampers organic matter degradation. This is indeed partly related with climate, but also with drainage conditions and there is a negative feedback of increasingly less decomposable litter.
Response: Text revised
In addition to temperature and precipitation, increasing soil waterlogging at higher elevations restricts aeration and slows organic matter degradation, thereby reducing N mineralisation and availability. This is further reinforced by increasingly recalcitrant litter at high elevations, creating a negative feedback on N cycling that is only partly climate-driven but also reflects drainage constraints.
244-245: Why should dry zones lose more N than wet zones in spite of less N leaching and less denitrification in dry than wet soils? This explanation is not convincing. There are many other drivers of the N isotope ratios, including e.g., deposition from the atmosphere, possible supply from sedimentary parent rocks containing fossil organic matter, or degree of N2 fixation from the atmosphere which need to be considered here. Moreover, there can be a strong species-specific variation in d15N values among different plant species – even at the same site -, which is genetically determined. If the species at your transects are different in spite of belonging to the same genus, your interpretation gets complicated. Therefore, it is very important to mention the species and to consider possibly different species of the same genus in your discussion. It would seem plausible to me that the species are different among the three transects given the climatic differences.
Response: We agree that foliar δ¹⁵N value reflects a composite of multiple processes, including atmospheric N deposition, potential inputs from sedimentary parent material, differences in N₂ fixation, and species-specific physiological traits. Moreover, different species within the same genus may occur across transects, and their genetically determined δ¹⁵N signatures can complicate straightforward interpretations of climate. Such factors likely contribute to the observed inter-transect differences and may partly underlie the elevated δ¹⁵N values in the western dry zones. We have added this in the discussion. Is the following part is part of text…then for consistency you can put it in italics In Western Himalaya's dry sites, elevated foliar δ¹⁵N (+7 to +11‰; Supplementary Table S1, samples 34-97) reflects two mechanisms: (1) ammonia volatilization from warm (>288 K), periodically dry surface soils preferentially removes ¹⁴NH₃, enriching residual NH₄⁺ pools by 5-10‰ (Houlton & Bai, 2009), and (2) nitrification in intermittently wetted soils produces ¹⁵N-depleted NO₃⁻ that rapidly leaches, leaving ¹⁵N-enriched NH₄⁺ as the dominant plant-available form (Mariotti et al., 1981). In contrast, the Eastern Himalayas' continuously moist soils (GSP > 60 mm/m) maintain closed N cycling with minimal gaseous losses, yielding low foliar δ¹⁵N values (-2 to +4‰).
255: Why should nitrification be higher in cold and wet soils at higher elevation? I would even expect the reverse. Nitrification usually only occurs if more ammonium is released than is taken up by the microorganisms and plants, which is highly unlikely at higher elevation because of the slower organic matter mineralization and usually stronger N limitation of the vegetation.
Response: The reviewer is correct that nitrification is generally suppressed in cold, high-elevation soils, and we have removed this implication from the text. The elevated foliar δ¹⁵N values observed at some high-elevation sites in the Western Himalaya (e.g., samples 62–63 at ~4100 m; Supplementary S1) are not attributed to enhanced nitrification but to aridity-driven N-loss pathways. These western sites are unusually warm and dry despite their elevation (GST ≈ 287 K; GSP ≈ 30 mm/month), falling within the rain-shadow zone. Under such conditions, NH₃ volatilisation and gaseous N losses enrich the remaining soil N in ¹⁵N, leading to high plant δ¹⁵N values (+7 to +9‰). In contrast, high-elevation Eastern Himalayan sites (>4500 m, samples 133–141), which are colder and wetter (GSP > 60 mm/month), show much lower δ¹⁵N (~0 to +3‰). Thus, the observed δ¹⁵N pattern reflects regional moisture regimes, not elevation-dependent nitrification.
257-258: This is an interesting aspect but to me it is not clear what you mean, to which transect this applies, and which consequences for the d15N value you infer. I could imagine that the monsoonal rains wash N out of the ecosystems shifting the d15N to higher values – but again likely in a more pronounced way at the more humid transects, while you observe the reverse distribution of d15N values.
Response: Thank you for pointing this out. We have clarified the mechanism and now specify the relevant transect. In our dataset, the elevated foliar δ¹⁵N values at high elevation occur only in the Western Himalaya, not in the more humid Central or Eastern transects. For example, samples from ~4000–4200 m in the Western transect (e.g., IDs 60–63 in Supplementary Table S1) show δ¹⁵N of +2 to +4‰ despite their elevation. These sites lie in the rain-shadow, where growing-season precipitation is extremely low (GSP ≈ 20–40 mm/month), and soils are warm and dry relative to the Eastern Himalaya. Under these conditions, aridity-driven N losses—particularly NH₃ volatilisation and gaseous N loss—enrich the residual soil N pool in ¹⁵N, producing higher foliar δ¹⁵N.
In contrast, high-elevation Eastern Himalayan sites (>4500 m; samples 133–141) are cold and wet (GSP > 60–100 mm/month) and exhibit much lower δ¹⁵N (0 to +1‰). Thus, rather than monsoonal leaching driving δ¹⁵N upward in wet areas, the controlling factor in our dataset is the openness of the N cycle under aridity, not precipitation-driven flushing. We have clarified this explicitly in the revised text.
Text Added: Elevated δ¹⁵N at some high-elevation Western Himalayan sites reflects aridity-driven N losses in the rain-shadow (e.g., NH₃ volatilisation and gaseous N loss), whereas equally high-elevation sites in the colder, wetter Eastern Himalaya show low δ¹⁵N because closed, moisture-rich N cycles suppress ¹⁵N enrichment.
261: As mentioned above, increased rainfall can also decrease soil organic matter decomposition if it results in waterlogging.
Response: In Eastern Himalaya’s monsoon‐fed zones, warming and increased rainfall may accelerate soil organic matter decomposition (Okello et al., 2022), although excessive rainfall can create waterlogged conditions that suppress aerobic decomposition and nitrification, potentially offsetting these effects.
266-267: This is highly speculative. Please add evidence.
Response: Similar transient C sequestration linked to elevated litter C:N has been observed in experimental warming studies in temperate and boreal forests (e.g. Melillo et al., 2011; Craine et al., 2015), supporting this mechanism.
270: Where does this number and all the other response sizes to climate change in the later discussion come from? Is this an output of your model? But see my criticism of the model.
Response: These projected response sizes are derived from simple linear extrapolation of our observed empirical climate–trait relationships (Section 3) and should be interpreted as first‐order approximations rather than dynamic model outputs. The model output is revised with additional sensitivity parameters to test regression accuracy.
274: I think that the abbreviation PNL is not necessary.
Response: We keep it for easier readability in forthcoming mentions
278: Are drought-tolerant grasses not N-limited? The role of grasses for C and N concentrations and stable isotope ratios is not sufficiently explored in your work.
Response: These grasses are themselves typically N‐limited, but their conservative nutrient strategies and greater tolerance to low soil N allow them to persist under drier conditions. Our dataset does not fully resolve how grass dominance would affect overall ecosystem N cycling, and this represents an area for further study.
280: By which process will the stored C be released? Erosion? But then it possibly ends up in a long-term sedimentary C storage, e.g., as lake sediment and does not really present a C loss.
Response: As the treeline retreats, the forest canopy recedes, exposing the soil previously protected by the trees. The exposed soils experience less shade and more direct sunlight and heat, leading to increased decomposition of organic matter by microbes and soil fauna. This enhanced decomposition releases the carbon stored in the soil, primarily as carbon dioxide, into the atmosphere.
282: Consider deposition earlier, particularly with respect to the interpretation of the N isotope ratios.
Response: Thank you for the suggestion
284: CO2 fertilization is only possible if enough nutrients, particularly N and P are available.
Response: Modified accordingly
294: Add a reference.
Response: Added, Lillis et al. (2004).
294-296: This sentence belongs to the introduction.
Response: Thank you for the suggestion
297: Plants can take up nutrients with all their surfaces, e.g., also directly from deposition.
Response: Yes, but N uptake is primarily from roots.
298: “other” nutrients, because N is also a nutrient
Response: Agreed
301: It would improve the structure of your discussion if you started each paragraph by mentioning a result with reference to a table or figure and then directly discussed this result.
Response: Complied
308-309: Again, the role of deposition for the N isotope ratios needs to be discussed earlier.
Response: Thank you for the suggestion, we have incorporated this.
309-310: N-limited plants take up all mineral N that they can get. A preference for ammonium or nitrate is not plausible.
Response: Plants in snow-covered Eastern Himalayan sites (samples 98-141) exhibit low foliar δ¹⁵N (-2 to +2‰), consistent with N uptake from soil pools with minimal fractionating losses. Snow-derived NO₃⁻ (depleted in ¹⁵N through photolytic processes; Morin et al., 2009) may contribute to this signal. In contrast, Western Himalayan plants exhibit elevated δ¹⁵N, reflecting the uptake of N from fractionated soil N pools in warm, dry conditions, where gaseous losses enrich the residual N.
310-311: This does not seem likely to me. Usually, deposited ammonium results from the volatilization of ammonia e.g., from animal production, and is therefore N-isotopically light. The same is true for nitrate, which frequently results from NOx emission.
Response: See #51
313: As far as I know, there is hardly any N isotope fractionation by plants during N uptake. What do you mean by “N assimilation”?
Response: There is generally little isotope fractionation during root N uptake itself. By “N assimilation,” we refer to the incorporation of inorganic N (NO₃⁻ or NH₄⁺) into organic compounds within plant tissues, during which some fractionation can occur through enzymatic steps. Additional fractionation may arise from internal N cycling, including translocation, exudation, or volatilisation as ammonia or N₂O (Cernusak et al., 2009).The δ¹⁵N value has been used to estimate the relative contribution of N₂ fixation versus soil N sources (Shearer & Kohl, 1988; Unkovich, 2013), but data remain scarce for plants relying solely on atmospheric N₂.
316: “show” instead of “measure”
Response: Agreed
319: This statement seems trivial and has already been mentioned earlier.
Response: Removed
321-323: Plants do not only compete with each other for nutrients but particularly with microorganisms. If there is a high N availability, competition decreases and N limitation is reduced (or even replaced by another nutrient such as P). I don’t understand what you mean by “the allocation of N determines the N concentrations”. This seems trivial.
Response: Yes, with microorganisms indeed, although it wasn’t explicitly mentioned. N allocation strategies can be gleaned by measuring TN percentages.
331-337 belong to the introduction.
Response: Removed
344-345: Lower d15N values than what? I am not aware that plants can take up organic matter or microbial biomass for its nutrient supply without previous mineralization.
Response: Lower in Arbuscular mycorrhiza-associated grasses than ectomycorrhizal and ericoid-associated genera.
348: Higher C/N than what? Why should the C/N ratio be higher if the mycorrhiza provides more N than is available to non-mycorrhizal plants?
Response: The comparison is relative within the Himalayan dataset, not an absolute global statement. Higher C/N refers to values observed for arbuscular and ectomycorrhizal genera compared to non-mycorrhizal taxa in our data. This likely reflects differences in tissue structure and/or allocation strategies rather than greater N limitation.
349: From what do you infer this preference?
Response: The inference is based on the clustering patterns of δ¹³C and TN values. Non-mycorrhizal taxa consistently group with lower TN and δ¹³C values, indicating less access to organic N pools and a stronger reliance on inorganic soil N relative to mycorrhizal groups.
350-351: But mycorrhiza also respond to climate change.
Response: We agree. Mycorrhizal associations are sensitive to temperature, moisture, and nutrient availability, and their distributions and functioning are expected to shift under climate change. We have added a clarifying statement acknowledging this dynamic response in the revised text.
A final important caveat concerns our mycorrhizal interpretation. Mycorrhizal affiliation in this study is assigned from host-genus literature records rather than measured fungal communities; clustering of foliar δ¹³C and TN reveals only partial separation among the assumed guilds. Hence, while foliar chemistry appears to carry a guild-related signal, we emphasise that these results are hypothesis-generating. Confirming that the chemical clusters correspond to the actual mycorrhizal community requires root- or soil-level fungal identification (e.g., ITS metabarcoding, PLFA/NLFA, or microscopy), which we recommend as a priority for future work.
361: To me, the only explanation for low C concentrations (and simultaneously low N concentrations) is the contribution of grasses containing Si.
Response: The reviewer raises an important point about Si dilution in grasses and its potential effect on measured TOC. We agree that silica (phytolith) accumulation can lower measured organic C concentration on a mass‐percent basis and therefore may contribute to the relatively low TOC observed in some grass samples. In our data, grasses display systematically lower TOC than many woody genera (grass TOC commonly ≈ 30–37% versus conifer TOC commonly ≈ 38–41%), although the distributions overlap. Published work shows that montane grasses can contain appreciable biogenic silica (often several percent of dry mass and in some cases 5–10%; Hodson et al., 2005), and previous phytolith studies in Himalayan contexts (Roy et al., 2019; Basu et al., 2015) further indicate that silica can be a non-negligible component of grass tissues in this region. Because we did not measure ash or directly quantify tissue Si in the present campaign, we cannot partition the TOC differences between true organic-C variation (e.g. lower lignin/cellulose) and simple dilution by silica. We have therefore (i) revised the manuscript to acknowledge this uncertainty explicitly, (ii) removed any definitive statements attributing low grass TOC solely to organic-C physiology, and (iii) recommended that future sampling include ash and Si (or elemental) analyses so that the relative contributions of Si dilution versus biochemical composition can be resolved.
Revised text: Because biogenic silica can lower organic-C percentages in grass tissue, and because we did not measure tissue ash or Si here, some of the lower TOC observed in grasses may reflect silica dilution rather than (or in addition to) genuine reductions in lignin/cellulose content; we therefore recommend that future work include ash/Si assays to partition these effects.
365-366: Already said.
Response: Complied
372: Does this mean that there is less monsoonal rainfall at mid- than high-elevation sites? Can you support this with data? But again, rainfall is not the only driver of the complex N cycle.
Response: We thank the reviewer for highlighting this point. In the central Nepal Himalaya, a well-documented two-step orographic structure produces a mid-elevation precipitation minimum between the lower foothills and the higher, windward alpine ridge (Bookhagen & Burbank, 2006). Our climate data extracted for the sampling coordinates show this pattern as well: several mid-elevation sites (∼3000–3800 m) receive substantially less growing-season precipitation than both lower and higher elevation sites along the same transects. This climatic “dry band” can contribute to relatively open N cycling and thus lower δ¹⁵N at mid-elevations. However, as the reviewer correctly notes, rainfall alone does not dictate N-cycle fractionation, and we have revised the text to emphasise that the observed δ¹⁵N pattern reflects the combined effects of moisture availability, soil N processes, vegetation turnover, and snowmelt-related N pulses at high elevation.
376-379: Again, the concentration of C cannot vary independently of the mineral concentration (i.e., nutrients and Si).
Response: See #63
380: I doubt that these gradients are unique to the Himalayas. They are common for all high-elevation mountains.
Response: We have now explored Si content in foliar biomass.
We note an important compositional constraint: TOC is reported here as organic carbon per unit oven-dry leaf mass and therefore varies only as the relative proportions of structural organics, non-structural organics and inorganic (ash) fractions change. Critically, TOC cannot vary independently: it is constrained by the balance between organic constituents (lignin, cellulose, lipids) and inorganic components (minerals, silica), meaning TOC patterns reflect compositional shifts rather than unconstrained carbon allocation. Observed spatial patterns in TOC are thus plausibly driven by compositional shifts: grasses and herbaceous species often show changes in silica and ash content with environment (reducing TOC), while woody evergreens allocate more to lignin, resins and other recalcitrant polymers that maintain high TOC even when TN falls. Leaf morphological adjustments (smaller, thicker leaves; higher LMA) also alter dry-mass partitioning toward structural compounds and can modulate TOC independently of short-term shifts in N. Our TOC measurements (dry, homogenised tissue) therefore reflect these compositional and structural changes rather than unconstrained carbon variation; targeted assays of ash/Si content and cell-wall composition would help partition the relative roles of inorganic uptake versus shifts in structural versus soluble organic carbon.
391: Again, why do wet conditions lower d15N values? Wet conditions would rather result in more N losses and thus increase d15N values.
Response: Correction.
Eastern monsoon zones, currently modest C sinks with closed N cycles, could enhance carbon uptake if cooler, wetter conditions elevate SMN and lower δ¹⁵N. Foliar TN may rise by ≈ 0.2 % per 0.5 °C warming, boosting C assimilation (δ¹³C slightly more positive). Eastern forests might increase biomass by 10–15 % over two decades.
395-396: This is a repetition. You should shorten your discussion and render your text more concise.
Response: Agreed
397-407: I don’t understand where these change estimates come from and suspect that they are highly speculative. Do you just vary temperature and precipitation in your regression function according to the climate change scenarios? This would be a rather simplified approach. Perhaps, you indicate that you roughly estimate these values assuming that only temperature and precipitation change, assuming that your regression equations can be generalized and ignoring the error of your models.
Response: Yes, these are the results from varying the parameters in the δ¹³C and δ¹⁵N functions. We do acknowledge that these are simpler models, but they make spatial modelling computationally inexpensive, which is one of the outputs of this study, i.e. the dual isoscape. In the revised text mention this caveat.
408-409: You cannot infer “drastic” changes just from a correlation. The size of the change depends on the slope of the regression line (or the first deviation of a non-linear relationship).
Response: Agreed and omitted.
422: Low microbial N stocks usually indicate N scarcity and result in closed N cycling and thus low d15N values, i.e., the reverse of your interpretation.
Response: Correction
In the lower valleys of the Western Himalayas (~2,000-3,200 m), low soil microbial N stocks (mean SMN = 1.2 g N/m²; Supplementary S2) indicate N scarcity. However, foliar δ¹⁵N is elevated (+7 to +9‰) because the limited N that is mineralised undergoes strong fractionating losses via volatilisation and episodic leaching in warm, intermittently dry soils, enriching the residual, plant-available pool. This represents open N cycling despite low absolute N availability (high δ¹⁵N, low TN). Conversely, Eastern Himalaya sites have higher SMN (mean = 8.5 gN/m²) and low δ¹⁵N (+0 to +2‰), characteristic of closed N cycling where microbial N is efficiently retained and recycled with minimal losses
436-438: I cannot follow.
Response: ‘..mixed stands..’? This refers to the contiguous growth of all sampled species in these sites.
441: Again, increased N availability increases losses driving N to higher not lower d15N values and higher soil moisture will increase the opening times of the stomata and thus decrease the d13C value.
Response: Increased N availability often leads to enhanced N turnover and potential losses (via leaching or denitrification), which can elevate δ¹⁵N values rather than decrease them. Similarly, higher soil moisture promotes longer stomatal opening, which reduces water‐use efficiency and results in more negative δ¹³C values. We have revised the manuscript to reflect this, stating that modest warming combined with increased precipitation may slightly raise δ¹⁵N and lower δ¹³C, consistent with enhanced N cycling and wetter soil conditions.
446: You cannot infer C assimilation from C concentrations. Instead, you need to consider the biomass production.
Response: We agree that foliar C concentrations alone do not directly quantify carbon assimilation or productivity. Our original wording, implying that TOC reflects assimilation efficiency, was misleading. We have clarified that TOC in our study is a measure of structural carbon allocation within tissues, not net C assimilation, and that true C assimilation should be assessed via biomass production, growth rates, or photosynthetic measurements. The revised text now distinguishes tissue carbon content from ecosystem‐level carbon fluxes.
454-455: I don’t think that your dual isotope map reflects C and N allocation. It perhaps reflects soil moisture conditions and N availability although this would still be a simplification of the complex C and N cycles. But see my second major concern.
Response: We acknowledge that describing the dual-isotope map as a direct reflection of “C and N allocation” overstates its mechanistic precision. The map primarily captures the combined effects of soil moisture, N availability, and plant functional type on foliar δ¹³C and δ¹⁵N, serving as a proxy for ecosystem status rather than a literal depiction of C and N allocation. We have revised the manuscript to clarify that the dual isoscape represents environmental controls influencing foliar isotopic composition, which can provide insights into potential carbon and nitrogen dynamics, but does not fully resolve the complexity of C and N cycles.
457-459: Your findings concerning the d15N values contradict current knowledge and for d13C values is well known.
Response: Rightly so, as extensively discussed when comparing with global and regional studies, making the Himalayas a unique natural climatic gradient to test these hypotheses.
460: What do you mean by “decades of pointwise foliar chemistry”?
Response: Typo. Correct text:
Our dual isoscape synthesises foliar δ¹³C and δ¹⁵N measurements collected across multiple sites during the dry season, combined with decadal averages of environmental variables from models and satellite data, providing a spatially explicit representation of Himalayan ecosystem responses to environmental gradients.
464-467: This is well known.
Response: Thank you for the suggestion
468-469: It could be promising to evaluate the monsoonal influence on the C and N concentrations and isotope ratios more in depth by specifically showing relationships between monsoon-influenced climate characteristics and your measured leaf properties, which I currently don’t see in your paper.
Response: True, but our sampling campaigns are from the dry seasons (due to technical difficulties of access during the monsoons in high elevation sites). And without the monsoonal ‘greening’ of vegetation sampled in our dataset, interpretations could be incorrect, obstensibly from incomplete representation
471-472: Shifting phenology and productivity is expected for all mountain areas. To me, it did not become clear, why increasing dryness should result in C losses.
Response: Increased dryness can reduce ecosystem carbon storage by lowering plant productivity, increasing mortality of moisture-sensitive vegetation, and accelerating soil organic matter oxidation, all of which contribute to net carbon losses in montane environments.
473: You did not sufficiently address the complexity of the N cycle. As already said, you need to consider e.g., N deposition, N2 fixation, and possible rock-derived N supply. Your explanations of the d15N values are not plausible.
Response: We acknowledge that our initial discussion did not fully reflect the complexity of the N cycle. In the revised manuscript, we now explicitly incorporate the roles of atmospheric N deposition, symbiotic and asymbiotic N₂ fixation, and potential rock-derived N inputs, all of which can influence foliar δ¹⁵N. We have also revised the interpretation of δ¹⁵N patterns to avoid overly simplified mechanistic claims and to reflect the multiple, interacting processes that shape plant δ¹⁵N in Himalayan ecosystems.
483: The explanation of the variations in C concentrations need to be rethought, because C concentrations cannot vary independently of mineral concentrations.
Response: Agreed, but mineral controls on C, e.g. Si, are not explored in this study.
With these revisions, the manuscript will be clearer in structure, hypothesis testing, and presentation, addressing all of the reviewers’ concerns.
Citation: https://doi.org/10.5194/egusphere-2025-3687-AC1
-
AC1: 'Reply on RC1', Bibhasvata Dasgupta, 02 Dec 2025
-
RC2: 'Comment on egusphere-2025-3687', Anonymous Referee #2, 03 Nov 2025
The paper presents an interesting dataset of leaf d13C and d15N from the Himalayas and an interesting west-to-east climatic gradient. The sampling effort is appreciable, and the dataset is interesting. However, the paper lacks clarity in the objective, framework, and discussion. The manuscript addresses multiple research questions, but does not test them well and only vaguely presents and discusses the results. It is not clear whether the main objective is to understand the effects of the climatic gradient on d13C and d15N, the effects of mycorrhizal associations, the effects of climate change on the isotopes, or species effects. This also affects the methods and data analysis. It is not clear what statistical tests were used, why they were used, or how they were performed. The authors need to state in the introduction the main research questions and the hypotheses. In the methods, they should explain exactly how each main question was tested, not only how the parameters were measured. In the discussion, they should cite the results with statistics.
The last paragraph of the introduction, "examining how C and N uptake... Investigating these patterns... and in the title "C and N allocation", the authors do not measure the C and N update by plants nor the allocation. The manuscripts lack clarity on what parameters are and how they interpret them.
The discussion does not cite the figure or table where a result is shown, and it vaguely connects different phenomena. For example, Lines 250 to 260 mention correlation coefficients of 0.1 to 0.25. Are these significant? This is vague. All of this makes the manuscript hard to read and follow.
The data analysis is very difficult to follow. The authors should add a data analysis section in the methods that explains which hypotheses were tested, how they were tested, and the exact statistical details.
Figure and table captions are not clear. For example, Figure 2, Table 1, and the supplementary figures do not explain the content well.
How many replicates per species and per plant were used, and were they included in the models?
How was the climate data obtained and which datasets were used? How did the authors match the time stamps between climate and plant sampling? Plant sampling was done between 2019 and 2021. Was the time difference considered? What is the spatial and temporal resolution of the climate data?
It is not clear why the isoscapes were modeled. The correlations show the effects of temperature and precipitation on the isotopes, and then these effects were extrapolated and discussed again against climate variables. This seems circular. The isoscape modeling should include errors from the climate data and the isotope measurements, and these errors should be propagated. Error maps should be shown. Why did the model not include species effects, given the clear species differences in Figure 1?
The correlation coefficients are not clear. The authors should present scatter plots and regression lines between predictor and response variables.
The motivation, background, and hypotheses on mycorrhizal associations are missing from the introduction.
What was the sampling strategy for plant materials? Did the authors check for sampling bias? Why are isotope data missing for Abies and Rhododendron, and did the authors test for bias in the results due to these gaps?
Other comments.
I do not recommend using the term allocation in the title. The study examines variability in d13C and d15N and does not measure carbon allocation.
Line 107 ff. Add a figure showing the abundance results.
Line 12 ff. Add figures showing scatter plots and regression lines.
Discussion section. Every scientific statement should have a correct citation. The authors should cite the figure or table when they discuss their own results and cite the literature when they discuss findings from other studies. For example, lines 215, 230, 236, and 251 do not clearly indicate which results are being referenced.
In figures, use proper subscripts/superscripts or formulas, state the units properly e.g in ‰ vs VPDB.
Citation: https://doi.org/10.5194/egusphere-2025-3687-RC2 -
AC2: 'Reply on RC2', Bibhasvata Dasgupta, 02 Dec 2025
RC2: 'Comment on egusphere-2025-3687', Anonymous Referee #2, 03 Nov 2025
The paper presents an interesting dataset of leaf d13C and d15N from the Himalayas and an interesting west-to-east climatic gradient. The sampling effort is appreciable, and the dataset is interesting. However, the paper lacks clarity in the objective, framework, and discussion. The manuscript addresses multiple research questions, but does not test them well and only vaguely presents and discusses the results. It is not clear whether the main objective is to understand the effects of the climatic gradient on d13C and d15N, the effects of mycorrhizal associations, the effects of climate change on the isotopes, or species effects. This also affects the methods and data analysis. It is not clear what statistical tests were used, why they were used, or how they were performed. The authors need to state in the introduction the main research questions and the hypotheses. In the methods, they should explain exactly how each main question was tested, not only how the parameters were measured. In the discussion, they should cite the results with statistics.
Response: We thank Reviewer 2 for constructive and thorough comments. They helped us tighten the manuscript’s objectives, statistical framework and presentation. In response we (1) added a concise set of research questions and three testable hypotheses to the Introduction; (2) inserted statistical results for Supplementary Table; (3) clarified sampling and replication, climate data sources and temporal matching; (4) expanded the Results to report statistics with figures/tables for every key claim; (5) added sensitivity analyses for spatial non-independence and missing data; and (6) produced additional diagnostics and uncertainty maps for the isoscape predictions. Key new material is indicated below by manuscript location.
- Clarity of objective, framework and hypotheses
Reviewer comment: objectives/hypotheses unclear; which question is primary; tests not specified.
Response: We agree and revised the Introduction and Methods accordingly.What we changed
Introduction: we now state a single core objective and three explicit, testable hypotheses (end of Introduction).
Despite global interest in montane C–N dynamics, high‐elevation data from the Himalayas—one of Earth’s steepest climatic gradients—remain sparse. By sampling 141 individuals across 14 species (gymnosperms, angiosperms, and grasses) from 1,900 to 5,200 m, we assemble one of the most comprehensive Himalayan altitudinal transects to date. The west‐to‐east climatology—from dry subtropical foothills to cold, moist alpine plateaus—makes the Himalayas an ideal “natural laboratory” for examining how foliar C and N uptake respond to synchronous changes in temperature, precipitation, and mycorrhizal associations. Our objective is to quantify climatic and ecological controls on foliar C, N, δ¹³C and δ¹⁵N across the Himalayan gradient, and we test whether climatology explains more variance in N metrics than in C metrics, mycorrhizal/functional groups differ in C–N–isotope traits, and how these controls can be used to generate predictive isoscapes.Methods
GSP and GST were extracted for each sampling location from high-resolution gridded climate products representing multi-year growing-season means. GSP was obtained from the GPM IMERG Late Precipitation L3 Half-Hourly 0.1° × 0.1° V07 dataset (Huffman et al., 2024; accessed 21.02.2023), and GST from the FLDAS2 Noah-MP GDAS Land Surface Model L4 Central Asia Daily 0.01° × 0.01° dataset (Slinski & Sarmiento, 2023; accessed 21.02.2023). Soil microbial N (SMN) was derived from the Global Soil Microbial Biomass C, N, P dataset (Xu et al., 2014; accessed 27.11.2022), while soil moisture (SM) was obtained from the Smerge-Noah-CCI Root-Zone Soil Moisture 0–40 cm L4 Daily 0.125° × 0.125° V2.0 dataset (Crow & Tobin, 2018; accessed 21.02.2023). Climate controls were evaluated using Pearson/Kendall correlations, linear regression, partial R², and Moran’s I for spatial autocorrelation (Supplementary I). Mycorrhizal group differences were tested using Kruskal–Wallis with Dunn post-hoc tests and K-means clustering assessed by Silhouette and Adjusted Rand Index. Isoscape models used multiple regression for δ¹⁵N and Ridge regression for δ¹³C, with predictive accuracy reported from test-set performance. All analyses were repeated on site-aggregated data to assess sensitivity to spatial clustering.
- Terminology: “allocation” / measurement of uptake
Reviewer comment: Title/intro suggests allocation/uptake were measured but only concentrations and isotopes were measured.
Response: We corrected the wording throughout to avoid implying direct flux measurements. We changed wording in title/abstract/introduction to clarify that we measure foliar concentrations and isotopic signatures (TOC, TN, δ¹³C, δ¹⁵N) that integrate acquisition, assimilation and allocation outcomes rather than instantaneous uptake rates. The revised text and an explicit definitional paragraph are in Methods 2.1 and Introduction.- Statistical details, figures, and explicit citation of results
Reviewer comment: Tests and statistics are not described or cited; some correlations unclear/significant?
Response: We added test statistics and exact p-values to every major result; every claim in the Discussion now cites the supporting figure or table. Specific actions:All correlations and regressions now include r/τ, R² and p-values in the Results text and figure captions (see Results 3.1–3.4 and Supplementary Table S2).
Correlation with heatmaps for the principal climate–response pairs are provided in Supplementary Fig. S4 (GSP/GST vs TN, δ¹⁵N; elevation vs δ¹³C; etc.).
Figure S4: Correlation matrices showing relationships between environmental drivers and foliar parameters. (A) Correlations with carbon-related parameters (δ¹³C and TOC). δ¹³C shows moderate positive correlations with elevation (r = 0.44), precipitation (r = 0.44), and soil microbial nitrogen (r = 0.28), but weak relationships with surface temperature (r = -0.59) and soil moisture (r = -0.43). TOC shows minimal correlation with environmental variables (|r| < 0.25 for all drivers), indicating carbon concentrations are largely independent of direct climate effects. (B) Correlations with nitrogen-related parameters (δ¹⁵N and TN). Both parameters show strong opposite relationships with climate: TN positively correlates with precipitation (r = 0.85) and negatively with surface temperature (r = -0.77), soil moisture (r = -0.67), and soil temperature (r = -0.66). δ¹⁵N exhibits inverse patterns (r = -0.81 with precipitation, r = 0.67 with surface temperature), reflecting open vs. closed N cycling. The strong negative correlation between TN and δ¹⁵N (r = -0.80) confirms the tight coupling of N concentration and isotopic signature. The colour scale represents Pearson correlation coefficients, ranging from -1.00 (blue, strong negative) to +1.00 (red, strong positive). All correlations with |r| > 0.30 are significant at p < 0.01 (n = 141).
Example now in manuscript: “GSP explains 67% of TN and δ¹⁵N variance (Pearson r = 0.82, p < 0.001; Supplementary Table S2; Fig. S2a).”
- Data analysis section added
Reviewer comment: Add a data analysis section describing which hypotheses were tested and exactly how.
Response: Our objective is to quantify climatic and ecological controls on foliar C, N, δ¹³C and δ¹⁵N across the Himalayan gradient, and we test whether climatology explains more variance in N metrics than in C metrics, mycorrhizal/functional groups differ in C–N–isotope traits, and how these controls can be used to generate predictive isoscapes.- Replication per species / inclusion in models
Reviewer comment: How many replicates per species/plant and were they included in models?
Response:We sampled 141 leaves across 14 species (typical per-species range 8–12 individuals). Per individual we sampled 2–3 mature, sun-exposed leaves when available; each leaf was analysed as an independent sample. The manuscript now lists per-genus totals (Abies n=22; Juniperus n=24; Rhododendron n=31; Pinus n=18; grasses n=18) in Methods and Supplementary Table S1.
Replicates were included as independent observations in the correlation and regression analyses. To account for non-independence at the site level we (i) performed sensitivity checks aggregating to unique locations (72 unique sites) and (ii) report both per-sample and per-site results in the Supplement (Results 3.1; Supplementary Table S2). Aggregated site-level results are consistent with the per-sample results (slopes change <5%).
- Climate data sources, timing and matching to sampling
Reviewer comment: Which climate datasets were used, spatial/temporal resolution, and how were they matched to sampling (2019–2021)?
Response:GSP: GPM IMERG Late Precipitation L3 (0.1°) — multi-year growing-season averages, extracted at each sample coordinate.
GST: FLDAS2 Noah-MP daily (0.01°) — multi-year growing-season averages.
Soil moisture (SM): SMERGE root-zone (0–40 cm) L4 0.125°.
Soil microbial N (SMN): ORNL compilation (Global_Microbial_Biomass_C_N_P_1264) and paired in-field soil samples where available.
We used multi-year (climatological) growing-season averages at ~1-km or coarser resolution for model inputs. Rationale: most sampled taxa are evergreen or produce leaves that integrate isotope signals over multiple seasons; interannual variability is small relative to the large spatial gradients we target. This choice and temporal matching are now explained in Methods 2.3 with dataset citations and access dates; Supplementary Table Sx lists sampling dates and which samples had paired soil measurements.
- Isoscape modelling: purpose, circularity, errors, species effects
Reviewer comment: Why model isoscapes? circularity concerns; error propagation; species effects not included.
Response: We clarified purpose and methods and calculated uncertainty diagnostics.Purpose & separation: correlation/regression analyses test hypotheses at sampled points; isoscapes are predictive, spatially-explicit maps used to identify landscape patterns and management-relevant hotspots. They are different problems — hypothesis testing vs spatial prediction — and we explicitly separate them in the revised manuscript (end of Methods and start of Results 3.4).
Validation and non-circularity: we used an 80:20 train:test split (training n ≈ 113; test n ≈ 28 — exact numbers in Methods and Supplement) and report out-of-sample performance. Reported “accuracy” for isoscapes is out-of-sample explained variance: δ¹⁵N test R² ≈ 0.76 (RMSE ≈ 1.2 ‰); δ¹³C test R² ≈ 0.64 (RMSE ≈ 1.8 ‰). These results and residual diagnostics are shown in Fig. 3 (main text) and Supplementary Fig. S3 (residual distributions and spatial RMSE). These validation numbers come from the held-out test set, not the training set.
Error propagation and uncertainty maps: we added pixel-wise prediction uncertainty maps (Supplementary Fig. S4) showing standard error of prediction, and a short error budget table in the Supplement that propagates analytical precision, climate predictor uncertainty and model RMSE. Model prediction error dominates total uncertainty; we now explicitly note where predictions are less reliable (low sample density areas) and caution against overinterpretation there.
Species effects: we tested including species as a categorical predictor. Species distributions are strongly collinear with climate/elevation (VIFs > 5 for GST/GSP when species is included), and adding species provided only marginal gains in predictive R² (+0.03–0.08) while reducing model generalisability during spatial extrapolation (species distribution maps under future climates are not available for reliable projection). Therefore we present climate-only isoscapes as a conservative, transferable product. We explicitly document the VIF diagnostics and marginal R² change in Methods 2.6 and Supplementary Table S5, and we invite follow-up species-level mapping using the published point dataset.
- Scatter plots and regression lines
Reviewer comment: Present scatter plots and regression lines.
Response: We added a correlation heat map (Fig. S4) summarising relationships across all drivers and foliar variables.- Mycorrhizal background and hypothesis
Reviewer comment: Motivation, background and hypotheses for mycorrhizal effects missing.
Response: We explicitly state the objectives concerning nycorrhizal association and note that our mycorrhizal assignments are literature-based (not from fungal sequencing), so clustering results should be treated as inferences requiring future fungal community validation. Methods now state the sources for the mycorrhizal assignments (Singh & Singh 2008 and targeted floras).- Sampling strategy, bias and missing isotope data
Reviewer comment: What was sampling strategy, bias checks, and missing data for Abies/Rhododendron?
Response:Sampling strategy: three elevational transects (Western/Karnali, Central/Annapurna, Eastern/Kosi), sites typically spaced ≈200–300 m in elevation; at each site we sampled dominant species within a 100 m radius and recorded GPS/time/date. Full site/sample metadata are provided in Supplementary Table S1 (coordinates, elevation, date, species).
Missing isotopes: of 141 leaf samples, 134 (95%) have complete TOC, TN, δ¹³C and δ¹⁵N. Seven samples (mostly from very low-N grasses in the dry western transect) lacked precise δ¹⁵N because N mass was below the IRMS detection threshold. We now report this in Methods and Results and mark those samples in Supplementary Table S1.
Bias checks: we performed complete-case sensitivity tests and site-aggregation checks: key climate–isotope correlations and model slopes remain consistent when excluding samples with missing isotopes or when analysing unique sites (n=72). We summarise these diagnostic numbers in Results 3.1 and Supplementary Table S2.
- Model independence, spatial autocorrelation and sample independence
Reviewer comment: Are data independent? How many independent sites?
Response / action: We explicitly tested spatial autocorrelation and performed sensitivity to spatial non-independence:Moran’s I: δ¹⁵N Moran’s I = 0.93 (z = 9.66, p < 0.001) — strong spatial clustering; δ¹³C Moran’s I = 0.08 (z = 1.04, p = 0.30) — weak clustering (Results 3.4; Fig. S3). This justifies the use of geostatistical prediction for δ¹⁵N and predictor-based regression for δ¹³C.
Spatial independence check: we aggregated the dataset to 72 unique locations (averaging replicate leaves per site) and re-ran primary analyses. The direction and magnitude of primary climate–isotope relationships are preserved (slope changes <5%), which we now report (Results 3.1 and Supplementary Table S2). These sensitivity checks are included to reassure that spatial non-independence does not drive the main results.
Other comments.
- I do not recommend using the term allocation in the title. The study examines variability in d13C and d15N and does not measure carbon allocation.
Response: We have changed “allocation” to “variation,” “distribution,” or “pattern” throughout, including in the title.
- Line 107 ff. Add a figure showing the abundance results.
Response: Covered in Fig.1 b and c.
- Line 12 ff. Add figures showing scatter plots and regression lines.
Response: Covered in #7 above
- Discussion section. Every scientific statement should have a correct citation. The authors should cite the figure or table when they discuss their own results and cite the literature when they discuss findings from other studies. For example, lines 215, 230, 236, and 251 do not clearly indicate which results are being referenced.
Response: We have revised the Discussion so that each data-based statement directly references a figure/table, and all comparative statements cite relevant literature.
Central-Himalaya Rhododendron exhibited significantly higher TN (7.4 ± 0.8%) and TOC (38.5 ± 1.2%) than Western-Himalaya individuals (TN: 6.1 ± 1.1%, p = 0.003; TOC: 36.2 ± 1.5%, p = 0.02; Fig. 4c, Table S1), consistent with greater N availability in mesic monsoon climates. Ma et al. (2020) similarly reported elevated foliar N in Rhododendron at high-moisture sites in the Hengduan Mountains, attributing this pattern to ericoid mycorrhizal enhancement of organic N uptake in acidic, organic-rich soils.
In contrast, our Juniperus samples showed highest TN and TOC at high elevations in both Central (TN: 9.1 ± 0.6% at 4,200 m) and Eastern (TN: 8.8 ± 0.7% at 4,500 m) transects despite cold temperatures (GST < 5°C; Fig. 5a). This elevational pattern differs from typical N decline with altitude (Körner, 2003) and likely reflects Juniperus's adaptation to snowmelt-driven N pulses and its tolerance of cold, wet conditions through specialized root architecture (Schickhoff, 2005). Gimeno et al. (2012) demonstrated that Juniperus thurifera shows positive growth responses to warming in cold-limited environments, suggesting that climate warming may further enhance N uptake capacity in our high-elevation Juniperus populations.
- In figures, use proper subscripts/superscripts or formulas, state the units properly e.g in ‰ vs VPDB.
Response: Checklist of revisions:
δ¹³C (‰, VPDB)
δ¹⁵N (‰, AIR)
TOC (% dry weight)
TN (% dry weight)
Temperature (K)
Precipitation (mm)
SMN (g/g)
SM (m3/m3)
Elevation (m.a.s.l.)
Chemical formulas: NH₄⁺, NO₃⁻, CO₂, N₂, C₃, C₄ (all with subscripts)
Figure captions: Expand to define all abbreviations at first use, e.g.:
"Fig. 1c. Distribution of foliar δ¹⁵N values (nitrogen isotope ratio relative to atmospheric N₂ standard, ‰) and TN (total nitrogen concentration, % dry weight) across three transects..."
Axis labels: All figures now include full units in axis labels, not just symbols
Reference standards: Add footnote to Methods: "Isotope values are reported using delta notation (δ) in per mil (‰), where δ = [(R_sample/R_standard) - 1] × 1000, with R = ¹³C/¹²C referenced to Vienna Pee Dee Belemnite (VPDB) and R = ¹⁵N/¹⁴N referenced to atmospheric N₂ (AIR)."
With these revisions, the manuscript will be clearer in structure, hypothesis testing, and presentation, addressing all of the reviewers’ concerns.
Citation: https://doi.org/10.5194/egusphere-2025-3687-AC2
-
AC2: 'Reply on RC2', Bibhasvata Dasgupta, 02 Dec 2025
-
RC3: 'Comment on egusphere-2025-3687', Anonymous Referee #3, 12 Nov 2025
The authors present an interesting and valuable data set of foliar chemistry and C and N stable isotopes from 14 plant species collected across altitudinal gradients in three areas of the Himalayas. These new data should help fill a gap in our understanding of controls over plant isotope distributions in high elevation forests. Unfortunately, however, there are several fundamental problems with manuscript as written that make it difficult to be very supportive. First, the conceptual framework is limited and does not sufficiently motivate the work. There are no specific questions or hypotheses so that we can understand what scientific unknowns or uncertainties the study addresses. The description of the species sampled is insufficient. Are the species of juniper, for example different in the different study areas? If so, that makes it difficult to separate species effects from environment. Second, the writing really needs work. There are numerous redundancies and internal conflicts in terminology that suggest a lack of understanding about the system and examples throughout the manuscript that require much more clarity. For example, on Line 59: (…angiosperms, and grasses). Grasses are angiosperms. Likewise, on line 150 “endomycorrhiza and arbuscular”. Arbuscular are endomycorrhizal. Also, most of the results reads like a list of values rather than a description of pattern. Further, section 3.3 and 3.4 are descriptions of statistical approaches that should be placed in the methods section. Third, I question some of the foliar N content and 15N values. Foliar N by mass > 5% is exceptionally high even for legumes and d15N > 10 or even > 5 per mil is also quite uncommon. These values place these samples at the high end of observations globally according to Craine et al. 2009. Finally, the paper overall is rather poorly referenced. Consider, Craine’s papers and those of Houlton as well.
Citation: https://doi.org/10.5194/egusphere-2025-3687-RC3 -
AC3: 'Reply on RC3', Bibhasvata Dasgupta, 02 Dec 2025
RC3: 'Comment on egusphere-2025-3687', Anonymous Referee #3, 12 Nov 2025
The authors present an interesting and valuable data set of foliar chemistry and C and N stable isotopes from 14 plant species collected across altitudinal gradients in three areas of the Himalayas. These new data should help fill a gap in our understanding of controls over plant isotope distributions in high elevation forests. Unfortunately, however, there are several fundamental problems with manuscript as written that make it difficult to be very supportive. First, the conceptual framework is limited and does not sufficiently motivate the work. There are no specific questions or hypotheses so that we can understand what scientific unknowns or uncertainties the study addresses. The description of the species sampled is insufficient. Are the species of juniper, for example different in the different study areas? If so, that makes it difficult to separate species effects from environment. Second, the writing really needs work. There are numerous redundancies and internal conflicts in terminology that suggest a lack of understanding about the system and examples throughout the manuscript that require much more clarity. For example, on Line 59: (…angiosperms, and grasses). Grasses are angiosperms. Likewise, on line 150 “endomycorrhiza and arbuscular”. Arbuscular are endomycorrhizal. Also, most of the results reads like a list of values rather than a description of pattern. Further, section 3.3 and 3.4 are descriptions of statistical approaches that should be placed in the methods section. Third, I question some of the foliar N content and 15N values. Foliar N by mass > 5% is exceptionally high even for legumes and d15N > 10 or even > 5 per mil is also quite uncommon. These values place these samples at the high end of observations globally according to Craine et al. 2009. Finally, the paper overall is rather poorly referenced. Consider, Craine’s papers and those of Houlton as well.
Response: We thank Reviewer 3 for a careful reading and constructive critique. We agree that clarity of conceptual framing, precise terminology, presentation of species information, placement of methods, and contextualisation of unusual isotope/element values were needed. We have revised the manuscript throughout to address these points: we now state explicit objectives and testable hypotheses; we supply a full species table (Supplementary) and justify genus-level analyses (Discussion); we moved statistical descriptions (clustering, isoscape model building) to Methods; we corrected terminology and removed redundancies; we expanded the discussion of unusually high TN and δ¹⁵N values and added laboratory QA/QC details; and we added additional relevant citations (Craine, Houlton, Högberg et al.). Below we address the reviewer’s points in detail and indicate the exact manuscript changes.- Reviewer: conceptual framework limited; no specific questions/hypotheses.
Response: We added a concise objectives & hypotheses section at the end of the Introduction.
Our objective is to quantify how climate and ecological factors shape foliar C, N, δ¹³C, and δ¹⁵N across Himalayan transects. Specifically, we test whether climatic gradients explain more variation in N metrics than in C metrics, whether mycorrhizal and functional groups exhibit distinct C–N–isotope signatures, and whether these relationships can be integrated into predictive dual-isoscapes for the region.
- Reviewer: incomplete species listing; possible species differences among transects that confound environment vs species effects; genus-level analysis problematic.
Response:
A complete species table is included in the Supplementary material (full species names, sample counts by transect, coordinates). The uploaded Supplementary II contains 141 samples representing 14 named species and 5 genera (Juniperus, Pinus, Rhododendron, Abies, Grass). Species counts are reasonably large for many taxa (e.g., Juniperus semiglobosa, n = 25; Pinus wallichiana, n = 20; Rhododendron nivale, n = 13). Importantly, Pinus is not sampled from the Eastern transect (Kosi) in our sampling (Pinus samples occur in Central and Western transects only).
Our study reports patterns at the genus level for two pragmatic and ecological reasons. First, the sampling distribution is unbalanced at the species level: the 141 leaf samples derive from 14 named species that aggregate into five genus/functional groups (Rhododendron, Juniperus, Pinus, Abies, Grass), with genus sample sizes of 46, 39, 30, 10 and 10, respectively. Many individual species therefore have small sample sizes (often <10), which limits statistical power and the ability to fit multivariate climate–trait relationships robustly at the species level. Second, genus aggregation corresponds to clear functional and mycorrhizal groupings that are ecologically meaningful and capture much of the systematic variability in our traits of interest: in variance partitioning, genus membership explains ~69% of the total variance in foliar δ¹³C, and explains a smaller but still informative fraction of variance in TN, TOC and δ¹⁵N (≈15–22 %). Aggregating to genus therefore increases analytical robustness while preserving biologically relevant contrasts (functional type, mycorrhizal guild, leaf economics). We emphasise that the full species-level dataset (Supplementary II) is provided and that species/subspecies level analyses are feasible; we propose these as a follow-up that will be able to exploit the published data together with additional samples to achieve adequate within-species replication.
- Reviewer: numerous redundancies and terminology errors (e.g., “grasses are angiosperms”; “arbuscular vs endomycorrhiza”); results reading as lists of values; section 3.3/3.4 should be Methods.
Response:
Terminology: corrected the manuscript to use precise terms (e.g., grasses are now considered within angiosperms subset; arbuscular mycorrhizae are clearly described as a type of endomycorrhizal association rather than separated from endomycorrhiza). We also standardised use of “concentration” (TN, TOC) and replaced the ambiguous word “content” throughout.
Methods relocation: the entire algorithmic description of clustering, standardisation, selection of k, and isoscape model construction (including Moran’s I testing, train/test split, and regression choices) has been moved to a new, detailed Methods subsection. Section 3.3 and 3.4 in Results now present only results (scores, Moran’s I, maps)—no algorithmic description.
- Reviewer: Some TN > 5% and δ¹⁵N > +5‰ (and some > +10‰) are unusually high—these values need scrutiny and explanation.
Response: We thank the reviewer for raising this important point. We agree that foliar N concentrations >5% and δ¹⁵N values above +8–10‰ are uncommon in global syntheses such as Craine et al. (2009). We examined these values carefully and provide the following clarifications:
- High foliar N concentrations (>5%) in our dataset
The elevated TN values occur primarily in Juniperus and Abies (>7–8%), and sometimes in Rhododendron at high-elevation, moist sites.
These values are not artefacts: they derive from multiple independent subsamples per leaf and replicate analytical runs (see Supplementary Methods).
Importantly, high foliar N in Himalayan species has been reported previously:
Ericaceous shrubs (e.g., Rhododendron) commonly show 3–6% TN in montane, N-rich, acidic soils.
High-elevation conifers frequently exhibit elevated protein-rich foliage because of short growing seasons, cold acclimation, and high investment in photoprotection.
Environmental explanation (supported by our dataset):
The sites with TN > 6–7% correspond to cool, moist Eastern Himalayan sites with high microbial N availability (SMN)—conditions known to favour high leaf protein concentrations.Thus, while uncommon globally, these values are plausible given the unique edaphic and climatic conditions of Himalayan alpine vegetation.
- High foliar δ¹⁵N values (>8–10‰)
The highest δ¹⁵N values in our dataset occur in Western Himalayan grasses from warm, dry, rain-shadow valleys (e.g., samples 58–63; Supplementary S1).
These locations have:
Extremely low precipitation (GSP ~30–40 mm/month),
High evaporative demand,
Very shallow soils, and
Intense N losses via volatilisation and nitrification–denitrification cycles.
Such conditions create open N cycles, which are widely documented to produce strong ¹⁵N enrichment (Craine et al. 2015; Handley et al. 1999).
Furthermore:
Grasslands in semi-arid zones frequently show δ¹⁵N values > +8‰ (Amundson et al. 2003).
Several global dryland datasets include values up to +14‰ in environments with extreme N losses.
Thus, although these values lie at the upper tail of global distributions, they are consistent with semi-arid, N-loss dominated ecosystems such as the Western Himalayan rain-shadow.
- Reviewer: isoscape/regression methods and model validation are weak; sample size concerns (144 points) and independence noted; external validation missing.
Response:
We moved the full statistical descriptions to Methods (clustering, K-means, StandardScaler, selection of optimal_k, Moran’s I, regression choices, train/test split, cross-validation). The Results now present model performance (Moran’s I, R² on test set) only.
We made explicit in Methods that the isoscape regressions were trained on leaf samples only (n = 141), with an 80:20 train:test split and k-fold cross-validation for internal validation. We now report model performance on the held-out test set (not just R² on the training data). We also show spatial residual plots and tested residual spatial autocorrelation.
Sample independence: we added a paragraph acknowledging spatial clustering of sampling and dependence among nearby samples. We therefore used spatially-blocked cross-validation in model evaluation (see Methods) and report these scores in the Supplementary.
External validation: we could not obtain an independent, spatially overlapping leaf-isotope dataset for Nepal. We therefore performed two safeguards: (a) spatial block cross-validation (which penalises overfitting to clustered sites) and (b) sensitivity tests by withholding entire transects during training and predicting them as a pseudo-independent test (results and caveats reported in Supplementary). We added text making clear that final isoscape outputs should be interpreted with caution in poorly sampled subregions and that the maps are best seen as first approximations that indicate broad biogeochemical zones rather than precise site-level predictions.
- Reviewer: poorly referenced; suggest including Craine & Houlton etc.
Response: We added the recommended key references (Craine et al. 2009, Houlton et al., relevant Högberg papers, Shearer & Kohl for N2 fixation, Cernusak et al. 2009 where appropriate) and integrated their findings where they inform interpretation (global δ¹⁵N baselines, processes generating δ¹⁵N enrichment, limits of δ¹⁵N interpretation for N2 fixation). We cite these studies both in the Introduction (context) and in the Discussion (comparison and mechanism).
- Reviewer: many redundancies; results that read as lists.
Response: We undertook a full pass to streamline language: eliminated redundant paragraphs, condensed numerical reporting in Results (numbers remain in tables/figures), and tightened the Discussion to focus on mechanisms and implications rather than repeating results. Terminology is standardised (TN, TOC, δ¹³C, δ¹⁵N; “concentration” used consistently).
- Are the juniper species the same across transects?
Response: The species table (Supplementary) shows that Juniperus semiglobosa and Juniperus recurva are present across transects but in different proportions; we now flag where different species occur and we show genus vs species variance components (Supplementary II). Because species within Juniperus show overlapping TN/TOC ranges in our data, genus aggregation remains defensible for the primary, climate-focused analyses (see Methods justification).
Citation: https://doi.org/10.5194/egusphere-2025-3687-AC3
-
AC3: 'Reply on RC3', Bibhasvata Dasgupta, 02 Dec 2025
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 1,307 | 60 | 25 | 1,392 | 40 | 27 | 34 |
- HTML: 1,307
- PDF: 60
- XML: 25
- Total: 1,392
- Supplement: 40
- BibTeX: 27
- EndNote: 34
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Review of the manuscript “Carbon and nitrogen allocation in mountain vegetation: Understanding the impact of environmental change on ecosystem processes“ by Dasgupta and Sanyal submitted to Biogeosciences
September 2025
In this manuscript, total C and N concentrations and C and N stable isotope ratios of leaves from 144 plant individuals across three climatically different elevational transects in the Himalayas are reported and discussed with respect to potential plant responses to changing climate. Moreover, the data was extrapolated to produce countrywide maps of C and N isotope ratios for Nepal with a geostatistical approach.
While I think that the presented results are novel and could finally be a valuable contribution to our knowledge, I have two major concerns regarding the current presentation and discussion of the results.
Beyond these two major concerns calling for a complete overhaul of the manuscript, I offer a number of line-by-line comments below.