Reply to Anonymous Referee #1 

General comments 

This study deals with effects of plant composition on soil nitrogen turnover in shrub tundra, a relevant topic considering the ongoing changes in vegetation composition and nutrient turnover associated with climate change in the arctic. It is a carefully designed and conducted study with clear objectives and finalized appearance. The motivation for the study as well as interpretation of results are well-argumented and referenced, and all the content is relevant, making it a nice and compact read. At places, some additional explanation would be good, as specified below. 

Reply: We thank reviewer 1 for their encouraging assessment of our manuscript. We are pleased that the study’s relevance, design, and clarity came through, and we appreciate your positive remarks on the motivation, interpretation, and overall structure. We have considered your suggestions for additional clarification and have addressed them point by point in our responses below. We hope the revisions improve the manuscript and meet your expectations. 

Major/intermediate comments 

1. Some additional info about plant removal treatments would be needed: 

It would be good to clarify the association between the ErC/ErM and AN/MN vs. dominant and rare treatments. Where dominant species mainly ErC/ErM or AN/MN plants, or en even mixture of the two types? How about rare? Knowing this might be relevant for interpreting the results, so it would be good to comment this in the text and perhaps add to Supplementary Table S1. 

Reply: We thank the reviewer for this insightful comment. The Dominant and Rare treatments were designed to include an even mixture of mycorrhizal types (ErM/EcM and AM/NM) to allow us to disentangle the effects of species richness from those of mycorrhizal association. We agree that clarifying this in the manuscript will help readers better interpret the results. 

Data show that both the Rare and Dominant treatments included a relatively balanced removal of species across mycorrhizal types. For example, in the Rare treatment, species removals included 6 ErM, 5 EcM, 7 AM, and 6 NM species. Similarly, the Dominant treatment included the removal of 7 ErM, 5 EcM, 7 AM, and 10 NM species. While the Dominant treatment removed slightly more NM species than the Rare treatment, both treatments maintained a mix of mycorrhizal types. 

Regarding the species remaining after treatment, Dominant plots contained by average 12 species, with 8 AM/NM species (31% cover) and 5 ErM/EcM species (43% cover). Rare plots contained by average 10 species, with 8 AM/NM species (7% cover) and only 1 ErM/EcM species (<1% cover). These differences in cover reflect the natural dominance of certain species but do not indicate a systematic bias in the design of the treatments. 

We will update Supplementary Table S1 to include this information and revise the sentence “Rare and dominant plant species were determined by reducing species richness to approximately 50 % of the community species pool while maintaining a mixed mycorrhizal plant community.“ on Line 119-120 for clarification. 

Lines 119-120 will read: 

The Dominant and Rare species removal treatments were designed to include a relatively even mixture of species representing different mycorrhizal types (ErM, EcM, AM, and NM). This design allowed us to separate the effects of species richness from those of mycorrhizal association. While the exact number of species removed varied slightly between treatments, both included a balanced representation of mycorrhizal types. 

Further, it would be worth commenting, how large plant biomass reduction was associated with each treatment. I would assume that this was highly variable between treatments, particularly between Rare (larger reduction) and Dominant (higher reduction). This reduction of plant biomass would be associated with reduction of labile C input and plant N uptake, which both may be important for soil N cycling. The focus of this study was in quality (= changes in plant composition), many aspects of C and N cycling are driven by simple quantity and this should be acknowledged. 

Reply: We appreciate the reviewer’s point regarding the importance of biomass quantity in driving soil C and N cycling. While we did not quantify aboveground biomass reductions directly, our experimental design was conceptually grounded in the mass-ratio hypothesis (Grime 1998), which posits that dominant species influence ecosystem processes primarily through their biomass-scaled traits, whereas rare species exert influence more through functional identity than quantity. 

Thus, the contrast between dominant and rare species in our study can be interpreted as a test of these two mechanisms: 

• For dominant species, biomass loss likely represents a key mechanism driving changes in soil processes. 

• For rare species, the effects are more likely tied to species identity and trait novelty, rather than biomass per se. 

We will clarify this conceptual framing in the manuscript and acknowledge that while biomass was not measured, it is implicitly central to the theoretical contrast we aimed to test. 

1. A few points could be clarified regarding the 15N method and results: 

The study design would have allowed determination of DNRA rate but was not determined (line 210->). Was this decided prior to running the model because it was assumed that it is unimportant, or was it found irrelevant based on the modelling results? 

Reply: Indeed, the tracing model would allow for quantifying DNRA rates. The process was excluded prior to running the model, as the measured ¹⁵N data showed no enrichment in the ammonium pool following ¹⁵N labeling of the nitrate pool. 

Reply to Anonymous Referee #2 

This study experimentally manipulated Arctic plant communities to test the effects dominant mycorrhizal associations as well as dominant vs rare plant communities. The writing is generally excellent, background well organized, and hypotheses clearly define. I appreciate that some of the results were explained as surprising in context of the hypotheses. 

Reply: We thank reviewer 2 for the constructive feedback. We are glad to hear that the experimental approach, organization of the background, and clarity of the hypotheses were well received. We also appreciate your recognition of our efforts to interpret unexpected findings within the context of our initial expectations. Your comments were encouraging and helpful, and we have addressed the specific points raised in your comments below. 

Intro: 

One main point I think could use addressing in the logic and setup of the study is that it’s a bit hard to understand the specific treatments. Specifically, why EcM/ErM were lumped as opposed to AM/NM, and how these two treatments are more or less similar to the dominant and rare treatments. I would imagine that the shrubs tend to collocate (EcM and ErM) and then the grasses. Was the natural tendency for EcM and ErM to be found next to one a reason why there was no AM/EcM or AM/ErM treatment? 

Reply: One of the motivations behind our groupings was to reflect the large-scale patterns in carbon and nitrogen dynamics identified by Averill et al. (2014) and Read & Moreno (2003), as well as the small-scale patterns described by Giesler et al. (1998) and Björk et al. (2007). EcM and ErM were grouped based on their shared functional traits—particularly their saprotrophic capabilities and ability to access organic nutrient pools in nutrient-poor soils. In contrast, AM and NM plants were grouped due to their association with faster nutrient cycling and preference for more mineral-rich soils. Tundra communities offer a unique advantage in this context, as they naturally encompass all major mycorrhizal associations within a single plant community, making them ideal testbeds for experimental manipulation. By altering these associations, we aim to disentangle the ecological processes underlying both large- and small-scale patterns. 

To ensure that observed effects are attributable to changes in mycorrhizal associations rather than shifts in species richness, we also implemented treatments that varied in species dominance and rarity. These treatments included a mixture of mycorrhizal types, allowing us to control for species composition while simultaneously addressing a broader ecological question: the relative importance of dominant versus rare species in driving ecosystem processes. 

We will consider a way to make this clearer in the objectives paragraph of the introduction. 

Methods: 

I understand that to sample in each of the plots for a Before After Control Impact study would have created additional disturbance and could have affected belowground N cycling. However, I think that some kind of spatial accounting for the blocks is necessary given the finding that blocks had an effect on multiple variables. The supplemental plots are great. Did you consider using fixed effects in your main model that may have differed between block, such as landscape position that can affect longterm patterns in soil temperature and moisture? The topographic wetness index is one I would suggest that would take into account relative differences in elevation.  I see in the supplement plot S3 that there is a difference in elevation (albeit a small one) from the highest to lowest plot (elevation y axis could use units), but this corresponds to differences in temperature, and to some degree soil moisture. From the PC scores in Fig S2, it looks like PC3 may follow an elevational/temperature trend. 

Reply: We agree that spatial heterogeneity, even within relatively small field sites, can influence environmental variables such as temperature and soil moisture. We did consider including additional fixed effects that varied by Block, such as elevation or other topographic features like the TWI. However, our experimental design included uneven replication across treatments, which limited our statistical power and increased the risk of overfitting if too many fixed effects were included in the models. To ensure model stability and interpretability given our sample size, we prioritized parsimony in our fixed effect structure. We addressed the influence of spatial variation by including Block as a random effect in our main GLMMs. Because block showed significant effects for several response variables, we also fitted supplementary models with Block as a fixed effect to further explore its influence. 

All plots were situated within a relatively small area (~2400 m2), with plots within blocks separated by only a few meters and blocks by tens of meters. While small differences in elevation corresponded to differences in soil temperature and, to some extent, VWC, we interpret these as natural small scale spatial variability rather than systematic topographic or landscape-scale effects. Because we measured VWC directly for each plot, we felt that incorporating a landscape-scale proxy such as the TWI, whose predictive power has been shown to be modest and highly dependent on DEM resolution (Riihimäki et al. 2021), would not substantially improve our model or interpretation.  

We do appreciate the suggestion and acknowledge that TWI or similar indices may be more useful in broader-scale studies. 

The experimental design seems well thought out and thorough. One consideration is the effect of clipping more or less vegetation in the different treatments and how this would affect water surface evaporation, evapotranspiration, and ultimately the amount of moisture in the soil that could affect N cycling. There seem to be differences in field moisture, and I wonder if the authors could briefly speak to how this disturbance could have affected processes belowground. Could even be 1-2 lines in the discussion. 

Reply: To minimize the influence of evaporation and evapotranspiration on our treatments, we enclosed each plot with a 1 µm mesh cloth and positioned them on a gentle slope. This design facilitates lateral water movement across plots, helping to buffer against localized water loss. As a result, we observed minimal treatment effects on our water-related variables, suggesting that evaporation and evapotranspiration are unlikely to have significantly influenced our results. We will add a short sentence at line 514: 

"Furthermore, the minimal treatment effects on water-related variables suggest that evaporation and evapotranspiration had limited influence on our results."

Small suggestions 

The background information is well written, and often there are many citations for sentences containing multiple clauses. I would suggest moving citations so that ones for soil C content or nutrient cycling come right after the respective clause (lines 40-45 for example) to match citations with the info. It may make readability a bit harder, so up to your discretion. 

Reply: Thank you for this thoughtful suggestion. We agree that aligning citations more closely with specific clauses can improve clarity in some cases. However, we also considered the potential trade-off with readability and flow. After reviewing the relevant section, we decided to retain the current citation placement for now, but we are happy to revise this if the editor feels that closer alignment would improve the manuscript. 

Line 68 – Starting with AM fungi being less common in Arctic systems does not sound compelling to include as a dominant type in this experiment. You could reword to start with EcM and ErM tending to be dominant in Arctic systems, but AM still prevalent and likely still affecting N cycling etc despite the lower abundance. 

Reply: Thank you for the suggestion, which we will incorporate. 

Line 68 will read: 

"Ectomycorrhizal (EcM) and ericoid mycorrhizal (ErM) fungi tend to dominate  Arctic ecosystems (Michelsen et al., 1998; Soudzilovskaia et al., 2017; Steidinger et al., 2019), whereas arbuscular mycorrhizal (AM) fungi are considered less common due to low cold tolerance but still prevalent (Kilpeläinen et al., 2016; Kytöviita, 2005; Ruotsalainen and Kytöviita, 2004; Wang et al., 2002)."

Line 70. “Different mycorrhizal types” could be instead “These three mycorrhizal types” as you introduce them already and don’t include orchid mycorrhizae. 

Reply: Agree we will incorporate this suggestion. 

Line 86. I would restructure the topic sentence a bit for succinctness, with something like: “ We aimed to determine the relative effects of functional (mycorrhizal) and structural (rare, dominant) diversity on soil N cycling.” 

Reply: Thank you for the suggestion, which we will incorporate. 

Line 90. Would remove “thus.” I’m a fan of thus and therefore, but doesn’t seem necessary here. And should the hypotheses be in past tense? 

Reply: Agree, that “thus” can be removed. However, we chose to present our hypotheses in the present tense in the Introduction to reflect their role in framing the study rationale, in line with guidance from Josh Schimel’s book Writing Science: How to Write Papers That Get Cited and Proposals That Get Funded. 

I like how the methods has a data analysis section for each component. In the isotope tracing section, I am curious why the rates are relativized by per g C, and not per g N (assuming this is bulk). I don’t have experience with these models, and C and N tend to increase together, but there can be interesting differences in CN. Forgive my ignorance if this is common practice. 

Reply: In the literature, gross N transformation rates have both been presented on a per g C and per g N basis, besides the most common per soil dry weight. From our point of view, the per g C basis is more appropriate, as we assume that in most soil (heterotrophic) microorganisms are limited by C, and not N. Hence, the rates expressed per g C provide information on if rates differ beyond those explained by the limiting resource for microbial activity, and indirectly for nitrification. 

Line 238. How did you validate models exactly? Visually or with additional fit statistics? 

Reply: We validated model assumptions using the DHARMa package (line 238), which simulates scaled quantile residuals to assess model fit. We examined residual vs. fitted plots and QQ plots and used DHARMa’s tests for uniformity and outliers to identify potential violations of non-normality and heteroscedasticity. We will revise the methods to clarify this approach. 

Line 238 will read: 

"We validated model assumptions using the DHARMa package (Hartig, 2022), which simulates scaled quantile residuals. Model fit was assessed through residual vs. fitted plots, QQ plots, and DHARMa’s tests for uniformity and outliers to detect deviations from normality and heteroscedasticity."

Line 380. The explanation of the findings is compared well to the literature. However, for clarity, I would suggest the topic sentence be more like the last sentence of the paragraph so the reader knows where the paragraph is going. I found it to be a bit winding and unclear what the main take away was. 

Reply: Our topic sentence reads "As hypothesized, we found the highest gross N mineralization rates in the EcM/ErM treatment, but unexpectedly, the treatment with only dominant species in the plant communities also exhibited high rates.", which we believe shows what the paragraph is about. The last sentence "Thus, our findings suggest that mycorrhizal status, particularly EcM/ErM associations, plays a more significant role in shaping gross N cycling dynamics than species dominance alone." concludes our findings. 

Section 4.2. The focus on background info about mycorrhizal effects that your study does not support takes away from the findings you do have on how the dominant community affects functional groups. Also, towards the end near line 455, there’s a cool line about how AM/NM had similar composition to dominant plots. I feel like this is worth highlighting towards the beginning of this paragraph, especially since you found differences in the dominant treatment. 

Reply: Thank you for the suggestion, we have rephrased the paragraph according to the suggestions.  

First paragraph in section 4.2 will read:  

"Despite the distinct roles of mycorrhizal fungi in N cycling (Castaño et al., 2023; Hobbie and Högberg, 2012; Tedersoo et al., 2020), the AM/NM and EcM/ErM plots did not differ in N-cycling gene abundances. However, altering plant composition revealed functional differences. Notably, our Dominant community plots—despite having a similar plant composition to the AM/NM plots—showed lower abundances of AOA and NIS functional groups and reduced gross nitrification rates. This may reflect stronger plant competition for NH₄⁺ (Hayashi et al., 2016) or reduced microbial reliance on NH₄⁺ (Hobbie & Hobbie, 2006; Schimel & Chapin, 1996). In contrast, nitrification gene abundances in Rare community plots were comparable to Control plots, despite lower gross nitrification rates. Since plant species richness was similar across Dominant and Rare treatments, our results suggest that dominant species traits—rather than richness—may drive ecosystem function, echoing findings in the ecological literature (Grime, 1998; MacGillivray et al., 1995). These traits appear to differ or be suppressed in the AM/NM community, suggesting that even a minor presence of EcM/ErM plants in an AM/NM-dominated community can shift how plant traits influence N dynamics in Arctic soils."

Line 460-480 This paragraph on the nitrifier community and N-cycling gene abundances seems a bit long for having no treatment effects. Does the last sentence about microbial traits and environmental pressures contradict your other findings and the next paragraph 4.3? Why do these treatments in the first place if the different plant community combinations have no effect? 

Reply: We agree that the section is currently long given the lack of strong treatment effects, and we will revise it for conciseness. 

Because plant and mycorrhizal communities can influence soil microbial composition and nutrient cycling, we believe the treatments are valid. The limited treatment effects on the functional guilds of nitrifiers do not rule out plant influences but indicate that abiotic constraints, such as temperature and moisture, may override biotic drivers in this context. By examining the abundance of different groups of nitrifier communities, we could conclude that guilds with resource-efficient ammonia and nitrite oxidation strategies are favored irrespective of treatment, which gives further support to our findings of a conservative N cycle in these Arctic soils. This is not contradicting the findings discussed in 4.3, as the genetic potentials reflect more long-term capacity and the ecology of the present nitrifier guilds, while the measured rates show the activity of the time of sampling.