Seasonal dynamics of stem N2O exchange follow the physiological activity of boreal trees

The role of trees in the nitrous oxide (N2O) balance of boreal forests has been neglected despite evidence suggesting their substantial contribution. We measured seasonal changes in N2O fluxes from soil and stems of boreal trees in Finland, showing clear seasonality in stem N2O flux following tree physiological activity, particularly processes of CO2 uptake and release. Stem N2O emissions peak during the vegetation season, decrease rapidly in October, and remain low but significant to the annual totals during winter dormancy. Trees growing on dry soils even turn to consumption of N2O from the atmosphere during dormancy, thereby reducing their overall N2O emissions. At an annual scale, pine, spruce and birch are net N2O sources, with spruce being the strongest emitter. Boreal trees thus markedly contribute to the seasonal dynamics of ecosystem N2O exchange, and their species-specific contribution should be included into forest emission inventories.

Boreal forests are widespread, where nitrous oxide has been largely overlooked. The authors spent a lot of time to measure annual changes in soil and stem nitrous oxide fluxes of the common boreal tree species spruce, birch, and pine. Particularly, the greenhouse gas fluxes were measured in the winter. I approve the authors worked hard and pushed this topic forward and stored useful data. However, I have the following concerns.
Stem nitrous oxide fluxes have been previously widely investigated. This work is a case investigation and lacks a novelty. When I am a reader, I cannot be impressed by the results.
What is physiological activity? What are physiological parameters? Soil water content, soil and air temperatures, and photosynthetically active radiation? To my knowledge, the authors measured environmental factors but did not measure tree physiological parameters. I do not think the authors studied the physiological activity and its direct relationships with seasonal nitrous oxide fluxes.
The purpose of this study is seasonal changes in stem nitrous oxide fluxes. Why are almost half of figures on carbon dioxide? I suggest that the figures on carbon dioxide can be deleted.
This work did not study the mechanisms of nitrous oxide production and emissions. The authors provided too much speculative discussion without evidence. I suggest that the discussion can be largely shortened and improved. -We want to thank the reviewer for the suggestions to focus in on and reinforce the main message of the manuscript. We have now carefully read through and analysed what we consider to be the manuscript's main message. In the revised version of the manuscript we have carefully highlighted the following points. For the first time, we present 1) a comprehensive assessment of yearly budget and seasonal and monthly pattern of N 2 O flux dynamics in main representatives of boreal tree species and soil. Also, we report 2) high variability in tree stem N 2 O dynamics among the different tree species, and that 3) N 2 O stem emission were closely related to the physiological activity of trees and ecosystem -mainly to stem CO 2 efflux and gross primary production (GPP). We also want to highlight 4) the evidence of the trees' ability to take up N 2 O. Finally, we estimate 5) the contribution of tree stem N 2 O flux to the net ecosystem N 2 O exchange. We improved the manuscript accordingly. These main results are now carried through the entire manuscript as well as carefully discussed. We believe that modification of the manuscript based on the comments of all reviewers improved the clarity of the main messages of the manuscript (please, see the proposed modifications throughout the manuscript in response to all reviewers' comments). -Discussion on the observed high variability in tree stem N 2 O dynamics among the different tree species has been re-worded and improved in many places of the revised version of the manuscript (e.g. page 5, rows 192-216, page 6, rows 257-282). -We elaborated in the Results (page 6, rows 237-249) that the stem N 2 O fluxes of pine, spruce, and birch trees were not significantly different at annual scale between trees growing on plots characterized with different soil water content, and, as soil moisture was not found to be among the driving variables of tree stem N 2 O fluxes, we can confidently conclude that soil moisture is not critically affecting stem N 2 O fluxes. Hence the site type difference in our study area does not play an important role, as stated in page 6, rows 250-256. -We added information about tree species composition of boreal forests in Finland, and discussed the role of spruce trees as the greatest emitter of N 2 O in Finland and in Europe generally (page 6, rows 257-282). -Furthermore, in order to relate our results to a more general/global context, we added brief discussion on the topic of how representative are the studied N 2 O flux dynamics within boreal and temperate regions (N limited vs. N "saturated" forest ecosystems) (page 7, rows 290-301). -Finally, we edited our overall conclusions in the manuscript to improve the clarity of the main message/results of our study (i.e. answering of the research questions stated at the end of the introduction section; see page 7, rows 302-310). -We added a new Supplementary Fig. 5 showing the topographic wetness index (TWI) of the studied forest site. We newly estimated the contribution of tree stem N 2 O fluxes from the studied plots characterized with different soil moisture to the total forest fluxes (page 6, rows 250-256): "Based on the topographic wetness index (TWI) at the site, the dry plot represents 48%, the moderately wet plot 37%, and the wet plot 11% of the forest (remaining 4% accounting for standing water, Supplementary Fig. 5). Thus, we estimate that the annual emissions from the wet and moderately wet plots together represent ca 50% and the emissions from the dry plot ca 50% of the total forest fluxes, respectively. As we have demonstrated that the tree stem N 2 O fluxes are not controlled by soil water content at the annual scale, we confidently can conclude that the site type does not play a critical role in stem N 2 O fluxes." -The issue of the inter-site variability in tree stem N 2 O fluxes is newly, briefly discussed on page 4-5, rows 184-191: "We speculate that the species variability in N 2 O exchange (Figs. 1, 5) might be explained by spatial variability of N 2 O concentration in soil, which is more pronounced under lower soil VWC. Under such conditions, N 2 O sources are more diverse due to simultaneously running aerobic and anaerobic N turnover processes leading to production and consumption of N 2 O. At dry conditions, therefore, root depth and distribution seem to play a more important role, species specificity is more pronounced, and differences among individual trees having different N 2 O sources available also are more prominent. This hypothesis should be confirmed by further research." What is the applicability of the results in the rest of the range of the boreal forest? -We want to thank the reviewer for this important comment. We have now added discussion on the representativeness of our measurement site within Finland and generally in the boreal region (see pages 6-7). Based on an analysis of the climate statistics between Finland (our site) and the Canadian boreal forest region, we find very similar seasonality (min and max temperatures) and mean annual temperature. Although the climate statistics of Finnish and Canadian boreal forests as well as seasonality in temperatures are very similar, upscaling from one representative study site to the whole boreal forest region seems too uncertain. We write this in the paper, but we also want to state that based on our data it is extremely difficult to extrapolate to other boreal forest regions (page 7, rows 290-301) as follows: "We have demonstrated that N 2 O emissions from tree stems are driven by physiological activity of the trees and by ecosystem activity, showing higher emissions during the active growing period and variation between uptake and emissions during the dormant season. Although our study may well be applicable to large upland forest areas in the boreal zone, which are typically nitrogen (N) limited (Högberg et al., 2017), our findings may not apply directly in N-affected central European or American forests known to exhibit elevated soil N 2 O emissions due to higher soil N content and faster N turnover rates (Aber et al., 1998;Butterbach-Bahl et al. 2002;Kreutzer et al., 2010). The N status of a forest directly influences soil N 2 O concentration, which has been shown to be a good proxy for N 2 O transport via the transpiration stream of trees . Until more studies and process understanding emerge, the global strength of N 2 O emissions from trees will remain largely unknown and could possibly be estimated by, for example, adding a fixed percentage (e.g. 10%) to the forest floor N 2 O emissions to represent N 2 O emission from trees." (see also response to the next reviewer's comment) -Although we considered that adding a comparison of weather statistics between the measurement site and Canadian and Siberian boreal forests is not necessary, we are presenting them here in support of our decision: Long-term weather statistics at our measurement site (Hyytiälä, Finland), with mean annual temperature of 3.5 °C and annual precipitation of 711 mm , and those in the boreal forest region in Canada ( -It is our understanding that the N 2 O exchange of trees is currently not considered in either global models or any forest ecosystem process models. This is because the research topic is very new and the number of published papers dealing with trees and N 2 O dynamics is rather small. Also, this is the first study to demonstrate a strong seasonality, which would be relevant to any process models and also should be acknowledged in the global models. We feel that this research field is still too new to include these data into global N 2 O models, partly because we lack a process for understanding the N 2 O exchange of trees. Our study plays a critical role in adding to understanding of the drivers (i.e. physiological activity of the trees and the ecosystem) of N 2 O fluxes of tree stems. Including the dependencies of tree stem N 2 O fluxes and stem CO 2 efflux, ecosystem evapotranspiration, and GPP into N cycling process models, or more likely into statistical models, could serve as the first steps towards estimating the rate of N 2 O emissions from trees. Nevertheless, estimation of regional or global N 2 O emissions from trees would require understanding of regional drivers of tree-N 2 O dynamics. An intensive European research project on N oxide emissions from European forest ecosystems (NOFRETETE) has concluded that the regional estimates of N 2 O emissions from European forest soils depend on the nutrient status, N deposition, and soil type, while at local scale the soil N 2 O emission dynamics were driven by soil moisture and temperature (Pilegaard et al. 2006 My only concern is the number of figures. Ten figures in this manuscript is quite a lot, I recommend to shift Fig. 4 to 6 to a supplementary section, as the outcome of the statistical analysis is also given in the text.
-We want to thank the reviewer for constructive comments that certainly have improved the manuscript. The manuscript is of high quality and deserves publication in a high-ranking journal like NCMSS. Therefore, I recommend publication of this manuscript with minor revisions.

Minor comments
At many places (Lines 39,44,221,250,408,416]  -We want to thank the reviewer for commenting on the possible role of mycorrhizal fungi in the soil N cycling processes and N 2 O uptake / production. We have modified the discussion accordingly to account for a possible role of mycorrhizal fungi in the N turnover and N 2 O exchange, and we highlighted the strong rhizospheric effects on soil N turnover processes (page 3-4, rows 120-141).
Although laboratory measurements show that high N 2 O soil concentrations are reflected in high stem N 2 O emissions , in an earlier study at the same site as our study, soil N 2 O concentrations were reported to be rather low, close to ambient concentrations, and that production and consumption processes take place simultaneously (Pihlatie et al., 2007). Pihlatie et al., (2007) report the lowest N 2 O concentration (net N 2 O uptake) during spring and elevated topsoil N 2 O concentration from mid-summer towards autumn. -The roots of spruce are traditionally assumed to grow mostly close to the soil surface (e.g. Konôpka et al. 2010, Puhe 2003, whereas the roots of birch grow in both vertical and horizontal directions (Huikari 1959). Hence, it is possible that the roots of Norway spruce are mostly located in the topmost organic layer where most of the soil N 2 O production and consumption also takes place (Pihlatie et al., 2007).This may be reflected in the tree stem fluxes, especially if the trees transport N 2 O from the soil to the atmosphere, as supported by our data and suggested in laboratory studies . -If we assume that N 2 O is mostly produced in the soil and transported via the trees to the atmosphere, we would not expect to see high burst N 2 O emissions from tree stems due to the rather low N 2 O production rates in the soil. We have to keep in mind, however, that we did not quantify the N cycling microbial activities in the trees, in tree wood, or on the surface of the trees and hence we cannot say anything of their potential role in the tree-N 2 O-exchange. -From previous research at the site, we know that the site is N limited, meaning that there are negligible amounts of nitrogen available for microbial N turnover processes (Pihlatie et al., 2007;Korhonen et al., 2013). Ambus et al. (2006) measured gross nitrification and denitrification rates and N 2 O production pathways of forest soils within Europe, and they showed that at the Hyytiälä SMEAR II site the soil has very low nitrification activity and that most of the soil N 2 O is produced via denitrification. Assuming this N limitation for the plants and microbes, the available NH 4 + or NO 3 ions for microbial N 2 O production are limited, and hence no bursts in soil N 2 O production are expected. This is supported by the earlier field measurements studies (Pihlatie et al., 2007;Korhonen et al., 2013).
-We present new and novel findings of strong seasonality in tree stem N 2 O dynamics of mature trees within boreal forests. We measured tree stem N 2 O fluxes from three dominant tree species in combination with a multitude of environmental and tree physiological variables. This unique data allowed us to link the N 2 O fluxes to their drivers. For the first time, we also demonstrate how boreal trees act during the dormant season and how this is reflected in the annual N 2 O budget. -The mechanism of N 2 O transport from the soil through the trees via transpiration stream has been demonstrated in laboratory conditions (Rusch & Rennenberg 1998;Machacova et al. 2013). In the field, N 2 O exchange of tree stems have been studied during growing seasons (Machacova et al., 2016;Machacova et al., 2017;Wen et al., 2017), whereas no other studies have captured the N 2 O dynamics also during the dormant season, which can cover several months of the year in the boreal region. Accordingly, there is a lack of data and research on the seasonal variation and driving factors of the seasonality. In addition, results from laboratory studies cannot be directly applicable in the field. Trees in natural N-limited forests have not previously been expected to participate in forest ecosystem N 2 O exchange because the soil N 2 O production is negligible, and, assuming the only source of N 2 O from tree stems is that being transported from the soil, the available N 2 O to be transported is minimal. In our study, we present the first findings that trees in nutrientpoor environments take part in N 2 O exchange. We further show that this N 2 O exchange has a strong seasonality, which is similar, albeit different in magnitude, among the three boreal species studied. -All these facts lead us to believe that the results presented are innovative.
-We addressed this concern of the reviewer in several parts of the manuscript by strengthening the message, as suggested also by Reviewer 1. These adjustments are visible, for example, on Page 2, rows 54-66 and Page 6, rows 237 onwards.
What is physiological activity? What are physiological parameters? Soil water content, soil and air temperatures, and photosynthetically active radiation? To my knowledge, the authors measured environmental factors but did not measure tree physiological parameters. I do not think the authors studied the physiological activity and its direct relationships with seasonal nitrous oxide fluxes.
-Tree stem CO 2 exchange, hereinafter and in the manuscript expressed as stem CO 2 efflux, is one indicator of tree physiological activity inasmuch as the net stem CO 2 efflux is a result of stem respiration, CO 2 diffusion due to water transport from the root zone to the atmosphere via the transpiration stream of the trees, and CO 2 re-fixation on the stem surface. Hence, the dynamics of stem CO 2 efflux directly reflect the respiratory activity of the tree stems and partly also the transport of water from the soil to the canopy. As a water soluble gas, CO 2 can be transported via the transpiration stream and part of the transported CO 2 then diffuses through the stem against a concentration gradient between the water in the wood and the atmosphere.
-As described above, the CO 2 efflux of tree stems is not only stem respiration but also CO 2 transported from the root zone (Aubrey & Teskey 2009;Hölttä & Kolari, 2009;Bloemen et al. 2013). Aubrey and Teskey (2009) suggest that a substantial part of below-ground autotrophic respiration (of trees) is transported from the root zone through the stems and emitted to the atmosphere from tree stems. Hölttä and Kolari (2009) show based on field data and process modelling that the stem CO 2 effluxes are comprised of both stem respiration and transport via xylem sap. Typically, CO 2 efflux measured at the bottom of tree stems, as in our study, underestimates the stem respiration due to transport of produced CO 2 through the transpiration stream, whereas CO 2 efflux measurements in the upper part of the tree stems overestimate the stem respiration due to CO 2 accumulation in the tree stems. The proportion of stem CO 2 effluxes originating from stem respiration and xylem transport cannot be generalized, however. As N 2 O has good water solubility comparable to that of CO 2 , we can state with confidence that xylem transport is equally possible for N 2 O, and hence the N 2 O emitted from tree stems can originate from internal N 2 O production and diffusion of N 2 O dissolved in the transpiration stream, which has been demonstrated in the laboratory . Strong seasonality drives the sap flow and transpiration rates of trees, and hence, as demonstrated in our study, it can be expected to have a strong effect on the stem N 2 O exchange.
-Gross primary productivity (GPP) and evapotranspiration are indicators of ecosystem activity. GPP is derived from net ecosystem exchange (NEE) and modelled ecosystem respiration (Reco) (GPP = −NEE + Reco). NEE is the ecosystem-scale net CO 2 exchange of the forest, which mainly consists of tree CO 2 uptake in photosynthesis. Reco is derived from night-time NEE measurements and hence indicates the net respiration of the forest, which consists mostly of soil respiration and up to 50% of tree and ground vegetation respiration during summer months ). These are measured above the forest canopy and largely indicate the activities of the trees. Hence, we can confidently use these ecosystem measures as indicators of tree physiological activity. -We modified the paragraph stating our study objectives to define which parameters we understand to be indicators of tree physiological and ecosystem activity in our study (page 2, rows 67-83). Furthermore, we emphasize this issue throughout the manuscript, e.g. page 3, rows 115-119.
The purpose of this study is seasonal changes in stem nitrous oxide fluxes. Why are almost half of figures on carbon dioxide? I suggest that the figures on carbon dioxide can be deleted.
-We appreciate the concern of the reviewer that the focus of the manuscript should be more on the actual N 2 O exchange of trees. We have carefully considered the importance of all the figures in the manuscript and we suggest a solution that moves some of the figures to the supplementary material. -We nevertheless do regard it as critically important to show the stem CO 2 efflux dynamics to the reader, because stem CO 2 efflux was one of the two most important drivers of the stem N 2 O exchange. -We have now moved Figs 2, 8 and 10 to the supplementary material, as they show the monthly fluxes of CO 2 from stem and forest floor (Fig. 2), effects of vegetation and dormant season to the tree stem CO 2 fluxes (Fig 8), and the annual CO 2 fluxes in tree stems and in forest floor (Fig. 10). This work did not study the mechanisms of nitrous oxide production and emissions. The authors provided too much speculative discussion without evidence. I suggest that the discussion can be largely shortened and improved.
-We agree with the reviewer that this study is not a process study conducted in controlled laboratory conditions, which would enable detailed analysis of the N 2 O exchange processes. Unfortunately, mature trees cannot be taken into the laboratory, nor can the environmental conditions in the field be controlled to such extent that certain processes could be pinpointed. An additional limitation in the field in such a long-term measurement station as Hyytiälä SMEAR II is that the use of stable isotopes is forbidden in order to avoid contamination of the site for future studies. We nevertheless present a comprehensive data set of a newly identified N 2 O source in boreal forests, namely N 2 O fluxes from boreal trees, and we relate the N 2 O fluxes to a unique combination of tree physiological and environmental variables measured simultaneously at the site. The novelty of the study consists in the fact that there are no previous measurements of seasonality in tree N 2 O dynamics anywhere in the world. Another strength is that we link the field-scale tree N 2 O fluxes to a series of simultaneously measured variables, which allow studying N 2 O exchange process drivers and links to the physiological activity of the trees. -As suggested, we have carefully analysed the manuscript and removed statements that we consider overly speculative. We strongly believe, however, that our findings of strong seasonality in N 2 O emissions from boreal trees, as well as the links between stem N 2 O fluxes and stem CO 2 efflux, evapotranspiration, and GPP are unique findings and show for the first time that tree-N 2 O dynamics at field scale are driven by the activity of the trees and the ecosystem as a whole. To highlight this novelty, and also in responding to a request of Reviewer 1, we have strengthened the main message of our study. In addition, we have improved the description of the environmental variables with respect to the indicators of tree physiological activity (stem CO 2 efflux, evapotranspiration, GPP) (e.g. page 2, rows 67-83). We want to thank the reviewer for all his valuable comments which definitely highlight important aspects of N 2 O production and mechanisms in plants and mycorrhizal fungi. We have now carefully read through and point out these aspects in the discussion.

a) Contribution of mycorrhizal fungi to the tree-derived N 2 O emissions
We have added a discussion concerning the possible direct and indirect role of mycorrhizal fungi in production of N 2 O in rhizosphere and therefore also in plant uptake of N 2 O and its emission into the atmosphere (page 4, rows 153-163). Detailed discussion of this complex topic would go beyond the main message of our article.
The response to my last point -the absence of N2O burst events during CO2 burst events-was answered in the correspondence letter with regard to site-specific parameters at the research station.
Moreover, respiration is a common parameter for biological activity. Increased biological activity-for example caused by an increase in temperature-goes along with an increase in respiration. This is why respiration (of CO2 fluxes) are often presented along with other data like N2O of CH4 fluxes. The robust relationship between respiration and plant-derived N2O emissions is documented in Lenhart et al. (2015) for cryptogamic covers and in Lenhart et al. (2018) for higher plants. Please add this information to the manuscript.

b) Robust relationship between respiration and plant-derived N 2 O emissions
We added information on strong relationship between CO 2 efflux and plant-derived N 2 O emissions from literature (incl. new citations, thanks the Reviewer for the references) and placed it in the context of our results (page 3, rows 118-127).
I think that investigating the mechanism of nitrous oxide production and emissions are far beyond the scope of this study. Investigating mechanisms of N2O production and transport processes deserves 15N  Hope you find this perspective useful.
We would like to thank the reviewer for pointing out the aspect of the key message and its clarity improvement. Our key message is that the newly detected seasonal dynamics of stem N 2 O exchange of boreal trees follow their physiological activity. a) We revised the title of our manuscript as follows: Seasonal dynamics of stem N 2 O exchange follow the physiological activity of boreal trees. We considered also the formal requirements on title with word limit of 15 words. b) We reorganised and modified the text of the abstract to highlight our key message and to improve the clarity of the main message to readers. c) We added a final paragraph in the introduction section summarising the major results of our study and highlighting the key message of the study (page 2-3, rows 86-92; also in accordance with formal requirements of Nature Communications). We would like to thank the reviewer for all the comments. We pointed out in our previous responses to the reviewers' comments, why the CO 2 exchange of tree stems can actually be used as an indicator of tree physiological activity. We also state in the introduction of the manuscript that the "stem CO 2 effluxes as well as ecosystem gross primary productivity (GPP) and evapotranspiration were considered as indicators of physiological activity" (page 2, rows 78-80). Thus, we believe that it is clearly stated in the manuscript what was measured, and to what parameters we refer to with the physiological activity of trees.

Comments from the editor -article templates
At the same time, when revising your paper, please consider our article templates for the main text: -The title which must be under 15 words, with no punctuation.
The title was shortened and modified in relation to Reviewer 4`s comments.
-The abstract (less than 150 words). It should include the background and context of the work, 'Here we show' or an equivalent phrase, and then the major results and conclusions of the paper. It must not contain references.
We improved the clarity and structure of the abstract, also according to Reviewer 4`s comments.
-Introduction (<1000 words), which must include the background and rationale for the work. The final paragraph should be a brief summary of the major results and conclusions. The results of the current study should only be discussed in this final paragraph.
We included a final paragraph (page 2-3, rows 86-92) to introduction part summarising the major results of our study and highlighting the key message of the study.
-Results, which must be split into subheaded sections, ensuring that the subheadings are no longer than 60 characters including spaces. Subheadings should contain no punctuation.
We shortened the first subheading in the result section (page 3, row 95).
-Discussion, without subheadings. The final paragraph of the Discussion should be your concluding paragraph.
Our paper contains one section "Results" presenting our results together with their discussion. In April, we asked the editor of Nature Communications Dr. Eithne Tynan concerning the format requirements and Results and Discussion section. She gave us the following answer: "Indeed our style marks separate "Results" and "Discussion" sections. However, we allow authors to just have a "Results" section and have discussion interweaved with the description of results. So we don't mandate that you have a "Discussion" section per se, but the section should be called "Results" not "Results and Discussion", and have the appropriate subheadings." The final paragraph of the Results is indeed our concluding paragraph (page 8, rows 331-338).
-Methods, which must be split into subheaded sections, ensuring that the subheadings are no longer than 60 characters including spaces. There is no word limit for the Methods section.
Our method section meets the above mentioned requirements.