On behalf of the authors, we would like to express our gratitude to the reviewer for their time. In the following text, the reviewer's comments are in italic, followed by our response:

In general, I find the manuscript interesting. The methodology is sound, and I see merit in the study. However, I am concerned that the manuscript is, at times, overly lengthy and difficult to follow, which makes it hard to read overall. In several sections, critical information is either difficult to locate or entirely absent.

Response: Thank you for your positive assessment of our methodology and the overall merit of the study. We agree that certain sections were overly lengthy and will revise the manuscript to improve clarity and focus.

I believe the authors could present more concisely what I see as the key result of this study: why and how evapotranspiration (ET) and gross primary production (GPP) decouple under anomalously dry conditions.

Response: Thank you for this helpful observation. While we did not originally describe our findings using the term “decoupling,” we agree that the different responses of ET and GPP during the drought year are one of the important results. In particular, GPP remained relatively stable while ET dropped, which led to a strong increase in EWUE. We will make this point clearer in the revised version.

Moreover, the Discussion section often repeats similar sentence structures (e.g., “These findings/results highlight that…”) and reiterates basic, well-established principles of ecosystem functioning. This repetition detracts from the overall readability and does not add new insights.

Response: We appreciate this observation and will revise the Discussion to avoid redundant phrasing and unnecessary repetition of well-known concepts.

The authors conduct numerous comparisons with other broadleaved forest ecosystems, which, I assume, are mostly not riparian systems. They attribute differences in net ecosystem exchange (NEE) or other variables to various factors such as soil nutrient availability or vaguely defined climatic variability. These comparisons sometimes feel overly detailed and only loosely connected to their own results. I recommend refining this section to focus on comparisons that directly support their findings.

Response: Thank you for this comment. Our aim with the comparisons was to provide context for the magnitude of carbon fluxes observed at our site, particularly given the scarcity of published NEE values for riparian alder forests. While we agree that most comparison sites are not riparian systems, we selected broadleaf forests and hemiboreal or boreal forests to offer a general sense of how our site fits within a broader range of ecosystems. In the revised version, we will refine this section by focusing on the most relevant comparisons and making clearer links between the observed differences (e.g., in ER or GPP) and site-specific drivers such as soil fertility, stand age, and hydrology.

Although I understand that specific data on other GHG fluxes (e.g., CH₄ and N₂O) are not available for the site under study, I believe the authors should at least qualitatively discuss these fluxes. Making educated assumptions about their potential roles in riparian ecosystems would strengthen the manuscript’s conclusion that riparian alder forests could contribute meaningfully to climate mitigation through carbon sequestration.

Response: Thank you for this suggestion. While other GHGs were not the focus of this manuscript, the measurements of CH₄ and N₂O were conducted at the same site in 2018 and 2019 (and published in Mander et al., 2021,2022), and we agree that including a brief discussion of these gases strengthens the interpretation of riparian alder forests in the climate mitigation context. The site was found to be a net methane sink in 2018-2019 and a source of N₂O. However, the N₂O emissions were relatively small and offset only 1–6% of the net CO₂ uptake (when expressed in CO₂-equivalents). We will summarize these findings briefly in the revised Discussion to provide a more complete picture of the site’s potential for climate change mitigation.



Despite these issues, I find the paper interesting and within the scope of Biogeosciences. It should be considered for publication after thorough revisions. I have included specific comments below, but I would like to emphasize that the authors should carefully revise the arguments, logic, and structure, particularly in the Results and Discussion sections, to improve readability and clarity in the next version.

Response: We thank the reviewer for this encouraging evaluation and will work to improve the clarity and scientific focus of the revised manuscript accordingly.

Specific comments:

Introduction:



The conclusions of Paragraphs 1 (l.26) and 2 (l.36) could be more clearly distinguished. As written, both highlight the need for monitoring with respect to drought response and carbon sink capacity, but without clearly separating their specific focuses (i.e., short-term drought effects vs. long-term C-sink function).

Response: We agree that the conclusions of these two paragraphs were not clearly distinguished. In the revised manuscript, we will rephrase these paragraphs to better separate their main messages. The first paragraph will focus on the long-term role of forests as carbon sinks and the factors that can weaken or reverse it over time. The second paragraph will introduce drought as a short-term but increasingly common disturbance that can significantly affect carbon and water exchange in forests. We believe this adjustment will help clarify the structure and motivation of the Introduction.

Line 50: You state that "few studies" exist, but provide no citations. Does this imply that no studies have been published yet? Please clarify or provide supporting references.

Response: Thank you for this observation. Our intention was to emphasize the lack of ecosystem-scale studies on carbon and water fluxes in alder forests, particularly those using the eddy-covariance method. To our knowledge, no EC-based estimates of carbon fluxes or water use efficiency (WUE) have been published for alder forests at the time of the manuscript preparation. The only ecosystem-scale carbon balance study we are aware of is Uri et al. (2017), which used a biometric approach to estimate net ecosystem production (NEP) across an alder chronosequence. We will clarify this point in the revised manuscript and cite Uri et al. to better support our statement.

Line 54: It would be helpful to briefly explain what is meant by "the traditional C budgeting method" to provide context for readers unfamiliar with the term.

Response: Thank you for this suggestion. In the revised manuscript, we will briefly explain that the “traditional C budgeting method” refers to an approach based on biometric measurements, such as stand biomass, production and litterfall, to estimate NEP, rather than direct flux measurements such as eddy-covariance.



Line 53: This paragraph is somewhat unclear. You suggest that climatic factors outweigh stand age in importance, yet both old and young stands are described as carbon sources. Does this imply differing climatic conditions between the sites? Please clarify. If the implication is that the sites differ in climate, that should be explicitly stated. Furthermore, without detailed knowledge of the study by Uri et al. (2019), one could infer that its findings, based on “nutrient-rich former agricultural land” (l.319), may not be broadly applicable to typical riparian alder forests, which are unlikely to share these conditions. While this may not be the case for your study, the question arises whether your results are representative or overly site-specific. Consider expanding this paragraph or the relevant discussion section to clearly position your site within the broader context of existing research, particularly when findings from other studies appear to diverge.

Response: Thank you for such a detailed comment. We agree that our use of the term “climatic factors” was misleading. What we actually meant was that interannual differences in weather during the measurement years had a stronger influence on NEP than forest age (as stated by the authors in Uri et al 2017). We will revise the wording in the manuscript to clarify this.

Regarding the site representativeness: our study site is located on former agricultural land, while the mature site in Uri et al. was established on former grasslands. We will clarify this distinction in the manuscript. We will also briefly reflect on the implications for generalising our results.

Lines 60f: If you cite your previous research as a foundation here, please briefly summarize its key findings. This paragraph is currently difficult to follow. It doesn’t explain how forests responded to a heatwave, nor why that is relevant to the current manuscript—aside from the mention that water fluxes were not considered previously. Consider including a paragraph that outlines previous findings and highlights the open questions your study aims to address. Then, consider reformulating your objectives for increased precision. In particular, Objective 2 (“different soil moisture regimes and WUE”) may already be included in Objective 1 (“quantify… water exchange… under varying hydroclimatic conditions

Response: Thank you for this suggestion. We agree that the paragraph describing our previous work was unclear and will revise it to more clearly summarize the main findings and their relevance to the present study. We will also refine the study objectives to reduce redundancy and improve clarity, as suggested.



Methods:

Figure 1: The overview figure is very informative. As there is free horizontal space, consider including an additional ground-level photo, perhaps of the instrumentation setup or the canopy. This would help everyone unfamiliar with the ecosystem visualize the site.

Response: Thank you for this thoughtful suggestion, we will search for a good photo.

Line 79: Minor detail—consider removing the term “total” since you’re only showing the 80% footprint. Out of curiosity: do the remaining 15% correspond to the river and the water bodies?

Response: We agree that the use of the term “total” was misleading in this context and will remove it in the revised manuscript. Regarding the remaining ~15% of the footprint, yes, it’s the river, but also some birch and spruce trees and a narrow section of the adjacent clear-cut. The clear-cut and spruce trees are located near the outermost edges of the footprint and are unlikely to significantly influence the flux measurements.

Lines 79f: It would be useful to include information on variation in stand height, stem diameter, etc. From the image in Figure 1, the stand appears relatively uniform. Quantifying this would strengthen your argument.

Response: Thank you for your suggestion, we will try to expand this section with additional information.

Lines 105f: This statement is surprising, as the manuscript has not yet introduced the discrepancy between soil and EC fluxes. Please clarify or provide context earlier.

Response: We agree that the reference to a discrepancy between soil and ecosystem-scale fluxes is abrupt without prior context. In the revised manuscript, we will briefly introduce this point earlier in the paragraph, clarifying that previous chamber-based measurements at the same site (Mander et al., 2022) showed differences compared to EC-based estimates, likely due to unaccounted advection. This motivated us to use the advection filtering.

Lines 122f: Consider briefly explaining your rationale for using the nighttime flux partitioning method (e.g. over the daytime method).

Response: We chose the nighttime-based flux partitioning method because it relies on directly measured ecosystem respiration, is widely used and appropriate for our relatively flat site with low nighttime advection after filtering. We will briefly include this rationale in the Methods section to clarify our choice.

Lines 129f: This paragraph feels a bit too short. A rationale for the analyses would be helpful. If it’s too lengthy for the Methods section, it could be placed earlier near the objectives

Response: We agree. In the revised version of the manuscript, we will add a brief explanation of why we chose these specific parameters, including a short rationale for the growing season length analysis.

Around Equation 1: It appears there’s an issue: the equation references ERday in the text, but this term doesn’t appear in the equation itself. Additionally, please explain how canopy light use efficiency was calculated and which variables were used. Lastly, please clarify what modifications were made to the model and cite its original source, not just studies that have used it.

Response: The ERday was indeed incorrectly mentioned in the explanatory text around Equation 1, it will be removed. The canopy light use efficiency (α) was obtained as a fit parameter of the rectangular hyperbolic light-response curve (Equation 1), which relates gross primary production (GPP) to incoming global radiation (Rg). This model was applied using a 5-day moving window on daytime half-hourly values with Rg > 15 W m⁻². The curve-fitting was performed using non-linear least squares, and only results with statistically significant fit parameters (p < 0.05) and R² > 0.5 were retained.

The model applied is a form of the original rectangular hyperbola by Michaelis and Menten (1913), without the ERday term, since it was already accounted for during the flux partitioning. We chose this form over the non-rectangular variant (e.g., as used by Musavi et al., 2017; Chen et al., 2023) because it resulted in a greater number of valid fits across all years in our dataset. We will expand the corresponding section in the revised manuscript.

Lines 143f: How exactly are start and end of the growing season defined - using a relative or absolute GPP threshold? Consider also explaining why growing season length is relevant to your analysis. Later (e.g., l.176), you note that growing season lengths do not differ significantly between years, but it’s unclear whether or how statistical tests were applied here. Please clarify. If the differences are statistically insignificant and not central to your main conclusions, you might consider shortening this section.

Response: The start and end of the growing season (GS) were determined using a curve-fitting approach, not an absolute or relative GPP threshold. Specifically, we applied the double-logistic fitting method to daily GPP values, following Gonsamo et al. (2013). The inflection points of the fitted curve were used to define the start and end of the GS.

We agree that the relevance of GS length to the analysis could have been made clearer. It was included to support the interpretation of interannual variation in C fluxes, particularly in relation to spring onset and autumn decline, which influence the total period of active photosynthesis and hence annual GPP and NEE. However, since the GS lengths in 2018 and 2019 differed by only nine days (179 vs. 170), and the 2017 GS onset was not captured due to the measurement start date, we did not apply formal statistical tests for GS length. We acknowledge that this could have been stated more clearly.

As suggested, we will consider shortening this section in the final manuscript revision, since GS length differences are small and not central to the main conclusions regarding interannual variability in fluxes.

Results:

Figure 2 (and all other figures): It looks like standard color palettes were used. Please ensure the color schemes are accessible to readers with color vision deficiencies. If needed, use color-blind-friendly palettes or add alternative line styles. Also, clarify whether panel d) represents net radiation or incoming radiation. Minor suggestion: you might consider omitting the year legend repetition across all panels as readers can refer back easily once it's introduced.

Response: We acknowledge the importance of accessibility in figure design, and we will revise the colour schemes across all figures to use color-blind–friendly palettes. The current colours are indeed the standard MATLAB palette, which appears to pass basic accessibility tests, but we agree that improvements can be made to enhance clarity for all readers. Regarding panel (d) in Figure 2, it represents incoming global radiation (Rg). We will fix it in the revised version. We also appreciate and agree with the suggestion regarding legend repetition.

Line 144: Consider using a different abbreviation for the correlation coefficient than "rs," as it could easily be mistaken for stomatal conductance commonly used term in flux studies.

Response: We used “rₛ” to denote Spearman’s correlation coefficient, following common statistical notation. However, we agree that this could lead to confusion with stomatal conductance in the context of flux studies. In the revised manuscript, we will replace rₛ with “ρ” to avoid ambiguity.

Lines 251f.: This paragraph is difficult to follow. It relies heavily on the brief mention of partial correlation analysis back in line 144, which readers are unlikely to remember without very close reading. Please reintroduce the purpose, method, and results of this analysis in a self-contained way here. A visual representation such as a scatterplot of the residuals  (just one of many possibilities) might help as well.  Currently, readers will likely jump to Figure 5 and find your interpretation hard to align with what’s shown. In this context, consider moving key results from the appendix table 2 into the main text, possibly in a visually more appealing way.

Response: We agree that this paragraph would benefit from greater clarity, and we will revise it to briefly reintroduce the purpose and method of the partial correlation analysis. We will also incorporate the key coefficients from Appendix Table 2 directly into the figure and/or accompanying text.

Line 264: The statement that “SWC was the leading driver…” is a bit surprising given that Fig. 5c2 doesn’t clearly support this. Is this conclusion mainly due to radiation?

Response: The conclusion is based on the partial correlation analysis controlling for radiation, where SWC showed the strongest correlation with GPP across all years. We will clarify this in the text to avoid confusion with the visual impression from Fig. 5c2.

Table 2: As mentioned earlier, please clarify the analysis conducted here. Also, since the table isn’t referenced in the main text, its purpose and contribution are unclear—please address this.

Response: We will clarify the purpose and method behind the analysis in Table 2 and reference the table in the main text. The table supports the interpretation of light and temperature response curves and will be better integrated into the Results section accordingly.

Line 269: You state that GPPmax was “slightly lower” in 2019, but Table 2 shows a difference of ~0.07, which is negligible. Please clarify this wording or interpretation.

Response: We agree with this comment and will revise the wording

Discussion

Line 315: Introducing NEP here is slightly confusing. Consider converting their values to NEE for consistency and improved readability.

Response: We retained NEP here to remain consistent with the original paper. However, we agree this may cause confusion, as NEP is not used elsewhere in the manuscript. Accordingly, we will convert these values to NEE for clarity.

Line 330: You suggest that rapid SWC fluctuations might reduce annual ER, shouldn’t this be testable based on your dataset, or is something missing?

Response: Thank you for this valuable suggestion. While rapid SWC fluctuations may influence decomposition and thus ER, our dataset does not allow direct testing of this hypothesis. Soil and air temperatures, strong drivers of ER, likely mask the subtler effects of SWC variability, and the absence of heterotrophic respiration measurements prevents us from isolating decomposition rates from total ecosystem respiration. We acknowledge this limitation and will highlight it in the revised manuscript.

Table 3: In the entry for the Swedish spruce forest (Lindroth et al., 2020), the GPP value is missing a minus sign. Overall, the formatting in Table 3 is inconsistent. Some entries include ± values, others list single values, and some show a range (“…”). There are also inconsistencies in spacing and in the use of decimal places versus integers. A general reformatting would help improve clarity. As currently formatted, it’s unclear whether you’re showing interannual variability or uncertainty, or whether “…” denotes a range. Please clarify.

Response: We agree that the formatting of Table 3 requires improvement for clarity and consistency. We will carefully revise the table to include missing signs, standardize the presentation of uncertainty or variability, and harmonize decimal places and spacing.

Line 337: Small note: The current sentence structure implies a direct connection between your results and the cited studies. Consider rephrasing for example: “Similar to other studies (e.g., Xy et al., Yz et al.), we observed that seasonal ET patterns were shaped by…” or omit the references if they are not directly aligning with your results.

Response: We agree with this comment and will revise the wording

Line 338: This statement is confusing. Your growing-season data (e.g., Fig. 3) shows ET is reduced during the drought year while GPP remains largely stable.

Response: Our reference to the close alignment refers to the overall seasonal patterns of GPP and ET as illustrated in Figure 4, rather than a direct indication of their quantitative coupling. However, this suggestion highlights the value of examining their correlation. We will consider including a brief analysis of the seasonal correlations between GPP and ET and discuss the findings in the revised manuscript.

Line 340: The sentence starting with “The mid-season decline in ET…” feels awkward, shouldn’t your analysis directly address and explain this pattern?

Response: We agree with this comment and will revise the wording, since our analysis confirms the pattern.

Line 350: A reported precipitation surplus in this riparian setting is surprising, given typically high evaporation. This is quite interesting, consider elaborating further.

Response: We appreciate the reviewer’s interest in this observation. While riparian systems are typically associated with high evapotranspiration, the apparent precipitation surplus during the 2018 drought likely reflects reduced plant water use due to stomatal closure under high VPD and soil moisture depletion. However, in the absence of runoff or drainage measurements, we cannot fully quantify the water balance, and our interpretation remains speculative. Furthermore, the ET fluxes presented were not corrected for lack of energy balance closure – that is, we did not adjust latent and sensible heat fluxes to match available energy, as measurements of net radiation and ground heat flux were not available. Although this may lead to an underestimation of absolute ET, we consider the interannual comparisons to remain valid, given that the methodological approach was applied consistently across years. We will clarify these limitations and elaborate on this point in the discussion in the revised manuscript.

Line 373: Just a curiosity in this context: how significant is soil water depletion between spring and summer in riparian systems? A brief discussion could be insightful.

Response: Soil water depletion between spring and summer in riparian systems can vary considerably depending on groundwater connectivity, precipitation patterns, and vegetation water use. In systems with strong hydrological connectivity to groundwater, depletion may be minor; however, under drought conditions or in systems with limited lateral or vertical recharge, significant drawdown can occur. At our study site, we observed a clear seasonal decline in topsoil moisture during summer, suggesting that even in this riparian setting, soil water depletion was substantial under dry conditions. We will briefly discuss this in the revised manuscript.

Line 377: You note that both GPP and ER decreased by 300 g C—is this the same amount for each, and does that mean NEE?

Response: Thank you for pointing this out, we will revise the wording for better clarity

Lines 394f: The discussion around the lack of a legacy effect and its occurrence in other ecosystems feels somewhat lengthy and secondary to your core findings—but I may be missing the relevance.  You demonstrate that physiological stress was present but relatively moderate, and that the ecosystem adapted and recovered quickly. This might be the key takeaway here. For readers less familiar with Nordic ecosystems, it might help to contextualize the severity of the drought in climatological terms. For example., was it a 10-year drought, 50-year event, etc.?

Response: We agree that the recovery is a particularly interesting finding, as previous studies have reported stronger drought impacts in the year following the event. While we did not observe such a legacy effect here, possibly because one year may be too short to detect lasting impacts, we think that the ecosystem’s rapid recovery is a key result. We will revise this section to be more concise and focused. We also thank the reviewer for highlighting the missing information on drought severity, which we will add in the revised manuscript. 

On behalf of the authors, we would like to express our gratitude to the reviewer for their time. In the following text, the reviewer's comments are in italic, followed by our response:

The study quantifies ecosystem-atmosphere carbon (C) exchange and evapotranspiration of a fertile riparian alder forest established on former agricultural land in Estonia by statistically analyzing nearly three years of eddy-covariance (EC) measurements. The data from the mature (ca. 40 yrs of age) hemiboreal forest covers the European 2018 heatwave, giving an opportunity to assess the response of alder forest to extreme hydroclimatic variability. The studied ecosystem shows reduced gross-primary productivity (GPP) and ecosystem respiration (ER), and improved ecosystem water-use efficiency (EWUE) during the most intensive drought period. On annual timescale, however, the net ecosystem productivity (NEP) was strongest during the drought year, and no significant carry-over effects on C exchange were observed in the first post-drought year. This suggests the alder ecosystem C balance is resilient to droughts, and that compensatory mechanisms (e.g. earlier growing season start etc.) during the year can have stronger impact on annual C sink than relatively short-term hydrological extremes.

Standard EC methodologies are used throughout, and the study design, data curation and applied flux gap-filling and partitioning methods seems sound. The only exception is that energy balance closure should be shown for each of the three growing seasons to increase the confidence on the low evapotranspiration (ET) values reported.

Response: Thank you for this valuable suggestion. We agree that assessing energy balance closure is important to support the interpretation of the observed low ET. Unfortunately, a rigorous evaluation is not possible due to limitations in our measurement setup - we lack direct measurements of net radiation and ground heat flux. However, we will try to provide an approximate closure assessment for all three years using available data. This analysis will be included in the appendix of a revised manuscript.

The study aims to reveal the inter-annual variability between hydrometeorologically contrasting years, particularly the effect of soil moisture content (SWC) and atmospheric dryness (vapor pressure deficit VPD) on ecosystem C flux components and water use characteristics. The study provides rather unique dataset from European hemiboreal alder ecosystem, and the scope fits that of Biogeosciences.

The main problems with the current manuscript (MS) are: 1) The statistical analysis applied are not well-suited to address and separate the impacts of SWC and VPD on other variability, whether due to seasonal cycle or due to correlations of these soil and atmosphere dryness-metrics with other environmental variables.  2) Because of this, the MS is too descriptive, and Discussion contains too many vague arguments that are not backed up with in-depth analysis or literature. Combined with some overly detailed (!) and repetitive parts in the Results section, this makes the MS a bit frustrating to read and it is hard to gasp the key points.

Overall, there is potential and the study can be a useful addition to the literature. However, additional analysis is needed to better reveal the short-term response of the ecosystem to progressing 2018 drought, and subsequent recovery.  The discussion can also be easily improved by better usage of literature to interpret the observed changes via physiological and biogeochemical knowledge. Some concrete suggestions of potential analysis are given below in the Detailed comments. I do not expect the authors to do all of them but provide them rather as ideas how to strengthen the analysis.

Response: We would like to thank the reviewer for their detailed and constructive feedback. We agree that in its current form, the manuscript is overly descriptive in some parts and could benefit from a focused, more causal analysis. In response to the suggestions, we plan to implement the following major revisions (with further minor improvements detailed in the section below):

1.    Energy balance closure and validation of ET and EWUE estimates:

We will assess energy balance closure using available (limited) data and include this analysis in the appendix. Additionally, we will review literature on ET from similar forest types and on alder leaf-level water use efficiency and stomatal control.

2.    Soil moisture thresholds:

We will calculate the soil saturation ratio and examine whether fluxes as well as GPPsat and EWUE exhibit threshold-type responses. This will allow us to identify physiologically meaningful drought conditions and use them to define ‘water-limited’ versus ‘non-limited’ regimes.

3.    Gs–VPD relationships under varying soil moisture:

We will implement the suggested analysis of surface conductance (Gs) response to VPD, stratified by soil moisture or soil saturation ratio levels, following the framework of Oren et al. (1999). This will provide insight into stomatal control under contrasting water availability.

4.    EWUE recalculation and clustering:

We will recalculate EWUE using GPP and ET from dry-canopy days during the active growing season. To isolate soil moisture effects, this analysis will be conducted within clusters of similar radiation and VPD conditions, thereby minimizing confounding effects of atmospheric demand.

5.    GPPsat under progressive drought:

We will subset the GPPsat data to include only periods with ample light and full canopy development, and assess how it evolves under progressive soil drying. This will allow us to more directly capture short-term drought impacts on photosynthetic capacity.

6.    Legacy effect:

We will make additional analysis by comparing flux responses across the three years under similar environmental conditions (e.g., soil moisture, VPD, temperature). If flux suppression persists in 2019 despite favourable conditions, this would indicate the presence of a legacy effect rather than just typical environmental variability.

7.    Results section revision:

When rewriting the results section, we will merge overlapping content (particularly in Sections 3.3 and 3.4) and reduce unnecessary repetition. We will structure the results more clearly around the key physiological responses and their environmental drivers.

Detailed comments:

L16: unclear to which time period reported ecosystem WUE represents.

Response: The reported values in lines 14-17 are annual values. The wording will be improved in the revised manuscript

L19: what is ‘in response to H2O limitation’? Do you mean response to VPD, soil moisture availability or the combined effect?

Response: Here, we meant soil moisture

L23-24: Natural climate solutions were not focus of the study and not addressed at all

Response: We agree, this part will be omitted from the revised manuscript.

L31-33: Sentence is vague and has no information; rephrase

Response: We agree, that the original sentence is too vague. In the revised manuscript, we will improve the wording to make it more specific and informative

L55: C sinks, while a young…

Response: Thank you, we will improve the wording

L64-65: Here and especially in results and discussion, the authors should pay more attention on the relevant timescales of the responses. Throughout the paper, it is often very unclear whether annual, seasonal or short-term variability (e.g. how fluxes respond to progressing soil water limitations, and how they recover after rainfall) is discussed. This is a major issue, and should be better addressed in the revised version

Response: Thank you for your observation. In the revised manuscript, we will pay extra attention to the timescales under study

L67: avoid buzzwords ‘nature-based solution for climate mitigation’ OR significantly deepen the discussion on the potential (i.e. impact, scalability) of using alder forests to improve land C sink on former agric. lands / and or to optimize riparian zone management.

Response: We appreciate your suggestion. In earlier drafts, we considered discussing the potential of alder forests as a nature-based solution for climate mitigation. However, as the manuscript evolved, the focus shifted more strongly toward characterizing the impact of environmental factors on water and carbon fluxes. To maintain a clear and consistent narrative, we decided to omit broader implications related to land-use policy and climate mitigation. The revised manuscript will reflect this.

L71: The site is on former agricultural land, so land-use history may have strong effect on soil C storage and thereby ER?

L78-84: Information on ecosystem leaf-area index (LAI), and site land-use and forest management history are missing. Are the above- and belowground C stocks quantified elsewhere?

Response: We agree with both comments above and acknowledge that land-use history can strongly influence soil carbon storage and ecosystem respiration. In the revised manuscript, we will try to expand the information of the site’s land-use history.

L99: Be consistent with terms NEP (used in introduction) and net ecosystem exchange (NEE)

Response: We agree with this comment and will use NEE throughout the revised manuscript.

L114-115: Bad sentence; I assume these percentages represent the data coverage?

Response: Thank you for pointing this out. The sentence was unclear, and we agree that clarification is needed. The percentages refer to the proportion of half-hourly NEE values that remained after quality control. We will revise the sentence to clearly reflect that these values indicate the data coverage for NEE following quality filtering.

L132: Calculation of EWUE needs more details. Was it estimated on 30min basis or from accumulated fluxes? What are the relevant timescales reported? Does it represent dry-canopy conditions or all conditions, and what are the impact of this choice on (physiological) interpretation of the results? If comparing the effect of soil drought on EWUE, shouldn’t you cluster the data into similar radiation and VPD conditions?

Response: Thank you for your valuable suggestions. Our initial approach to calculate EWUE at different time scales led to indeed a more descriptive paper and complicated the analysis of environmental drivers’ impact. Thus, we plan to implement the following steps in the revised manuscript:

1. EWUE will be re-calculated using daytime aggregated values of GPP and ET for the active canopy period, omitting rainy days to avoid biases due to canopy wetness affecting ET measurements.

2. To analyse the impact of drought on EWUE, we will follow your suggestion and cluster the data into bins with similar radiation and VPD conditions

L139 (eq. 1): Definition of ER_day is not relevant here

Response: The ERday was indeed incorrectly mentioned here and will be removed.

L143-144: The growing season was defined using GPP; thus it corresponds to carbon uptake period.

Response: In this context, we used the term “growing season” as defined in the original methodological paper by Gonsamo et al. (2013), which is based on GPP dynamics. Our intent was to capture the period of active photosynthesis, corresponding to canopy activity. The term “carbon uptake period” can refer either to the period of nonzero GPP or negative NEE (i.e., net carbon uptake), making it more ambiguous in this context.

L149-151: Sentences are unclear, and I wonder if using constant activation energy (E0) is consistent with the nighttime flux-partitioning method used?

Response: We thank the reviewer for this observation. We agree that the original wording was unclear. To clarify: in our application of the Lloyd and Taylor equation, the activation energy parameter (E₀) was estimated separately for each year, while the reference temperature (Tref) was set to 15 °C, and T0 was kept constant at -46.02 °C. This approach is consistent with the standard implementation of the nighttime flux-partitioning method. We will revise the sentence to more clearly describe this and to avoid confusion regarding the parameterization.

L160 Figure 2 panel (f): add cumulative precipitation on right y-axis; current barplot is not informative alone.

Response: We appreciate this comment and will improve the figure accordingly.

L176-178: The study focuses on inter-annual variability. What controlled the IAV of growing season length; is early onset of growing season related to high air temperature?

Response: We agree that higher air temperature is likely the primary driver of earlier growing season onset. However, our dataset is limited in this regard: measurements in 2017 began relatively late, so we effectively have reliable estimates of growing season start only for two years. This sample size is too small to draw statistically robust conclusions on the controls of inter-annual variability in growing season length. We will clarify this limitation in the revised manuscript.

L217-218: Example of vague text: what is meant by late autumn, winter and early spring?

Response: Thank you for this example. In the revised manuscript we will try to improve the wording and be more precise.

L220 & L222-223: What time periods the daily average NEE refers to? Is this information necessary for the study goals?

Response: We agree that daily average NEE values do not add essential information for the study’s main goals. Therefore, it will be omitted in the revised manuscript to improve clarity and focus.

L224-227: Example of unnecessary repetition of figures. Please interpret the figures causally using e.g. Fig. 2 instead of repeating their content. Same concerns to large extent whole section 3.3; that there is seasonal cycle in C uptake and ER is not particularly new. Consider merging Sect 3.3 and 3.4 to better link the changes in ecosystem fluxes to their drivers, to reduce repetition and to improve the clarity.

L233-239: Link ET variability to weather variability and plant phenological stage (LAI development) rather than repeat the figure in text.

Response: Thank you for these suggestions. We agree, that the results section in its current state has a lot of unnecessary repetitions. For the revised manuscript, we will merge the suggested sections to explain the changes in carbon and water fluxes together with their drivers.

L251-251: At which timescale and period? Core growing season or throughout?

To demonstrate the significance of stomatal control on GPP (and NEE) further, compute surface conductance (Gs) from measured ET, cluster it to conditions with ample light and show the dependency of Gs to VPD. You may see different shape of Gs-VPD curve or dropping reference conductance (Gs_ref) when soil is dry? Oren et al. (1999; https://doi.org/10.1046/j.1365-3040.1999.00513.x ) model Gs/Gs_ref = – m * ln(VPD), where Gs_ref is reference conductance at VPD=1kPa and m~0.6 provides theoretical grounds to compare the observed dependency.

Response: We thank the reviewer for this valuable and constructive suggestion. We agree that calculating surface conductance (Gs) from measured ET and examining its dependency on VPD under ample light conditions, along with soil moisture stratification, will provide important mechanistic insights into stomatal control of carbon and water fluxes.

Following the Oren et al. (1999) framework, we will compute Gs, cluster the data accordingly, and analyze the Gs-VPD relationship, including potential shifts in reference conductance under dry soil conditions. This analysis will be included in the revised manuscript.

Similarly, consider showing how e.g. GPPmax varies during progressing drought.

Response: Thank you, the corresponding analysis will be added to the revised manuscript

L253: rs is not defined

Response: rs here is a partial correlation coefficient. Will be added to the revised text. 

L259 Figure 5: Nice figure but interpreting the responses of NEE, components and ET (y-axis) to single environmental factors (x-axis) is complicated because you consider the whole growing season, meaning that e.g. high soil moisture conditions represent spring and autumn (Fig. 2). Same concerns temperature and VPD responses.

As you have evaluated GPPmax in a moving window, consider sub-setting the data so that you include only ‘the stable summertime when the canopy is fully developed’. This will enable better insights on the role of VPD and soil moisture as controls of ecosystem behavior?

I also suggest you explicitly show the response of NEE, GPP, ER and EWUE to VPD and soil moisture over 2018 (and maybe for other years as well) while selecting only conditions with ample light and temperature constrained to a narrow range (i.e. avoiding extremes?). Also, omit rainy periods.

Response: Thank you for this helpful observation! We agree that including the entire growing season in Figure 5 complicates the interpretation of flux responses to single environmental drivers. To address this, we will revise the analysis and figures by subsetting the dataset to include only the stable summertime period. Additionally, we will explicitly show the response of NEE, GPP, ER, and EWUE to VPD and soil moisture, focusing on the drought year 2018, subsetting ample light, narrow temperature ranges, and non-rainy periods to isolate the effects of water availability and atmospheric demand. 

L272: What is meant by ‘annual photosynthetic capacity’?

Response: As mentioned in the methods section, annual photosynthetic capacity is 95th percentile of each year’s GPPsat. However, these values seem meaningless for the overall discussion, so they will be omitted.

L277 Figure 6: Also here seasonal variability and short-term variability are mixed in panel b-d. What is the main message of this figure? Can it be improved e.g. by showing different years with different symbols and adding day of year as a color scale?

L289 Figure 7: Again seasonal variability overshadows responses to drought? In last panel, high EWUE occurs during rainy days in 2018. Are you sure this is not an artifact of underestimated ET measurement when the canopy is wet?

Response: Thank you for these constructive comments. We agree that in both Figure 6 and Figure 7, seasonal variability and short-term variability are currently confounded, making it difficult to isolate the effects of drought or other environmental drivers. To address this, we will revise the analysis by subsetting the data to to include only the stable summertime period without rain, as suggested in the earlier comments above. This filtering will help isolate physiological responses from seasonal trends and avoid artifacts caused by low ET values under wet-canopy conditions.

L301-311: This part would benefit significantly from separating seasonal cycle from more short-term drought impacts. Strengthen the arguments by use of literature.

Response: We agree that this section, as currently written, does not sufficiently distinguish between seasonal trends and short-term responses to drought. In the revised manuscript, we will reorganize this part and strengthen the interpretation by including supporting references.

L315-320: There is nearly order-of-magnitude difference in the net C sink of these two alder forests. Different land-use history is one plausible reason, but this is presented as a hypothesis as no references are given? At L323 it is noted that ER of the studied ecosystem is lower than comparable boreal and hemiboreal forests. Can this be due to the land-use change and depleted soil C storage – what does the literature tell us?

Response: Thank you for pointing this out. We agree that the role of land-use history and its impact on carbon cycling was underrepresented in the original manuscript, despite its obvious relevance. In the revised version, we will expand this part of the discussion and include supporting literature. 

L329-330: Two issues: 1) heterotrophic respiration was not quantified and therefore the argument is not backed up with the data. 2) GPP shows significant reduction in 2018 during the dry period (Fig. 4 & Fig 2) compared to other years. On annual / growing season scale GPP was not reduced, likely because of larger uptake in early warm spring season?

Response: We agree that the interpretation regarding suppressed heterotrophic respiration during the drought year is speculative. This statement will be revised or removed to avoid overinterpretation. We also acknowledge that GPP did decline during the peak drought period in 2018. It was indeed compensated by enhanced uptake in early spring, resulting in little change in total growing season GPP. We will revise the text.

L330-331: Rapid fluctuations in SWC… this is pure speculation the effect was not addressed.

Response: The sentence regarding rapid SWC fluctuations dampening decomposition will be removed since it was not tested in the current study.

Table 3: For the Danish beech forest (Soroe), cite original reference Pilegard & Ibrom (2020, https://doi.org/10.1080/16000889.2020.1822063) rather than Lindroth et al. (2020) drought synthesis

Response: Thank you for this comment, we will fix it in the revised manuscript

L340-346: The data (Fig. 5c4) indicates ET has bell-shaped but scattered response to soil water content (SWC). At the wet end (high SWC) this does not mean excess soil water content or limited oxygen availability would be restrict transpiration, as the conditions with high SWC cluster to early/late growing season days when evaporative demand is low (i.e. low available energy and VPD).

Response of stomatal conductance and thereby transpiration rate (proportional to LAI x gs x VPD) to soil moisture is typically highly non-linear and it would be interesting to see how this manifests itself in the data. If you see clear threshold-type response, that could be used to cluster data to ‘no drought’ vs. ‘water-limited’ regimes, to explore how Pmax, Gs, EWUE etc. differ when soil water content is limiting?

In practice: subset data for ample light, no rain, fully developed canopy LAI etc. and show Gs, GPP, ET, … vs. SWC, or preferably ‘soil saturation ratio’, i.e. SWC/porosity where porosity ~ upper percentiles of observed SWC.

This is an example of how to move from ‘descriptive interpretation of the flux timeseries’ into more physiologically relevant impact-analysis. See also earlier suggestion on additional analyses towards this direction.

Response: Thank you for this comment and your suggestions. We agree, that the seasonal variability overshadowed the impact of SWC. We will follow your suggestion to calculate soil saturation ratio and assess if we get any visible threshold-type response.

L347-352: Can/should you comment on the role of lateral water flows? You study a riparian forests so I assume those can be important for soil moisture dynamics especially in early growing season, leading to delayed depletion of SWC and thus mitigating for late summer drought stress. It is interesting that you still see such a strong drought response in ET.

Response: Although our site has a gentle slope of about 1%, lateral water flows can still be relevant, specifically for the riparian ecosystem. Such low-gradient areas promote subsurface lateral flow and shallow groundwater movement rather than rapid overland flow. These lateral flows can help maintain soil moisture by redistributing water within the soil profile and delaying drought stress, especially during the early growing season. However, since we lack direct measurements of lateral flows at our site, we can only briefly acknowledge their potential contribution to soil moisture dynamics in the revised manuscript.

L353-355: Argument on increasing transpiration may be true but remains fully speculative as ET partitioning was not done.

Response: Thank you for this important point. We acknowledge that without ET partitioning, statements regarding changes in transpiration remain speculative. While the separation of transpiration and evaporation components were out of the scope of the current study, we base our interpretation on the general understanding of ecosystem water flux responses to drought and stomatal regulation, supported by literature. We will clarify this in the revised manuscript to avoid overstating the conclusions.

L356-363: The EWUE in your study is very high, as shown by comparison with other forests. This is either due to surprisingly tight stomatal control of alder, or due to underestimated ET. Did you check the energy balance closure and evaluate whether the reported (low!) ET values are plausible compared to other forests in similar climate conditions? Can plant trait databases or publications on leaf-level water use efficiency provide support to your interpretation that water use of alder is extremely conservative, i.e. leaf-level IWUE = A/gs is high?

Response: Thank you for the valuable comment and suggestions! Unfortunately, in the absence of Rn and G measurements, rigorous energy balance closure assessment is impossible.  However, we will try to provide an approximate closure assessment for all three years using available data. This analysis will be included in the appendix of a revised manuscript. We will also explicitly acknowledge this limitation in the main text. Additionally, following your advice, we will review the literature for comparable ET values under similar climatic conditions and investigate available plant trait databases and studies on leaf-level water use efficiency of alder to better support our interpretation of the observed high EWUE.

L379: Here and throughout the MS: consider how many significant digits to report taking into account typical uncertainties

Response: We agree, and we will fix it in the revised manuscript

The current analysis is not well suited to detect the legacy effects of the 2018 drought  

L395: Why this would be a legacy effect and not just a typical response to environmental variability (e.g. VPD and soil moisture) over 2019 growing season?

Response: We agree that the current analysis is insufficient to clearly distinguish a legacy effect from typical responses to environmental variability during the 2019 growing season. In the revised manuscript, we will compare flux responses across the three years under similar environmental conditions (e.g., soil moisture, VPD, temperature). If flux suppression persists in 2019 despite favourable conditions, this would indicate the presence of a legacy effect rather than just typical environmental variability.

L397-398: What is meant by recovery phase? Was the ER higher due to environmental conditions, of because of there was excess of undecomposed young litter from the dry 2018 year? Argument is just handwaving.

Response: Thank you for your valuable observations. In the revised manuscript, we will improve the analysis (as detailed in our response to the previous comment) to provide a stronger basis for interpreting the observed patterns. However, due to the absence of litterfall measurements, we can only speculate on the potential contribution of excess undecomposed litter from the 2018 drought to the elevated ER in 2019. We will clarify this limitation and avoid over-speculation in the discussion.

L415-417: sentence needs backing from the literature reference.

Response: references will be added in the revised manuscript

L433: “extreme conditions” --> during limited water availability (or soil drought)?

Response: Thank you for this suggestion, we will improve the wording accordingly.