| 02 Sep 2026
Biogenic volatile organic compound (BVOC) emissions from boreal forest floors during the first years after wildfire: effects of stand age and management
Abstract. Boreal forests emit a wide range of biogenic volatile organic compounds (BVOCs) that influence atmospheric chemistry and climate through oxidation reactions and the formation of secondary organic aerosols. These emissions originate from plant metabolism, microbial activity, and storage pools such as litter, and vary in response to environmental conditions and disturbances. Wildfires, the dominant natural disturbance in boreal ecosystems, substantially alter ecosystem structure and carbon cycling, yet their effects on BVOC emissions during the early post-fire years remain poorly understood.
Here, we quantified BVOC emissions during the first years after wildfire across gradients of stand age, fire severity, and post-fire management. BVOC emissions were strongly controlled by tree mortality and management practices. High-severity fire in young stands resulted in persistently low emission rates and a shift away from monoterpene dominance, likely reflecting limited biomass and substrate availability. In contrast, sites with surviving mature trees exhibited elevated emissions dominated by monoterpenes, driven by litter-derived storage pools. Salvage logging caused a transient increase in emissions in the first year after fire due to accumulated needle debris, followed by a rapid decline as litter inputs were depleted. Conversely, reforestation led to increasing emissions over time.
These results demonstrate that early post-fire BVOC dynamics are regulated by the interaction between disturbance severity, stand age and management strategies, leading to distinct emission trajectories in the early recovery phase. Such shifts in BVOC emission magnitude and composition may have important implications for post-disturbance atmospheric chemistry and climate feedbacks. Our findings highlight the need to account for disturbance legacy effects and management practices when representing BVOC emissions in large-scale models.
The manuscript by Jaakkola et al. investigates BVOC emissions from Swedish boreal forest floors during the very early stage of post-wildfire succession, with particular interest in the effects of stand age, fire severity, and post-fire forest management strategies. The measurements and data were collected during campaigns conducted in 2019- 2012, that requires a lot of efforts I believe. The repeated measurements over several years collected valuable information on post-fire BVOC emissions and dynamics, which fills the remaining study gap.
However, I have several concerns regarding the experimental design and/or statistical analyses, comparability of observations across years and sites and soil layers, and interpretation of the emission patterns. In particular, the study design makes it difficult to separate the effects of fire severity, stand age, and management practices, while several mechanistic explanations appear to be stronger speculation than the evidence presented. The issues need to be addressed to improve the interpretation of the results and ensure that the discussions and conclusions are supported by the data. My major comments are seen below.
In Results and Discussion, the authors emphasize litter-derived storage pools as the main driver of BVOC emissions during the early stage of post-fire recovery. The authors suggest that initially elevated emissions are driven by fresh needle litter, followed by a decline as litter inputs are depleted and a later increase associated with vegetation recovery. Litter biomass, other BVOC storage pools, and root-associated BVOC production were not measured. The suggested sequence of litter-derived emissions, litter depletion, and vegetation-driven recovery would be better presented as a possible explanation rather than a general recovery pattern demonstrated by this study. The many underlying processes/ related factors were not quantified; therefore, the observed emission patterns cannot alone support the proposed ecological mechanisms. The absence of pre-fire BVOC measurements in this study should be kept in mind, and that limits the conclusions.
Minor comments:
Abstract
Line 11“Wildfires, the dominant natural disturbance”, what about “ a major natural disturbance”
Line 12: Although the authors acknowledge the limited understanding of BVOC emissions during early post-fire recovery, they do not clearly identify the specific scientific questions that remain unresolved.
Line 13: What specific time is referred to by “during the first years after wildfire”
Line 13: Please further clarify how stand age, fire severity, and post-fire management were defined and characterized in this study. Whether the reported emission patterns are supported by a chronosequence approach, which is important for interpreting post-fire emission trajectories.
Line 17: “…driven by litter-derived storage pools…” this is a too strong conclusion. This have to be checked from study methods and results.
Lines 15 -17: I would prefer to see the amounts of BVOC emssions in abstract.
Line 23: The implications for atmospheric chemistry and climate feedbacks would benefit from a clearer connection to the increasing frequency of forest fires under ongoing climate change.
Introduction
Lines 35 – 40: I suggest moving this sentence to a different part of the Introduction, as it seems somewhat disconnected from the above and below text. The discussion of BVOC emissions and composition across different soil layers would fit better with the content in Lines 41–50, where these topics are more closely related.
And the sentence in Lines 35 - 40 needs to be rephrased for scientific accuracy. In particular, the relationship between BVOC oxidation, secondary organic aerosol formation, and sunlight scattering should be clarified.
Lines 41 – 50: What is the role of plant roots and understory vegetation in forest floor BVOC emissions?
Lines 45–48: I don’t understand these. The explanation in these sentences is somewhat unclear. “Understory vegetation can act as a sink for BVOCs emitted from the mineral layer.” Why are the BVOCs taken up by vegetation originating from the mineral layer? Second, the previous sentence states that understory vegetation acts as a BVOC sink, but then the later sentence explains the reduced emissions through adsorption to organic matter and microbial consumption. These are different processes, and the relationship between them needs to be clarified.
Line 55: The distinction between fire intensity and fire severity should be clarified.
Line 57: Fluxes of CO2, what do they represent: soil respiration or something else? Please revise this sentence. And please consider focusing more directly on the effects of fire severity on BVOC emissions and forest floor processes, and why CO2 fluxes matter to the main focus of this study.
The Introduction would benefit from a stronger explanation of why fire severity and post-fire management practices are expected to influence BVOC emissions differently. In particular, the potential effects of different management strategies on vegetation, litter availability, and soil conditions should be briefly introduced to better establish the rationale for the study.
Line 70: BVOC emissions do not necessarily increase with biomass recovery, as they are also influenced by changes in vegetation composition, litter availability, and soil processes. This statement is quite general, and I think the later discussion is better, as it is supported by details and previous studies.
Lines 72 - 79: “Because forest floor BVOC emissions are largely driven by litter storage pools ...” The authors repeatedly emphasize that forest floor BVOC emissions are largely driven by litter storage pools. However, litter is not the only source of forest-floor BVOC emissions. Even during this early post-fire period, other vegetation, such as grass and herbaceous plants, may rapidly present after fire and contribute to BVOC emissions.
If this statement is based on observations from recently burned forests, I suggest clearly specifying that, in the earlier paragraph, it applies to the initial years after fire. Otherwise, the authors should avoid overgeneralizing the role of litter storage pools.
Lines 77 – 81: a couple of references are needed.
Line 91: The first hypothesis seems inconsistent with the earlier discussion suggesting that understory vegetation can act as a BVOC sink. If vegetation removal is expected to reduce BVOC emissions, please clarify whether understory vegetation is assumed to act primarily as a BVOC source under the conditions examined in this study.
Line 92: The second hypothesis assumes the organic layer may act as a sink for BVOCs, you have stated that the understory vegetation sinks BVOCs emitted from the mineral layer. The potentially contrasting roles of the organic layer and understory vegetation as BVOC sources or sinks should be more clearly distinguished.
Materials and Methods
Line 121: Please spell out the abbreviations when they are first introduced.
Line 122: Typo: should be “…dominated by vascular plants at LM…”.
Lines 121 -123: Referring to the 3rd hypothesis, “BVOC emissions increase with increasing vegetation cover.” You seem to have observed vegetation coverage in 2022, the last year conducted measurements, have you observed coverage from each measurement year for validating the increasing vegetation coverage?
Section 2.2, Lines 131–134: I am confused: “At each study site, ten collars were installed in May 2019. Two collars per site were selected for BVOC measurements of the organic layer ,….To assess emissions from the mineral soil, two additional collars were established at each site ….” How were the ten collars accounted for?
Section 2.2, Lines 131–134: The description of the experimental setup is somewhat confusing. Ten collars were initially installed at each site, but only two were selected for organic-layer BVOC measurements, and two additional collars were established for mineral-soil measurements. What were the remaining eight original collars used for? Please clarify how the collars were allocated and used in the study.
Lines 131 -135: the text does not match the contents in Table1.
Line 134, how did the organic layer remove before measurements? Please provide more details on the removal of the organic layer.
Line 147: Have you also monitored PAR conditions during measurements? That is shown in Figure 2.
Line 147: What is the chamber surface area?
Lines 150-151: Please clarify the sampling flow rate through the adsorbent tubes and explain why three tubes were used for each collar. Were the three samples collected as sequential replicates, and how were they treated in the subsequent data analysis? How long was the sampling time?
Line 158, “In total, 2014 samples were collected…” I think you mean 214.
Section 2.3: Some information on the BVOC sampling procedure is repeated between the end of Section 2.2 and Section 2.3, particularly regarding the use of three adsorbent tubes and replicate samples. I suggest consolidating the sampling details in Section 2.3 to avoid repetition and improve the organization of the Methods.
Line 168: Please check the abbreviation “TC-GC-MS”. Should it be “TD-GC-MS”, as TD for thermal desorption?
Line 169: Should the device information be served to “TC-GC-MS”.
Line 181: Please provide the equation used to calculate BVOC emission rates.
Lines 185- 190: I suggest bringing together all information related to environmental measurements in this section. For example, chamber air temperature measurements were already described in Line 147, whereas PAR, air temperature, and relative humidity measurements are introduced here.
Line 199: High BVOC concentrations do not necessarily indicate measurement errors, particularly in post-fire environments where emissions may be spatially heterogeneous and episodic. Please explain how the thresholds were determined, why these observations were considered invalid, and whether the excluded samples were concentrated in particular sites or years. Please also explain how the seven compounds were selected.
Line 210: Why regression models of log-transformed concentrations have been used for examining temporal trends in BVOC emissions?
Line 212: How were the compound-specific LODs converted into the corresponding emission-rate detection thresholds?
Lines 210 – 215: Please provide the full specification of the Tobit regression models, including the fixed and random effects, if applicable. Given that BVOC emissions were repeatedly measured from the same collars across years, please clarify how the non-independence of these observations was accounted for in the statistical analysis. And whether the three sequential adsorbent-tube samples were treated as independent replicates in the regression analysis.
Results
Section 3.1:
The authors interpreted that BVOC emissions from exposed mineral soil as the contribution of the mineral soil layer is not entirely accurate.
Whether the differences between soil-layer treatments were statistically evaluated. As the reported emission rates show substantial variability, particularly for exposed mineral soils.
Lines 217–220: “Overall, emissions increased over time, with a strong dominance of monoterpenes across all years.” The authors reported that BVOC emissions from the organic layer increased annually, whereas emissions from the exposed mineral soil decreased in 2021 before rising sharply in 2022.
Line 229: …with an increased contribution…
Lines 219, 220-221, 224, 231 & 238: You repeatedly state that monoterpenes dominated the BVOC emissions in organic and mineral soil layers in the UM site.
Lines 220- 232: I suggest the author rephrase the sentences for better logical order, the current version is difficult to follow because the sentences regarding organic layer emissions here and there, the same for the mineral soil layer emissions. The information in results would be easier to follow if the related content appears together with good, logical order.
Lines 231 – 233, 238–240: “...indicating a generally larger contribution from mineral soils.” is potentially misleading. The measurements represent net emissions from intact forest floors and exposed mineral soils, rather than the independent contributions of the two soil layers. Please clarify this distinction. Higher emissions from exposed mineral soil do not directly demonstrate BVOC uptake by the organic layer, particularly as the pattern was not consistent across all years.
Line 240: I disagree with “the organic layer reduces BVOC fluxes to the atmosphere”. The measurements were conducted over the mineral layers (collars without forest floor); there was no organic layer over the mineral layer, and the mineral layer was exposed by soil scarification, as that is what I understood from the previous sections. It is ok to say the emissions from the organic layer are less than those from the mineral layer, but I don’t think your measurements and results can prove that the organic layer reduced BVOC fluxes to the atmosphere.
Lines 241–243: “Wildfire severity strongly influenced BVOC emissions” seems too strong, given that fire severity, stand age, and management practices are not fully independent in the study design and study sites. Please distinguish the observed differences among sites from effects that can be directly attributed to fire severity. Also, “lower emission rates” would be more appropriate than “reductions”, as pre-fire emissions are not available.
Lines 244 – 246: Comparing the 2019 emissions at SHM with those measured at UM in subsequent years may be misleading.
Lines 247–251: The differences between HM and SHM they may reflect site-specific characteristics as well as management practices. Please avoid attributing these differences solely to salvage logging without further supporting evidence.
Lines 252 -259: The increase in BVOC emissions at HY in 2022 is described as a recovery pattern. However, an increase in emissions does not necessarily indicate ecosystem recovery. In Section 3.3 reports no significant temporal trend in total emissions at this site.
Lines 260- 254: WereS there any particular compounds or individual measurements driving the large difference between 2020 and 2021 at the LM site? Or did environmental variability cause such a big difference observed at the LM site?
Lines 271 -272: This sentence partly repeats the sentences in Section 3.1.
Lines 270 -274: You firstly stated that BVOC composition varied substantially across the sites and years. But in the following sentences, you stated the similarity at sites UM and LM.
Line 273: “Figures 31, S4a” please refer to the correct figure.
Lines 276 -281: The changes in BVOC composition are mainly presented as relative proportions. However, an increase in the relative proportion does not necessarily mean an increase in the absolute emission rate of that compound class. Please distinguish changes in compound contributions from changes in their absolute emission rates, particularly at sites where total BVOC emissions varied substantially between years.
Line 278: “In 2019 ... in 2019” repeated phrase in one sentence.
Lines 282–284: Please check the references to Figure 4 panels. According to the figure caption, panels b and c represent organic layers, whereas panels f and g represent mineral layers. The panel references in this sentence appear to be reversed.
Lines 305, 311: No significant trends in BVOC emissions at sites HY and SHM. The emissions at the two sites declined in 2020 and increased in 2022. It is unclear whether a significant temporal variation in the emission, and whether a linear temporal trend is appropriate for these sites.
Lines 300 - 305: “Several compounds contributed to this increase” and “These declines were driven by decreases”; rephrase them.
Lins 310 – 313: The sentence needs to be rephrased.
Discussion
Lines 325–332: The climatic control on BVOC emissions is possible, but the statistical analysis does not directly support it. I suggest that the relationships between environmental variables and BVOC emissions be evaluated.
Lines 337 - 340: Since no measurements were available from the unburnt reference site in 2019, direct comparisons with unburnt conditions are only possible for the subsequent years. In addition, differences in stand characteristics should be considered when interpreting this result.
Line 341 – 344: The comparison between SHM emissions in 2019 and the mean emissions at UM during 2020–2022 is difficult to interpret because of the substantial interannual variability at the reference site. Please reconsider the relevance of this comparison
Line 346: “depend strongly on forest stand characteristics and management”, your results do not fully support that. Instead, your results suggest that post-fire BVOC emissions on the forest floor varied substantially among forest stands with different characteristics and forest management strategies.
Line 350 – 353: I am not fully convinced that the organic layer acts as a sink for BVOCs emitted from the underlying mineral soil. As described in Section 2.2, the organic layer was removed to measure mineral soil emissions. However, higher emissions from exposed mineral soil do not directly demonstrate a sink effect, as removing the organic layer may alter soil conditions and BVOC production or consumption processes. Although microbial consumption and adsorption within the organic layer may reduce BVOC emissions, these processes were not directly quantified. I suggest interpreting the differences in net BVOC emissions between intact forest floors and exposed mineral soils more carefully, rather than concluding that the organic layer acts as a BVOC sink.
Additionally, a general comment for the Discussion. I suggest being more careful when stating that a hypothesis is supported based on observations from only one site or selected years. The conclusions should reflect the overall patterns observed across the study sites and years, as well as any inconsistent results. If a hypothesis is supported only under specific conditions, this should be clearly stated rather than generalized to the entire study.
Line 360: Since the sink effect was not directly quantified, it is difficult to determine whether increased BVOC production by litter inputs and warmer weather exceeded the assumed BVOC sink effect.
Lines 367–375: Root-derived BVOC production and mycorrhizal activity were not measured concurrently with BVOC emissions, a demonstrated explanation.
Lines 376–380: Have you measured plant growth or root development? Also consider other potential explanations for the observed emission increase.
Lines 383–385: This statement seems too strong, as the independent effects of fire severity, stand age, and management practices cannot be clearly distinguished in the current study design. In addition, the study did not directly quantify or report vegetation mortality, litter inputs, or organic matter availability.
Lines 391–401: The proposed role of needle litter in driving the elevated initial emissions and subsequent decline is a possible mechanism, but it is not directly demonstrated in your results. The dominant compounds, α-pinene and 3-carene, do not confirm a litter-derived origin.
Lines 396 – 400: needle litter depletion after 2020 was directly observed or quantified?
Line 407–414: BVOC emissions have not shown an increase over the measurement years; this does not necessarily indicate a lack of ecosystem recovery.
Lines 415–419: This conclusion is too strong. Other sources, including remaining vegetation, roots, and microbial processes, may also contribute.
Lines 420–422: I am not fully convinced that the observed increases at HY and SHM provide sufficient support for Hypothesis 4. T
Table 1: I am wondering at what age the forest stands are defined as young and what age is for mature forest stands; please clarify that in the related section.
Also, in Table 1, total FF depth and charred FF depth: how do you use them for characterizing the forest stands? Are they characteristics for classifying fire severity?
Fire severity appears to be classified primarily based on tree mortality. Please clarify whether other indicators, such as organic layer consumption, charring depth, or vegetation damage, were also considered.
Table 1: Please double-check the reported charred forest floor (FF) depth at SLM (0.11 ± 1 mm), as the value and its precision seem unusual.
Figures 3–4: Since the study aims to compare fire severity and post-fire management effects, I suggest including the low-severity sites (LM and SLM) in the main figures rather than presenting them only in the Supplementary Material. This would make comparisons across all study sites easier to follow.
Figure 5: The y-axis uses a logarithmic scale, and the figure caption claims the BVOC emission rates are presented on a log scale.
Table S2: The last column provides compound-specific LODs, but the authors mentioned that different analytical devices were used in 2019 and 2020–2022; the two devices cannot have the same set of compound LODs.
The LODs are reported in ng, a mass unit, not gas-phase concentrations or emission rates.