High variations of BVOC emissions from Norway spruce in boreal forests

Hakola, Hannele; Taipale, Ditte; Praplan, Arnaud; Schallhart, Simon; Thomas, Steven; Tykkä, Toni; Helin, Aku; Bäck, Jaana; Hellén, Heidi

doi:https://doi.org/10.5194/acp-2022-478

Preprints

https://doi.org/10.5194/acp-2022-478

Preprints

28 Jul 2022

| 28 Jul 2022

Status: this preprint was under review for the journal ACP but the revision was not accepted.

High variations of BVOC emissions from Norway spruce in boreal forests

Hannele Hakola, Ditte Taipale, Arnaud Praplan, Simon Schallhart, Steven Thomas, Toni Tykkä, Aku Helin, Jaana Bäck, and Heidi Hellén

Abstract. The biogenic volatile organic compound (BVOC) emission rates of Norway spruces published vary a lot. In this study we combined published Norway spruce emission rates measured in boreal forests (Meeningen et al., 2017; Bourtsoukidis et al., 2014a, 2014b; Hakola et al., (2003, 2019)) and added our new, unpublished emission data from southern (SF) and northern Finland (NF). Standardized summer monthly mean emission potentials of isoprene vary from below detection limit to 7 µg g^-1_(dw) h^-1, and monoterpene (MT) and sesquiterpene (SQT) emission potentials 0.01–3 µg g^-1_(dw) h^-1 and 0.03–2.7 µg g^-1_(dw) h^-1, respectively. In this study, we found much higher SQT emissions from Norway spruces than measured before and on average SQTs had higher emission potentials than isoprene or MTs. The highest monthly mean SQT emission potential 13.6 µg g^-1_(dw) h^-1 was observed in September in southern Finland.

We found that none of the younger (33–40 years) trees in Hyytiälä, southern Finland, emitted isoprene, while one 50-year-old tree was a strong isoprene emitter. However, this could not be confirmed at other sites since all measured small trees were growing in Hyytiälä, so this could also be due to the same genetic origin. On average, older trees (>80 years) emitted about ten times more isoprene and MTs than younger ones (<80 years), but no clear difference was seen in SQT emissions. SQT emissions can be more related to stress effects.

As shown here for Norway spruce, it is possible that the emission factor of SQTs is significantly higher than what is currently used in models, which may have significant effects on the prediction of formation and growht of new particles, since the secondary organic aerosol (SOA) formation potential of SQTs is high and this may have significant effects on the formation and growth of new particles. Due to high secondary organic aerosol (SOA) formation potentials of SQTs the impact on SOA formation and mass could be even higher.

Received: 05 Jul 2022 – Discussion started: 28 Jul 2022

Download & links

Hannele Hakola, Ditte Taipale, Arnaud Praplan, Simon Schallhart, Steven Thomas, Toni Tykkä, Aku Helin, Jaana Bäck, and Heidi Hellén

Status: closed

RC1:
'Comment on acp-2022-478', Anonymous Referee #1, 17 Aug 2022
Summary: Hakola et al present a comparison of old and new BVOC emission measurements from Norway spruce across various regions in the European boreal forest. They also model particle formation from the emissions data. They found that sesquiterpene emissions are generally higher than predicted in MEGAN. They also show variability in whether Norway spruce is an isoprene-emitter. Detailed BVOC emission measurements represent an important contribution to the field. However, the research approach and analysis presented in this paper are not ready for publication for reasons outlined below. The research presented contains fundamental flaws in justification, clarity, and technical approach.

Major Concerns

The paper lacks adequate context in the introduction. The authors start with a discussion of OH reactivity, but that is not the focus of the paper. One of the highlights of the research is probably the modeling of aerosol formation and growth, but there is no justification presented in the introduction for this research. The introduction was generally very difficult to read due to lack of logical flow and many grammatical errors.

The paper lacks clear research questions and/or objectives. They establish there is some uncertainty about BVOC emission rates from Norway spruce due to variation in previous measurements, but the new measurements really don’t address that. They are also highly variable. There is no justification for the modeling work. The manuscript would be strengthened by clearly outlining research questions/objectives and aligning the figures with those questions/objectives. As written, it lacks clear purpose.

The experimental section is very difficult to follow. It goes back and forth between discussing the new measurements and the old measurements. Even in the section titled, “2.1 New, unpublished emission measurements” the old measurements are discussed (see lines 190-191). Also, this should be labeled section 2.2, not 2.1. And the section labeled 1.1 in the experimental section should be 2.1.

The introduction states that the study is investigating if growing location or age of tree can explain variation in emission measurements (line 87), but this isn’t clearly addressed with the analysis. Furthermore, it does not appear that they had enough replicates of both old and young trees across all sites to even address this question. Addressing this question requires measurements from both old and young trees at multiple sites, which they do not have. Thus, there is a disconnect between the methods and one of the only clear objectives stated in the paper. The research objectives need to align with the sampling approach, but that is not the case.
Citation: https://doi.org/10.5194/acp-2022-478-RC1
- AC1:
  'Reply on RC1', Hannele Hakola, 13 Oct 2022
  We thank reviewer 1 for valuable comments and have tried to take them into account when rewriting the introduction and experimental sections.
  The paper lacks adequate context in the introduction. The authors start with a discussion of OH reactivity, but that is not the focus of the paper. One of the highlights of the research is probably the modeling of aerosol formation and growth, but there is no justification presented in the introduction for this research. The introduction was generally very difficult to read due to lack of logical flow and many grammatical errors.
  
  The introduction has been rewritten.
  The paper lacks clear research questions and/or objectives. They establish there is some uncertainty about BVOC emission rates from Norway spruce due to variation in previous measurements, but the new measurements really don’t address that. They are also highly variable. There is no justification for the modeling work. The manuscript would be strengthened by clearly outlining research questions/objectives and aligning the figures with those questions/objectives. As written, it lacks clear purpose.
  
  The purpose of this study is to find out if the existing measurement data could help finding representative emission rates for emission inventory and atmospheric impact research. We try to seek for example if the age of tree or a growing location would affect emissions and should be taken into account in atmospheric modelling research. This has been written in the introduction.
  The experimental section is very difficult to follow. It goes back and forth between discussing the new measurements and the old measurements. Even in the section titled, “2.1 New, unpublished emission measurements” the old measurements are discussed (see lines 190-191). Also, this should be labeled section 2.2, not 2.1. And the section labeled 1.1 in the experimental section should be 2.1.
  
  The section has been partly rewritten, also numbering has now changed.
  The introduction states that the study is investigating if growing location or age of tree can explain variation in emission measurements (line 87), but this isn’t clearly addressed with the analysis. Furthermore, it does not appear that they had enough replicates of both old and young trees across all sites to even address this question. Addressing this question requires measurements from both old and young trees at multiple sites, which they do not have. Thus, there is a disconnect between the methods and one of the only clear objectives stated in the paper. The research objectives need to align with the sampling approach, but that is not the case.
  
  The introduction has been rewritten taken into account the questions raised by the reviewer. The effect of growing location and the age of trees is discussed in chapter 3.2.6. It is true that we do not have replicates at all sites, the young trees were only measured in southern Finland in Hyytiälä, but 3 older trees were measured at both sites. The results were variating a lot.
  The study was meant to be done with the existing data, therefore the sampling approach is not always best possible.
  
  Citation: https://doi.org/10.5194/acp-2022-478-AC1
RC2:
'Comment on acp-2022-478', Anonymous Referee #2, 01 Sep 2022

The manuscript reports an important new viewpoint, especially on the emission of sesquiterpenes from Spruce and the potential consequences on new particle formation and SOA growth. Unfortunately, the manuscript is written kind of carelessly, like in a very big rush. There are obviously rearranging issues with non-fitting subsection numbering. At many places, the text jumps back and forth, or, terms are defined after they are already used. Explanations needed in the methods description appear in the results or even discussion as footnotes etc. All that lead to the point that it is not clear what the aim of this paper is! Maybe this is also the reason that the paper’s title is very superficial.

The measurement and data part would need a more clear focus how all goes together. At the moment it’s hard to assess if there are repeated measurements, how many? Some are repeated after longer time to check for changes by age etc. The quality of these data can not be assessed at the moment and therefore it remains unclear if the differences found are really effects or occurred by chance.

The modelling part would need some clarifications. Did you model monthly or daily time steps? Or both. Which model was used? Basically the first part of the presented equations are a temperature driven parameterisation according to Guenther, as stated. This part somehow deals with the scaling of the measured emission factors to scale them by temperature for comparison with published and older data, as I get it. I kind of understand that MEGAN was used to create another set of emissions that is processed with the second part of equations given for the aerosol growth predictions. In general, the modelling part does not give any grading on the quality of the model or models used in relation to the input parameters. Especially as there are many assumed relations like constant values per month in some input data, constant specific leaf mass, scaling factors between HOM and SOA yields and so on.

Specific comments:

Line 91, which models/equations were used to make this aerosol formation and growth calculations from measurements? The one you present or MEGAN?

Line 175ff, need more details here. First, do you use a “trap” or a “cold trap”? According the procedure described, you sample into a “trap” filled with adsorbents keeping the temperature high enough to avoid condensation. Then, after that, the sample is pushed to the thermodesorber part, that, usually means a heating up and then cooling down below zero before entering finally the column in a GC-MS system. Using a “cold trap” directly would mean, to my knowledge, to use low temperatures while sampling.

Line 215ff, here, it’s a bit puzzling because you describe the emission factor with a unit of ng per g dry weight and hour, I guess. Then in the next section you use the emission factor with micro gram per square meter and second. The conversion factor comes very very late (line 637) in the footnotes of table 5. I would leave some information already here that this conversion is needed to get on to the next step, the model.

Line 250ff, in the section, you describe the constraints, i.e., data used for the parameterisations. It is very puzzling which time scale was used, daily or monthly or both? You use “monthly median”, “daily maximum” and “monthly median of daily maximum” to express these scales but there is no clear description how the model is using them?

Line 303, the temperature is linked with the reaction rates in table 1. It’s better mentioned there in the caption, not together with the chemical reaction schemes.

Line 325ff, here, you present the gamma values, and that these have multiple values possible as I understand, they are further numbered as well and state to present HOM, or as SOA yields. However, that makes the need to have another constant (2.2) in to account for. It may be ok, but somehow this lacks a clear reasoning and description of how such changes in model parameters will impact the model’s output.

Line 358ff, it’s a bit strange to define CC after it is used in an equation, that makes the reader searching for it in subsequent equations. Usually, if given in that notion, this is a “side-parameter” and do not need and have an own equation number. It would be beneficial to introduce it before, i.e., renumber also the equations.

Line 450, usually, box plots use medians, as stated later in the caption also. I would somehow rephrase and skip the “Monthly mean” at the beginning of the caption.

Citation: https://doi.org/10.5194/acp-2022-478-RC2
- AC2:
  'Reply on RC2', Hannele Hakola, 13 Oct 2022
  We thank reviewer 2 for valuable comments and have corrected our manuscript accordingly as shown below.
  The manuscript reports an important new viewpoint, especially on the emission of sesquiterpenes from Spruce and the potential consequences on new particle formation and SOA growth. Unfortunately, the manuscript is written kind of carelessly, like in a very big rush. There are obviously rearranging issues with non-fitting subsection numbering. At many places, the text jumps back and forth, or, terms are defined after they are already used. Explanations needed in the methods description appear in the results or even discussion as footnotes etc. All that lead to the point that it is not clear what the aim of this paper is! Maybe this is also the reason that the paper’s title is very superficial.
  
  Introduction, experimental sections have been partly rewritten to clarify the manuscript. The purpose of this study is to find out if the existing measurement data could help finding representative emission rates for emission inventory and atmospheric impact research. We try to seek for example if the age of tree or a growing location would affect emissions and should be taken into account in atmospheric modelling research. This has been written in the introduction.
  The measurement and data part would need a more clear focus how all goes together. At the moment it’s hard to assess if there are repeated measurements, how many? Some are repeated after longer time to check for changes by age etc. The quality of these data can not be assessed at the moment and therefore it remains unclear if the differences found are really effects or occurred by chance.
  
  This section has been rewritten and no clear differences were found.
  The modelling part would need some clarifications. Did you model monthly or daily time steps? Or both. Which model was used? Basically the first part of the presented equations are a temperature driven parameterisation according to Guenther, as stated. This part somehow deals with the scaling of the measured emission factors to scale them by temperature for comparison with published and older data, as I get it. I kind of understand that MEGAN was used to create another set of emissions that is processed with the second part of equations given for the aerosol growth predictions. In general, the modelling part does not give any grading on the quality of the model or models used in relation to the input parameters. Especially as there are many assumed relations like constant values per month in some input data, constant specific leaf mass, scaling factors between HOM and SOA yields and so on.
  
  Thank you for pointing out the specific information which is lacking! The time step for each module in the model is 60 s and the model was simulating 1 d at a time. For the modelling work, we used the model presented in Taipale et al. (2021) as also stated on L230 in the originally submitted manuscript. The model does currently not have a name, but your comment underlines the need for the model to get named. This action will be taken in a modelling paper which is currently under preparation (Taipale, D., in preparation, 2022). Just to clarify: it is correct that we used Guenther parameterisations to predict the emissions of the VOCs, but the full version of MEGAN was not used, and in the manuscript we have also not claimed this. All equations related to prediction of VOC emissions are already provided in the originally submitted version of the manuscript. The individual processes included in the model (subsections of Sect. 2.4) are aimed to imitate our best mechanistic understanding of those processes. The descriptions of the individual processes have been evaluated separately in earlier studies (references are provided in the subsections of Sec. 2.4). The model’s ability to reproduce canopy scale emissions of VOCs and the influence of organic compounds on aerosol formation and growth in a Scots pine forest has furthermore been tested by constraining and validating the model with observations from the SMEAR II station (the Station for Measuring Ecosystem-Atmosphere Relations II) in Hyytiälä, Finland (Taipale, D., in preparation, 2022).
  Taipale, D.: Impact of biotic and environmental stress and perturbations of Scots pines on formation and growth of atmospheric aerosol particles – a modelling study with quantitative estimates, in preparation, 2022.
  Clarifications and elaborations have been added to Sect. 2.4.
  Specific comments:
  
  Line 91, which models/equations were used to make this aerosol formation and growth calculations from measurements? The one you present or MEGAN?
  
  We used the model presented in Taipale et al. (2021).
  At the relevant line (i.e. at the end of the Introduction) we have emphasised that the modelling work was carried out using the model presented in Taipale et al. (2021).
  Line 175ff, need more details here. First, do you use a “trap” or a “cold trap”? According the procedure described, you sample into a “trap” filled with adsorbents keeping the temperature high enough to avoid condensation. Then, after that, the sample is pushed to the thermodesorber part, that, usually means a heating up and then cooling down below zero before entering finally the column in a GC-MS system. Using a “cold trap” directly would mean, to my knowledge, to use low temperatures while sampling.
  
  The trap was kept at 20-25 C degrees, so this is not a cold trap. We have deleted the word cold.
  Line 215ff, here, it’s a bit puzzling because you describe the emission factor with a unit of ng per g dry weight and hour, I guess. Then in the next section you use the emission factor with micro gram per square meter and second. The conversion factor comes very very late (line 637) in the footnotes of table 5. I would leave some information already here that this conversion is needed to get on to the next step, the model.
  
  A mentioning of this has been added to Sec. 2.4.1.
  Line 250ff, in the section, you describe the constraints, i.e., data used for the parameterisations. It is very puzzling which time scale was used, daily or monthly or both? You use “monthly median”, “daily maximum” and “monthly median of daily maximum” to express these scales but there is no clear description how the model is using them?
  
  Thank you for pointing this out! The model was simulating 1 d during every month. This was deemed sufficient, since the aim was not to simulate one specific year or location, but instead ilustrate and investigate a potential effect that high variations of emission potentials can have on new particle formation. Thus, in the case of [O3], we calculated monthly median values and used those values (one for each month) as input to the model. In the case of [OH], we first calculated the maximum concentration during each day in each month, and then we calculated the monthly median of those maxima and used that as input to the model. Since it is unreasonable to assume that the concentration of OH is constant throughout the day, the concentration of OH was calculated – within the model – to depend on the availability of light. Thus, when there was no light, the concentration of OH was zero, and when the daily maximum light was reached, the concentration of OH reached the value of the monthly median of the daily maximum OH concentration calculated using the proxy by Petäjä et al. (2009). And so on. We have now added the information that the model was simulating 1 d during every month to Sec. 2.4 to clarify the matter.
  Line 303, the temperature is linked with the reaction rates in table 1. It’s better mentioned there in the caption, not together with the chemical reaction schemes.
  
  You are completely correct and we have now changed it accordingly.
  Line 325ff, here, you present the gamma values, and that these have multiple values possible as I understand, they are further numbered as well and state to present HOM, or as SOA yields. However, that makes the need to have another constant (2.2) in to account for. It may be ok, but somehow this lacks a clear reasoning and description of how such changes in model parameters will impact the model’s output.
  
  The gamma values are all treated the same in the model. They origin from HOM and SOA yields, simply because robust HOM yields do not exist for all considered reactions. So that’s why we have considered SOA yields as well. As also stated in the manuscript, we have divided the reported SOA yield by a factor of 2.2 such that they correspond to HOM yields, because SOA yields represent mass yields, while HOM yields represent molar yields. We use a factor of 2.2 under the assumption that every reacted C10H16 forms C10H16O10. Molar yield = n(C10H16O10)/n(C10H16) = 1 = 100%. Mass yield = n * m(C10H16O10)/(n * m(C10H16O10)) = 296/136 = 2.2 = 220%. We have tried to clarify the text further and added more reasoning.
  Line 358ff, it’s a bit strange to define CC after it is used in an equation, that makes the reader searching for it in subsequent equations. Usually, if given in that notion, this is a “side-parameter” and do not need and have an own equation number. It would be beneficial to introduce it before, i.e., renumber also the equations.
  
  Thank you for the comment! We agree with you that since it is a “side-parameter”, there is no need for it to have its own equation and in the original manuscript, we have already in words written what it is, which is sufficient for the reader to reproduce the equation. So we have now taken the equation out.
  Line 450, usually, box plots use medians, as stated later in the caption also. I would somehow rephrase and skip the “Monthly mean” at the beginning of the caption.
  
  This has been rephrased
  
  Citation: https://doi.org/10.5194/acp-2022-478-AC2

Status: closed

RC1:
'Comment on acp-2022-478', Anonymous Referee #1, 17 Aug 2022
Summary: Hakola et al present a comparison of old and new BVOC emission measurements from Norway spruce across various regions in the European boreal forest. They also model particle formation from the emissions data. They found that sesquiterpene emissions are generally higher than predicted in MEGAN. They also show variability in whether Norway spruce is an isoprene-emitter. Detailed BVOC emission measurements represent an important contribution to the field. However, the research approach and analysis presented in this paper are not ready for publication for reasons outlined below. The research presented contains fundamental flaws in justification, clarity, and technical approach.

Major Concerns

The paper lacks adequate context in the introduction. The authors start with a discussion of OH reactivity, but that is not the focus of the paper. One of the highlights of the research is probably the modeling of aerosol formation and growth, but there is no justification presented in the introduction for this research. The introduction was generally very difficult to read due to lack of logical flow and many grammatical errors.

The paper lacks clear research questions and/or objectives. They establish there is some uncertainty about BVOC emission rates from Norway spruce due to variation in previous measurements, but the new measurements really don’t address that. They are also highly variable. There is no justification for the modeling work. The manuscript would be strengthened by clearly outlining research questions/objectives and aligning the figures with those questions/objectives. As written, it lacks clear purpose.

The experimental section is very difficult to follow. It goes back and forth between discussing the new measurements and the old measurements. Even in the section titled, “2.1 New, unpublished emission measurements” the old measurements are discussed (see lines 190-191). Also, this should be labeled section 2.2, not 2.1. And the section labeled 1.1 in the experimental section should be 2.1.

The introduction states that the study is investigating if growing location or age of tree can explain variation in emission measurements (line 87), but this isn’t clearly addressed with the analysis. Furthermore, it does not appear that they had enough replicates of both old and young trees across all sites to even address this question. Addressing this question requires measurements from both old and young trees at multiple sites, which they do not have. Thus, there is a disconnect between the methods and one of the only clear objectives stated in the paper. The research objectives need to align with the sampling approach, but that is not the case.
Citation: https://doi.org/10.5194/acp-2022-478-RC1
- AC1:
  'Reply on RC1', Hannele Hakola, 13 Oct 2022
  We thank reviewer 1 for valuable comments and have tried to take them into account when rewriting the introduction and experimental sections.
  The paper lacks adequate context in the introduction. The authors start with a discussion of OH reactivity, but that is not the focus of the paper. One of the highlights of the research is probably the modeling of aerosol formation and growth, but there is no justification presented in the introduction for this research. The introduction was generally very difficult to read due to lack of logical flow and many grammatical errors.
  
  The introduction has been rewritten.
  The paper lacks clear research questions and/or objectives. They establish there is some uncertainty about BVOC emission rates from Norway spruce due to variation in previous measurements, but the new measurements really don’t address that. They are also highly variable. There is no justification for the modeling work. The manuscript would be strengthened by clearly outlining research questions/objectives and aligning the figures with those questions/objectives. As written, it lacks clear purpose.
  
  The purpose of this study is to find out if the existing measurement data could help finding representative emission rates for emission inventory and atmospheric impact research. We try to seek for example if the age of tree or a growing location would affect emissions and should be taken into account in atmospheric modelling research. This has been written in the introduction.
  The experimental section is very difficult to follow. It goes back and forth between discussing the new measurements and the old measurements. Even in the section titled, “2.1 New, unpublished emission measurements” the old measurements are discussed (see lines 190-191). Also, this should be labeled section 2.2, not 2.1. And the section labeled 1.1 in the experimental section should be 2.1.
  
  The section has been partly rewritten, also numbering has now changed.
  The introduction states that the study is investigating if growing location or age of tree can explain variation in emission measurements (line 87), but this isn’t clearly addressed with the analysis. Furthermore, it does not appear that they had enough replicates of both old and young trees across all sites to even address this question. Addressing this question requires measurements from both old and young trees at multiple sites, which they do not have. Thus, there is a disconnect between the methods and one of the only clear objectives stated in the paper. The research objectives need to align with the sampling approach, but that is not the case.
  
  The introduction has been rewritten taken into account the questions raised by the reviewer. The effect of growing location and the age of trees is discussed in chapter 3.2.6. It is true that we do not have replicates at all sites, the young trees were only measured in southern Finland in Hyytiälä, but 3 older trees were measured at both sites. The results were variating a lot.
  The study was meant to be done with the existing data, therefore the sampling approach is not always best possible.
  
  Citation: https://doi.org/10.5194/acp-2022-478-AC1
RC2:
'Comment on acp-2022-478', Anonymous Referee #2, 01 Sep 2022

The manuscript reports an important new viewpoint, especially on the emission of sesquiterpenes from Spruce and the potential consequences on new particle formation and SOA growth. Unfortunately, the manuscript is written kind of carelessly, like in a very big rush. There are obviously rearranging issues with non-fitting subsection numbering. At many places, the text jumps back and forth, or, terms are defined after they are already used. Explanations needed in the methods description appear in the results or even discussion as footnotes etc. All that lead to the point that it is not clear what the aim of this paper is! Maybe this is also the reason that the paper’s title is very superficial.

The measurement and data part would need a more clear focus how all goes together. At the moment it’s hard to assess if there are repeated measurements, how many? Some are repeated after longer time to check for changes by age etc. The quality of these data can not be assessed at the moment and therefore it remains unclear if the differences found are really effects or occurred by chance.

The modelling part would need some clarifications. Did you model monthly or daily time steps? Or both. Which model was used? Basically the first part of the presented equations are a temperature driven parameterisation according to Guenther, as stated. This part somehow deals with the scaling of the measured emission factors to scale them by temperature for comparison with published and older data, as I get it. I kind of understand that MEGAN was used to create another set of emissions that is processed with the second part of equations given for the aerosol growth predictions. In general, the modelling part does not give any grading on the quality of the model or models used in relation to the input parameters. Especially as there are many assumed relations like constant values per month in some input data, constant specific leaf mass, scaling factors between HOM and SOA yields and so on.

Specific comments:

Line 91, which models/equations were used to make this aerosol formation and growth calculations from measurements? The one you present or MEGAN?

Line 175ff, need more details here. First, do you use a “trap” or a “cold trap”? According the procedure described, you sample into a “trap” filled with adsorbents keeping the temperature high enough to avoid condensation. Then, after that, the sample is pushed to the thermodesorber part, that, usually means a heating up and then cooling down below zero before entering finally the column in a GC-MS system. Using a “cold trap” directly would mean, to my knowledge, to use low temperatures while sampling.

Line 215ff, here, it’s a bit puzzling because you describe the emission factor with a unit of ng per g dry weight and hour, I guess. Then in the next section you use the emission factor with micro gram per square meter and second. The conversion factor comes very very late (line 637) in the footnotes of table 5. I would leave some information already here that this conversion is needed to get on to the next step, the model.

Line 250ff, in the section, you describe the constraints, i.e., data used for the parameterisations. It is very puzzling which time scale was used, daily or monthly or both? You use “monthly median”, “daily maximum” and “monthly median of daily maximum” to express these scales but there is no clear description how the model is using them?

Line 303, the temperature is linked with the reaction rates in table 1. It’s better mentioned there in the caption, not together with the chemical reaction schemes.

Line 325ff, here, you present the gamma values, and that these have multiple values possible as I understand, they are further numbered as well and state to present HOM, or as SOA yields. However, that makes the need to have another constant (2.2) in to account for. It may be ok, but somehow this lacks a clear reasoning and description of how such changes in model parameters will impact the model’s output.

Line 358ff, it’s a bit strange to define CC after it is used in an equation, that makes the reader searching for it in subsequent equations. Usually, if given in that notion, this is a “side-parameter” and do not need and have an own equation number. It would be beneficial to introduce it before, i.e., renumber also the equations.

Line 450, usually, box plots use medians, as stated later in the caption also. I would somehow rephrase and skip the “Monthly mean” at the beginning of the caption.

Citation: https://doi.org/10.5194/acp-2022-478-RC2
- AC2:
  'Reply on RC2', Hannele Hakola, 13 Oct 2022
  We thank reviewer 2 for valuable comments and have corrected our manuscript accordingly as shown below.
  The manuscript reports an important new viewpoint, especially on the emission of sesquiterpenes from Spruce and the potential consequences on new particle formation and SOA growth. Unfortunately, the manuscript is written kind of carelessly, like in a very big rush. There are obviously rearranging issues with non-fitting subsection numbering. At many places, the text jumps back and forth, or, terms are defined after they are already used. Explanations needed in the methods description appear in the results or even discussion as footnotes etc. All that lead to the point that it is not clear what the aim of this paper is! Maybe this is also the reason that the paper’s title is very superficial.
  
  Introduction, experimental sections have been partly rewritten to clarify the manuscript. The purpose of this study is to find out if the existing measurement data could help finding representative emission rates for emission inventory and atmospheric impact research. We try to seek for example if the age of tree or a growing location would affect emissions and should be taken into account in atmospheric modelling research. This has been written in the introduction.
  The measurement and data part would need a more clear focus how all goes together. At the moment it’s hard to assess if there are repeated measurements, how many? Some are repeated after longer time to check for changes by age etc. The quality of these data can not be assessed at the moment and therefore it remains unclear if the differences found are really effects or occurred by chance.
  
  This section has been rewritten and no clear differences were found.
  The modelling part would need some clarifications. Did you model monthly or daily time steps? Or both. Which model was used? Basically the first part of the presented equations are a temperature driven parameterisation according to Guenther, as stated. This part somehow deals with the scaling of the measured emission factors to scale them by temperature for comparison with published and older data, as I get it. I kind of understand that MEGAN was used to create another set of emissions that is processed with the second part of equations given for the aerosol growth predictions. In general, the modelling part does not give any grading on the quality of the model or models used in relation to the input parameters. Especially as there are many assumed relations like constant values per month in some input data, constant specific leaf mass, scaling factors between HOM and SOA yields and so on.
  
  Thank you for pointing out the specific information which is lacking! The time step for each module in the model is 60 s and the model was simulating 1 d at a time. For the modelling work, we used the model presented in Taipale et al. (2021) as also stated on L230 in the originally submitted manuscript. The model does currently not have a name, but your comment underlines the need for the model to get named. This action will be taken in a modelling paper which is currently under preparation (Taipale, D., in preparation, 2022). Just to clarify: it is correct that we used Guenther parameterisations to predict the emissions of the VOCs, but the full version of MEGAN was not used, and in the manuscript we have also not claimed this. All equations related to prediction of VOC emissions are already provided in the originally submitted version of the manuscript. The individual processes included in the model (subsections of Sect. 2.4) are aimed to imitate our best mechanistic understanding of those processes. The descriptions of the individual processes have been evaluated separately in earlier studies (references are provided in the subsections of Sec. 2.4). The model’s ability to reproduce canopy scale emissions of VOCs and the influence of organic compounds on aerosol formation and growth in a Scots pine forest has furthermore been tested by constraining and validating the model with observations from the SMEAR II station (the Station for Measuring Ecosystem-Atmosphere Relations II) in Hyytiälä, Finland (Taipale, D., in preparation, 2022).
  Taipale, D.: Impact of biotic and environmental stress and perturbations of Scots pines on formation and growth of atmospheric aerosol particles – a modelling study with quantitative estimates, in preparation, 2022.
  Clarifications and elaborations have been added to Sect. 2.4.
  Specific comments:
  
  Line 91, which models/equations were used to make this aerosol formation and growth calculations from measurements? The one you present or MEGAN?
  
  We used the model presented in Taipale et al. (2021).
  At the relevant line (i.e. at the end of the Introduction) we have emphasised that the modelling work was carried out using the model presented in Taipale et al. (2021).
  Line 175ff, need more details here. First, do you use a “trap” or a “cold trap”? According the procedure described, you sample into a “trap” filled with adsorbents keeping the temperature high enough to avoid condensation. Then, after that, the sample is pushed to the thermodesorber part, that, usually means a heating up and then cooling down below zero before entering finally the column in a GC-MS system. Using a “cold trap” directly would mean, to my knowledge, to use low temperatures while sampling.
  
  The trap was kept at 20-25 C degrees, so this is not a cold trap. We have deleted the word cold.
  Line 215ff, here, it’s a bit puzzling because you describe the emission factor with a unit of ng per g dry weight and hour, I guess. Then in the next section you use the emission factor with micro gram per square meter and second. The conversion factor comes very very late (line 637) in the footnotes of table 5. I would leave some information already here that this conversion is needed to get on to the next step, the model.
  
  A mentioning of this has been added to Sec. 2.4.1.
  Line 250ff, in the section, you describe the constraints, i.e., data used for the parameterisations. It is very puzzling which time scale was used, daily or monthly or both? You use “monthly median”, “daily maximum” and “monthly median of daily maximum” to express these scales but there is no clear description how the model is using them?
  
  Thank you for pointing this out! The model was simulating 1 d during every month. This was deemed sufficient, since the aim was not to simulate one specific year or location, but instead ilustrate and investigate a potential effect that high variations of emission potentials can have on new particle formation. Thus, in the case of [O3], we calculated monthly median values and used those values (one for each month) as input to the model. In the case of [OH], we first calculated the maximum concentration during each day in each month, and then we calculated the monthly median of those maxima and used that as input to the model. Since it is unreasonable to assume that the concentration of OH is constant throughout the day, the concentration of OH was calculated – within the model – to depend on the availability of light. Thus, when there was no light, the concentration of OH was zero, and when the daily maximum light was reached, the concentration of OH reached the value of the monthly median of the daily maximum OH concentration calculated using the proxy by Petäjä et al. (2009). And so on. We have now added the information that the model was simulating 1 d during every month to Sec. 2.4 to clarify the matter.
  Line 303, the temperature is linked with the reaction rates in table 1. It’s better mentioned there in the caption, not together with the chemical reaction schemes.
  
  You are completely correct and we have now changed it accordingly.
  Line 325ff, here, you present the gamma values, and that these have multiple values possible as I understand, they are further numbered as well and state to present HOM, or as SOA yields. However, that makes the need to have another constant (2.2) in to account for. It may be ok, but somehow this lacks a clear reasoning and description of how such changes in model parameters will impact the model’s output.
  
  The gamma values are all treated the same in the model. They origin from HOM and SOA yields, simply because robust HOM yields do not exist for all considered reactions. So that’s why we have considered SOA yields as well. As also stated in the manuscript, we have divided the reported SOA yield by a factor of 2.2 such that they correspond to HOM yields, because SOA yields represent mass yields, while HOM yields represent molar yields. We use a factor of 2.2 under the assumption that every reacted C10H16 forms C10H16O10. Molar yield = n(C10H16O10)/n(C10H16) = 1 = 100%. Mass yield = n * m(C10H16O10)/(n * m(C10H16O10)) = 296/136 = 2.2 = 220%. We have tried to clarify the text further and added more reasoning.
  Line 358ff, it’s a bit strange to define CC after it is used in an equation, that makes the reader searching for it in subsequent equations. Usually, if given in that notion, this is a “side-parameter” and do not need and have an own equation number. It would be beneficial to introduce it before, i.e., renumber also the equations.
  
  Thank you for the comment! We agree with you that since it is a “side-parameter”, there is no need for it to have its own equation and in the original manuscript, we have already in words written what it is, which is sufficient for the reader to reproduce the equation. So we have now taken the equation out.
  Line 450, usually, box plots use medians, as stated later in the caption also. I would somehow rephrase and skip the “Monthly mean” at the beginning of the caption.
  
  This has been rephrased
  
  Citation: https://doi.org/10.5194/acp-2022-478-AC2

Hannele Hakola, Ditte Taipale, Arnaud Praplan, Simon Schallhart, Steven Thomas, Toni Tykkä, Aku Helin, Jaana Bäck, and Heidi Hellén

Viewed

Total article views: 817 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
599	187	31	817	18	21

HTML: 599
PDF: 187
XML: 31
Total: 817
BibTeX: 18
EndNote: 21

Views and downloads (calculated since 28 Jul 2022)

Month	HTML	PDF	XML	Total
Jul 2022	108	24	8	140
Aug 2022	76	27	3	106
Sep 2022	27	13	2	42
Oct 2022	38	16	4	58
Nov 2022	34	9	0	43
Dec 2022	19	7	0	26
Jan 2023	14	0	14
Feb 2023	22	12	2	36
Mar 2023	31	3	0	34
Apr 2023	12	3	0	15
May 2023	18	11	2	31
Jun 2023	11	3	0	14
Jul 2023	10	9	1	20
Aug 2023	22	4	0	26
Sep 2023	26	8	3	37
Oct 2023	22	5	0	27
Nov 2023	15	3	3	21
Dec 2023	18	3	1	22
Jan 2024	23	0	23
Feb 2024	17	14	0	31
Mar 2024	23	11	2	36
Apr 2024	13	2	0	15

Cumulative views and downloads (calculated since 28 Jul 2022)

Month	HTML	PDF	XML	Total
Jul 2022	108	24	8	140
Aug 2022	76	27	3	106
Sep 2022	27	13	2	42
Oct 2022	38	16	4	58
Nov 2022	34	9	0	43
Dec 2022	19	7	0	26
Jan 2023	14	0	14
Feb 2023	22	12	2	36
Mar 2023	31	3	0	34
Apr 2023	12	3	0	15
May 2023	18	11	2	31
Jun 2023	11	3	0	14
Jul 2023	10	9	1	20
Aug 2023	22	4	0	26
Sep 2023	26	8	3	37
Oct 2023	22	5	0	27
Nov 2023	15	3	3	21
Dec 2023	18	3	1	22
Jan 2024	23	0	23
Feb 2024	17	14	0	31
Mar 2024	23	11	2	36
Apr 2024	13	2	0	15

Viewed (geographical distribution)

Total article views: 853 (including HTML, PDF, and XML) Thereof 853 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 10 Apr 2024

Short summary

Norway spruce is one of the main tree species growing in the boreal area. We show that volatile organic compound emission potentials and compound composition vary a lot. We have investigated if e.g. growing location or age of a tree could explain the variations. Recognizing this observed large variability in spruce BVOC emissions (precursors for new particle formation processes), we also tested the consequences of this variability in simulations of aerosol formation.


Total:	0
HTML:	0
PDF:	0
XML:	0