Decrease in radiative forcing by organic aerosol nucleation, climate, and land use change

Organic nucleation is an important source of atmospheric aerosol number concentration, especially in pristine continental regions and during the preindustrial period. Here, we improve on previous simulations that overestimate boundary layer nucleation in the tropics and add changes to climate and land use to evaluate climate forcing. Our model includes both pure organic nucleation and heteromolecular nucleation of sulfuric acid and organics and reproduces the profile of aerosol number concentration measured in the Amazon. Organic nucleation decreases the sum of the total aerosol direct and indirect radiative forcing by 12.5%. The addition of climate and land use change decreases the direct radiative forcing (−0.38 W m−2) by 6.3% and the indirect radiative forcing (−1.68 W m−2) by 3.5% due to the size distribution and number concentration change of secondary organic aerosol and sulfate. Overall, the total radiative forcing associated with anthropogenic aerosols is decreased by 16%.

chamber are the same as those which form particles in the upper troposphere either, which is where, in this paper, most pure biogenic NPF is happening. Indeed, one would expect higher volatility HOMs such as those the authors produce to be able to stick together at low upper tropospheric temperatures. Consequently, not simulating autoxidation might be a reasonable approximation, in the end, when one considers that the alternative, of trying to simulate it using yields from pure alpha-pinene ozonolysis experiments, is also certainly not ideal.
Ultimately there are two approaches here: here, the authors use chemical reactions that are explicitly modelled in MCM and therefore should happen in the atmosphere, to produce molecules which may or may not form new particles. Before, we used a representation of chemical reactions that happen in chamber experiments but are not guaranteed to happen at the same rate in the atmosphere, to make molecules that we could be more confident can form particles. The approaches are different -it is too early to say which is better -and it is valuable that the authors have explored the one that we didn't explore already.
To make things clearer, I think the authors should emphasise explicitly somewhere that they do not simulate autoxidation. Also, I think it would be useful to show the effect of the new chemical mechanism separately from the effect of the temperature dependence the authors introduce.

Temperature dependence
In the methods, the authors say "The organic nucleation rates were multiplied by exp(-(T-278)/10). (Dunne et al, 2016)" This temperature dependence is important, and will substantially affect the results of this study. When we used this same T dependence for a sensitivity study in Gordon et al (2016), we found that the cloud albedo forcing was reduced to +0.14Wm-2 from 0.22Wm-2. The 0.14Wm-2 (a 17% reduction in forcing, rather than a 27%) is stated in the main text of Gordon et al (2016) and is much more similar to the results of the present manuscript, although in Gordon et al it is only the forcing from pure organic NPF; in this manuscript the forcing effect of H2SO4+HOM NPF is also included.
Unfortunately I have no evidence for this temperature dependence. It is only an educated guess. On page 31 of the Dunne et al SI I wrote that it has "no physical basis", maybe looking back, I should have restated this in the Table S7 caption (sorry!). We didn't have any experimental data with which to get this T dependence at the time.
The authors should state explicitly that the T dependence that they used is not supported by published experimental evidence and they should explore the sensitivity of their results to the temperature dependence with new simulations. Maybe also worth referring to Yu et al (ACP 2017). This useful study considers the possible T dependence of the NPF rate, though not of the HOM production mechanism (if one assumes autoxidation is the main mechanism for HOM formation, it has a strong T dependence which would counteract to some unknown extent the T dependence of the NPF rate).
2. Particle growth L127: "The global average mass of sulfuric acid in the nucleation and Aitken mode particles in newSOA is larger than that of organics. Thus, condensed sulfuric acid and water are responsible for most of the growth of small newSOA particles after organics drive the formation of new particles, consistent with some laboratory and field measurements (27,28)." Put crudely, if a molecule has low enough volatility to form new particles, it should also be able to grow them. Of course the dependences on vapour concentration can be different, but ultimately I think that if NPF is dominated by organics and not H2SO4, growth should also be dominated by organics and not H2SO4, with H2SO4 making only a minor contribution. The only situation (apaart from strange chemical reactions) in which this would not be true is if the nucleating organics had lower average volatility than H2SO4 and H2SO4 was more abundant in the gas phase. This seems unlikely, though not impossible, and the possibility is not supported by the references cited by the authors. The reference 27 documents chamber measurements of growth rates due to H2SO4, and reference 28 is atmospheric observations of organic NPF. Neither of these papers say that newly formed organic clusters grow mostly by condensation of H2SO4. On the other hand, Lawler et al (GRL 2018) show that 90% of the mass of newly formed particles in Hyytiala of around 40nm in diameter is organic. I think the statement of the authors points to an inconsistency between the formation of nucleating organics in their model and the formation of the organics that grow the particles (or perhaps the NPF rate at low temperatures is too high compared to the binary H2SO4 formation rate). This possibility should at least be acknowledged in the text. Further, at line 132, the concentration of SVOC over the Amazon is surely much higher than the concentration of H2SO4, not really comparable.
The authors acknowledge implicitly that this has large consequences for their results, at line 166 and 174. It seems that about half the increase in newSOA number concentrations in PD compared to PI is due to increased NPF rates and half is due to condensation of H2SO4. The increased newSOA in the PD compared to the PI may be exaggerated if organics are insufficiently able to grow nucleation-mode particles (especially in the PI). This would suppress the PD-PI indirect forcing. Conversely, too high organic formation rates might exaggerate the PD-PI indirect forcing. Ideally, the authors should do a sensitivity simulation, along the lines of allowing more SVOC to condense onto smaller particles, and some discussion of this possible shortcoming would also be helpful.

Model evaluation
While it is clear from comparing Fig.1 and SI Fig 2 that the model does a reasonable job of reproducing observations in the very challenging Amazon environment, the comparison should be made more precise by using the radii of the Aitken mode to cut off the simulated size distribution at 10nm and 20nm as appropriate for an explicit comparison. At the moment, the simulated number concentration is clearly higher than the observed number concentration, but this may not be true if the particles below the CPC cut off size are excluded. Then I think the authors should put the model and measurement data on the same plot in the main text.
The authors show that their model reproduces the Amazon, where organic NPF is not seen. However, what about places where NPF is observed to be much higher in the boundary layer than binary H2SO4-H2O-ion NPF would predict, like Hyytiala in Finland, or various places in the USA (e.g. Yu et al, ACP 2015)? Does the model reproduce NPF rates observed at Hyytiala and/or other polluted boundary layer sites strongly influenced by BVOCs? And what about (e.g.) northern Finland, Canada or Siberia, where one might also expect pure organic NPF?

Minor comments
Abstract: "Organic nucleation" (also line 53) is a bit vague; please specify "nucleation involving either organics alone or organics with sulphuric acid" once in the abstract.
Towards the end it is not clear whether climate and land use change cause an enhancement of 3% on the DRF independently of organic NPF or only because of organic NPF.
Line 56 "BVOC nucleation" is also unclear -this was NPF of BVOCs and H2SO4.
Line 62: See major comment: please could the authors mention here that the assumption that HOM yields are instantaneous is, to a very good approximation, what is observed in chamber studies, and in particular in the chamber studies from which their NPF parameterisation is derived (e.g. Ehn et al, 2014, Kirkby et al, 2016. Exocyclic monoterpenes react differently with ozone and OH than endocyclic monoterpenes, so about 40% of monoterpenes (see SI of Jokinen PNAS 2015) mostly react with OH and mostly don't form HOMs. We tried to replicate this in Gordon et al (2016) and it did make quite a sizeable difference compared to just using alpha-pinene, though it is by no means the most important uncertainty. This might be worth discussing in your "implications". Line 248: "Organic nucleation results in accumulation mode number concentrations of newSOA of 1.9 and 1.4cm-3." These very small numbers are a bit surprising, given the 10x larger increase in CDNC and also the substantial change in supersaturation. Presumably this means some of the Aitken mode newSOA is activating, or the activation doesn't happen in the PBL. It would be helpful to know the mean diameters of the modes involved and the indirect effect of all aerosol to understand this better. The CDNC changes also seem small given the large AIE. Scott et al (ACP 2014, Table 2) found that a PI change of 35cm-3 in CDNC or 38% led to an AIE of -0.95Wm-2. This is a 30% lower AIE than your result, for a 3x higher change in CDNC. It would be good to try to understand what is driving this, though I understand it may be difficult and beyond the scope of the paper. Please could you state your relative increase in global mean CDNC? This may be enough to give us some idea.

Methods:
Please state the horizontal and vertical resolution of the model. Please cite the source for the binary H2SO4 NPF mechanism. Do you track the sink of newSOA due to coagulation? I would expect this is important compared to dry/wet deposition in the Amazon biomass burning (dry) season. Please specify the size ranges (particle diameters) that the nucleation, Aitken and accumulation modes are designed to cover. Please state the hygroscopicity (Kappa or van t'Hoff factor) assumed for SOA. Please state the origin of the updraft speeds used in the offline activation calculation; I couldn't immediately see this in ref. 43.
The overall formation rates seem very high in the Amazon. Are the formation rates at very low temperatures allowed to exceed the kinetic limit for HOM-HOM intermolecular collisions?
Are the radiative effects of newSOA calculated by taking the difference between a simulation with the NPF mechanism switched on and a simulation with the NPF mechanism switched off. Or are these effects all calculated from the same simulation, by ignoring and then not ignoring the newSOA modes in the activation/radiative transfer scheme? Main claims/findings of the paper are that organic nucleation is an important source of aerosols in the pristine continental areas and during the preindustrial period, and that the increased emissions of a-pinene and its oxidation products (nucleating and forming SOA) are enhancing due to the climate and land use changes. Together these events: organic nucleation, land use and climate, lead to an overall 15 % reduction in radiative forcing when applied together with present day anthropogenic aerosol in the tropics. This paper also demonstrates reproductions of vertical profile of condensation nuclei over the Amazon forest. I only have more general comments to this work rather than going into details: -The subject of the study is very complex and the study seems novel at least considering most parts of the work. I'm pleased to see that the effect of climate and land use is included in this work with particle formation happening above the planetary boundary layer. As organic (pure biogenic) nucleation has only been observed and published a few times (Kirkby et al., Nature = lab study and Bianchi et al., Science = field study), models are not accounting it in yet. Even if I consider the methods and results to be novel, I feel that there are too many misinterpretations and misunderstanding with the resent experimental work done in the field of nucleation studies and HOMs. I also find referencing to previous work inadequate and somewhat misleading, especially considering the referencing of experimental publications. Thus, I cannot advise publication in the current form.
-I found that it was difficult to understand if all the work is concerning the vertically integrated values or upper troposphere/PBL conditions only. This should be clarified throughout the manuscript.
-Applying experimental information: The paper claims that the Amazon is the most important region producing new organic particles and has the highest organic nucleation rate primarily because of the high emissions of α-pinene and the high production of HOMs from α-pinene oxidation. My knowledge is that the Amazon is known for its extremely high concentrations and emissions of isoperene, not a-pinene. Even though both a-pinene and isoprene emissions are estimated to increase due to land use and temperature, isoprene is still clearly dominating the BVOCs in the Amazonas. It should be noted in the manuscript that the large fraction of SOA forming HOMs have very different formation pathways, isoprene is mostly oxidized by OH-radicals to form HOM and monoterpenes produce HOM from ozonolysis (e.g. Ehn et al., 2014), OH-radicals (E.g. Berndt et al., 2016) and NO3-radicals (E.g. Ayres et al., 2015, Yan et al., 2016. Major changes in the oxidation pathways could cause very different results in nucleation, SOA formation and load. Lots of HOM related work, production mechanisms, yields of different HOMs from various monoterpenes, sesquiterpenes and isoprene (SOAS campaign, Alabama) has been published since the initial Ehn et al, 2014 paper.

2
We thank editor and two reviewers very much for dedicating their time to read our manuscript and present important comments. We carefully studied these comments and revised the manuscript widely. In summary, we updated the chemical mechanism in our model to reduce uncertainty based on reviews' comments. We also added model validations and sensitivity experiments in the supplementary. Replies to these comments are listed point by point as below.

Reviewers' comments:
Reviewer #1: Review of "Organic aerosol nucleation, climate and land use change: Decrease in radiative forcing".

Hamish Gordon, 6/8/18
This article explores the implications for atmospheric particle number concentrations and radiative balance of the mechanisms for new particle formation (NPF) from highly oxidised organic molecules (HOMs) alone, and HOMs with sulphuric acid, first quantified at the CLOUD experiment. The authors produce HOM-like molecules in a different way to the previous implementation of the The manuscript addresses an important topic, and I believe that, taken as a whole, it will be a sufficiently significant advance on previous work to be appropriate for Nature Communications once my comments below are addressed. It is well written, previous literature is discussed appropriately, and I was pleased to be asked to review it. My "major" comments should not require a huge amount of extra work to address.
Major comments 1. NPF processes and model uncertainties As well as implementing pure organic new particle formation in a different (and more sophisticated) host model, the authors make two key changes to the new 4 particle formation process in their model compared to Gordon et al (2016). The first is to use an explicit chemical mechanism for HOM production which does not involve autoxidation, instead of trying to simulate autoxidation in the model. The second is to use a temperature dependence for the NPF rates as their baseline, which was introduced as a sensitivity study by Gordon et al (and was only ever intended as a sensitivity study, as it has no physical basis).

Chemical mechanism
Line 293: "Most experiments indicated that HOMs are produced through H-shift and peroxy radical autoxidation". The basic postulate presented by the authors at the process level seems to be that this doesn"t happen in the atmosphere.
This hypothesis seems quite unlikely to be strictly true, given that the chamber and field-measured mass spectra for HOMs are quite similar (Schobesberger et al 2013). In terms of individual molecules, the chamber and field spectra are certainly more similar to each other than they are similar to the mass spectrum of HOMs the authors produce using the MCM model.
That said, the yield of HOMs via the autoxidation of a-pinene could well be quite different to the yield we measured in the CLOUD chamber and used in Gordon et al, 2016. NOx, RO2 (e.g. from isoprene), and HO2 are all possible candidates to suppress this yield, as the authors also point out. Furthermore, there"s nothing to guarantee the molecules that form particles at 5°C in the CLOUD chamber are the same as those which form particles in the upper troposphere either, which is where, in this paper, most pure biogenic NPF is happening. Indeed, one would expect higher volatility HOMs such as those the authors produce to be able to stick 5 together at low upper tropospheric temperatures. Consequently, not simulating autoxidation might be a reasonable approximation, in the end, when one considers that the alternative, of trying to simulate it using yields from pure alpha-pinene ozonolysis experiments, is also certainly not ideal.
Ultimately there are two approaches here: here, the authors use chemical reactions that are explicitly modelled in MCM and therefore should happen in the atmosphere, to produce molecules which may or may not form new particles.
Before, we used a representation of chemical reactions that happen in chamber experiments but are not guaranteed to happen at the same rate in the atmosphere, to make molecules that we could be more confident can form particles. The approaches are different -it is too early to say which is better -and it is valuable that the authors have explored the one that we didn"t explore already.
To make things clearer, I think the authors should emphasize explicitly somewhere that they do not simulate autoxidation. Also, I think it would be useful to show the effect of the new chemical mechanism separately from the effect of the temperature dependence the authors introduce.
Re: We emphasized our explicit chemical mechanism is quite different with the assumption of fast autoxidation used in your studies (around Line 65). We also noted that your model applied HOMs yield and nucleation parameter purely based on CLOUD experiment, which may not reflect what happens in the atmosphere exactly (around Line 69). We added a discussion of the effects of using the temperature dependence of new particle formation rates in supplementary information Section S3.1 for a result from a sensitivity test without temperature dependence. We found the aerosol number concentration will peak around 1.5 km 6 in the Amazon when the model without temperature dependence is used. At the same time, there is still a small peak of aerosol number concentration in the upper troposphere around 12 km in the Amazon. As a result, the sensitivity experiment results in a 5.3% increase of total radiative forcing of anthropogenic aerosol compared to the result calculated in the base model in the main text. Re: The HOMs condensed on the new particles is able to grow them up. However, the concentration of HOMs is much less than the concentration of H2SO4 even if in the tropics, so that HOMs is not able to grow them as quickly as H2SO4 does.
There may be much higher concentration of HOMs in the real atmosphere, but we currently have limited knowledge of the explicit chemical reactions that lead to the production of extremely low volatility HOMs. In our model, the growth of newSOA by organics is dominated by the partitioning of SVOC. Notably most of the partitioning of SVOC takes place on particles with a large mass of pre-exisiting organics rather than on the smaller particles. As the result, organics don"t have a large contribution to the growth of very small newSOA particles, but contribute a lot to the growth of larger newSOA particles. Our model results indicate that the mass of organics in tropics always accounts for >80% of the total mass of newSOA particle with diameter >50nm, which agrees with the measurements in Hyytiala.
We added a sensitivity test to examine the effect of different mechanisms for uptake of organics on newSOA in supplementary information section S3.2. We assumed IEPOX, glyoxal and methylgloxal are also able to condense on newSOA, whereas they are only taken up by new sulfate particles in the base model in the main text. In this sensitivity test, the low-volatility products formed from IEPOX, glyoxal and methylgloxal take part in the growth of newSOA. As a result, organics dominate the growth of newSOA in the tropics. The mass concentration of organics on newSOA in each mode is larger by a factor of 4-7 than that of sulfuric acid in the tropics. For Line 132 (now around Line 140), we meant the concentration of organics condensed on newSOA is comparable to that of sulfuric acid on newSOA, but not the concentration of SVOC and sulfuric acid in the atmosphere of Amazon. We have changed that sentence to "the concentration of organics on the newSOA is comparable to that of sulfuric acid".

Model evaluation
While it is clear from comparing Fig.1 and SI Fig 2 that  of HOMs in Finland, Canada is also much less than that in the Amazon, even in the summer. The correlation coefficient between monthly simulated number concentration of newSOA and observed aerosol number concentration at the three sites in Finland and one site in Canada are all higher than 0.6, which indicates the model is able to reproduce the pattern associated with the seasonal variation of aerosol number concentration in regions strongly influenced by BVOCs.

Minor comments
Abstract: "Organic nucleation" (also line 53) is a bit vague; please specify "nucleation involving either organics alone or organics with sulphuric acid" once in the abstract.
Re: We have specified with "pure organic nucleation and heteromolecular nucleation of sulfuric acid and organics" in the beginning of abstract. Also, added "involving either pure organics or organics with sulfuric acid" around L58 Towards the end it is not clear whether climate and land use change cause an enhancement of 3% on the DRF independently of organic NPF or only because of organic NPF.
Re: In the latest version, the DRF is increased by 18% as a result of only organic nucleation. Climate and land use change alone cause an enhancement of 6.3% on the DRF. The schemes for the difference between with and without climate and land use change in the PI both include organic nucleation. We have changed the discussion of these two differences in that section (around L222~L231). The description is hopefully much clearer now.
Re: "new particle formation from BVOC and sulfuric acid" has been added around L61. Re: We have changed that sentence to "previous model studies applied empirical or semi-empirical fixed and instantaneous HOM yields. These studies assumed that autooxidation occurs rapidly and produces HOMs after the initial oxidation step based on the CLOUD project experiments." around L66-68.
Exocyclic monoterpenes react differently with ozone and OH than endocyclic monoterpenes, so about 40% of monoterpenes (see SI of Jokinen PNAS 2015) mostly react with OH and mostly don"t form HOMs. We tried to replicate this in Gordon et al (2016) and it did make quite a sizeable difference compared to just using alpha-pinene, though it is by no means the most important uncertainty. This might be worth discussing in your "implications".
Re: Thank you for suggestion. We have added "Exocyclic monoterpenes like βpinene mostly react with OH and do not form HOMs, so while excluding the 13 production of HOMs from exocyclic monoterpenes adds some uncertainty, it is not likely to be large compared to other uncertainties." in the implications and discussion around L298.
Line 117: it would be worth comparing to    Please cite the source for the binary H2SO4 NPF mechanism.
Re: The mechanism for binary sulfuric acid-water nucleation used in our model is based on Vehkamaeki et al. (2002). This has been added to the methods section (around Line 341).
Do you track the sink of newSOA due to coagulation? I would expect this is important compared to dry/wet deposition in the Amazon biomass burning (dry) season.
Re: Yes, we calculated the coagulation of newSOA in the model, but we did not keep track of the totals. Our simulation results are able to reproduce the peak of bOC in the Amazon during the dry season (July-September). The coagulation should be important sink of newSOA when the aerosol number concentration is high.
Please specify the size ranges (particle diameters) that the nucleation, Aitken and accumulation modes are designed to cover.
Please state the hygroscopicity (Kappa or van t"Hoff factor) assumed for SOA.
Re: We assumed a Kappa factor of 0.14 to describe the hygroscopicity of SOA.
When sulfuric acid is coated on newSOA or SOA is internally mixed with other aerosols, the Kappa factor of the mixed particle is calculated according to the volume fraction of composition and the individual constituent Kappa factors.
Please state the origin of the updraft speeds used in the offline activation calculation; I couldn"t immediately see this in ref. 43.
Re: The updraft speeds are calculated from the eddy diffusivity of the CAM model refer to Wang et al. 2009.
The overall formation rates seem very high in the Amazon. Are the formation rates at very low temperatures allowed to exceed the kinetic limit for HOM-HOM intermolecular collisions?
Re: The kinetic limit collision coefficient is calculated as ~3×10 -10 cm 3 s -1 which, for the maximum nucleation rate, results in a kinetically limited nucleation rate of 10 6 cm -3 s -1 when the temperature is 200K and the concentration of HOMs is highest. Thus, the kinetic limit nucleation rate is much higher than the maximum nucleation rate (~2 cm -3 s -1 ) in our simulation.
Are the radiative effects of newSOA calculated by taking the difference between a simulation with the NPF mechanism switched on and a simulation with the NPF mechanism switched off. Or are these effects all calculated from the same simulation, by ignoring and then not ignoring the newSOA modes in the activation/radiative transfer scheme?
Re: The radiative effects due to SOA calculated in this study are all caused by all SOA (newSOA and internally mixed SOA), not by only newSOA. These effects are the difference in radiation between with and without SOA. The other inputs to 17 the radiation calculation are all same. The difference in the schemes between with and without organic nucleation is the simulation with the mechanism of new organic particle formation switch on and off. We have described these at the end of Method section. Main claims/findings of the paper are that organic nucleation is an important source of aerosols in the pristine continental areas and during the preindustrial period, and that the increased emissions of a-pinene and its oxidation products (nucleating and forming SOA) are enhancing due to the climate and land use changes. Together these events: organic nucleation, land use and climate, lead to an overall 15 % reduction in radiative forcing when applied together with present day anthropogenic aerosol in the tropics. This paper also demonstrates reproductions of vertical profile of condensation nuclei over the Amazon forest.
I only have more general comments to this work rather than going into details: -The subject of the study is very complex and the study seems novel at least considering most parts of the work. I"m pleased to see that the effect of climate and land use is included in this work with particle formation happening above the -Applying experimental information: The paper claims that the Amazon is the most important region producing new organic particles and has the highest organic nucleation rate primarily because of the high emissions of α-pinene and the high production of HOMs from α-pinene oxidation. My knowledge is that the Amazon is known for its extremely high concentrations and emissions of isoprene, not apinene. Even though both a-pinene and isoprene emissions are estimated to increase due to land use and temperature, isoprene is still clearly dominating the  We have looked at all the paper you mentioned above about the chemical mechanism. Although there are many pathways to form HOMs, the issue is whether the formed HOMs lead to nucleation or not. Diacyl peroxide is suspected to be the one kind of HOMs to nucleate based on the evidence published in Ziemann (2002) we have found with any evidence that they contribute to nucleation and new particle formation. Pinic acid and diacyl peroxide form from the ozonolysis of α-pinene, and pinanediol and its oxidation product form from an oxidation product of the reaction of α-pinene and OH radical. We didn"t find any evidence for the nucleation of HOMs formed from the oxidation of isoprene. As the result, we updated the chemical mechanism used in our model to include the formation of these additional HOMs species (diacyl peroxide, pinic acid, pinanediol and its oxidation product). Further, we updated the nucleation scheme in the model to calculate organic nucleation based on these four kinds of HOMs. We believe the knowledge about the nucleation of HOMs is still limited. The pathways to form HOMs leading to nucleation have large uncertainty. We stated the uncertainty in the discussion section as "The calculation of SOA and RF is uncertain in part because of unclear chemical formation mechanisms. We currently have limited knowledge of the explicit chemical reactions that lead to the production of extremely low volatility HOMs, although there should be many ways that these form. We only included the HOMs with evidence to contribute to nucleation in the model now, but it"s possible that there are other species of HOMs will be identified in the future. The use of different HOMs leading to nucleation should be a large uncertainty to the RF estimation associated with organic nucleation." For the SOAS project you mentioned in the comments, we found the publications from this project mainly talk about the pathway to form organic nitrate, but do not mention organic nucleation. We updated the chemical mechanism to form organic nitrate in our model referring to Fisher et al. (2016) which was based on the SOAS project.
I thank the authors for the informative response to both of the reviews. All my questions and comments have been satisfactorily addressed in your point to point reply. I have no further comments concerning this manuscript at this point.