Dear Palmira Messina et al.,

While both referees found the manuscript interesting, they were also very critical on some key aspects of the paper. These included the derivation of emission factors for different plant functional types, and the validity of the leaf area index calculated by ORCHIDEE. The referee #2 was not satisfied with the revised version (see comments below). Thus I subject the paper to major revision, after which it will be reconsidered.

The derivation of EFs needs to be revised with more systematic approach. I agree with the referees that “qualitative and comparative method” sounds very subjective. As authors mention in their response to the referee #2 that the modification of EFs “represents an important part of the upgrade of the emission module”, being more systematic and transparent in the methodology is crucial. The comments by the referee #2 offer good guidelines in this: Take into consideration only primary publications (no reviews, no proceeding modeling studies), select the papers to be used (weed the papers for obvious outliers and otherwise suspicious data), calculate mean EF (or median, and possibly standard deviation). Finally describe the method and show which papers you have actually used.

According to the studies pointed out by the referee, the areas in which the ORCHIDEE and MODIS LAI deviate most (lower boreal, temperate, and tropics) are the ones where MODIS is thought to be most certain. Thus the impact of the deviation of LAIs should be addressed in a more rigorous way e.g. by running the model by forcing the LAI to match the MODIS values geographically and analyze the effect on the VOC fluxes.

Sincerely,
Janne Rinne


Comments by Referee #2:

I am still confused about the derivation of EFs. It is now said that the authors used a ‘qualitative and comparative method’, which generally means to me that there is no rule behind the selection. Later on it is stated that for isoprene they ‘principally use the most recent papers’, which would be okay if the given examples for larch (here 5 references between 2002 and 2012 are used, outliers disregarded) and also for BoBrDe (2003 – 2012) wouldn’t proof differently. The example of larch indicates another oddity here: the authors explain which papers were reviewed (10) and which were actually used to determine the arithmetic average (5). However, in the reference list of table 3 all the references which are discarded for good reasons are still listed. So are these (bad) references still used in the derivation of some of the other EFs?? Interestingly, for monoterpenes outliers are also ‘officially’ disregarded although ‘generally’ any paper is taken into account. Overall, I still cannot see which references are actually used for EF derivation and the explanations give the impression of justifying a selection that has done with no clear rules in mind. I can understand the difficulties in determining ‘objective’ values and I hope not to be overly picky here being educated by the authors that ‘there is no universal agreement on parameterization’ but why don’t disregard any clear outliers as well as any reference with unclear sources or which refers to measurements that are already considered? In Table 3, the references used for each PFT should be clearly differentiated, possibly with numbers given behind the respective EF value.

From my given options to proceed with the LAI problem in the simulations (1 Improve the LAI simulations, 2 Improve the argumentation to a degree that the reader can accept ORCHIDEE simulations as equally likely as MODIS data, or 3 Run all simulations with MODIS derived values only), the authors choose no. 2. The arguments are 1) that also MODIS data don’t provide the ‘real’ LAI, and 2) that it is the intention of the article to study the weaknesses in LAI modelling, so that they can be resolved in a future model version of ORCHIDEE. I am a bit surprised about the reasoning given that the Maignan et al. study which investigated LAI and suggested improvements already 5 years ago seemingly was not considered. Indirectly supporting Maignan et al., Fang et al. 2013 found that the largest uncertainties of MODIS data are in the Australian southwest, the Rocky Mountains and in lower boreal regions while the tropics and temperate regions are relatively ‘certain’. Unfortunately these are the regions that deviate most with ORCHIDEE simulations and are most important for BVOC emissions. So I only can support the view of the authors that the weaknesses of LAI simulations have to be analysed and improved but I think that this should be done in a more differentiated way than with a mere global reduction/increase of simulated LAI – and much preferably in a separate study preceding the BVOC analysis.
By the way: The referenced MODIS data are indicated to be obtained from Yuan et al. 2011, which however, only supplies a tool for improving MODIS data and doesn’t provide a global dataset. So this doesn’t seem to be the correct reference.

Leaf age, drought stress and CO2 activity modifications are explained but I couldn’t find the give reference (Sinderalova et al. 2014) in the reference list. I guess it is Sindelarova et al. 2014. In this publication the CO2 effect is determined to increase isoprene emission by 2.7 percent based on the concentration of the year 2003. It is indeed probably less if the average during the simulation period is concerned.

Fang, H., et al. (2013), Characterization and intercomparison of global moderate resolution leaf area index (LAI) products: Analysis of climatologies and theoretical uncertainties, J. Geophys. Res. Biogeosci., 118, 529–548, doi:10.1002/jgrg.20051.

Sindelarova K, Granier C, Bouarar I, Guenther A, Tilmes S, Stavrakou T, Müller JF, Kuhn U, Stefani P, Knorr W. 2014. Global dataset of biogenic VOC emissions calculated by the MEGAN model over the last 30 years. Atmospheric Chemistry and Physics, 14: 9317-9341.

Dear Palmira Messina et al.,

I find the manuscript much improved. I only have one question:

Page 19, lines 12-14: “a) First, we select only papers that provide EFs in the proper units (μgC g-1 h-1) and for standard conditions such as defined in ORCHIDEE (PAR = 1000 μmol m-2 s-1, temperature = 30 °C), unless the information needed to convert the EFs accordingly was given.” In here I guess you mean have only considered EFs given per leaf biomass, not per leaf area? The EFs given in molar units or in grams of compound, or per hour, are trivial to convert to desired unit. This could be written more explicitly.

Sincerely,
Janne Rinne