Emissions of monoterpenes from new Scots pine foliage: dependency on season, stand age and location and importance for models

Abstract. Models to predict the emissions of biogenic volatile organic compounds (BVOCs) from terrestrial vegetation largely use standardised emission potentials derived from shoot enclosure measurements of mature foliage and usually assume that the contribution of BVOCs from new conifer needles is minor to negligible. Extensive observations have, however, recently demonstrated that the potential of new Scots pine needles to emit several different BVOCs can be up to about 500 times higher than that of the corresponding mature foliage. Thus, we build on these discoveries and investigate the impact of previously neglecting enhanced emissions from new Scots pine foliage on estimates of monoterpene emissions and new atmospheric aerosol particle formation and their subsequent growth. We show that the importance of considering the enhanced monoterpene emission potential of new Scots pine foliage decreases as a function of season, tree age and latitude, and that new foliage is responsible for the majority of the whole tree's foliage emissions of monoterpenes during spring time, independently of tree age and location. Our results suggest that annual monoterpene emission estimates from Finland would increase with up to ~ 25 % if the emissions from new Scots pine foliage were explicitly considered, with the majority being emitted during spring time where also new particle formation has been observed to occur most frequently. We estimate that our findings can lead to increases in predictions of the formation rates of 2 nm particles during spring time by ~ 75–275 % in northern Finland and by ~ 125–865 % in southern Finland. Likewise, simulated growth rates of 2–3 nm particles would increase by ~ 65–175 % in northern Finland and by ~ 110–520 % in southern Finland if the enhanced emissions of monoterpenes from new Scots pine foliage were explicitly considered. Our findings imply that we need to introduce a more comprehensive treatment of the emissions of BVOCs from new coniferous foliage in biogenic emission models.


conducted from the end of June, while the detection of the emissions from new foliage was only started in the end of July. As the measurements were performed sporadically, only seasonally averaged potentials have been provided. The authors found that new needles have a higher potential to emit monoterpenes than mature needles by a factor of two, which is comparable to what is used in Guenther et al. (2012). However, these measurements did not cover the vital spring season. Taipale et al. (2011) and Rantala et al. (2015) measured the ecosystem scale flux continuously from April/May to September during four years. In both studies, the micrometeorological measurements were conducted on the same ~50 year old Scots pine forest at the SMEAR II station (Station for Measuring Ecosystem-Atmosphere Relations). The canopy, within an area with a radius of 200 m, is made up by Scots pine (~75 %), Norway spruce (~15 %) and deciduous species (~10 %), mainly silver birch (Mäki et al., 2019). The potential of the forest to emit monoterpenes per ground area was in both cases shown to significantly decrease from spring and over the summer (Taipale et al., 2011;Rantala et al., 2015). Since the pines in that region carry about 2.5 needle age classes (Ťupek et al., 2015), the foliage mass is approximately 40 % less in the spring than later in the season (i.e. about August onwards). Hence, the conclusion by Taipale et al. (2011) and Rantala et al. (2015) is further amplified if the potential to emit is considered per foliage mass.
If a model utilises rather static needleleaf development combined with only slightly higher emission potentials of new than mature needles, the influence of new coniferous foliage to canopy BVOC emissions is predicted to be minor (Guenther et al. 2012). However, though the mass of new foliage is very small in the beginning of the growing season, correspondingly larger emission potentials of new foliage during spring time would change the conclusion of the contribution of new Scots pine foliage to Scots pine canopy BVOC emissions. In order to obtain a complete understanding of the formation of new aerosol particles, it is especially crucial to investigate this importance of new Scots pine foliage to ecosystem BVOC emissions during spring time, since that is the time of year that new particle formation has been found to be most frequent (Vehkamäki et al., 2004;Dal Maso et al., 2005, 2008Manninen et al., 2010;Vana et al., 2016).
We investigated the importance of considering the contribution of enhanced constitutive emission potential of new Scots pine foliage on the whole tree's emission potential. We examined this as a function of season, stand age and location in Finland, utilising published emission rates by Aalto et al. (2014) and models to predict the seasonal and yearly growth of Scots pine foliage. In order to analyse the potential underestimation of regional emissions when the enhanced emissions from new foliage is not accounted for, we upscaled our results to answer how many Gg of carbon could be underestimated in the predictions of constitutive monoterpene emissions from Finland. Finally, we estimated how this underestimation impacts forecasts of formation and growth of new small particles. Our ultimate objective was to investigate and answer whether we need to introduce a more comprehensive treatment of the emissions of BVOCs from new coniferous foliage in biogenic emission models.

Yearly development of Scots pine needle mass
The yearly development of Scots pine needle mass was calculated for southern and northern Finland, by considering the total amount of needle age classes present in the stand and the maximum stand needle biomass. Hence, we defined that the stands carry 2.5 and 5.5 needle age classes in southern and northern Finland, respectively, which is based on observations from Finland Wang et al., 2013;Ťupek et al., 2015). A maximum stand needle biomass of 5000 kg ha -1 , which is representative for southern and middle Finland (Ilvesniemi and Liu, 2001), was used for southern Finland, while 3500 kg ha -1 , which is representative for a relatively poor site in Lapland , was used for northern as a lower estimate of the impact of the emission of monoterpenes from new foliage to the total stand emission. Finally, it is assumed that needle mass development follows a sigmoidal form (e.g. Mäkelä, 1997). Since tree foliage growth models usually omit simulating the growth of very young trees (e.g. Hari et al., 2008;Minunno et al., 2019), because of their low relevance with respect to e.g. biomass production, we likewise only modelled the growth of trees aged ≥10 years. The maximum stand needle mass in southern Finland is reached at the same time as the observed canopy closure at the SMEAR II station, Hyytiälä, southern Finland (e.g. Hari and Kulmala et al., 2001). It is assumed that the maximum is reached in northern Finland 15 years later, due to slower forest growth in the north (Fig. 1a). Since the stand foliage mass is higher in southern than northern Finland, and since fewer needle age classes prevail in the south, both the mass of new needles and the mass of senescing needles are significantly higher in southern than northern Finland (Fig 1b, Fig. 1c). The mass of new needles is calculated as: where is the growth of new needles during year i (kgC), is the maximum needle mass during year i (kgC) and G i N m i N S i N is senescence during year i. After canopy closure, and thus: where is needle longevity in the two locations. Since the foliage production rate is high in young stands (derivative of I j Fig. 1a), the fraction of new needles to the total stand needle mass is also higher in young than mature pine forest stands ( Fig. 1d).

Seasonal development of Scots pine needle mass
The seasonal development of Scots pine needle mass was modelled with the CASSIA growth model (Schiestl-Aalto et al., 2015), where the daily growth of tree organs is driven by environmental variables, mainly temperature. Scots pine needles start elongating in spring simultaneously with the shoot, but shoot length growth is completed approximately one month before the growth of needles finishes. The model considers two parameters, which need to be estimated for the location of interest. Those are: time of growth onset and time of growth cessation. CASSIA has previously been parameterized using growth data measured in 2008 at the SMEAR II station, and the model has been shown to successfully predict the growth of needles (Schiestl-Aalto et al., 2015). We used this parameterization of time of growth onset and time of growth cessation to predict the seasonal development of Scots pine needles in southern Finland, while the corresponding growth in northern Finland was predicted utilising needle growth measurements conducted at the SMEAR I station in Värriö, Finnish Lapland, during the 2017 growing season. Furthermore, the model considers needle length by the end of the growing season as a yearly varying parameter. This parameter can be modelled if needed, but as the final needle length was measured at both stations during years 2009-2011, we used the measured values. Additionally, the length of the needle primordia (i.e. the needles inside the bud) was set to 1 mm, and it was assumed that needle length is proportional to needle biomass (Aalto et al., 2014;Schiestl-Aalto et al., 2015 combining the behaviour shown in Fig. 2a with total stand needle mass values from Fig. 1a. The seasonal behaviour is also in accordance with observations (Rautiainen et al., 2012) before needles fall off. The fraction of new needles out of total stand needle mass for Scots pine stands of different ages growing in southern and northern Finland is provided in Fig. 2c.
This has been calculated by combining the behaviour shown in Fig. 2a with new stand needle mass values from Fig. 1c.

Emissions of monoterpenes
We utilised measured emission rates of monoterpenes and chamber temperatures described and published in Aalto et al. (2014), hence we refer to Aalto et al. (2014) for details on the measurement set-up. In brief, the shoot exchange of monoterpenes was measured with an automated gas-exchange enclosure system and analysed by PTR-QMS (Proton Transfer Reaction -Quadrupole Mass Spectrometer) from a ~50 year old Scots pine tree located at the SMEAR II station during 2009-2011. Only periods with data from both new and mature needles were considered. Since our analysis focused on emission potentials, we did not include exactly the same data as Aalto et al. (2014), because we were limited by occasional breaks in the measurements of chamber temperature. The emission rates were standardised by Eq. (5) in Guenther et al.
(1993) (T s = 30 °C, β = 0.09 °C -1 ) in order to compare to literature values. Thus, the presented potentials cannot be directly compared with and implemented into MEGAN (see e.g. Langford et al., 2017).
The ratios of the emission potential of new needles to the emission potential of mature needles for the growing seasons in 2009-2011 are presented in Fig. 3. The subfigures in Fig. 3 have been cut due to clarity, but the excluded outliers are compiled in Table A1 together with information about the total amount of data points considered per one week average.
As seen from the figure and also concluded by Aalto et al. (2014), new Scots pine needles have a much greater potential to emit monoterpenes than mature needles. The difference in the potential to emit decreases throughout the season, but lasts until the lignification of the shoot is finalised. Hence, young needles continue to have a higher potential to emit monoterpenes than mature needles until the end of August / beginning of September (Fig. 3f). Figure 3 also illustrates why continuous measurements of VOC emissions are needed for providing sound emission potentials; (1) there is a large spread in the emission rates, even when standardised, thus having only a few measurement points might lead to biased emission potentials, and (2) emission rates, and hence potentials, are seasonally dependant, which has been shown already earlier for Scots pine, but also for other tree species (e.g. Hakola et al., 2001Hakola et al., , 2006Wang et al., 2017;Karl et al., 2003;Komenda and Koppmann, 2002). Additionally, it is clear that temperature is not always sufficient in explaining short term fluctuations, as there are large variations in the emission potentials within the one-week averages.
The uncertainty on annual global emissions of monoterpenes into the atmosphere is estimated to be around a factor of three (Lamb et al., 1987;Guenther et al., 2012). This uncertainty originates from the used emission algorithm, biomass densities, land use distributions and emission potentials. About 15-25 % of the uncertainty is attributed to emission potentials (Lamb et al., 1987;Guenther et al., 2012). With this in mind, we present the monoterpene emission potentials of new and mature Scots pine needles, calculated based on Aalto et al. (2014), together with literature values, in Fig. 4. Literature values are included so that we have a larger basis to draw conclusions on. The literature values, which have also been standardised to 30 °C, represent different measurement years, locations, tree ages, needle ages, and measurement techniques (see Table A2). The requirement for including a study was that either the emission had been standardised to 30 °C or it was possible to (re)standardise it using the information provided in the paper. If the emission was not already standardised, a value of = 0.09 °C -1 was used as this is the most commonly used value in the literature for monoterpenes, though is known to vary and new foliage, respectively, during spring, though not as large as Aalto et al. (2014). However, such a seasonal pattern is not detected in all studies (e.g. not in Janson, 1993 andHakola et al., 2006). Räisänen et al. (2009), who provide emission potentials of new and mature needles, individually, show that the potential of new needles to emit monoterpenes is twice as high as that of mature needles. This is based on measurements from August-September, and is in accordance with findings by Aalto et al. (2014), who show that the difference in the potentials of the two needle age classes is about a factor of two in August (Fig. 3f).
By far most literature values, which are based on enclosure measurements, are reported to be within ~0.1 -2.3 g g -1 h -1 . This also includes the entirety of emission potentials of mature needles based on Aalto et al. (2014). A few points range up to ~6 g g -1 h -1 , while only one measurement point results in a potential of ~15 g g -1 h -1 (when data based on Aalto et al. (2014) is not considered). These few high potentials are based on measurements during spring and autumn on branches where both new and mature foliage were present, or in one case, only mature needles (Ruuskanen et al., 2005). The exceptionally high value of ~15 g g -1 h -1 originates from one measurement point of a mature shoot carrying buds (Tarvainen et al., 2005). The smallest reported potentials (~0.1 g g -1 h -1 ) are of new needles in the end of the growing season, and based on measurements by Aalto et al. (2014). The reported emission potentials of Scots pine seedlings are found in the lower end of the range (~0.2-0.9 g g -1 h -1 ), even though up to half of their needles are current year generation. However, the emissions from the seedlings were measured in the laboratory or in a research garden, and thus it is possible that the plants emit differently than plants growing in the field.
Five papers report ecosystem scale fluxes of Scots pine forests. Rinne et al. (2000) provide an ecosystem scale emission potential that is within the range reported from enclosure measurements (1.2 g g -1 h -1 ), while Rinne et al. (2007) and Räisänen et al. (2009) report values that are slightly higher than the general range (2.5 and 2.9 g g -1 h -1 ). The potential by Räisänen et al. (2009) is reported as a seasonal average (July -mid September) and is notably higher than the potentials based on Aalto et al. (2014) during the same time period. Canopy scale emission potentials by Taipale et al. (2011) and Rantala et al. (2015), which both measured in SMEAR II during separate years, are in a very good agreement with each other, though the micrometeorological method was different. Both studies observe a clear diminishment in the forest's potential to emit throughout the summer. The potential during April was, however, found to be less than during the summer months (Rantala et al., 2015), which can possibly be attributed to the fact that the potential represents the entire month of We calculated the importance of new Scots pine foliage on total canopy monoterpene emission potential using the means of the weekly medians of the monoterpene emission potentials from 2009-2011 (based on Aalto et al. (2014)). In our investigations, we also considered the minima and maxima of the weekly medians of the monoterpene emission potentials from the three measurement years (Fig. 5). The premise is that this is representative for southern Finland. In order to approximate the influence of new Scots pine needles in northern Finland, we assumed that the potentials of needles to emit monoterpenes are similar in southern and northern Finland, but that they depend on timing of foliage growth. Since the foliage growth onset at the SMEAR I station is delayed by two weeks of that seen at the SMEAR II station, also the monoterpene emission values -both for mature and new foliage -were delayed accordingly (Fig. 5). Since needle growth has been observed to end about 1 week earlier in northern than southern Finland (Fig. 2), the seasonally dependent emission potentials of northern Finland have been modulated likewise, thus, the emission potentials have been "squeezed" to fit the more intensive, but (~ three weeks) shorter period of growth in the north (Fig. 5). The presumption that the potential of the foliage to emit monoterpenes is similar in southern and northern Finland is supported by previous investigations on Scots pine (Tarvainen et al., 2005) and silver birch (Maja et al., 2015) in Finland. Finally, we assumed that all mature needles have the same potential to emit monoterpenes independent of their needle age class. Though Scots pine foliage preserves its ability to emit monoterpenes after a completed growing season (Vanhatalo et al., 2018), we only focus on the period of growth, as our interest lies in the difference that new and mature foliage presents. This difference diminishes by the end of the growing season, as the potentials to emit are then similar for all needle age classes.
In our analysis, we compared the canopy emission potential resulting from Aalto et al. (2014) with a canopy emission potential that assumes that the emission potential of current year needles is enhanced in a similar manner as in Guenther et al. (2012). This "MEGAN style" canopy emission potential has been calculated as: where new,MEGAN and F bud are the emission potential and fraction of new foliage before needle elongation properly starts, respectively, while growing,MEGAN and F new are the emission potential and fraction of new foliage during the period with a significant needle elongation rate, respectively. mature,MEGAN and F mature are the emission potential of mature foliage and fraction of mature foliage, respectively. Using the coefficients from Guenther et al. (2012, Table 4) that describe the relative emission rates of buds, growing and mature foliage, Eq. (3) can be reformulated to: which can be shortened to: since we did not consider periods with senescing needles. In our calculations, mature is from Fig

Scots pine forest stand coverage in Finland
We utilised the coverage of Scots pine forests in Finland of different tree age classes ( Fig. 6)  not accounted for, since no data is available. The coverage of Scots pine on forest land is 6.064×10 6 ha in southern Finland and 6.867×10 6 ha in northern Finland (Finnish Statistical Yearbook of Forestry 2014). In our calculations, we assumed that there is an even distribution of trees of all ages within each tree age class (Fig. 6). Hence, within the first tree age class (1-20 years), we excluded 45 % of the stand area, as it is assumed to be covered by trees aged 1-9 years.

The emission potentials of new and mature Scots pine foliage as a function of season
Though the emission potential of new foliage is high, the corresponding biomass is low. Hence, in order to investigate the importance of new foliage to the whole tree's foliage emission potential, the products of the emission potentials of new ( new ) and mature ( mature ) foliage, respectively ( Finland, while the corresponding contribution is ~60 -75 % in northern Finland, though at times it could be even higher. Though the new foliage biomass increases as the season progresses, the very high new foliage emission potential collapses in the beginning of the summer (Fig. 5), and the importance of the emissions from new Scots pine foliage therefore decreases as a function of the season (Fig. 7). The contribution of new Scots pine foliage to the whole tree's emissions decreases with tree age (Fig. 7), because the proportion of new foliage of the total stand foliage mass decreases with an increase in tree age ( Fig. 2c). Likewise, new foliage accounts for a larger fraction of the total Scots pine monoterpene emissions in southern than in northern Finland (Fig. 7), where needles are preserved for a longer time (Fig. 2c).

The importance of new foliage to the whole Scots pine tree's foliage emission potential
The canopy emission potentials ( new × F new + mature × F mature ), as a function of season for trees of various ages and locations, are compared, in Fig. 8, to (1) the emission potentials of mature foliage ( mature , Fig. 5c), as several widely used models (e.g.
LPJ-GUESS and ORCHIDEE) assume that the monoterpene emission potential is independent of needle age, and (2) canopy emission potentials that assume that the emission potentials of current year needles are enhanced in a similar manner as in Guenther et al. (2012) (see Sec. 2.3 for how this was calculated). We did not directly compare our canopy emission potentials to the potentials utilised in global BVOC models, as they do not use the same values, they do not utilise tree species specific, but instead plant functional type specific emission potentials, and often they assume some dependency on light. The underestimation of the whole Scots pine tree's needle emission potential caused by not considering the enhanced potential of new foliage is displayed in Fig. 8g-r. Models will greatly underpredict canopy emissions during the first ~2.5 months of the growing season in southern Finland if they assume that the monoterpene emission potential is independent of needle age or that the emission potential of new foliage is enhanced in a similar manner as in Guenther et al. (2012) (Fig. 8g-i, m-o). The underestimation will be less severe for predictions of emissions from northern than from southern Finland (e.g. up to a factor of ~7 vs ~29 for 10 year old forest), and more severe for younger than older stands (e.g. up to a factor of ~29 vs ~19 for 10 vs ≥50  of Scots pine growing in southern and northern Finland is less than a factor of 2.5 and 2, respectively. Values below the reference lines in Fig. 8g-r are caused by higher measured emission rates from mature than from current year needles (Aalto et al., 2014) at the end of the growing season. Assuming that the emission potential of new needles is enhanced as in Guenther et al. (2012) will only lead to a neglectable increase in the Scots pine canopy monoterpene emission potential (Fig.   8).
Canopy scale emission potentials by Taipale et al. (2011) and Rantala et al. (2015), derived from continuous micrometeorological flux measurements of a ~50 year old pine forest in SMEAR II, are included in Fig. 8c,i,o for comparison. Please be aware that the measured canopy, within an area with a radius of 200 m, is only covered by ~75% Scots pine (and ~25% other tree species), thus our results cannot be directly compared to Taipale et al. (2011) and Rantala et al. (2015), but these two studies provide the most suitable observations for validation of our results. We refer to Table A2 in the Appendix for details on how these potentials (per ground area) have been converted ( The reported canopy scale emission potentials agree very well with our suggested whole tree foliage emission potentials and the agreement is much better than that between Taipale et al. (2011) or Rantala et al. (2015) and assuming that the emission potential is independent of needle age or that the potential of new foliage is enhanced as in Guenther et al. (2012). Our enclosure-derived canopy emission potential overestimates the canopy micrometeorological-derived potential by a factor of ~1.6 during May, and then slightly underestimates it during the summer. The overestimation can partly be due to interannual variations in emission rates and seasonal foliage mass development, and partly due to plant-to-plant variations (as rates by Aalto et al. (2014) were conducted on one tree). An underestimation during summertime is expected, since the emission potentials by Taipale et al. (2011) and Rantala et al. (2015) consider all sources of monoterpenes in the ecosystem, and not only Scots pine foliage. These additional sources include at least Scots pine stems, forest floor, understory vegetation, Norway spruce (15 % of the stand) and deciduous species (~10 %) (Bäck et al., 2010;Aaltonen et al., 2011Aaltonen et al., , 2012Vanhatalo et al., 2015;Mäki et al., 2019).

Effects of stand age and season on the underestimation of the whole Scots pine tree's foliage emission potential
The underestimation of the whole Scots pine tree's needle emission potential caused by not considering the enhanced potential of new foliage, is presented in Fig. 9 as a function of tree age, for southern and northern Finland separately. The ranges in the underestimation are provided in Table A3. The underestimation has been calculated individually for the spring and for the full season, since new particle formation events have been shown to occur more frequently during March -May in both southern and northern Finland (Vehkamäki et al., 2004;Dal Maso et al., 2005Manninen et al., 2010;Nieminen et al., 2014;Vana et al., 2016). Hence, in our calculations, spring starts at the same time as emissions from new foliage is observed and lasts until the end of May, while the full season naturally includes the entire measurement period.
Trees aged less than 10 years are excluded from our analysis, as it might not be reasonable to extrapolate conclusions extracted from emission rate measurements of ~50 year old trees to very young trees. For example, Komenda and Koppmann (2002) showed that the emission potential of a 40 year old Scots pine tree was about five times higher than that of 3-4 year old seedlings. It should though be mentioned that measurements of seedlings were conducted in laboratory conditions, thus the difference in emission potential between seedlings and mature trees might be less. The underestimation caused by not considering the enhanced emissions from new foliage during the entire growing season in southern Finland is similar to not accounting for the greater emissions from new needles during the spring in northern Finland, especially in the cases of younger Scots pine tree stands. An additional important conclusion from Fig. 9 is that it seems that neglecting the age of the stand only leads to a minor error if the longevity of needles is short (max ~8 %), but to a larger error if more needle age classes prevail (max ~20 %). This is because the relative proportion of new needles in stands that carry more needle age classes varies more between individual stands of different ages (Fig. 2c). Tree age is not usually considered specifically in BVOC models, instead only the biomass and/or leaf area index is/are included.
The spring time differences in emission potentials lead to uncertainties in predictions of monoterpene emissions that are much greater than what has been estimated by Lamb et al. (1987) and Guenther et al. (2012). These investigators have estimated that the uncertainty on annual global emissions of monoterpenes into the atmosphere could be around a factor of three in total, with about 15-25 % of that uncertainty attributed to emission potentials (Lamb et al., 1987;Guenther et al., 2012). Guenther et al. (2012) emphasis that these uncertainties are estimated for annual global emissions, thus the uncertainty can be much greater for specific times and locations. Though the emissions from Scots pine species have been extensively measured, emissions during spring time have only relatively recently received more appropriate attention, thus spring time Scots pine BVOC emissions are currently not well represented in models and they are therefore connected with a larger-than-average uncertainty.

National level impacts caused by omitting the enhanced emissions from new Scots pine foliage
About 12.931⨉10 6 ha in Finland, i.e. ~43 % of the total land area in Finland, is covered by Scots pine forests (Finnish Statistical Yearbook of Forestry 2014). Hence, the underestimation of not considering the emission potential of new Scots pine foliage (Fig. 9) is upscaled to Finland in Fig. 10. This has been estimated by (1) calculating the mean of the underestimation shown in Fig. 9 within the respective tree age classes provided in Fig. 6, and (2) normalising the product of the mean foliage biomass (Fig. 1a) within each tree age class (Fig. 6) and the stand area within each tree age class (Fig. 6).
For this calculation, we have assumed that there is an even distribution of trees of all ages within each tree age class, and we have excluded the fraction of trees younger than 10 years old. Hence, it is assumed that there is no underestimation connected with the emission potential of Scots pine forest aged less than 10 years. The results presented in Fig. 10 only refer to underestimations in the emission potentials of Scots pine dominated areas and not to a general emission potential that would be representative for the entire Finland and hence also consider e.g. Norway Spruce and various deciduous species.
The national scale uncertainty is controlled by the uncertainty connected to trees aged ≥50 years, because the majority of trees in Finland are older than 50 years and their foliage mass is larger than that of younger trees. Thus, it seems largely unnecessary to include a tree age dependant emission potential for global annual calculations of BVOC emissions. However, an exclusion will lead to an error of up to 20 % in simulations of specific locations.

Emission potentials used in models
We emphasize that, in this study, we have not investigated of the emission potential, which are used in models, can lead to an underestimation. As ecosystem scale flux measurements become increasingly available, such data is progressively being incorporated into biogenic VOC emission models. This is fortunate, since such measurements capture the entire emissions from the ecosystem. Unfortunately, such measurements are most often conducted in summer. Thus, if the potentials they produce are not modulated by the seasons in models, a similar underestimation persists.
According to Guenther (2013), the emission potentials of Needleleaf Evergreen Boreal Trees in MEGAN v2.1 are based on enclosure and canopy micrometeorological measurements and landscape inverse modelling of various boreal forest species. However, almost all measurements of Scots pine utilised for compiling the monoterpene emission potential are enclosure measurements (Guenther, 2013). Results by Taipale et al. (2011) and Rantala et al. (2015) are not considered in MEGAN v2.1, at least in the latter case due to its (more) recent publication date. Micrometeorological measurements by Rinne et al. (2000Rinne et al. ( , 2007 and Räisänen et al. (2009) are considered (Guenther, 2013), but these measurements were mainly conducted during summer time. The monoterpene emission potential of the boreal needleleaf evergreen tree type in ORCHIDEE is extracted from the corresponding emission potentials used in Guenther et al. (2006Guenther et al. ( , 2012, and otherwise exclusively from literature on enclosure measurements when Scots pine is concerned (Messina et al., 2016). LPJ-GUESS by far mostly considers enclosure measurements for construction of their emission potentials, but as in the case of MEGAN, also ecosystem scale fluxes from Rinne et al. (2000) are used (Schurgers et al., 2009).
Monoterpenes are not the only atmospherically relevant VOCs that are being emitted in substantially greater quantities from new than mature Scots pine needles (Aalto et al., 2014). For example, Aalto et al. (2014) showed that the emission of methanol, acetone and 2-methyl-3-buten-2-ol from developing needles can contribute with up to about 50, 35, and 75 %, respectively, of the whole tree foliage emission in case of a ~50 year old Scots pine stand. It is also likely that emerging foliage of other conifers evergreen tree species would have a significantly higher potential to emit VOCs than its corresponding mature foliage. Thus, it should be reconsidered how the emission of all atmospherically important VOCs from new evergreen conifers foliage should be included in models.

Impacts on monoterpene emission predictions from Finland
The error of not accounting for new foliage monoterpene emissions in the canopy's emission potential translates directly into the predicted emission rates as emission potentials are multiplied with various activity factors in models in order to produce the emission rates (e.g. Guenther et al., 2006Guenther et al., , 2012. Thus, under the same environmental conditions and foliage mass or leaf area index, a change in the emission potential leads to a proportional change in the predicted emission rate (F): We investigated how many Gg of monoterpenes the emissions from Finland could be underestimated, if biogenic emission models only consider the emissions from mature foliage. For this analysis, we utilised Eq. (5) in Guenther et al. (1993) and considered the tree age (i) and time (j) dependant foliage mass per area (M, Fig. 2b) and the tree age dependant Scots pine stand area (A, Fig. 6): SMEAR I (9 m, Aalto et al., 2019b) stations. In our calculations, it is assumed that the temperature of all needles equals the ambient temperature, which is a reasonable assumption for low density canopies (Pier and McDuffie Jr., 1997;Martin et al., 1999;Zweifel et al., 2002;Leuzinger and Körner, 2007). T s and β are the same as in Sec. 2.3. Eq. (7) considers our suggested canopy scale emission potentials (Fig. 8) and our emission potential of mature needles (Fig. 8)

Impacts on predictions of new particle formation and growth
BVOCs, and especially monoterpenes, have been shown to participate in the formation (Kulmala et al., 1998Donahue et al., 2013;Riccobono et al., 2014;Schobesberger et al., 2013) and growth Riipinen et al., 2012) processes of the climatically important secondary organic aerosol particles in the atmosphere. As already stated earlier, the frequency of new particle formation events in boreal forests have been observed to be highest during spring time. We, therefore, The formation of neutral 2 nm sized clusters, J 2 , from sulfuric acid (H 2 SO 4 ) and oxidised organic compounds can be expressed as follows (Paasonen et al., 2010): where K s1-3 are kinetic coefficients. The condensational growth rate, GR, of 2-3 nm particles can be calculated as follows : where CC is the concentration of condensable vapours, which we assume to be the sum of sulfuric acid and organics. We assume that the molar mass of organics is four times higher than that of sulfuric acid  and hence we can write: Changes in the formation and growth rate depend on the absolute concentrations of sulfuric acid and oxidised organics.
Hence, we have calculated the impact on formation and growth rates utilising sulfuric acid concentrations of and concentrations of organic condensables of , which are reasonable ranges according 0 cm 1 À 1 Á 10 6 À3 cm 1 À 5 Á 10 7 À3 to measurements of sulfuric acid and estimates based on observations of growth rates, respectively (Paasonen et al., 2010).
The increase in the formation and growth rates are calculated in a similar manner as in the case of the emissions: (Y1-Y2)/Y2⨉100 %, where Y1 = emission, formation or growth rate considering the emission potential of both new and mature needles, and Y2 = emission, formation or growth rate considering only the emission potential of mature needles. In our calculations, we assume that simulations including the emission potential of both new and mature Scots pine foliage would lead to concentrations of organic condensables in the range of . Thus, [org] is decreased by a factor cm 1 À 5 Á 10 7 À3 of 2.8 (northern Finland) and 6.6 (southern Finland) in the calculations of the formation and growth rates using only the mature foliage emission potential. The resulting changes in the formation and growth rate are presented in Table 3 and illustrated in Fig. 11. Models would predict significantly higher formation and growth rates of small new particles during spring time, if they considered the enhanced emissions from new Scots pine foliage. Since the increase in emissions of monoterpenes would be highest in southern Finland, also the induction in the simulated new particle formation and growth would be greatest there. The scale of the enlargement largely depends on the ratios of concentrations of sulfuric acid and organics originating from monoterpene oxidation. Hence, the increases in the predicted formation and growth rates are modest at high [H 2 SO 4 ]/[org], but still greater than the uncertainty connected to the instrumentation used to obtain the rates Wagner et al., 2016;Kangasluoma and Kontkanen, 2017) and the uncertainty related to the calculation of these rates (Yli-Juuti et al., 2011). At low [H 2 SO 4 ]/[org] (e.g. ⅕ ⨉ 10 -1 cm -3 ), J 2 would be predicted to be ~10 times larger in southern Finland, when also considering the enhanced emissions from new foliage, while the corresponding growth rate would be ~6 times greater. Such increases in the predictions of new particle formation and growth would severely impact climate change predictions.

Conclusions
We have investigated the importance of considering the enhanced monoterpene emission potential of new Scots pine foliage on the whole tree's emission potential as a function of season, stand age and location. As methods, we used several years of continuous measurements of the emission rates of monoterpenes from new and mature Scots pine foliage, and growth models to predict the seasonal and yearly development of Scots pine needles. We found that the importance of the emissions Scots pine foliage is responsible for the majority of the whole tree's foliage emissions of monoterpenes, independently of tree age and location. We show that neglecting the specific age (but not biomass or leaf area index) of the stand at most leads to an error of ~20 % in simulations of specific locations. We demonstrate a good agreement between our whole tree foliage emission potentials, which account for the emissions from developing foliage, and monoterpene emission potentials derived from measured ecosystem scale fluxes of a Scots pine dominated forest. We also show that this agreement is much better than between the ecosystem scale-derived emission potentials and the emission potential of mature Scots pine foliage or the whole tree potential when it is assumed that the emission from new foliage is enhanced in a similar manner as in MEGAN v2.1.
Our results suggest that the emission of monoterpenes from Finland is underestimated by ~27 Gg monoterpenes / year, which corresponds to a very significant fraction of the total monoterpene emissions predicted from Finnish forests. The underestimation is especially severe during spring months where new particle formation is most frequent. Thus, the implications of our findings can lead to increases in the predictions of formation and growth rates of small particles during spring time in northern Finland by ~75-275 % and ~65-175 %, respectively, and in southern Finland by ~125-865 % and ~110-520 %, respectively. We conclude that new Scots pine foliage should be accounted for in biogenic emissions and atmospheric models, especially when simulating the spring season, either using separate enclosure measurements of new and mature foliage or by utilising ecosystem scale emissions conducted during spring time. We cannot make conclusions about the importance of new foliage of other tree species, but our findings calls for future investigations on other evergreen needle species. Author contributions. JA developed and calculated the yearly needle mass growth, PS-A calculated the seasonal needle mass development and wrote the corresponding method section, while DT conducted the remaining calculations. DT prepared the paper, with contributions from all co-authors.
Competing interests. The authors declare that they have no conflict of interest.  Tree Physiol., 22, 1125-1136, 10.1093/treephys/22.15-16.1125, 2002.   Table A2). Literature values, which have been re-standardised to 30 °C, represent different years and locations (see Table A2). "New", "mature", "bud", "seedling" and "ecosystem" indicate that the emissions were measured from either new or mature needles, from buds or seedlings or as an ecosystem scale flux. A "?" indicates that no information was provided about the age of the measured needles, but it does not include measurements from seedlings nor the entire ecosystem. The added error bars to literature values are those that the respective authors reported. range for the measured period, which is illustrated by the box in the figure. We refer to Table A2 for further details about the literature values used.   1.8 times that of mature needles, respectively (see Sec. 2.3), while "Mature needles" presume that the emission potential is independent of needle age. Canopy emission potentials for a ~50 year old Scots pine forest derived from micrometeorological flux measurements by Taipale et al. (2011) and Rantala et al. (2015) are included for comparison in c .
Ranges of the whole foliage emission potential are not included in this figure due to clarity, instead we refer the reader to  Figure 9. The underestimation of the whole Scots pine tree's needle emission potential caused by not considering the enhanced potential of new foliage, presented as a function of tree age. The underestimation has been calculated as: (the integral of "other study" -the integral of "This study") / the integral of "This study", where "other study" is either "MEGAN style" or "Mature needles" and the integrals are the areas under the curves presented in Fig. 8. The underestimation has been calculated for the spring and for the growing season separately and for both southern (S.F.) and northern (N.F.) Finland.
Ranges in the underestimation are not indicated in the figure due to clarity, but they are provided in Table A3.   These values are only for Scots pine and calculated using the total annual monoterpene emissions given in Kellomäki et al. (2001) Table 3. Observed ranges in the concentrations of sulfuric acid (H 2 SO 4 ) and condensable organics (org) together with the differences in the formation rate of 2 nm clusters (J 2 ) and growth rate of 2-3 nm particles (GR) when the increased emission potential of new Scots pine foliage is considered in addition to the emission potential of mature foliage, and compared to situations where only the emission potential of mature foliage is included. All values are for spring time, while the resulting differences (ΔJ 2 and ΔGR) are provided for northern and southern Finland, individually. The concentrations of condensable organics (org) predicted for northern and southern Finland, using only monoterpene emissions from mature foliage, are assumed to be 2.8 times (northern Finland) and 6.6 times (southern Finland) less than the observed concentrations.
[  -Table A3. The range in the underestimation of the whole Scots pine tree's needle emission potential caused by not considering the enhanced potential of new foliage (as shown in Fig. 9), presented for selected tree ages. The lower boundaries in the ranges have been calculated using the upper boundaries for the emission potential of mature needles and the lower boundaries for the emission potential of new needles (both from Fig. 7). Likewise, the upper boundaries in the ranges have been calculated using the lower boundaries for the emission potential of mature needles and the upper boundaries for the emission potential of new needles (both from Fig. 7). The ranges have been calculated as: (the integral of "other study" -the integral of "This study") / the integral of "This study", where "other study" is either "MEGAN style" or "Mature needles". The ranges in the underestimation are provided for the spring and for the growing season separately and for both southern (S.F.) and northern (N.F.) Finland.