Measurements were made at East Grange, Fife, Scotland (latitude–longitude (WGS84) 56∘5′21′′ N, 3∘37′52′′ W), elevation 45–60 m, one of the 16 SRF trial locations established by Forest Research (Harrison, 2010; Stokes, 2015). Soil type and texture at the site is surface-water gley and sandy silty loam, respectively, containing 4.9 % clay, 53.0 % silt and 42 % sand (Drewer et al., 2017; Keith et al., 2015). In 2010, the ex-agricultural site was planted with a single block of 40 randomised tree species plots and 8 control plots. Each plot (20m×20m) consisted of a single species containing 200 trees with a 2m×1m spacing arrangement (Harrison, 2010). Ten species were planted, and the two broadleaved species with the best survival and growth rates across the trials in the first 6 years, hybrid aspen (Populus tremula L. × P. tremuloides Michx.) and Italian alder (Alnus cordata Desf.), were selected for the measurements here, along with Sitka spruce (Picea sitchensis (Bong.) Carr, produced by vegetative propagation) (McEvoy, 2016; Parratt, 2018). After initial establishment of the young saplings, the site remained unmanaged. Branch and forest floor sampling chambers were installed in single south-facing plots of each species.

Meteorological data were collected from an unplanted plot in the middle of the site between May 2018 and July 2019. Minimum and maximum soil temperature (T107, Campbell Scientific, Shepshed, Leics, UK), air temperature and relative humidity (HMP45C, Campbell Scientific) were monitored hourly. In addition, photosynthetic active radiation (PAR, SKP 215 quantum sensor, Skye Instruments, Llandrindod Wells, UK) was measured at the same site every 5 min. Monthly means and ranges are provided in Fig. S1 in the Supplement. Occasional power failure at the site led to some missing data. For the modelling of BVOC emissions using the Pocket MEGAN 2.1 excel beta 3 calculator (Guenther et al., 2012) the missing PAR and mean temperature data were replaced by measurements from the Easter Bush site of the UK Centre for Ecology & Hydrology lying 45 km to the south-east (latitude–longitude (WGS84) 55∘51′44′′ N, 3∘12′20′′ W). A summary of the combined East Grange and Easter Bush data used in the model can be found in Fig. S2 in the Supplement.

The climate in east Scotland is colder, with fewer sunshine hours than in the south of England. To encompass these climate differences, meteorological data from Alice Holt Forest (51∘9′13′′ N, 0∘51′30′′ W), Hampshire, in southern England recorded during 2018 and 2019 were also used for the modelling and scaling up of the measured BVOC emission potentials from this study. A summary of the PAR and air temperature data for this field site is given in Fig. S3 in the Supplement.

Forest floor in this context includes soil, leaf litter, fallen small twigs/branches and flowers, understorey vegetation, microorganisms, and underground biomass that may all be sources of BVOC from the ground of the plantation. A static chamber method was used for the plantation floor enclosures. Polyvinyl chloride plastic soil collars (with a flange), 40 cm diameter×18 cm high, were installed per tree species plot prior to sampling (Asensio et al., 2007c, 2007b; Greenberg et al., 2012; Janson, 1993) and remained in the ground for the duration of the experiment. One or two collars were installed in 2017 and used during 2018. Additional collars were installed during 2018, resulting in a total of three soil collars per plot for the 2019 sampling. The collars were placed towards the centre of each plot to reduce the likelihood of plant debris from other plots contaminating them. Leaf litter and understorey vegetation were not removed from the collars prior to sampling to reflect actual changes in BVOC emissions with changes in the forest floor composition through the seasons.

A clear acrylic lid (with a foam lined flange), 40 cm diameter×22.5 cm high, was placed over the soil collar during sampling periods only, enclosing a total chamber volume of 51 L. The lid was sealed using clamps around the rim. A small 12 V axial fan (RS components Ltd, Colby, UK), 4cm×4cm×1cm, was attached to the chamber lid to mix the air inside the chamber (Janson, 1993). Samples of BVOC in the enclosed air were collected through PTFE tubing onto a 6 mm o.d. stainless-steel automated thermal desorption (ATD) cartridge (PerkinElmer, Waltham, MA, USA) packed with 200 mg Tenax TA 60/80 (11982 SUPELCO, Sigma-Aldrich, St Louis, MO, USA) and 100 mg Carbotrap 20/40 (20273 SUPELCO, Sigma-Aldrich) at a flow rate of 0.2 Lmin-1 using a handheld pump (210-1003MTX, SKC Ltd, Blandford Forum, UK). Samples were collected for 30 min after closure, equating to a total sample volume of 6 L. Pressure compensation was maintained through a small hole in the side of the chamber to prevent negative pressure inside the chamber and potential degassing of air from the soil pores. Ambient air samples were collected concurrently with the chamber sample in order to quantify BVOC emissions from the forest floor by difference. This is discussed further in Sect. 2.5.2. No ozone filter was used during sampling so amounts of some monoterpenes may have been reduced by reaction with ozone (Ortega et al., 2008). However, it has also been suggested that ozone may be lost by dry deposition onto the chamber walls in the first minute (Janson et al., 1999). Chamber air temperature (Electronic Temperature Instruments Ltd, Worthing, UK) and humidity (Fisherbrand™ Traceable™ humidity meter, Fisher Scientific, Loughborough, UK) were measured at the end of the 30 min sample collection period.

Volumetric soil moisture (ML3 ThetaProbe soil moisture sensor, Delta T, Cambridge, UK) was measured at three locations around each chamber, and soil temperature was measured at a single location at 7 cm depth close to but outside the soil collar to avoid disturbance of the forest floor. Both measurements were performed after sample collection to prevent perturbation of the ambient air sample.

A dynamic chamber method was used for branch enclosures. Three sample points were established per tree species plot and used to mount a removable flow-through acrylic chamber (Potosnak et al., 2013), 53 L in volume. The chambers were set up during each sampling visit and used to enclose a single branch, alternating between three similar branches per tree species. The branches were selected to be of similar size and in a similar position on the tree. All branches were approximately 1.5 m from the ground and in a south-facing position. Ambient airflow was delivered from an oil-free double-ended diaphragm pump (Capex V2, Charles Austen pumps Ltd, Surrey, UK) through PTFE tubing (Morrison et al., 2016; Purser et al., 2020) at a flow rate of 10 Lmin-1 to obtain the desirable air exchange rate of 4–5 min (Ortega and Helmig, 2008). In addition, the chamber contained a small 12 V axial fan (RS components Ltd, Colby, UK), 8cm×8cm×2.5cm, to ensure sufficient mixing of air inside the chamber.

After set-up, the branch enclosure was left for a period of 30 min to attain a steady state. Both the inside and outside of the enclosure were then sampled concurrently for 30 min at a flow rate of 0.2 Lmin-1 (total sample volume of 6 L) using a handheld pump (210-1003MTX, SKC Ltd, Blandford Forum, UK). In cases of low light levels, low temperatures or smaller volumes of foliage, the sampling time was sometimes extended (up to 60 min) to ensure sufficient sample was collected on the sample cartridge. Multiple sequential samples were taken over a given day. All enclosure sample tubes were stored in a fridge at 4 ∘C until analysis.

After BVOC sample collection, the leaves inside the chamber were counted, and a representative subsample of approximately 10 % of the total number of leaves on the measured branch were removed from a nearby branch. The leaves were dried at 70 ∘C until constant mass, typically after 48 h. In the case of the Sitka spruce, subsidiary branches were used. Measurements of chamber temperature and relative humidity (CS215, Campbell Scientific, Shepshed, UK) were made each minute during sampling. In addition, PAR (SKP 215 PAR quantum sensor, Skye Instruments, Llandrindod Wells, UK) was measured outside but next to the branch chamber, with measurements made every minute. The chambers had 85 % transparency to PAR (400–700 nm), so the measured PAR values were correspondingly adjusted to represent the illumination conditions inside the chamber.

As no substantial isoprene emissions were observed during an initial assessment, only monoterpenes were quantified from the forest floor. Monoterpene emissions from the forest floor (Ffloor) were calculated as micrograms (µg) of carbon for a given compound per ground surface area (µgCm-2h-1) using Eq. (1), where Csample is the concentration of a monoterpene inside the chamber (µgCL-1), Cambient is the concentration of a monoterpene in the ambient air outside the chamber (µgCL-1), A is the area of forest floor inside the chamber (m2), V is the volume inside the chamber and t is the sampling duration (min). 1Ffloor=Csample-Cambient×V×60A×t In some cases, the concentration in ambient air was larger than inside, resulting in a negative emission value, i.e. a net uptake.

The isoprene or monoterpene emission (Fbranch) from an enclosed branch was calculated as micrograms of carbon (C) for a given compound per leaf dry mass basis, µg C gdw-1 h-1, using Eq. (2), where f is the flow rate through the chamber (Lmin-1), and m is the dry mass (g) of foliage inside the chamber. 2Fbranch=Csample-Cambient×fm Isoprene emissions have previously been shown to be controlled by both light and temperature and were standardised to 30 ∘C and 1000 µmolm-2s-1, respectively (Guenther et al., 1993). Mean chamber air temperature and PAR for each period of sample collection were therefore used to standardise the measured Fbranch emissions for isoprene (Eqs. 3–5) and monoterpenes (Eq. 6) to facilitate comparison between this study and previous literature. The algorithms developed in Guenther et al. (1993) are subsequently referred to as G93.

The standardised isoprene emission rate Fisoprene at 30 ∘C and 1000 µmolm-2s-1 PAR is a function of the measured emission Fbranch, a term CL to correct for the effect of light and a term CT to correct for the effect of temperature Eq. (3). 3Fisoprene=FbranchCL×CT The light-correction term CL is calculated from Eq. (4), where α=0.0027 and CL1=1.066 are empirical coefficients in G93, and L is the experimentally measured mean PAR (µmolm-2s-1) during sampling. 4CL=αCL1L1+α2L2 The temperature-correction term CT is calculated using Eq. (5), in which the terms CT1 (95 000 Jmol-1), CT2 (230 000 Jmol-1) and TM (314 K) are all empirically derived coefficients from G93. R is the molar gas constant 8.314 JK-1mol-1, T is the mean air temperature (K) during sampling and Ts is the standardised temperature of 303.15 K, equivalent to 30 ∘C. 5CT=exp⁡CT1(T-Ts)RTsT1+exp⁡CT2(T-TM)RTsT Monoterpene emissions from branch chambers, Fbranch, were standardised to temperature based on the calculations from G93 using Eq. (). Ts is the standard temperature (303 K), and T is the mean air temperature during sampling. Fmonoterpene is the standardised monoterpene emission rate (µgCgdw-1h-1), and Fbranch is the measured monoterpene emission rate (µgCgdw-1h-1). 6Fbranch=Fmonoterpeneexp⁡β(T-Ts) Standardised isoprene and monoterpene emission rates from sequential samples calculated for a given day were then averaged to give a single standardised branch emission rate per tree species per measurement day. In addition, daily measurements were grouped into seasons to give a standardised emission potential per season, Fb_season.

Italian alder (Alnus cordata) emitted very low amounts of isoprene, ranging between <0.0005–0.035 (standardised 0.017–0.037 µggdw-1h-1) depending on season (Table 4), comparable with previous standardised emission rates reported as <0.1–3 (0.09–2.64 µggdw-1h-1) (Calfapietra et al., 2009). The equivalent median and interquartile ranges for the data collected during this study can be found in Table S4 in the Supplement.

Mean measured emissions for total monoterpene ranged between 0.041–0.393 (standardised 0.073–1.5µggdw-1h-1), with higher emission rates during spring and summer 2018 than in 2019. The major monoterpenes emitted were d-limonene, α-pinene, β-myrcene and β-pinene, which were consistently emitted through the spring and summer (Fig. 1). No previous data for total or speciated monoterpene emission rates from Italian alder could be found in the literature. However, other alder species have also been reported to be low emitters of monoterpenes and to emit slightly more monoterpenes than isoprene. Studies that report similar low levels of total monoterpene standardised emissions from alder include 0.8 from grey alder (Hakola et al., 1999), 0.13 from black (or common) alder (Aydin et al., 2014) and 1–2 µggdw-1h-1 from green alder (Alnus rugosa) (Isebrands et al., 1999). For speciated emissions, 3-carene, β-phellandrene, β-ocimene, p-cymene and sabinene have also been reported to be emitted from Alder spp. (Aydin et al., 2014; Copolovici et al., 2014; Hakola et al., 1999; Huber et al., 2000). Emissions of some monoterpenes such as β-myrcene are suggested to be induced by herbivory by aphids (Blande et al., 2010). However, since no data on the composition of monoterpenes under laboratory studies in the absence of herbivory are available for Italian alder, it is difficult to know which, if any, of the monoterpenes measured in our field study may have been induced by previous herbivory.

Measured isoprene emissions from hybrid aspen ranged from 0.053 to 21 (standardised 0.19–23 µgCgdw-1h-1) (Table 4). No measurements were made during autumn senescence or in winter on the bare branches. Measured emissions were lower in spring for the newly emerged leaves compared to summer (Fig. 1). As noted in Sect. 3.1, the onset of spring at the field site was earlier in 2018 compared to 2019. European aspen (Populus tremula) measured in late spring (May) 2 weeks after bud burst has also previously been reported to have a lower emission rate than in summer (Hakola et al., 1998). Isoprene emission rates made on leaves (not branches) on aspen in spring in the boreal forest were also reported to be a third of the emission rate measured in the middle of summer (Fuentes et al., 1999). In our study, the hybrid aspen plantation showed signs of stress thought to be associated with lower rainfall and soil moisture locally during summer 2018, causing a yellowing of leaves and early leaf shedding in July. It is widely accepted that isoprene emissions increase with increases in temperature and PAR (Guenther et al., 1991; Monson and Fall, 1989) but that under stress during drought isoprene can be emitted at much higher rates than usual, only to eventually decline as resources are depleted in the leaves (Brilli et al., 2007; Seco et al., 2015). However, standardised isoprene emissions measured during this study on green aspen leaves did not differ between the two years, 2018 (23) and 2019 (22 µgCgdw-1h-1), despite the signs of stress in 2018 noted above. The standardised isoprene emissions for hybrid aspen reported here were much lower than those previously reported for European aspen, 51 (i.e. 45 µgCgdw-1h-1) (Hakola et al., 1998).

Total monoterpene emissions measured for hybrid aspen ranged from 0.040–0.20 (standardised 0.082–0.259 µgCgdw-1h-1), with substantially higher emissions occurring in summer 2018 (Table 4, Fig. 1). Increased emissions for some monoterpenes have been shown to be predominately driven by increases in temperature (Guenther et al., 1991). In particular d-limonene, the major monoterpene emitted here, was found to correlate with an increase in temperature, comparable to elevated temperature experiments for European aspen (Hartikainen et al., 2009). However, total monoterpene emission rates were an order of magnitude lower in summer during our study, which is closer to the findings of Brilli et al. (2014) from a SRC plantation of poplar and in contrast to the 4.6 (4.1 µgCgdw-1h-1) reported for European aspen by Hakola et al. (1998). d-Limonene, α-pinene, carene and β-phellandrene collectively accounted for 50 %–95 % of the total measured monoterpene emissions, although the composition for different days was highly variable (Fig. 1). Emissions of α-phellandrene peaked at 27 % of total monoterpenes measured in April when catkins were present but were otherwise <13 % (except on 6 June 2018).

Previously studies on European aspen report monoterpene emissions of 3-carene, limonene, α-pinene, trans-ocimene, eucalyptol, β-myrcene, and sabinene (Aydin et al., 2014; Hakola et al., 1998; Hartikainen et al., 2009) and on hybrid aspen (Populus tremula–Populus tremuloides) report α-pinene, β-pinene and β-ocimene, (Blande et al., 2007), although differences between clones were noted.

Mean measured isoprene emissions from Sitka spruce ranged from 0.031 (standardised 0.14) in winter to 5.9 (standardised 15.0 µgCgdw-1h-1) in summer (Table 4), which are comparable to the range of previously reported emissions from UK field measurements, 0.005–1.48 (standardised 0.88–14.1 µgCgdw-1h-1) (Street et al., 1996). Standardised isoprene emissions were lower in spring than summer during both years in our study (Fig. 1). Standardised isoprene emissions in summer 2018 (15.0) were more than twice those in summer 2019 (6.8 µgCgdw-1h-1), likely reflective of the wetter and cooler conditions in 2019. However, laboratory measurements using trees acclimatised at a constant laboratory temperature of 20 ∘C and PAR of 1000 µmolm-2s-1 for a week prior to sampling showed emission rates similar to summer 2018 emission rates, 13.4 (11.8 µgCgdw-1h-1) (Hayward et al., 2004). The measured isoprene emissions in our study declined dramatically at higher chamber temperatures, > 31 ∘C, despite the high PAR levels. An optimum of 33 ∘C for isoprene emissions from Sitka spruce was noted by Street et al. (1996), although a higher optimum of 39 ∘C was suggested by Hayward et al. (2004) based on a laboratory study. We therefore suggest that Sitka spruce trees acclimatised under field conditions in Scotland with variable day and night temperatures and light levels may have a lower optimum temperature than observed under laboratory conditions. The previous suggestion that Sitka spruce reaches maximum emissions of isoprene at a low level of PAR of 300 µmolm-2s-1 (Hayward et al., 2004) was difficult to confirm under field conditions as high PAR values were correlated with high temperatures (Fig. 2). However, it is worth noting that the majority of field emissions collected by Street et al. (1996) align well with the emissions measured at lower PAR and temperature in this study (Fig. 2).

Total monoterpene emissions measured from Sitka spruce peaked on the 29 April 2019 (9.5 µgCgdw-1h-1), coinciding with the new shoot extension growth on the branches (Fig. 1). Monoterpene emissions have shown to be present in spring in advance of isoprene emissions for Norway spruce (Picea abies) (Hakola et al., 2003). Overall, monoterpene emissions were generally higher in summer than in spring (Table 4). Total monoterpene emissions were still higher in 2018 (standardised 4.5) than in 2019 (2.3 µgCgdw-1h-1) even once standardised to 30 ∘C, which could indicate an increased release of monoterpenes in response to the drier warmer conditions. The total monoterpene emissions in 2019 are comparable to the previously reported total monoterpene emission of 3.0 (2.6 µgCgdw-1h-1) from a laboratory study (Hayward et al., 2004). Monoterpene emissions from Sitka spruce comprised predominately of β-myrcene, d-limonene, α-pinene and eucalyptol, collectively accounting for 83 %–97 % of total monoterpenes across all measurement days (Fig. 1).

β-Myrcene was the most abundant, consistent with the findings of Geron et al. (2000), and has been reported to be highest during spring in leaf oils, associated with new growth in this species, only to decline later in the growing season (Hrutfiord et al., 1974), but this was not evident during our study. d-Limonene emission rates reported during our study are comparable in size to Hayward et al. (2004), although it is not the dominant monoterpene as previously reported. Furthermore, other studies have also reported limonene to be present in smaller quantities than α-pinene and β-myrcene (Beverland et al., 1996; Hrutfiord et al., 1974). Monoterpene composition was generally consistent between measurements throughout our study even though different branches and trees were measured. This may reflect that the trees grown via vegetative propagation could be from a genetically similar source. However, the variability between the previous literature discussed here may point towards the potential for different chemotypes within Sitka spruce, as suggested by Forrest (2011) and similar to that of Norway spruce (Hakola et al., 2017) and Scots pine (Bäck et al., 2012). Norway spruce has also been found to be a significant emitter of sesquiterpenes (Hakola et al., 2017). Given the dominance of Sitka spruce plantations in the UK (and Ireland), the potential for variation within this species and the limited literature data on BVOC emissions, we suggest further measurements are needed at the branch and canopy level to fully assess the terpenoid species composition and their subsequent impact on air quality.

Negative fluxes for total monoterpenes were measured on two occasions, 4 July and 24 July. The highest total monoterpene emissions were observed on 18 October 2018 (18) and 7 June 2019 (24 µgCm-2h-1) (Fig. 3). Day-to-day variations were associated to some degree with changes in chamber temperature and soil moisture (Fig. 3). Seasonal variations in mean emissions were also apparent (Table 5). The forest floor acted as a sink for monoterpenes during summer 2018 when there was bare soil inside the collars. During summer 2019 vegetation grew inside the soil collars and resulted in the forest floor being a more substantial source of monoterpenes (Fig. 4). Monoterpene composition reflected the seasonal changes that occurred on the forest floor. The monoterpenes emitted in autumn (October 2018) were dominated by d-limonene, α-pinene, 3-carene and some β-myrcene, consistent with the composition of Italian alder foliage and attributed to the accumulation of leaf litter. However, the profile in June 2019 during the highest total monoterpene emissions showed significant emissions of γ-terpinene and α-phellandrene and likely reflects the changing understorey vegetation, hogweed species, growing inside the chamber collars and which was only present in the alder plantations. The particular species at East Grange was not identified, but Heracleum mantegazzianum (giant hogweed) has been determined to be a substantial γ-terpinene emitter (Matoušková et al., 2019). This highlights the importance of the specific understorey vegetation to the overall monoterpene flux composition.

The highest measured total monoterpene emissions, 9.18 and 5.83 µgCm-2h-1, occurred in July 2018 and were associated with the lowest soil moisture and warm temperatures. In contrast, negative monoterpene emissions were also observed in July (24th) and seem to be associated with an increase in soil moisture (Fig. 3). Overall spring (0.30) and summer (0.06 µgCm-2h-1) total monoterpene emission rates in 2019 (Table 5) were smaller by an order of magnitude than in spring (0.71) and summer (3.84 µgCm-2h-1) 2018. Higher rainfall during 2019 (S1) resulted in increased soil moisture (Fig. 3), which may have suppressed some monoterpene emissions (Asensio et al., 2007b). In addition, during 2018, litterfall started in July and peaked in October, by which time the canopy had lost all its leaves.

The composition of the monoterpene emissions from the forest floor during 2018 was similar to those measured from the branch chambers (Fig. 1) and was consistent between days. The main monoterpenes comprised α-pinene, β-pinene, camphene, d-limonene and 3-carene. The contribution from the floor of an aspen plantation has not previously been investigated, although soils taken from underneath aspen (Populus tremula) trees showed d-limonene as the predominant monoterpene, with a maximum emission of 15.9 µgCm-2h-1 under laboratory conditions (Owen et al., 2007). Quantifiable emissions of monoterpene from the leaf litter of American aspen (Populus tremuloides) exist (Gray et al., 2010), although they are not chemically speciated

Total monoterpene emissions measured from the Sitka spruce forest floor peaked during July 2018 (66.5 µgCm-2h-1) and coincided with the highest chamber temperatures and the lowest soil moisture readings (Fig. 3). The lowest measured emissions (0.03 µgCm-2h-1) were observed on 12 April 2018 when the temperature was lowest (7.5 ∘C, Fig. 3), suggesting soil moisture and temperature are likely interacting controlling variables of monoterpene emissions. In addition, there were clear seasonal differences when measurement days were grouped. Mean measured summertime emissions of total monoterpenes from the forest floor in 2018 were larger than those measured in 2019 (Table 5). Temperatures measured in the chambers were 3 ∘C degrees higher on average during 2018 compared to 2019, which could have contributed to the higher observed emissions, although soil moisture at 7 cm depth was not significantly different. The young Sitka spruce plantation had litter present all year round unlike in the deciduous species plantations, but the covering was sparse (Fig. 4) compared to a mature plantation. Total monoterpene emissions measured in summer 2018 (40.3 µgCm-2h-1) were slightly higher but similar in magnitude to the 33.6 µgCm-2h-1 (29.6 µgCm-2h-1) previously reported for the upper-most layers of the floor in a mature Sitka spruce plantation (Hayward et al., 2001). Norway spruce plantation have also been reported to have a slightly higher emission rate at 50 µgCm-2h-1 (Janson et al., 1999).

The monoterpene composition profile in 2018 was comparable to 2019 and consistent with the branch emissions recorded during our study, with the major emitted monoterpenes being β-myrcene, α-pinene, β-pinene, d-limonene and camphene. β-Myrcene accounted for a larger percentage, 20 %–50 %, of emissions in summer 2019 compared to only 5 %–10 % in summer 2018 (Table 5), although there is no obvious explanation for this difference.

Forest floor and branch emissions were sometimes measured on the same occasion, enabling calculation of the contribution of each source to the total monoterpene emissions of the plantation per square metre of ground (based on non-standardised data) (Fig. 5). In most cases, particularly in summer, emissions from the canopy dominated. For Sitka spruce, high monoterpene emissions from the plantation occurred when canopy emissions were high, which supports previous summertime observations on conifer species that the forest floor contributes little to the overall forest monoterpene emissions (Hayward et al., 2001; Janson, 1993). We found that in some instances, more often in spring when canopy foliage was sparse (alder and aspen) or dormant due to cold temperatures (spruce), the forest floor contributed the majority of the plantation monoterpene emissions. This trend was also reported for conifer species in the boreal forest (Mäki et al., 2019b).

For hybrid aspen the opposite was true, with the forest floor contributing more in the summer, as a result of understorey vegetation or early litter fall, contributing up to 40 % of the total monoterpene emissions of the plantation. In the Italian alder plantation the contribution was more mixed. Canopy emissions in late winter/early spring were only from the alder flowers (catkins). The low observed emissions at this time of year from the forest floor were likely caused by colder temperatures and high soil moisture. However, later in spring (April) monoterpene emissions came largely from the forest floor (90 %) as understorey vegetation began to grow and soil temperatures also increased. The canopy at this point was at the stage of leaf emergence when the foliage was sparse and so contributed little to the overall emissions. However, by summer just over half of the monoterpenes came from the canopy (now in full foliage), and the forest floor contributed around 40 % of the monoterpenes, related to the presence of understorey vegetation.

This section discusses modelled emissions of BVOC from the canopy per square metre of ground. The bottom-up approach of estimating BVOC emissions in this study using the chamber technique is useful for determining the contribution of different ecosystem components to BVOC emissions, but in this section emissions do not include modelled forest floor emissions. It is noted that forest floor processes are still being integrated into models in order to reliably capture the full complexity of the forest floor BVOC emissions for prediction purposes (Tang et al., 2019).

Mean standardised summertime emission factors for each tree species in Sect. 3.3 (derived using the simplified G93 algorithms) (Table 3) were adjusted on an hourly basis by the Pocket MEGAN 2.1 excel beta 3 calculator to derive hourly BVOC emissions per unit of ground area (Guenther et al., 2012). This allowed for a more advanced method of estimation of monthly and subsequent annual BVOC emissions from the canopy across two years (2018–2019) and two locations, East Grange (Scotland) and Alice Holt (England), for a given air temperature, PAR, and the influence of these parameters over the previous 24 and 240 h. In addition, changing LAI across the year (Table 2) had an influence on the biomass density of the canopy, which influenced the emission rate of BVOCs per unit area of ground. Similar to previous modelling studies (Ashworth et al., 2015; Zenone et al., 2016), standardised mean summertime measurements were used as the basis for this calculation.

Given the above, modelled mean diurnal canopy emissions of isoprene for hybrid aspen were calculated to be approximately 2, rising to a maximum of 7 mgCmground-2h-1 in July, the warmest month, across both years (Fig. 6a). These modelled emissions for the UK are broadly comparable to those reported from measured eddy covariance flux measurements above a hardwood forest, comprising primarily of aspen (Populus tremuloides and Populus grandidentata, LAI: 3.24–3.75) in Michigan, USA, and the boreal forest in Canada (predominantly Populus tremuloides, LAI: 2.4), where the mean summertime emissions are reported to peak at 11 and 6.87 mgCmground-2h-1 respectively (Fuentes et al., 1999; Pressley et al., 2006).

Mean total monoterpene emissions are 2 orders of magnitude smaller than isoprene (Fig. 6b) for hybrid aspen. Figure 6c and d highlight the difference in the relative magnitudes of emissions between the three SRF species. Mean emissions from the canopy of Italian alder for isoprene (0.002) and monoterpene (0.05 mgCmground-2h-1) were very small, and no above-canopy measurements could be found in the literature for comparison. For Sitka spruce mean canopy-scale emissions for July in Scotland were modelled to be 1.5 and 0.5 mgCmground-2h-1 for isoprene and total monoterpene respectively. There has only been one attempt in the UK to quantify BVOC directly above a Sitka spruce plantation (Beverland et al., 1996), where a relaxed eddy accumulation system was used and mean isoprene emissions were reported to be 0.146 mgCmground-2h-1 in a 24 h period in early July (temperature range 7–19 ∘C). These emissions are much lower than our model estimates, although it was reported that there were analytical difficulties with the micrometeorological techniques and limited data which could account for this disparity.

Table 6 shows the modelled annual BVOC emissions per hectare of plantation for each species for the two meteorological years (2018–2019) at East Grange in Scotland and for the contemporaneous meteorology experienced in southern England (at Alice Holt). The modelled annual fluxes of isoprene and total monoterpenes per hectare of Sitka spruce plantation averaged over the two contrasting years were roughly similar, at 13.8 and 15.7 kgCha-1yr-1, respectively. Hybrid aspen was modelled to emit only an average of 0.3 total monoterpene but much more isoprene (15.5 kgCha-1yr-1), whereas the model estimated that Italian alder emitted minimal isoprene (0.02 on average) but larger monoterpene emissions of 0.81kgCha-1yr-1.

It is worth noting that use of an mean summer flux could lead to a potential overestimation of emissions during other seasons and the subsequent total annual flux. Modelled isoprene emissions from Sitka spruce during 2018 for both East Grange and Alice Holt were higher than monoterpene emissions. In 2019, however, monoterpene emissions were more abundant than isoprene emissions using the East Grange meteorology data and of the same magnitude using the Alice Holt meteorology data. The lower PAR during 2019, which was more pronounced for East Grange than Alice Holt, limited the isoprene emissions. Monoterpenes were less affected as these were only temperature driven. The relative proportions of isoprene and monoterpenes in the atmosphere are important since they have differing effects on the formation and concentration of atmospheric pollutants such as ozone and secondary organic aerosol (SOA) (Bonn et al., 2017; Heinritzi et al., 2020). Long-term BVOC emission measurement above Sitka spruce plantations is needed to confirm this model observation.