Biogenic Volatile Organic Compound (BVOC) Emissions from Various Endemic Tree Species in Turkey

Compositions of biogenic volatile organic compound (BVOC) emissions from seven endemic tree species (Troy Fir, Uludag Fir, Cilician Fir, Oriental Sweetgum, Boz Pirnal Oak, Ispir Oak and Vulcanic Oak) in Turkey were determined. Field samplings were carried out in the forested areas using a specific dynamic enclosure system during the summers of 2011 and 2012. The selected branches of tree species were enclosed in a chamber consisted of a transparent Nalofan bag. The air-flows were sampled from both inlet and outlet of the chamber by Tenax-filled sorbent tubes in the presence of sunlight. Isoprene, monoterpenes, sesquiterpenes, oxygenated sesquiterpenes and other oxygenated compounds including sixty five BVOC species were analyzed with a GC/MS system. Temperature, humidity, photosynthetically active radiation (PAR) and CO2 concentrations were monitored both inside the enclosure and in ambient air. Calculated emission rates were normalized to standard conditions (1000 μmol/m s PAR and 30°C temperature). Ispir Oak, Oriental Sweetgum and Cilician Fir were the highest BVOC emitters with total normalized emission rates of 19.4 ± 19.2, 16.3 ± 16.1 and 15.5 ± 11.4 μg/g/h, respectively while Boz Pirnal Oak had the lowest emission rate of 0.84 ± 0.68 μg/g/h. Alpha-pinene, betapinene, beta-myrcene and limonene were the compounds dominating the monoterpene emission profiles while transcaryophyllene, isolongifolene, alpha-humulene and copaene were the prominent sesquiterpenes. Predominant oxygenated compounds were also found as eucalyptol, linalool-L and alpha-terpineol. As reported in the literature, coniferous and broad-leaved species were predominantly monoterpene and isoprene emitters, respectively. Oxygenated compounds were the third most prominent BVOC group and sesquiterpenes had relatively lower contributions for all species.


INTRODUCTION
Vegetation covering the landmasses releases several biogenic volatile organic compounds (BVOCs). BVOCs including terpenes (isoprene, monoterpenes and sesquiterpenes) and oxygenated compounds (alcohols, aldehydes, ketones, acetates) are important for atmospheric chemistry since they contribute to secondary organic aerosol formation and play an important role in the oxidative capacity of the atmosphere (Andreae and Crutzen, 1997;Fuentes et al., 2000;Niinemets et al., 2011;Sun et al., 2012). Furthermore, because of their high reactivities, BVOCs affect the chemical composition of the atmosphere as they photochemically react with nitrogen oxides (NO x ) and form tropospheric ozone (Fehsenfeld, 1992;Roselle, 1994;Simpson, 1995;Bonn and Moortgat, 2003;Fares et al., 2011).
BVOC emissions are mostly sampled by leaf/needle or branch enclosure measurements for the plant species either in laboratory or on-site (Komenda et al., 2002;Padhy and Varshney, 2005;Dominguez-Taylor et al., 2007;Joo et al., 2010;Fares et al., 2011;Matsunaga et al., 2011;Baghi et al., 2012;Matsunaga et al., 2013). In these studies, BVOC emission compositions of numerous plant species including common and several site-specific (endemic) species were determined, and the attention was generally focused on trees as the most substantial component of vegetation Tambunan et al., 2006;Räisänen et al., 2009;Fares et al., 2011).
In order to understand biochemical emission mechanisms and related controlling factors in detail, long-term and one or a few species based studies can be more reasonable (Harley et al., 1996;Kempf et al., 1996;Boissard et al., 2001). Stone pine (Pinus pinea) and Scots pine (Pinus sylvestris l.) are two common trees that were frequently investigated for their seasonal and diurnal BVOC emission patterns in several studies Street et al., 1997;Sabillon andCremades, 2001, Komenda andKoppmann, 2002;Räisänen et al., 2009). Several tree species were also included in some studies to determine their species-specific emission rates Padhy and Varshney, 2005;Calfapietra et al., 2009). For example, Geron et al. (2001) investigated 18 American oak (Quercus) species and 6 species from other genera. In Japan, Tambunan et al. (2006) also determined isoprene emissions of 42 indigenous species.
Getting knowledge on vegetative emissions of endemic species (that do not naturally grow in any other region) in a selected region may be important to provide information on local air composition. Japanese Cedar (Cryptomeria japonica) and Bamboo-leaf Oak (Quercus myrsinaefolia) were studied for their monoterpene and isoprene emissions by (Bao et al., 2008;Matsunaga et al., 2011;Matsunaga et al., 2012b;Matsunaga et al., 2013). Three endemic Mexican trees; Sacred Fir (Abies religiosa), Netleaf Oak (Quercus rugosa née) and Patula Pine (Pinus patula) were also investigated for their isoprene and monoterpene emissions by Dominguez-Taylor et al. (2007) to guide the future biogenic emission inventories for the Mexico City area. However, there is only one study in the literature that directly focused on BVOC emissions from forested areas in Turkey (Aydin et al., 2014). The objective of this study was to determine the normalized BVOC emission rates (at 1000 µmol/m 2 s PAR and 30°C temperature) of isoprene, monoterpenes, sesquiterpenes, oxygenated sesquiterpenes and oxygenated compounds from seven endemic tree species in Turkey. Seven tree species, i.e., Troy Fir (Abies nordmanniana subsp. equi-trojani), Uludag Fir (Abies nordmanniana subsp. bornmuelleriana), Cilician Fir (Abies cilicica subsp. isaurica), Oriental Sweetgum (Liquidambar orientalis), Boz Pirnal Oak (Quercus aucheri), Ispir Oak (Quercus macranthera subsp. syspirensis), Vulcanic Oak (Quercus vulcanica), were investigated by conducting on-site measurement campaigns in different regions.

Sampling Program
On-site samplings were conducted in the natural habitats of seven tree species between June 2011-August 2012. Three middle-aged (20-40 years) and healthy (i.e., no pest or aphid affected, having well-developed stem and branches, not obstructed by neighboring plants, favorable soil conditions and no other biotic/abiotic stress sources) trees were selected for sampling. BVOC samples were collected concurrently from two different branches of three selected trees of each species. Thus, BVOC emission rates of each species were calculated by averaging six values (3 trees × 2 branches). Measurements were conducted during photosynthesis of trees in the presence of sunlight (generally between 10:00 a.m.-14:00 p.m.) in summer months. All sampled trees were selected in open parts of the forest canopy to avoid underestimation of emissions due to the shaded leaves during the samplings. The collected samples were kept at 4°C until they were analyzed in the laboratory. Table 1 illustrates the sampling program including the studied species, sampling regions, sampling dates, and environmental conditions during the samplings. The sampling sites are also shown in Fig. 1.
Seven endemic tree species consisting of three coniferous and four broad-leaved trees were investigated. Turkey has four of the forty fir species known in the world which covers 2136 km 2 land area as pure fir forestlands, and three of these (Troy Fir, Cilician Fir, Uludag Fir) are among the endemics investigated in this study (GDF, 2010). The four broad-leaved species included three oaks and a sweetgum species. Oriental Sweetgum (or Turkish Sweetgum) is a rare endemic tree such that has a narrow distribution only about 5 km 2 , additionally it has a special oiled secretion (balsam or sweetgum oil) that is used as a medicinal product like the other sweetgum species (Celik et al., 1997;GDF, 2010). Oak (Quercus) species have the largest natural reserve in Turkey's forestland inventory including 18 species, 9 subspecies, 2 varieties and 7 natural hybrids (GDF, 2004).

Sampling Method
A specific dynamic branch enclosure system was used for sampling of BVOC emissions. Selected healthy branches of the trees were enclosed into transparent Nalofan bags (7 L). The air-flows were sampled from both inlet and outlet of the enclosure by Tenax-filled adsorbent sorbent tubes. Nalofan bags have been commonly used for air sampling due to their inertness. Some recent applications using Nalofan bag as a chamber include the sampling of odorous gas emissions (Dincer et al., 2006), volatile organic compounds (VOCs) emissions from an apple tree (Vallat et al., 2005), a fig tree (Grison-Pige et al., 2001), moss roses (Caisard et al., 2006), and maize (Hoballah et al., 2004). Temperature, humidity, CO 2 and PAR (photosynthetically active radiation) were continuously monitored and recorded both in the enclosure and in ambient air. Conditions in the enclosure were regulated to provide similar conditions to ambient air. Air circulation inside the enclosure was provided by consistent air inflow and outflows via two Teflon tubes connected to diaphragm pumps. The inflow air was purified by passing through three sequential columns containing silica gel, potassium iodide (KI) and activated carbon to remove humidity, ozone, and VOCs in ambient air, respectively. The operational scheme of the measurement system was given in Fig. 2. Both purified inflow and BVOC contained outflow rates were regulated by automatic mass flow controllers (MFCs) (Aalborg DFC26) at 5.83 L/min and 0.15 L/min, respectively. There were also two other Teflon tubes placed in the enclosure for by-passing the excess air and supplying air to the CO 2 analyzer at a flow-rate of 1 L/h. In order to avoid greenhouse effect in the enclosure under sunlight, cold water was circulated by a spiral Teflon tubing inside the bag using a circulating water bath. The integrated single humidity/temperature probe was also placed inside the chamber using a small cone-shaped case to protect the probe from direct sunlight. PAR sensor  has to be leveled horizontally to make accurate measurements. Since it was not possible to place the sensor inside the chamber in this way, PAR was monitored only outside the chamber. An experiment was conducted to determine the PAR attenuation inside the bag. One of the two identical PAR sensors was placed into a blank enclosure chamber while the second one outside the chamber. PAR values were monitored concurrently and recorded in one minute intervals for several hours covering a wide range of solar angles. During this experiment, average PAR values outside and inside the chamber were 1145 ± 158 and 1037 ± 137 µmol/m 2 s, respectively. Average attenuation rate of the bag was determined as 10 ± 2% and no systematic difference was observed with the varying solar angle. Based on this experimental extinction rate, the PAR values in the chamber were determined by multiplying those measured outside by a correction factor of 0.90.
Installation of the enclosure chamber was the most critical operation of the sampling that should be achieved without any damage/disturbance to branches and leaves since this may cause overestimated emissions. For the first two hours of the experiment, enclosure system was purged with purified ambient air to provide steady-state environmental conditions and to stabilize the emissions from possible initial disturbances during installation of the chamber. The purge time was selected as 2 h based on a recent study indicating that emissions are stabilized within this period (Oz, 2012). Following the purging, the air was sampled simultaneously from both the inlet and outlet lines for an hour. Tenax TA filled sorbent tubes were used to adsorb BVOCs from the enclosure. Enclosed branches were harvested at the end of the samplings to determine the dry foliage weight and leaf area. Dry foliage weight was obtained by drying the sample in an oven at 60°C for 48 hours while total leaf area was measured with a portable leaf area meter (Li-Cor LI-3000C). Using these data, both dry foliage weight and leaf area based emission rates were calculated.

Laboratory Analysis
Samples were analyzed for 65 compounds (isoprene, 21 monoterpenes, 15 sequiterpenes, 3 oxygenated sesquiterpenes, and 25 oxygenated VOCs) with a gas chromatograph (GC) (Agilent 6890N, Agilent) equipped with a mass selective detector (Agilent 5973 inert MSD, Agilent) and a thermal desorber (Tekmar, Aerotrap 6000). Samples were desorbed for 10 min at 240°C using helium flow at the rate of 40 mL/min. Internal trap temperature during sample desorption was kept at 35°C. The trap was desorbed for 1 min at 240°C. Then, it was baked for 10 min at 250°C. Valve oven and transfer line temperature of the thermal desorber was 200°C.
The chromatographic column was HP5-MS (30 m, 0.25 mm, 0.25 µm) and the carrier gas was helium at 1 mL/min flowrate and 37 cm/s linear velocity. The split ratio was 1:40. The inlet temperature was 240°C. Temperature program for VOCs was: initial oven temperature 40°C, hold 3 min, 40°C to 120°C at 5 °C/min, hold 5 min, 120°C to 240°C at 30 °C/min, hold 3 min. Ionization mode of the MS was electron impact (EI). Ion source, quadrupole, and GC/MSD  interface temperatures were 230, 150, and 280°C, respectively. GC/MS was operated at "scan" and "selected ion monitoring" modes simultaneously. Compounds were identified based on their retention times (within ± 0.05 minutes of the retention time of calibration standard), target and qualifier ions and were quantified using the external standard calibration procedure. Calibration standards in methanol including 65 compounds (isoprene, 21 monoterpenes, 15 sequiterpenes, 3 oxygenated sesquiterpenes, and 25 oxygenated VOCs, see also Table 2) were obtained from SPEX CertiPrep, UK. Six levels of VOC solutions were prepared in methanol as the calibration standards. Thermal desorption tubes used for calibration were loaded by spiking with 1 µL of the calibration standards following the US EPA TO-17 Method (U.S. EPA, 1999). The standard spike volume was selected based on the recent studies, indicating that 1 µL spike volume does not interfere with the analytes due to solvent effect (Kim and Kim, 2012;Kim et al., 2013). Then, the standard loaded tubes were run at specified conditions to calibrate the analytical system (Thermal desorber-GC-MS) (Odabasi et al., 2005;Dincer et al., 2006). In all cases linear fit was good with r 2 > 0.999.

Quality Assurance/Quality Control
Instrumental detection limits (IDL) were determined from linear extrapolation, based on the lowest standard in calibration curve and using the area of a peak having a signal/noise ratio of 3. The quantifiable VOC amounts were between 5.7 (p-cymene) and 295 pg (guaiol). Blank Tenax tubes were routinely placed in the field to determine if there was any contamination during sample handling and preparation. The method detection limit (MDL, pg) was defined as the mean blank mass plus three standard deviations. Instrumental detection limit was used for the compounds that were not detected in blanks. MDL ranged between 7.5 (trans-dihydro-b-terpineol)-1455 (crotonaldehyde) pg. Average VOC amounts in blanks were less than 5% of the amounts found in samples. Samples were blank-corrected by subtracting the average blank amount from the sample amount. MDL determined using an average sampling volume of 0.009 m 3 ranged between 0.0008 (trans-dihydrob-terpineol) and 0.16 (crotonaldehyde) µg/m 3 .
The system performance was confirmed daily by analyzing a midrange calibration standard. If the relative standard deviation from the initial calibration was < 10%, system was recalibrated. Analytical precision determined from three pairs of duplicate samples ranged between 2-5%.
Inlet concentrations of the chamber were close to blank concentrations indicating that activated carbon trap effectively removed the BVOCs from purge air. Use of the cold water circulation kept the temperature difference between the ambient air and bag within 5°C. Net photosynthesis rates during the samplings were also calculated using the leaf area, inlet and outlet CO 2 concentrations, and purge flow rate. Average rates ranged between -1110 and 27600 µmol CO 2 /m 2 h, indicating that photosynthesis was the dominant mechanism rather than respiration during the samplings.
For six samples, a back-up tube was connected in series with the sample tube during sampling from the enclosure to check if there was any breakthrough. Back-up tubes contained similar BVOC quantities as the blanks indicating that the breakthrough from the sample tubes was not a problem during sampling.

Calculation of Normalized BVOC Emissions Rates
In this study, emission rates (ε, µg compound/g h) of the investigated BVOCs were determined using the measured BVOC concentrations, air flow-rates, sampling period and dry weight (or surface area) of the enclosed leaves as given in Eq. (1) .
where, C outlet and C inlet are the concentrations (µg/m 3 ) measured at the outlet and inlet of the chamber, respectively; Q is the air flow-rate (m 3 /h) pumped into the chamber; t is the sampling period (h), and m is the dry mass (g) of enclosed leaves. Since the environmental conditions varied during samplings, all experimentally determined emission rates were normalized to standard conditions (30°C temperature and 1000 µmol/m 2 s PAR for isoprene and 30°C temperature for the remaining compounds). The temperature and light dependent algorithm by Guenther et al. (1993) for nonstoring species was used to normalize the BVOC emissions.
Monoterpene emission rates under experimental conditions were normalized depending on the enclosure temperature as follows (Guenther et al., 1993): where, ε mtp is the monoterpene emission rate under experimental conditions, ε mtps is the normalized emission rate under standard conditions (30°C), T is the enclosure temperature (K), T s is the standard temperature (303 K), β is the empirical constant and γ mtp is the correction factor. β has been determined previously from measurements of the emission rate as a function of temperature. It is typically around 0.09-0.10 for monoterpenes (Guenther et al., 1993;Guenther et al., 2006) and 0.17-0.25 for sesquiterpenes (Helmig et al., 2006;Matsunaga et al., 2011). In the present study, β was used as 0.10 for monoterpenes and oxygenated BVOCs, and 0.17 for sesquiterpenes and oxygenated sesquiterpenes since they are used as the default parameters for MEGAN model . Isoprene emission rate (ε isos ) under standard conditions (30°C and 1000 µmol/m 2 s PAR) is determined by correcting the emission rate (ε iso ) at any temperature and PAR with temperature (C T ) and light (C L ) correction factors (Guenther et al., 1993): Light (C L ) and temperature (C T ) based correction factors are formulated as: where, C L1 (1.066) and α (0.0027) are the empirical constants, L is te PAR value (µmol/m 2 s).
where, T is the enclosure temperature (K), R is the ideal gas constant (8.314 J/K/mol), T s is the standard temperature (303 K), T M (314 K), C T1 (95000 J/mol) and C T2 (230000 J/mol) are the empirical constants. A clear light dependency for several Mediterranean terpene-storing and non-storing species emitting large amounts of monoterpene emissions has been established in the literature (Loreto et al., 1996;Bertin et al., 1997;Staudt et al., 1997;Llusia and Penuelas, 1998;Penuelas and Llusia, 1999;Llusia and Penuelas, 2000;Owen et al., 2002;Keenan et al., 2009;Demarcke et al., 2010). In this study, the Guenther algorithm referred above was used to normalize the measured monoterpene emissions from all tree species since the recent studies have taken into account this approach (Baghi et al., 2012;Matsunaga et al., 2012a;Pokorska et al., 2012a, b;Helmig et al., 2013;Kajos et al., 2013;Matsunaga et al., 2013). Morever, there is not any information in the literature as to whether these endemic species are terpene-storing or non-storing.

Normalized BVOC Emission Rates
Emissions for 65 analyzed BVOCs were calculated under experimental conditions and then, normalized to standard conditions of 30°C temperature and 1000 µmol/m 2 s PAR or to 30°C only (see the section Calculation of Normalized BVOC Emissions Rates). Table 2 summarizes the average normalized emission rates of 65 BVOC species for seven endemic tree species in Turkey.
On the basis of total emissions, Ispir Oak, Oriental Sweetgum and Cilician Fir were the highest three BVOC emitters with the total normalized emission rates of 19.4 ± 19.2, 16.3 ± 16.1 and 15.5 ± 11.4 µg/g/h, respectively while Boz Pirnal Oak had the lowest normalized emission rate of 0.84 ± 0.68 µg/g/h. Isoprene was the prominent compound for Cilician Fir, Vulcanic Oak, Ispir Oak and Oriental Sweetgum while the others were mainly monoterpene emitters (Fig. 3). Monoterpenes dominated the emissions of three species, i.e., Boz Pirnal Oak, Troy Fir, and Uludag Fir. Boz Pirnal Oak had the highest monoterpene ratio of 93% in total BVOC emissions and Uludag Fir had the highest emission rate of 2.27 ± 1.26 µg/g/h. On the other hand, sesquiterpenes were a relatively less significant BVOC group as they contributed less than 0.8% to total emissions. Oxygenated compounds had higher percentages than sesquiterpenes for all tree species and even than isoprene for Boz Pirnal Oak. Furthermore, 15 and 11% of all emissions consisted of oxygenated compounds for Troy Fir and Uludag Fir, respectively.
The ratio of monoterpenes to sesquiterpenes was different among the species. The ratio ranged from 1 to 193 for Ispir Oak and Troy Fir, respectively. The composition pattern could be classified into three types; monoterpene-rich group (Troy Fir, Boz Pirnal Oak), sesquiterpene-rich group (Vulcanic

Characterization of BVOC Emissions
Alpha-pinene, beta-pinene, beta-myrcene and limonene are the most characteristic compounds specifying the general monoterpene emission profiles (Table 2 and Fig. 4). Limonene was emitted by all trees except by Cilician Fir, while beta-phellandrene was emitted only by Cilician Fir. Monoterpene emissions of Boz Pirnal Oak and Oriental Sweetgum mainly consisted of trans-beta-ocimene (77%) and sabinene (82%), respectively. As previously mentioned, Oriental Sweetgum is a featured tree with its unique essential oils and secretions that are used for medicinal purposes, therefore, sabinene might be a marker compound for its extracts. Even they showed lower contributions to total emissions, trans-caryophyllene, isolongifolene, alphahumulene and copaene are the prominent sesquiterpenes (Fig. 5). As shown in Fig. 6, predominant oxygenated compounds were eucalyptol, linalool-L and alpha-terpineol. However, no linalool-L emission was detected from Troy Fir.
Fir species are known as mainly monoterpene emitters with the dominant compounds like alpha-pinene, beta-pinene, limonene and camphene, and relatively low isoprene emissions were reported in previous studies (Moukhtar et al., 2006, Dominguez-Taylor et al., 2007. Two of the fir species (Uludag Fir and Troy Fir) investigated in this study clearly showed a similar tendency. Monoterpenes had the highest portion with nearly 50% in total BVOC emissions for these species (Fig. 3). Pokorska et al. (2012b) reported that alpha-pinene, camphene, beta-pinene and limonene emissions from Silver Fir (Abies alba) in Belgium were in the ranges of 0. 168-0.822, 0.158-1.026, 0.069-0.402 and 0.278-0.648 µg/g/h, respectively. In the present study, 0.17 and 1.42 µg/g/h of alpha-pinene, 0.86 and 0.31 µg/g/h of limonene, 0.47 and 0.34 µg/g/h of beta-pinene emissions were determined for Troy Fir and Uludag Fir, respectively. In contrast, Cilician Fir was determined as an isoprene emitter. Although it has a high isoprene ratio of 91%, it emits considerable amounts of monoterpenes as much as the other fir species. Similarly, Harrison et al. (2001) reported average annual isoprene and monoterpene emission rates of 18.4 ± 3.8 µg/g/h and 2.7 ± 1.1 µg/g/h, respectively for Bulgarian Fir (Abies borisii-regis) which is endemic to Balkans. Table 3 shows these comparisons in detail.
Oak species are strong isoprene emitters as broad-leaved trees. However, a considerable variation was observed within the reported values in the literature (see Table 3). In previous studies, researchers have commonly focused on isoprene emissions for oak species and therefore, rarely determined the emission rates for monoterpenes or any other BVOCs. Two of three studied oak species (Vulcanic Oak and Ispir Oak) exhibited the typical emitting characteristics of the genus. However, Boz Pirnal Oak was different with substantially lower isoprene emission (0.028 ± 0.026 µg/g/h). Although the monoterpene emission rate was also relatively low (0.78 ± 0.63 µg/g/h), its contribution (about 93%) to total BVOCs makes the tree a monoterpene emitter. Another oak species, Holly Oak (Quercus ilex) is one of the most widely investigated in previous studies Loreto et al., 1998a;Grote et al., 2006) due to its prominently high monoterpene emitting potentials.
Differences in BVOC emission characteristics of the congeners in this study might be resulted from anatomical variations and/or biotope conditions. Cilician Fir had geographically different habitat conditions. It grows on lime-soils at southern latitudes while the others prefer humid and acidic soils of northern latitudes. This might affect the BVOC emission compositions. Relatively high isoprene emissions for fir species were also reported by some other researchers (Harrison et al., 2001;Pokorska et al., 2012b). Rasmussen (1978) conducted a detailed survey on several tree species to determine their isoprene emissions. Silver Fir and Grecian Fir (Abies cephalonica) were shown to emit isoprene, however, any emission rate was not reported. Presently, there is no specified parameter that could be related to these unusual emitting behaviors. However, habitat conditions, growing stages, metabolic and enzymatic processes, unidentified stress factors or genetic constitutions might be effective Fall, 1999;Kim et al., 2005;Singh et al., 2007;Joó et al., 2010). On the other hand, as an evergreen tree, Boz Pirnal Oak is anatomically     (Steinbrecher et al., 2009) 0.007 ± 0.011 12.0 ± 12.7 n.r. n.r.  n.r. 21.7 ± 2.00 n.r. n.r.  n.r.
2 (Konig et al., 1995) Pendunculate Oak  (Steinbrecher et al., 2009) dissimilar to the other two studied deciduous oaks. It is a small shrub-like tree, reaching a maximum height of 5-6 meters and prefers low altitudes and high temperature zones. Therefore, emissions of Boz Pirnal Oak did not show similarities with Vulcanic Oak and Ispir Oak. Another oak species, Kermes Oak (Quercus coccifera) having similar characteristics with Boz Pirnal Oak, was also reported as a low BVOC emitter (Dasdemir et al., 2012). Likewise, Vulcanic Oak and Ispir Oak were also physically comparable with some massive oaks previously investigated in the literature such as Pedunculate Oak (Quercus robur) and Sessile Oak (Quercus petraea) with their typically highisoprene and low-monoterpene emission patterns (see Table  3). Furthermore, monoterpene, sesquiterpene and oxygenated compound compositions of these species were mainly consist of the same predominant compounds as illustrated in Figs. 4, 5 and 6.

Quercus robur
Based on their studies on several oak species, Loreto et al. (1998b) and Loreto (2002) have concluded that (i) oaks grown in the western hemisphere emit mainly isoprene; (ii) monoterpene emission is specific to evergreen oaks of the Mediterranean environment; (iii) another group of Mediterranean oaks does not emit isoprenoids in substantial amounts; and (iv) the differences in emission profiles generally matches the taxonomical classification of oaks based on anatomical or molecular markers. The emission profiles for different oak species observed in the present study were in close agreement with these findings.

CONCLUSIONS
BVOC emission rates from seven endemic tree species in Turkey were determined using a specific dynamic branch enclosure system. Five of seven species (Uludag Fir, Troy Fir, Ispir Oak, Vulcanic Oak and Oriental Sweetgum) showed the typical emitting behaviors of their genus. As commonly reported in the literature, isoprene and monoterpenes were predominant compounds for all species. The highest isoprene and monoterpene amounts were emitted by Ispir Oak and Uludag Fir with the normalized emission rates of 19.2 ± 19.0 and 2.27 ± 1.26 µg/g/h, respectively. Oxygenated compounds were the third most prominent group and sesquiterpenes had relatively lower portions with a maximum value of 0.080 ± 0.097 µg/g/h for Ispir Oak contributing 0.34% to its total normalized BVOC emissions. Cilician Fir and Boz Pirnal Oak had different emission compositions, similar to those that were also reported by several previous studies. It was also reported that the values determined even for same species in different studies varied substantially.
BVOC emissions of endemic tree species in Turkey were characterized for the first time in the present study. These specific emission rates could be utilized in the future BVOC inventories and air quality modeling studies. 15-17 (Harley et al., 1996) 68 n.r. n.r.
11 (Geron et al., 2001) Oriental Sweetgum Sum of the collected number of samples for studied two tree species (Q. robur and F. excelsior).