On the influence of VOCs on new particle growth in a Continental-Mediterranean region

Three different sites with different kinds of emissions to the atmosphere and also different pollution properties have been selected. The objective was to have three stations in three different places with different volatile organic compounds (BVOC, AVOC) emissions. These sites were aligned along the North-South axis and their location in the Madrid basin is shown in figure 1.

The Madrid region has a population of around 6.5 million people, and the main pollutant emission source is road traffic along with heating combustion devices in the cold months. The vehicular fleet has increased and changed in the last years, with more diesel than gasoline-powered vehicles in the fleet (a total of 5 million vehicles in 2020, 54.5% of which are diesel-powered and 43.6% gasoline, DGT (2021)). Emissions from light industry contribute in a lesser proportion and there is no heavy industry in the region.

According to the inventory of Madrid city air pollutant emissions (Ayuntamiento de Madrid, 2018), total VOC emissions are having a decreasing trend since 1999 to 2015, from 45 720 to 19 852 tons. In 2015, the main contributor group was 'Solvent and other product use' followed by 'Other sources and sinks (nature)' and 'Road transport', these two groups with similar emission volumes. Along the years (1999-2015), the natural sources have kept more or less constant while the other two groups have had strong reductions, 50 and 85% respectively. These reductions are expected to continue in the next years, so the VOC natural sources are going to increase their importance in the future.

The highest BVOC emissions were expected in the North of the Madrid region. The station site selected (40° 34' 1.07' N and 3° 44' 7.89' W, ∼725 m a.s.l.) was in Tres Cantos municipality, in an environmental education area located approximately 20 km north far from the Madrid city center. This municipality is in the north-eastern border of the preserved Mediterranean oak forest of El Pardo, which extends over an area of 170 km² very close to Madrid. The vegetation at the Tres Cantos site is a natural forest of typical Mediterranean species, Quercus ilex including some introduced stands of Pinus pinea. The BVOC emissions of this site have been previously characterized and documented in experimental studies (Núnez et al 2002, Plaza et al 2005). Because of these characteristics, it can be classified for this study as a rural site (R), with expected high BVOC and low AVOC emissions.

CIEMAT site (40° 27' 23.49' N and 3° 43' 31.77' W, ∼650 m a.s.l.) is located approximately 9 km north-northwest of the Madrid city center (figure 1). The experimental site is surrounded by two green areas (Casa de Campo Park to the southwest of the CIEMAT, the greatest peri-urban Park of Madrid, and Dehesa de la Villa Park, an urban park located to the northeast). Tree and shrub vegetation species at these areas are diverse including mostly Mediterranean species (Quercus ilex and Pinus pinea) and other common gardening species (Platanus hispanica, Retama sphaerocarpa, Populus alba, Quercus suber and others). This is the main difference between this urban park and the others, specifically with the Tres Cantos site. This site can be considered a priori a representative urban background site (UB), as it is not directly affected by any local pollution source. It receives BVOC and AVOC emissions under variable conditions. The driving meteorological conditions have been previously characterized in the region. In summer, the development of strong thermal convective activity and the influence of the mountains produce characteristic circulations and mixing layer development (Crespí et al 1995), producing mountain breezes that can transport airmasses from the city early in the morning and from green areas in the afternoon and evening.

The third site was located in Leganés, a town (ca.190 000 inhabitants) in the metropolitan crown at 11 km south-west from the Madrid city center. The selected site (40° 20' 23.2' N and 3° 45' 16.3' W, ∼660 m a.s.l.) was in the north of the town and located in an educational institute for adults, close to the railway, having several avenues with high traffic density and an urban park (Parque de la Chopera). Its characteristics and historical air pollution records (provided by the co-located station belonging to the Madrid regional monitoring network) make it be considered a priori representative of an urban traffic site (UT), typical of the Madrid metropolitan area, with expected low BVOC and high AVOC emissions.

A comprehensive description of the instruments used during the campaign can be found in table 1. Besides the instruments to identify the NPF and the possible VOC associated, some other pollutants were measured, for instance eBC or PM concentrations, to help to identify the airmass origin and, consequently, the aerosol nature.

In the three sites there was an SMPS to measure the particle number size distribution (PNSD) that, depending on its configuration, varied between 1 and 661.2 nm. In the Tres Cantos station (A), it was a nano SMPS-TSI (TSI DMA 3085 + CPC 3775). In CIEMAT (B) two different systems were installed: a TSI DMA 3081 + CPC 3775 and a TSI Nano Enhancer Model 3757 and a Differential Mobility Analyzer Model 3086. In Leganés site (C), it was a TSI DMA 3081 + CPC 3772. The SMPS at CIEMAT site belongs to the ACTRIS European Network (www.actris.eu), and so to SARGAN (Rose et al 2021), and followed its standard procedures for SMPS maintenance and data quality (Wiedensohler et al 2012). This SMPS has participated and been approved in intercomparisons in the TROPOS laboratory (World Calibration Center for Aerosol Physics). Three of the instruments are included in REDMAAS, the Spanish Network on Environmental DMAs (Gómez-Moreno et al 2015), and participated in its intercomparison campaigns, where the CPC and SMPS were checked and the DMA calibrated. The complete SMPS were also intercompared and were approved when the measurements were in the +/− 10% of the all SMPS average. The Nano Enhancer Model 3757 and Differential Mobility Analyzer Model 3086 belong to the TSI manufacturer and fulfilled quality test previously to the campaign, as well as an intercomparison with other instruments at CIEMAT before the start of the campaign.

In the three sites a Tapered Element Oscillating Microbalances (TEOM) was installed to measure particulate matter (PM₁₀ and PM_2.5) levels. In addition, at the CIEMAT site a GRIMM 1107 optical instrument measured the three fractions PM₁₀, PM _2.5 and PM₁. These instruments were verified against reference gravimetric methods. The TEOM showed a difference smaller than 15% for PM₁₀ and 35% for PM_2.5. The GRIMM 1107 had similar results for PM₁₀ and PM_2.5 (30 and 15%) and 25% for PM₁. These uncertainties are not very large to identify traffic or Saharan dust periods, which was the objective of them.

About the aerosol chemical composition, to determine the equivalent black carbon (eBC) mass concentration, a Magee AE 33 aethalometer equipped with a PM_2.5 cut off inlet was installed at every site. Complementing these data, at the CIEMAT site an Aerodyne Aerosol Chemical Speciation Monitor (ACSM) measured continuously the composition of the non-refractory submicron aerosol (NR-PM₁) species (Organics, NO₃, SO₄, NH₄ and Cl). All these instruments, but an aethalometer, belong to the ACTRIS Network and followed its standard procedures (Cuesta-Mosquera et al 2021, Freney et al 2019). The CIEMAT aethalometer has given a maximum deviation of 10.3% from the reference instrument. The aethalometer not belonging to ACTRIS Network was previously compared with the other two with good results, showing similar uncertainties to the CIEMAT instrument. The ACSM participation in the intercomparison campaigns has shown that this instrument is well below the uncertainty accepted (30%), being 9% for ammonium.

At the three sites, VOCs were measured by offline techniques. The volatile organic compounds were sampled by using Tenax^TM tubes and later analyzed in the laboratory by Gas Chromatography-Mass Spectrometry (GC-MS). The identification of the majority of the compounds has been made using certified calibration gas mixtures of 5 ppmV (Abelló-Linde, Spain). The Tenax cartridges were dopped at the same sampling flow (0.3 l min⁻¹) for a known time at 10, 15, 20 and 25 sec and analyzed using the same methodology as the field samples. The species that were not in the gas mixtures calibration cylinder had been identified by using direct injections to the EUPHORE smog chambers (Borrás et al 2017) by means of commercial compounds. Once the reference standard compounds have been introduced into the chamber, heating it and air pass through into the chamber, Tenax cartridges were sampled following the same procedure used in the field campaigns. Just in case the comparison with the NIST library is conclusive, the identification has been done to the corresponding isomer directly.

The sampling method in Tres Cantos and Leganés sites was automatic while in CIEMAT it was operated manually. The automatic method allowed sampling during nocturnal periods obtaining longer data coverage. Samples were taken using commercial cartridges supplied by Markes International Ltd (UK). The tubes were packed with approximately 200 mg of TENAX TA (2,6-diphenylene oxide). During the field campaign a 0.3 l min⁻¹ air flow passed through the cartridge for 8 or 12 h. After sampling, the cartridges were capped and stored at 4 °C in a portable fridge during the transport to the laboratory where a standard refrigerator was used until sample treatment. Analysis was performed before 4 weeks.

Desorption of compounds was carried out by thermal desorption (Markes) for 15 min at 250 °C in a GC port of an Agilent GC-MS (Santa Clara, USA) equipped with a HP-5MS column of 30 m × 0.25 mm I.D × 0.25 mm film. The chromatograph was programmed at 40 °C for 3 min, then ramped at a rate of 5 °C min⁻¹ to 100 °C, 5 °C min⁻¹ to 160 °C, finally, was held for 1 min The injection port was held at 250 °C and the transfer line from GC to MS was held at 300 °C. Samples were injected in splitless mode, using helium as carrier at flow of 1 ml min⁻¹. The EI-voltage was 70 eV, the ion source temperature at 200 °C and the quadrupole temperature at 100 °C. Full scan mode was used (m/z 45–350) to identify the most abundant ions of hydrocarbon compounds. The analytical method used allowed determining the concentration and identifying more than 50 VOC species, both BVOCs and AVOCs.

For regulated pollutant gases, Thermo Scientific TM 122 i-series gas analyzers were used. The model 17i for NO/NO₂/NO_x and 49i for O₃ in both stations: Tres Cantos and Leganés. In these two stations, the official procedures stablished in the European regulations have been followed in the maintenance, calibration and data treatment. It is known (Villena et al 2012) that possible NO_y interferences in the NO_x measurements can appear. These interferences could be more important in Tres Cantos station, a rural site with higher O₃ concentrations, than in Leganés station, an urban site with lower O₃ concentrations. The possible corrections for these interferences have not yet been included in the official data treatment, so they are not considered here. At the CIEMAT station they were measured by a DOAS spectrometer (OPSIS AR-500), which is regularly calibrated. These data, together with the eBC measurements, were useful to support the data analysis, to determine the possible anthropogenic origin of the pollutants. SO₂ is only measured in a few Madrid network stations as it is a pollutant with very low concentrations due to its emission inventory characteristics. In Madrid city, the average value was 6 μg m⁻³ (2.4 ppb) during 2020 (Ayuntamiento de Madrid 2021) and 1 μg m⁻³ (0.4 ppb) in the Madrid region during 2018 (Comunidad de Madrid 2019). Due to these low concentration values, in the range of the DOAS detection limit, those of the Madrid monitoring network stations measuring SO₂ can be used for this study. As it is an official network, it is mandatory to use the official procedures published by the Spanish Official State Gazette, where the maximum uncertainty accepted is 15%, so these measurements used are below that value.

The meteorological variables were obtained at CIEMAT by a permanent instrumented tower (54 m high) not affected by obstacles, such as trees or buildings, being representative of the local and regional meteorology. This tower measured atmospheric pressure (AP) on the ground level, relative humidity (RH) at 4 m above ground level (a.g.l.), solar irradiance (SR) and precipitation (P) at 35 m a.g.l. and wind speed and direction (WS and WD, respectively) at 54 m a.g.l. The temperature (T) is measured at two different heights, on the ground level (low-level temperature (LT)) and at 54 m a.g.l. (high-level temperature (HT)), the ΔT between both levels helped to interpret the atmospheric thermal stability situation. This meteorological station is calibrated twice per year according to standard procedures. Another small tower was installed at the Leganés site for pressure, temperature, wind speed and relative humidity. This station was maintained by the regional air quality network.

Vertical profiles of aerosols were measured at CIEMAT by means of a lidar system. It uses a pulsed Nd:YAG laser emitting at 1064, 532, and 355 nm, configured in a monostatic biaxial alignment pointing vertically to the zenith. The receiving line consists of a Newtonian telescope and wavelength separation unit with dichroic mirrors, interferential filters, and polarization cubes. The collected radiation is split into five channels allowing the detection of elastic signals at 1064, 532, and 355 nm and two Raman channels at 387 and 607 nm (nitrogen Raman-shifted signal from 355 and 532 nm, respectively). The optical set-up of the system yields a full overlap at about 300 m above the instrument. The lidar signal was registered in 1 min integrated time, with a vertical resolution of 3.75 m. Continuous measurements were performed during the campaign, providing 'quicklooks' of the atmosphere up to 12 km to assist in the identification of the atmospheric situation and the presence of aerosol-rich elevated layers.

To study the synoptic meteorological conditions during the campaign, sea level pressure maps at 00:00 and 12:00 UTC provided by AEMET (Spanish State Meteorological Agency, http://www.aemet.es/) have been analyzed.

Airmass back-trajectories arriving at Madrid (40.42°N, 3.68°W) were computed using the HYSPLIT (Hybrid Single Particle Lagrangian Integrated Trajectory) model (Rolph et al 2017, Stein et al 2015). 5 days (120 h) back-trajectories were calculated starting from heights of 500, 1500 and 3000 m a.g.l., at 12:00 UTC for every day of the campaign period.

The occurrence of Saharan dust outbreaks during the field campaign was studied using the NAAPS (Navy Aerosol Analysis and Prediction System) model (https://www.nrlmry.navy.mil/aerosol/). It combines data obtained by satellites, aerosol simulations and predictions at a global scale.

NPFs were identified visually from the surface plots of PNSDs and classified using the criteria defined by Dal Maso et al (2005). The Condensation Sink (CS), which determines how rapidly molecules can condense onto the already existing aerosol instead of generating new particles, was calculated for the three sites using the TSI-SMPS data (Lehtinen et al 2003, Kulmala et al 2001). This CS parameter has been calculated according to Lehtinen et al (2003) using the following equation:

$$\begin{eqnarray}&&CS=2\pi D\displaystyle \sum _{i}^{0}{\beta }_{M.i}{d}_{p}{N}_{i}\end{eqnarray} \tag{ 1 }$$

where D is the diffusion coefficient of the condensing vapor in the gas phase (in this study assumed that D is the H₂SO₄ diffusion coefficient, 0.104 cm² s⁻¹ (Lyman et al 1990)), β_M,i is the transitional correction factor (Fuchs and Sutugin 1971), dp is the particle diameter and N_i (cm⁻³) is the particle number concentration for each particle size discrete interval i. Tuovinen et al (2021) evaluated the uncertainty arising from the assumptions taken. The main uncertainty sources are: the mass accommodation coefficient (which is included in the transitional correction factor) is usually assumed to be unity, the condensing vapor is assumed to be sulfuric acid, and evaporation is assumed to be negligible. They conclude that if the condensing vapor is VOC, the CS is smaller. If the mass accommodation coefficient is smaller than unity, the CS can be significantly overestimated. CS at Tres Cantos site was not calculated due to the different measurement size ranges compared with CIEMAT and Leganés sites, as the values obtained would not be comparable.

Shrinkages were identified and classified following Alonso-Blanco et al (2017).

At CIEMAT, an estimation of the H₂SO₄ concentration was also calculated using the SO₂ concentration measured in the Casa de Campo station. This station belongs to the city of Madrid's air quality monitoring network and is very close to the CIEMAT site (at 4 km). It was not possible to estimate this concentration in the other two sites because of the lack of SO₂ measurements. This estimation was done following the semi-empirical proxy for sulfuric acid concentration (Mikkonen et al 2011, Petäjä et al 2009).

$$\begin{eqnarray}&&\left[{H}_{2}S{O}_{4}\right]=8.21\times {10}^{-13}\,\kappa \,Radiation\,{\left[S{O}_{2}\right]}^{0.62}\,{(CS\,RH)}^{-0.13}\end{eqnarray} \tag{ 2 }$$

where: $\kappa$ is the reaction rate constant, which is calculated according to equation (3) in Mikkonen et al (2011) and is scaled by multiplying it with 10¹² (m²W⁻¹s⁻¹); Radiation is global radiation (Wm⁻²), [SO₂ ] is the measured SO₂ concentrations·(molec·cm⁻³); CS is the condensation sink (s⁻¹) computed following equation 1; and RH is the relative humidity (%). Mikkonen et al (2011) gave some values for the uncertainty associated with the proxy, founding that the relative errors are 42%. The main source of this uncertainty seems to be the role of SO₂ and the uncertainties associated to its concentration measurements (Herrmann et al 2014). This H₂SO₄ can be neutralized by ammonia, amines and/or other bases. Lehtipalo et al (2018) showed that when NH₃, H₂SO₄ and HOMs (highly oxygenated molecules) were present simultaneously, high nucleation rates were obtained. Experiments at higher NH₃ concentrations than 200 pptv showed higher nucleation rates than similar experiments without added NH₃. This confirmed some CLOUD experiments (Bianchi et al (2012) that showed that when increasing the ammonia concentration, the nucleation rates increased until about 200 pptv, when the nucleation rates reached saturation. In previous campaigns (Reche et al 2015, Artíñano et al 2018), the ammonia concentration has been characterized in the Madrid region, finding for urban and urban background sites ammonia concentration well above that saturation value, one order of magnitude higher. Although not published, data obtained in the rural background show similar concentrations to the urban and urban background sites. In the three sites, the ammonia concentration will be in the range where the response to different concentrations is saturated.

Distinct atmospheric conditions occurred during the study period being the most important issues shown in figure 2 and S2-S4.

Zoom In Zoom Out Reset image size

The period of study was characterized at the synoptic scale by a low pressure gradient at the surface level and by the advection of southern air masses at upper levels. The presence of a low pressure system at the SW-W sector of the Iberian Peninsula (IP) in the periods 1–5 and 10–13 July triggered such atmospheric circulation. For this reason, Saharan dust contributions were produced over the central region of the IP during the periods 1–4 and 12–13 July (see figure S3 and shaded periods in figure 2). The lidar profiles (figure S1) show a mixed layer up to 1.5 km, and a second layer, between 2.5 and 4 km, probably due to the Saharan dust contribution during the period 1–5. Tenuous elevated layers appeared on the 10 July (figure S1A) and a stronger layer, between 3 and 4 km, on the morning of the 11 July, leading to very high mixed layer, up to 6 km, in the following days, according to the lidar instrument (figure S1B). In the transition period between both dust outbreaks, especially from 8 to 9 July, a rather unstable atmospheric situation was established due to the development of N-NE air mass flows. Slow and moderate air flows prevailed during the last days of the campaign due to the influence of the Azores high pressure system over the IP. Between the 15 and 18 July, the mixed layer detected by lidar (figure S1C) shows a daily evolution during these days, with another residual layer up to 3 km that produced a rapid development during the morning. The atmospheric conditions clearly affected the course of the meteorological surface measurements during the campaign.

The meteorology was measured at the CIEMAT tower along the campaign (figure 2). Temperatures oscillated between 20 and 40 °C, corresponding the less warm days to the periods 8–10 and 17 July. The highest ΔT was recorded during the 7 and 11 to 13 July nights (4.3 °C the 12 July night) indicating a high atmospheric stability and surface inversion. Rain events (figure 2) appeared only on 8 and 13 July and were associated with local summer storms. The relative humidity was low, mostly lower than 50%. The daily solar irradiance maximum was always above 800 W m⁻². Wind speed remained below 10 m s⁻¹ during all the campaign with very few exceptions (1 and 8 July). The days with the lowest wind speed were 7, 11, 12 and 17 July (figure 2). The wind rose for the period of the campaign shows that the air circulation pattern is the usual one in Madrid, with two dominant directions along the NE and SW axis aligned with the nearby Guadarrama mountain range (Plaza et al 1997). The NE wind usually occurs during mornings and the SW wind during afternoons and with lower wind speeds. In the other two sites the local meteorology was similar, taking into account they are more affected by local conditions like obstacles and near buildings. So, the typical wind rose observed in CIEMAT tower was slightly modified in the Leganes site where the wind speed was always below 4 m s⁻¹.

The complete series of synoptic surface maps, air mass back trajectories and dust forecast maps during the CRISOL field campaign can be found in the figure S2, S3 and S4 respectively in the supplementary information.

The number of detected NPF episodes events was higher in CIEMAT (6) and Leganés (5) than in Tres Cantos (2). At the beginning of the campaign (days 3–7th July), several NPF events were simultaneously observed at the three sites (figure 3), however, unlike Tres Cantos, in Leganés and CIEMAT the growth and formation rate of new particles was well defined. At the end of the campaign, some events appeared again but only at CIEMAT site (16–17th). The events and their classification (Dal Maso et al 2005) can be found in table 2. The different classes of NPF events at CIEMAT site were previously characterized (Gómez-Moreno et al 2011). Class I corresponds with events in which the particle growth rate can be determined or observed. In Class Ia the typical banana shape is clearly visible. It is mostly related with low wind speed, below 4 m s⁻¹, and with stable wind direction. Class II corresponds to a situation in which particle growth determination is not possible. It is related with higher wind speed (5–6 m s⁻¹) or the wind direction was rotating in the typical clockwise sense from northeast to west. Class Ib events had characteristics between those of classes Ia and II. In this case the banana shape is not completely clear, but the particle growth can be observed. All the events occurred around noon, when the solar radiation was maximum.

Zoom In Zoom Out Reset image size

During the period 3–7th, the weather was mostly stable; the wind speed was below 3 ms⁻¹ during most of the day except after the midday when it reached 5 ms⁻¹. During this period the wind direction followed a pattern varying from SE to SW. The relative humidity was low, always below 50% and reaching values below 20% in the warmest hours. However, on the period 16–17th, the weather conditions were rather different, with similar wind speed and relative humidity but directions followed a well-defined pattern starting NE and clockwise rotating until reaching SW.

The H₂SO₄ concentration estimated (Mikkonen et al 2011) for CIEMAT shows a clear dependence of the solar irradiance (SR), as can be seen in the figure 3(B). The condensation sink reached similar values all days in CIEMAT and Leganés sites (CIEMAT average CS 1.07 10⁻² s⁻¹, maximum 4.17 10⁻² s⁻¹ and minimum 2.36 10⁻⁴ s⁻¹; Leganés average CS 1.30 10⁻² s⁻¹, maximum 5.97 10⁻² s⁻¹ and minimum 4.97 10⁻³ s⁻¹). However, nucleation events only appeared in some cases, indicating that besides the possible H₂SO₄ participation in these events, the influence of other participants in the nucleation or in the subsequent growth, as it has been previously established for the potentially significant role of VOCs in the particle formation (Lee et al 2019), cannot be rejected.

In this section, the evolution of the most common pollutants with possible influence in the NPF is analyzed for the three sites. This evolution can be found in figure -S5. Going from north to south, in Tres Cantos, all the pollutants had the lowest concentrations, with PM₁₀ and PM_2.5 values increasing during some days (12 and 13 July) because of the Saharan dust outbreaks, which affected the entire region. The anthropogenic pollutant, NO_x, had very low levels, NO concentrations were one order of magnitude smaller than those observed at the urban traffic station (in average 3.3 ± 2.8 versus 11.9 ± 13.8 μg m⁻³ (2.6 ± 2.2 versus 9.5 ± 11.0 ppb); maximum values were 14.5 versus 80 μg m⁻³ (11.6 versus 63.5 ppb), respectively).

The CIEMAT station presented intermediate values for NO (average 3.5 ± 5.9 μg m⁻³ (2.8 ± 4.7 ppb) and maximum 86 μg m⁻³ (68.7 ppb)) but the NO₂ level was very high, reaching an average of 44 μg m⁻³ (23 ppb) and the maximum of 144 μg m⁻³ (75 ppb). Ozone behavior was clearly opposite to NO₂ and reaching values above 150 μg m⁻³ (75 ppb). High concentrations for both PM₁₀ and PM_2.5 were also observed at this site during the Saharan dust outbreaks, and lower concentrations for PM₁.

In the case of the urban station, Leganés, the PM₁₀ and PM_2.5 levels were similar to those recorded at CIEMAT as well as Ozone levels, but NO₂ concentrations were lower. It should be noted that this station is closer to the emission sources and the NO concentrations were high (up to 150 μg m⁻³ (120 ppb)).

Besides NO_x, another pollutant directly related with traffic is the equivalent black carbon (eBC). This parameter was monitored at the three sites. The highest eBC concentrations during the campaign were found in Leganés (1.9 ± 1.9 μg m⁻³; maximum 11.5 μg m⁻³) followed by CIEMAT (1.1 ± 0.8 μg m⁻³; maximum 4.6 μg m⁻³) and Tres Cantos (0.5 ± 0.3 μg m⁻³; maximum 1.4 μg m⁻³), with the same positive South-North gradient as that observed for the pollutant gases. As it has been mentioned before, Leganés and CIEMAT are influenced by road traffic emissions, especially Leganés, as it is an urban site and the station is close to several important avenues and streets. The eBC evolution at the three stations can be found in figure 4.

Zoom In Zoom Out Reset image size

AVOC and BVOC average concentrations (Table S2) were higher in CIEMAT-UB followed by Leganés-UT and Tres Cantos-R. Comparing the different measurement areas, the evolution along the campaign of the VOCs at the three sites shows a slightly different behavior, see figure 5. In Tres Cantos, the highest values for total VOCs were around days 8 and 10-11th. However, the BVOCs show a clear maximum around 8–9th related with the lower temperature and higher humidity, which allowed higher emissions from vegetation in this site. In CIEMAT, the maximum values for VOCs were reached 12th, 15th and 16th, with a smaller peak at the beginning of the campaign. In Leganés, the main peak was also the 12th, but no peak appeared later.

Zoom In Zoom Out Reset image size

AVOCs and BVOCs showed a similar daily pattern in each site, being higher in the morning hours in CIEMAT and Leganés coinciding with the morning rush hours and in Tres Cantos during afternoon and evening hours.

Figure 6 shows the time evolution of some specific VOC concentration at each site sampled by the Tenax^TM tubes. These VOCs have been selected because of their high concentration at each site. At Tres Cantos site, the main BVOC detected was Benzylalcohol, which presented the highest concentrations during the NPF events. During the days 10 and 11th, besides this compound, m-Cymene and Sabinaketone were also found at significant concentrations. Among the AVOCs, the highest concentrations were found for Benzene and Phenol. Other important compounds with high concentrations were: 1-Heptene, Butylacetate, Propylbenzene, 4-Ethyltoluene, Acetophenone, Decane, Undecane, Dodecane and Tetradecane. There are two groups, those in the aromatic group and those with a linear structure. This last group is for C10 and above.

Zoom In Zoom Out Reset image size

These two groups were also found in the other two stations, CIEMAT and Leganés, although in higher concentrations due to a higher traffic emission influence. In these two stations the daily evolution was similar. At CIEMAT, the most common BVOCs during the NPF events were α-Pinene together with Camphene, β-Pinene and Limonene. About the AVOCs, the most common compounds were similar to those in Tres Cantos station: Benzene, Toluene, Decane, Dodecane and 1,2,3-Triclorobenzene during the NPF events and Tetradecane, 4-Ethyltoluene and Trimethylbenzene during non-event days.

At Leganés, the main BVOCs were: m-Cymene, m-Propiltoluene, Benzilalcohol and Valerolactone, always with very low concentration. The most common AVOCs were the same as in the other stations: Benzene and Toluene during NPF days and Etilbenzene, Butilacetate, Propilbenzene, Decane, Dodecane, Tetradecane and others. This higher concentration for AVOCs can explain several NPF events observed at the site, especially Benzene and Toluene (Ling et al 2019).