Terpene emissions from boreal wetlands can initiate stronger atmospheric new particle formation than boreal forests

We deployed state-of-the-art instrumentation to Finnish wetland, Siikaneva (61°49’59.4“N 24°11’32.5“E, 162 m a.s.l.) where is located a class II ecosystem ICOS (European Integrated Carbon Observation System) station⁴⁰ and to SMEAR II station (Station for Measuring Ecosystem-Atmosphere Relations)⁴¹, in Hyytiälä (61°50’47.1“N 24°17’43.2“E, 181 m a.s.l.) and investigated all the relevant components that are known to influence the new particle formation. The observations were performed on 10th May–15th June 2016. We monitored direct VOC and CH₄ emissions from wetland and the concentrations of oxidation products of VOCs, SO₂, and O₃. We monitored concentrations and chemical composition of atmospheric clusters, aerosols, and air ions from the smallest sizes (0.5 nm) up to 40 nm approaching sizes which can be activated to CCN. As a reference, we utilized SMEAR II station in Hyytiälä, located 5 km east of these measurements. The SMEAR II station is monitoring over 1200 variables, including also the ones measured in the Siikaneva wetland.

The Hyytiälä site is a relatively homogeneous Scots pine stand surrounded by evergreen coniferous forests⁴¹, while the Siikaneva site is located in a pristine boreal fen. Peat started to accumulate in Siikaneva after the latest ice age about 9000 years ago and peat depth at the measurement site is approximately 4 meters^42,43. Siikaneva fen is characterized by relatively flat topography with a number of vegetation communities and some surface patterning featuring drier hummocks and wetter lawns.

The measurement site consisted of a small hut containing all the instrumentation, which was equipped with sampling inlets at heights of approximately 1.5 m and 3 m. The CI-APi-TOF and APi-TOF, NAIS, PSM, O₃ measurements were conducted with the inlet at 1.5 m, while all the meteorological, CH₄, CO₂, and VOC data were obtained at 3 m.

Data sets from the SMEAR II station at Hyytiälä can be obtained from the AVAA smartSMEAR website (https://avaa.tdata.fi/web/smart)⁴⁴. A detailed description of the SMEAR II station at Hyytiälä can be found elsewhere^41,45. Siikaneva station is part of ICOS (European Integrated Carbon Observation System) network that includes two classes of Ecosystem stations, referred to as Class 1 (complete) and Class 2 (basic) stations. They differ in costs of construction, operation, and maintenance due to the reduced number of variables measured at the Class 2 stations. Siikaneva station is classified as the class 2 ecology site.

Air temperature and relative humidity (RH) were measured with Rotronic HC2 sensor (Rotronic AG, Switzerland) at 2-meter height in Siikaneva. The air temperature was measured at 2 min and RH one minute time resolution. Photosynthetically active radiation (PAR) was measured once in a minute by a Li-Cor Li-190SZ quantum sensor (LI-COR, Inc., USA). Wind speed and direction were measured with Metek USA-1/Gill HS 50 anemometer at 3 meters height. The averaging period for all auxiliary measurements was 30 minutes.

VOC concentrations were measured with a proton transfer time-of-flight mass spectrometer (PTR-TOF, Ionicon) which consists of a proton transfer reaction ion source (PTR) and a TOF-MS⁴⁶. The PTR instrument is described in detail in literature^47,48 and only short description is given here. The PTR consists of a H₃O + ion source (hollow cathode discharge in water vapor) and a drift tube where protonated water is mixed with the sample and protons are transferred to the VOC species according to Eq. 1:

This charging mechanism works for VOCs with higher proton affinity than that of water, most atmospheric VOC fulfill this requirement⁴⁷.

The ionized VOCH⁺ are then passed to the TOF and the mass is determined with an accuracy of 20ppt and resolving power of 3000Th/Th. The VOC is identified using the accurate mass and the prior made calibration. The concentrations of VOCs can be computed from the calibration as the ratio of sample to reagent ion using equation Eq. 2:

where [H₃O ⁺] is the concentration of H₃O ⁺ in the absence of reacting neutrals, k is the reaction coefficient of the proton transfer reaction and t is the average time the ions spend in the reaction region⁴⁷. Product kt is obtained from calibration.

Terpene and isoprene emissions are depended on temperature and light⁴⁹. Accordingly, an increase in both concentrations is observed when approaching summer, indicating an increase in biogenic emissions (Supplementary note 6 and 7. Fig. S10-S12).

The chemical composition of air ions was measured with atmospheric pressure interface (APi) time of flight mass spectrometer⁵⁰ (APi-TOF, Tofwerk AG). The sample was driven to the instrument through 10 mm electropolished stainless steel tube with a flow rate of 6lpm. The sample was further introduced to APi through a critical orifice with a sample flow of 0.8 l min⁻¹, ions are transported into the TOF to determine their mass to charge ratio(m/Q). The ion beam is focused by two guiding quadrupoles and an ion lens assembly, in three separate differentially pumped chambers, leading into the TOF. The instrument has resolving power of >3000 Th/Th and mass accuracy <20ppm.

Second APi-TOF equipped with chemical ionization (CI) inlet was used to measure the concentration of highly oxidized organic molecules and sulphuric acid. The design of the CI-inlet is similar to one used earlier^51,52. The sample was drawn though a ¾” electropolished stainless-steel tube with a flow rate of 10 lpm. Nitrate ions are created by exposing clean air containing nitric acid to a soft x-ray radiation. This sheath flow is then introduced in an ion reaction tube concentric to the sample flow. Nitrate ions in the sheath flow are directed into the sample flow by means of an electric field. The interaction time between ions and sample gas is approximately 200 ms.

The signal of highly oxidized organic compounds (HOMs) is distributed over multiple mass peaks. Example of spectrum is shown in supplementary note 7. Marker peaks were selected to represent each HOM group (Table 1). Concentrations of HOMs were calculated by assuming collision limited charging and the same calibration coefficient (5·10¹⁰) as for sulphuric acid according Eq. 3

where, ζ – calibration coefficient, I_m – signal intensity of HOM marker, I_c – signal intensity of charger ion, n – number of marker compounds of corresponding HOM group.

Calibration for sulphuric acid was performed prior the campaign in the laboratory by oxidation of SO₂ with OH to produce the sulphuric acid⁵³. The hydroxyl radical is produced by UV photolysis of water vapor. The final H₂SO₄ concentration is calculated by a numerical model at the outlet of the calibration source. Comparison of this modelled concentration and the signals measured by CI-APi-TOF yields a calibration factor.

Comparison of Hyytiälä and Siikaneva sulphuric acid concentrations and diurnal patterns are shown in supplementary note 9, Fig. S17 and S18.

Atmospheric ions and total particles were measured with a Neutral cluster and Air Ion Spectrometer⁵⁴ (NAIS), an instrument for measuring mobility and size distribution. The range for electrical mobility for NAIS is 3.2-0.0013 cm² V⁻¹ s⁻¹ corresponding to a mobility diameter range of 0.8–42 nm. In total particle mode, the size distribution starts from 2 nm since the charger ions are indistinguishable at smaller particle size ranges^55,56.

For measurements of sub-3 nm particles, a Particle Size Magnifier (PSM) in series with a Condensation Particle Counter (CPC) was deployed. This setup enables the detection of single particles without charging the particles⁵⁷. PSM uses diethylene glycol for activating and growing the particles before entering the CPC. This enables the detection of particles as small as ∼1 nm in mobility diameter⁵⁷.

Growth Rates (GR) were determined using the 50% appearance time method^58,59. This method uses particle size distributions measured by NAIS or DMPS instruments to determine the time when half the concentration maximum is reached in different size bins. The growth rate was obtained by following the evolution of the 50% appearance times as a function of the size distribution during clustering events.

Condensation Sinks (CS) for both measurement sites, Hyytiälä and Siikaneva, were calculated using the particle size distributions measured with the NAIS and DMPS instruments. Since there were measured at ambient conditions a parameterization to correct for ambient hygroscopicity was used⁶⁰. The CS describes the sink for condensing vapors arising from the available surface area of pre-existing aerosols and can be determined using the Eq. 4:

where D is the diffusion coefficient of the condensing vapor, H2SO4 in our case, β_m, d_p refers to the transition regime correction⁶¹, while N_dp is the particle number concentration.

Condensation sink (CS) is usually calculated from DMPS data, since unlike NAIS (2 – 42 nm) it detects aerosol population up to 1000 nm. However, the DMPS system was not available at Siikaneva, and it was verified that calculating CS from NAIS data introduces larger systematic error than using CS calculated from Hyytiälä DMPS (Supplementary note 10, Fig. S19). The comparison of CS calculated from the NAIS shows a good linear correlation between the CSs determined for the two measurement sites (Supplementary note 10, Fig. S20). Error from using CS for Siikaneva calculated from Siikaneva NAIS compared to the Hyytiälä DMPS is larger and justifies the use of the CSs from Hyytiälä DMPS when determining formation rates at Siikaneva. Example clustering event measured simultaneously at two stations is depicted in supplementary note 11 in Fig. S21.

Formation rates (J_dp) were calculated for the Siikaneva measurement site based on the particle number size distribution, the coagulation sink and the determined GR⁶² with Eq. 5

Where J_dp is the formation rate of the cluster with diameter d_p while CoagS_dp is the coagulation sink arising from collisions with larger particles⁶³.

Greenhouse gas flux measurements, in this case carbon dioxide (CO₂) and methane (CH₄), were conducted with eddy covariance (EC) method at 3 meters height. CO₂ and water vapor (H₂O) flux measurements were done with high-frequency optical gas analyzer (LI- 7200, LI-COR Biosciences) and 3D sonic anemometer (USA-1, Metek GmbH; CSAT3 Campbell Scientific, Inc.). CH₄ flux was measured with CH₄ analyzer RMT-200 (Los Gatos Research Inc., Mountain View, California, USA) and 3D sonic anemometer (USA-1, Metek GmbH; CSAT3 Camp- bell Scientific, Inc.). CO₂ and CH₄ fluxes were measured every half an hour. Uncertain values of CO₂ and CH₄ fluxes were filtered out with friction velocity values less than 0.1 m s⁻¹. In Hyytiälä, net ecosystem exchange (NEE, CO₂ flux) was directly measured using Metek USA-1 anemometer and LI-COR LI-7000 gas analyzer.

Map product used to estimate the area of wetland was produced by European Space Agency Climate Change Initiative (ESA CCI) Land Cover project. The CCI-LC project delivers a consistent global land cover maps at 300 m spatial resolution on an annual basis from 1992 to 2015. The Coordinate Reference System (CRS) used for the global land cover databases is a geographic coordinate system (GCS) based on the World Geodetic System 84 (WGS84) reference ellipsoid.

The CCI-LC combined spectral data from 300 m and 1000 m resolution ENVISAT (MERIS) surface reflectance to classify land cover into 36 land cover types following the United Nations Land Cover Classification System (UNLCCS) legend^64,65. The whole archive of MERIS data was first pre-processed for radiometric and geometric corrections, cloud screening, and atmospheric correction with aerosol retrieval. An automated classification process, combining supervised and unsupervised algorithms, was then applied to the full-time series to serve as a baseline to derive land cover maps that were representative of three 5-year periods⁶⁶. The current updated product is provided in 1-year periods (ESA CCI⁶⁷).

From CCI-LC maps we calculated the extent of wetlands located north from 30°N to be 1.2 million km². Equivalent values reported in the literature are 3.88–4.08 million km^{2 34,35,36}. The reliability of land cover estimations based on remote sensing data depends on the sampling, preprocessing and interpretation of the data. The observed discrepancy with earlier studies is therefore reasonable. The area of wetland in boreal region (north form 50°N) is 0.5 million km².

The highest density of wetland in north from 50°N is located in Siberia (31–180°E, 3.8·10⁵ km²), especially in areas of Khanty-Mansi and Yamalo-Nenets regions (60–90°E, 2.2·10⁵ km²), and in Scandinavia (5–31°E, 6.2·10⁴ km²) (SI Fig. 9).

Source: Ecology - nature.com