University of Groningen Carbonyl sulfide, a way to quantify photosynthesis Kooijmans, Linda Maria Johanna

(1)

Carbonyl sulfide, a way to quantify photosynthesis

Kooijmans, Linda Maria Johanna

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date: 2018

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Kooijmans, L. M. J. (2018). Carbonyl sulfide, a way to quantify photosynthesis. University of Groningen.

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

(2)

5

S

OURCES AND SINKS OF CARBONYL

SULFIDE INFERRED FROM

ATMOSPHERIC OBSERVATIONS AT

THE

L

UTJEWAD TOWER

With plant uptake being its dominant sink, carbonyl sulfide (COS) is a promising tracer for the estimate of terrestrial ecosystem gross primary production (GPP); however, under-standing other minor sources and sinks, e.g. anthropogenic and soil, is also critical to the success of the approach. We infer the sources and sinks of COS using continuous in situ mole fraction profile measurements of COS at the 60-m tall Lutjewad tower (1 m a.s.l., 53°240N, 6°210E) in the Netherlands. We show that the diurnal cycles of COS mole fractions at all heights (7, 40 and 60 m) are mainly driven by nighttime uptake as a result of surface uptake of COS coupled with a shallow nocturnal boundary layer. We further determined the nighttime COS fluxes to be -3.0 ± 2.6 pmol m°2_s°1_{using the radon-tracer correlation} approach. We detected lower COS mole fractions from inland than from sea, which is likely driven by vegetation and soil uptake, and found no indications that the mud flats and salt marshes at the coast are a net sink or a net source. We observed enhancements of COS mole fractions on the order of 100 ppt (lasting a few days) to 1000 ppt (lasting a few hours) at three occasions, which coincided with ploughing of nearby agricultural fields, indicating that COS is produced within the agricultural soil. One spike of COS coincided with that of SF₆, suggesting an anthropogenic source; however, the exact origin could not be deduced. These results are useful for improving our understanding of the sources and sinks of COS, contributing to the use of COS as a tracer for GPP.

This chapter is in review as: Kooijmans, L. M. J., Scifo, A., Scheeren, H.A., Meijer, H.A.J., Mammarella, I., and Chen, H.: Sources and sinks of carbonyl sulfide inferred from atmospheric observations at the Lutjewad tower, in review, 2018.

(3)

5

5.1. I

NTRODUCTION

The interest in the budget of carbonyl sulfide (COS) has grown over the last decade due to the close relation of COS and carbon dioxide (CO₂) vegetative uptake. The two gases follow a similar uptake pathway from the leaf boundary layer up to the site of reaction in the plant (Stimler et al.,2010a). COS therefore provides a means to separate the concurrent uptake of gross primary productivity (GPP) and respiration flux of CO2 ((Montzka et al.,2007;

Campbell et al.,2008)). Those individual fluxes can otherwise not be measured directly at scales larger than the leaf scale. Besides the interest in COS as a tracer for GPP, COS is of interest in the stratosphere as it plays a role in the formation of the stratospheric sulfate aerosol layer, which has an overall cooling effect to the Earth’s climate.

On average, mole fractions of COS in the atmosphere range between 350 and 550 parts per trillion (ppt) globally. The vegetative uptake of COS is the largest sink in the atmospheric COS budget, followed by uptake by soils (Berry et al.,2013;Whelan et al., 2018). The main sources of COS are anthropogenic emission, the ocean, wetlands and biomass burning; however, the current COS budget has large uncertainties, and lacks of COS sources to balance the sinks, mainly due to uncertainties on the contribution of the tropical ocean and anthropogenic emissions (Whelan et al.,2018).

On average, mole fractions of COS in the atmosphere range between 350 and 550 parts per trillion (ppt) globally. The vegetative uptake of COS is the largest sink in the atmospheric COS budget, followed by uptake by soils (Berry et al.,2013;Whelan et al., 2018).. The main sources of COS are anthropogenic emission, the ocean, wetlands and biomass burning. Anthropogenic emissions of COS can be either direct emissions of COS (e.g. coal combustion, aluminum smelting, pigment and paper industry), or indirect through emissions of CS2(e.g. rayon production, agricultural chemicals and tire wear),

which can be oxidized to COS (Zumkehr et al.,2018). Unfortunately, the current COS budget has large uncertainties, and lacks of COS sources to balance the sinks, mainly due to uncertainties on the contribution of the tropical ocean and anthropogenic emissions (Whelan et al.,2018).

The long-term COS record presented by (Montzka et al.,2007) gave insight into the seasonality of COS mole fractions and showed that in the northern hemisphere it is largely influenced by uptake by the biosphere, and by oceanic emissions in the southern hemi-sphere. Those measurements were made using discrete flask samples (1 to 5 samples per month) that were analyzed by a gas chromatographic mass spectrometer. Recently, optical instruments that are capable of making high-frequency (1 to 10 Hz) in situ simul-taneous measurements of COS and CO₂ (Stimler et al.,2010b) became available, e.g. a quantum cascade laser spectrometer (QCLS). The availability of these instruments cre-ated opportunities to advance our understanding of the COS sources and sinks, through flux measurements using the eddy-covariance technique and soil and branch chamber measurements (e.g.Berkelhammer et al.,2014;Maseyk et al.,2014;Commane et al.,2015; Kitz et al.,2017;Wehr et al.,2017;Sun et al.,2018a;Yang et al.,2018), and through atmo-spheric mole fraction measurements within the continental and marine boundary layer (Commane et al.,2013;Belviso et al.,2016;Kooijmans et al.,2016;Lennartz et al.,2017).

This study aims to better understand the local to regional sources and sinks of COS using atmospheric mole fraction measurements of COS, CO2, carbon monoxide (CO)

and sulfur hexafluoride (SF6) at the Lutjewad station in the Netherlands. We analyze the

(4)

5.2.METHODOLOGY

5

101

at a boreal forest site in Finland, where the soil, branch and ecosystem COS fluxes have been extensively studied (Kooijmans et al.,2017;Sun et al.,2018a). Furthermore, with the measurements at the coast of the Netherlands we have the possibility to study COS mole fractions in air masses from distinct origins; that is, from the sea in the north and agricultural land with anthropogenic activity in the south. Simultaneous measurements of CO₂, CO and SF₆can help provide insights into the main source and sink contributions in the region. CO and SF₆are both tracers for urban and industrial emissions as their main sources are anthropogenic. Moreover, we investigate which processes influence atmospheric COS mole fractions and estimate nighttime COS ecosystem fluxes.

5.2. M

ETHODOLOGY

5.2.1. M

EASUREMENT SITES

Profile measurements were performed at the Lutjewad atmospheric monitoring station in the Netherlands (53°240N, 6°210E). The Lutjewad station is located at the north coast of the Netherlands in front of the Wadden Sea (largely consisting of tidal mud flats). The first kilometer towards the north is covered by salt marshes. Towards the south, the area is used for agriculture. Much of the land in the area is reclaimed from the sea with the use of dikes since the 15tthcentury. The agricultural land around the Lutjewad station has been reclaimed from the Waddensea in the 19thand early in the 20thcentury; therefore, the soil consists of clay that originates from the sea. The station is located next to the dike (which is 7 m high) of the Waddensea and consists of a 60 m tall tower. The area is sparsely populated: the closest village is Hornhuizen (ª200 inhabitants) at a distance of 1.3 km towards the south; the closest city is the city of Groningen (ª200.000 inhabitants) at a distance of 25 km towards the southeast. 10 km towards the west of the station is a small ferry port. Farmlands around the measurement station are planted with seed potatoes, sugar beets and winter wheat. 40 km towards the southeast is an aluminum smelting factory (Damco Aluminium; 53°1802100N, 6°580300E). Regionally, there are several aluminum facilities at 250 km distance in the German Ruhr-area (Trimet Aluminium, Hydro Aluminium), that may act as a source of COS.

Diurnal cycles of COS mole fractions in Lutjewad are compared with those from a boreal forest site at the Station for Measuring Forest Ecosystem–Atmosphere Relations (SMEAR II) in Hyytiälä, Finland (61°510_{N, 24°17}0_{E). That site is covered by forest in all}

directions apart from a few small grassland patches and an oblong lake at about 700 m towards the southwest. The dominating tree species is Scots pine (Pinus sylvestris). The nearest village is Juupajoki (ª2000 inhabitants) at a distance of 6 km towards the southeast. The site has a 125-m tall tower and the canopy height is 17 m.

5.2.2. M

EASUREMENTS OF

COS, CO

₂ AND

CO

A QCLS was used to measure mole fractions of COS, CO2, CO and H2O at the two sites

between August 2014 and February 2018 (Table 5.1). The data measured in Lutjewad are shown inKooijmans et al.(2016, Fig. 12) for the period between August 2014 and April 2015. The setup of the QCLS in Lutjewad and Hyytiälä is described in detail in Kooijmans et al.(2016) andKooijmans et al.(2017) respectively. In summary, the QCLS was sampling air from different heights (see an overview in Table 5.1) where the different

(5)

5

sampling lines (Synflex (Decabon) or Teflon) were switched with a multi-position Valco valve (VICI; Valco Instruments Co. Inc.). The sampling time differed per period (Table 5.1), depending on other measurements being done each hour, e.g., soil and branch measurements in Hyytiälä. A reference cylinder was measured every half hour to remove instrument drift and to calibrate the measurements to the common scales. The reference cylinders were calibrated against two NOAA/ESRL standards for COS (NOAA-2004 scale) and CO₂(WMO-X2007 CO₂scale) at the University of Groningen (Kooijmans et al.,2016). The measurements in Lutjewad had to be corrected for a leaking solenoid valve for the period between August 2014 and January 2015, which was done by comparing the CO2

measurements with that from a collocated cavity ringdown spectrometer (Picarro Inc. model G2401-m) and applying a similar dilution factor to all gas species (see details in Kooijmans et al.,2016). A target cylinder was measured once every hour in all periods except for the measurements in Lutjewad in January–February 2018. Kooijmans et al. (2016) gave an overview of all uncertainty contributions that are relevant for obtaining accurate and precise COS mole fractions; that is, the repeatability of the NOAA scale (2.1 ppt), calibration of reference standards and ambient air samples (2.8 ppt), water vapor correction (2.9 ppt) and measurement precision. The measurement precision (defined as the standard deviation over minute-averaged target cylinder measurements after drift correction with reference measurements) has changed over the years; an overview of the average precision is given for each measurement period in Table 5.1.

Field standards are calibrated against NOAA standards in the laboratory before and after each measurement period to test for drift in the cylinders. The COS mole fraction measurements of nine cylinders are available, and five cylinders changed less than 2.5 ppt year°1_{, two cylinders decreased by ª10 ppt year}°1_{and two cylinders decreased by ª30 ppt}

year°1_{. The four cylinders that drifted more than 10 ppt year}°1_{were not used as reference}

cylinders in the data processing. All of the cylinders were uncoated aluminum cylinders, which, according to experience at NOAA, are more prone to COS mole fractions drift than Aculife treated aluminum cylinders.

Besides the in situ measurements at Lutjewad, we also measured flasks that were sampled at the Lutjewad tower at 60 m between December 2013 and February 2016 with an average of four samples per month. 81 % of the flask samples were taken at noon. For a detailed description of the measurement procedure seeKooijmans et al.(2016). The flask measurements of COS mole fractions were used together with the in situ measurements in Lutjewad to construct a seasonal fit to the data.

5.2.3. M

EASUREMENTS OF

SF

6

For the analysis of SF₆, ambient air sampled at 60 m altitude is cryogenically dried and analyzed by an Agilent 6890N gas chromatograph equipped with a micro-Electron Capture Detector offering about seven measurements per hour (Van der Laan et al.,2009b). The ambient SF6mole fraction is calculated using working standards that are calibrated against

five primary standards (provided by NOAA-ESRL) linked to the World Meteorological Organization X2006 scale for SF₆. We estimate the overall measurement uncertainty to be < 0.1 ppt for SF₆.

(6)

5.2.METHODOLOGY

5

103

Table 5.1 | Measurement periods at the Lutjewad and Hyytiälä sites with an overview of the measurement heights,

sampling time and 1-minute measurement precision based on target cylinder measurements (apart from Jan.–Feb. 2018 in Lutjewad when no target measurements were made).

Location and Period Measurement heights [m]

Sampling time per height and

frequency

Precision [ppt]

Lutjewad, The Netherlands

Aug. 2014–Apr. 2015 7, 40, 60 Two times 8 min.,

every hour 5.3

Jan.–Feb. 2018 60 Two times 27 min.,

every hour

-Hyytiälä, Finland

Jun.–Nov. 2015 0.5, 4, 14, 23, 125 3 min., every hour 6.5 Apr.–Nov. 2016 0.5, 4, 18, 125 4 min., every 1.5

hours 9.1

Feb.–Jul. 2017 18, 125 2.5 min., every hour 8.5

5.2.4. S

EASONAL FIT

A nonlinear least squares fit was made to the 60 m COS mole fractions from Lutjewad, see Appendix Fig. A5.1. The shape of the fit is represented by a harmonics function after Thoning et al.(1989, eq. 1 therein). We used the highest available heights, such that the mole fractions are the least affected by local influences, and we selected only daytime data, such that the measured mole fractions are not influenced by the shallow nocturnal boundary layer. The seasonal fit of CO2(not shown) is based on continuous measurements

of a colocated cavity ringdown spectrometer in 2014 and 2015 in Lutjewad. For the seasonal fit of CO2we selected only data with wind direction from the north (wind direction < 30

° or > 260 °) to make sure that the data represent background air and are not affected by anthropogenic influences. This data selection was based on the wind direction analysis presented in Fig. 5.3.

5.2.5. N

IGHTTIME ECOSYSTEM FLUX IN

L

UTJEWAD

Nighttime fluxes of COS and CO₂are estimated for the Lutjewad area based on the radon-tracer method, similar to the calculation of nighttime fluxes in Hyytiälä by Kooijmans et al.(2017). Measurements of222Rn can be used to calculate fluxes of other gases because

222_{Rn is produced in the soil with a constant rate and it diffuses through the soil into}

the air. Once it is in the atmosphere, it is only affected by radioactive decay and by the effect of atmospheric mixing. The nighttime mole fraction of gases get either enriched (in the case of dominant sources) or depleted (in the case of dominant sinks) in a shallower nocturnal boundary layer compared to the daytime boundary layer. This means that, when the222Rn exhalation rate (FRn) is known, the surface fluxes of another gas—in this

case of COS (FCOS) and CO2(FCO2)—can be determined from the mole fraction changes

of the gas (¢COS and ¢CO2) over the night, relative to that of222Rn (¢222Rn): e.g., FCOS=

F_Rn¢COS/¢222Rn (Schmidt et al.,1996;Van der Laan et al.,2009a;Belviso et al.,2013). F_Rn was determined for the Lutjewad area in different measurement and modelling studies of which an overview is given inVan der Laan et al.(2016). In these studies, F_Rn varied

(7)

5

between 2.3 and 5.1 mBq°2 s°1. We will use the average over these studies, 3.7 mBq°2 s°1, with a standard deviation of 1.2 mBq°2s°1. The222Rn measurements in Lutjewad are made with an ANSTO dual-flow loop two-filter detector (Whittlestone and Zahorowski, 1998). Details about the measurement procedure are described in VVan der Laan et al. (2009a). Fluxes are only calculated for nights when at least 7 data points are available, where the R2values between222Rn and COS (CO₂) mole fractions are larger than 0.4 (0.5) and where the standard error of the flux (based on the uncertainty of the slope between

222_{Rn and COS or CO}

2mole fractions) is smaller than 4 pmol°2s°1(COS) and 1.5 µmol°2

s°1(CO2). Furthermore, the uncertainties of the radon-tracer method largely result from

the uncertainty of FRn. The flux uncertainty is therefore calculated as the quadrature sum

of the uncertainty on the slope and of FRn(1.2 mBq°2s°1).

300

350

400

450

Time of day, UTC + 1 [h]

COS [ppt] 300 350 400 450 300 350 400 450 0 3 6 9 12 15 18 21 Lutjewad − Sept. (a) 60 m 40 m 7 m 300 350 400 450

COS [ppt] 300 350 400 450 300 350 400 450 300 350 400 450 300 350 400 450 0 3 6 9 12 15 18 21 Hyytiälä − Sept. (b) 125 m 23 m 14 m 4 m 0.5 m 380 400 420 440 460

CO 2 [ppm] 380 400 420 440 460 380 400 420 440 460 0 3 6 9 12 15 18 21 Lutjewad − Sept. (c) 380 400 420 440 460

CO 2 [ppm] 380 400 420 440 460 380 400 420 440 460 380 400 420 440 460 380 400 420 440 460 0 3 6 9 12 15 18 21 Hyytiälä − Sept. (d)

Figure 5.1 | Diurnal variation of COS (top) and CO2(bottom) mole fractions in Lutjewad (left) and Hyytiälä (right).

Data are plotted as the median diurnal cycle per height. Data are shown for the month September (2015), which is the month where profiles are available for both sites. The medians are taken over 23–29 measurements per hour in Lutjewad and 26–27 measurements per hour in Hyytiälä.

(8)

5.3.RESULTS

5

105

5.3. R

ESULTS

5.3.1. D

IURNAL CYCLE OF

COS

AND

CO

₂ MOLE FRACTIONS

The COS mole fractions are lowest during the nighttime and closest to the surface, which is especially the case for the 0.5 m measurements at Hyytiälä (Fig. 5.1). The amplitude of the diurnal cycles of COS mole fractions is larger in Hyytiälä than in Lutjewad with a difference between maximum and minimum COS mole fractions of 32 ppt in Lutjewad (at 7 m) and 52 ppt in Hyytiälä (23 m), which is at comparable heights above the vegetation. On the other hand, the nighttime enhancement of CO₂mole fractions is larger in Lutjewad (37 ppm) than in Hyytiälä (21 ppm). The nighttime vertical gradients in Hyytiälä are strongest between 0.5 m and 4 m (-12.3 ppt m°1for COS and 6.61 ppm m°1for CO2), because that

is where turbulence is limited. Within the canopy, the gradients are relatively small (-1.0 ppt m°1 and 0.30 ppm m°1 between 4 m and 14 m, -0.41 ppt m°1 and 0.06 ppm m°1 between 14 m and 23 m for COS and CO₂, respectively). For both Hyytiälä and Lutjewad, the gradients are much smaller during the daytime. The daytime gradients are near zero in Lutjewad, but in Hyytiälä the vertical gradients between different heights persist during daytime, which is likely due to insufficient vertical mixing within the canopy. Furthermore, the early morning increase of COS mole fractions within the canopy takes place later than the decrease of CO2mole fractions, which is especially the case for the 0.5 m mole fractions

in Hyytiälä with a time delay of ª3 h.

− 40 − 30 − 20 − 10 0 10

COS

7m

−

da

ytime mean COS

7m [ppt] − 40 − 30 − 20 − 10 0 10 − 40 − 30 − 20 − 10 0 10 − 40 − 30 − 20 − 10 0 10 0 3 6 9 12 15 18 21 Lutjewad Dec−Jan Apr Aug−Sep Oct−Nov (a) − 40 − 30 − 20 − 10 0 10

COS

18m

−

da

ytime mean COS

18m [ppt] − 40 − 30 − 20 − 10 0 10 − 40 − 30 − 20 − 10 0 10 − 40 − 30 − 20 − 10 0 10 − 40 − 30 − 20 − 10 0 10 0 3 6 9 12 15 18 21 Hyytiälä Feb−Mar Apr−May Jun−Jul Aug−Sep Oct−Nov (b)

Figure 5.2 | Monthly average diurnal cycle of COS mole fractions minus the mean daytime mole fractions to

indicate the seasonal variation of the diurnal cycle amplitude. Data are plotted for Lutjewad (left) and Hyytiälä (right). The data show the 7 m mole fractions for Lutjewad and 18 m mole fractions for Hyytiälä. Mole fractions of 18 m (top of the canopy) rather than 23 m (above the canopy, comparable to 7 m in Lutjewad) are shown because the 18 m mole fractions cover a larger part of the year (February–November). The median difference between the 23 and 18 m mole fractions is 0.2 ± 1.4 ppt during daytime and 0.3 ± 1.0 ppt during nighttime for one month where both heights were sampled every hour (June 2017).

The largest diurnal variability occurs in the growing season (Fig. 5.2). In Lutjewad we didn’t measure mole fractions in the peak of the growing season, but still we observe that the largest diurnal variability occurs in the warmer months (April, August–September).

(9)

5

The nighttime drawdown of COS mole fractions that is larger in the spring and summer months is consistent with larger nighttime COS uptake by the ecosystem in those months in both Hyytiälä (Kooijmans et al.,2017) and Lutjewad (see below, Fig. 5.4).

5.3.2. S

OURCES AND SINKS BY WIND DIRECTIONS

Figure 5.3 shows the deviation of the COS, CO2and CO mole fractions from their seasonal

cycle for the Lutjewad site against wind direction. A negative (positive) deviation (e.g., COS7m– COSseas.< 0) is indicative of a sink (source) that is not represented by the seasonal

cycle. Typically, there is a difference in signals between daytime and strongly stable nights, especially for deviations of 7 m mole fractions (left plots in Fig. 5.3). No large COS deviations are observed for daytime data and weakly turbulent nights (when the temperature gradient between 60 and 7 m is lower than 0.75 °C), apart from a decrease of ª 15 ppt with wind from the east (see Fig. 5.3a-b). For nighttime data with strongly stable conditions we observe larger deviations from the seasonal cycle. For 7 m deviations we generally observe the largest depletions in COS from eastern wind and southwestern wind, which is with wind from inland (wind directions between 50 and 300º); however, no clear depletions are observed with wind directions from the south. For 60 m (right plots in Fig. 5.3) we also find COS to be depleted in eastern wind directions (Fig. 5.3b), and weakly from the southwest. COS mole fractions were substantially lower at all heights in a period of a few days between 1 and 8 September 2014 (not shown).

For CO2and CO, we observe elevations from the seasonal cycle for both day and night.

The elevations span the range of wind directions where air originates from inland. The CO₂mole fractions are further enhanced in the nighttime. The CO elevations are similar for day and nighttime, apart from a peak at 200 degrees, which is higher during strongly stable conditions.

Peaks do not necessarily point to larger sources or sinks in a certain direction, but could originate from a few nights with strongly stable conditions that drive large changes in mole fractions and that have a relatively large influence on the averages. The binned averages of the strongly stable nighttime conditions are more prone to such peaks because these data represent less data (332 data points) than the weakly stable nights (1269 data points).

In Section 5.4.3 we discuss the spatial distribution of COS sources and sinks based on these results.

5.3.3. E

STIMATE OF NIGHTTIME

COS

AND

CO

₂FLUXES

Figure 5.4 shows the nighttime fluxes of COS and CO2in Lutjewad based on the

radon-tracer method. Most of the derived COS fluxes are negative, implying COS sinks at the surface. Occasionally, there are positive fluxes, which coincide with periods in which we observe COS spikes after ploughing (see Section 5.3.4). The median nighttime COS flux is -3.0 ± 2.6 pmol m°2 _s°1_{(excluding the positive fluxes), with -2.9 ± 2.2 pmol m}°2_s°1

from August to November 2015 and -7.2 ± 2.8 pmol m°2s°1in April 2015. The nights with COS emissions have an average COS flux of +3.5 ± 2.1 pmol m°2s°1. Nighttime CO2fluxes

decrease from August to December, then increase in January and reach highest CO2fluxes

(10)

5.3.RESULTS

5

107 ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● 0 50 100 150 200 250 300 350 − 100 − 50 0 50 Wind direction [°] COS 7m − COS s e a s . [ppt] ● ● ● ● ● ● ●● ● ●● ● ● ●● ●● ●● ● 0 50 100 150 200 250 300 350 − 100 − 50 0 50 ● ● ● ● ● ● ●● ●●● ● ●● ● ● ● ● ● ● 0 50 100 150 200 250 300 350 − 100 − 50 0 50 7 m (a) ● ● ● Daytime

Nighttime: weakly stable Nighttime: strongly stable

● _● ●● ● ● ● ● ● ● ●● ● ●● ●● ● ● ● 0 50 100 150 200 250 300 350 − 100 − 50 0 50 Wind direction [°] COS 60m − COS s e a s . [ppt] ● _● ● ● ● ● ●● ● ●●● ● ● ●● ● ● ● _● 0 50 100 150 200 250 300 350 − 100 − 50 0 50 ● ● ● ● ● ● ●● ●●●● ●● ● ● ● ● ● ● 0 50 100 150 200 250 300 350 − 100 − 50 0 50 60 m (b) ● ● ●_● ● ● ●●_●● ● ● ●● ● ● ● ● ● ● 0 50 100 150 200 250 300 350 0 20 40 60 Wind direction [°] CO 2 − 7m − CO 2− se a s. [ppm] ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ● 0 50 100 150 200 250 300 350 0 20 40 60 ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● 0 50 100 150 200 250 300 350 0 20 40 60 (c) ● ● ●● ● ● ●_● ●● ● ● ●● ● ●● ● ● ● 0 50 100 150 200 250 300 350 0 20 40 60 Wind direction [°] CO 2 − 60m − CO 2− se a s . [ppm] ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ● _{● ●} ● 0 50 100 150 200 250 300 350 0 20 40 60 ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● 0 50 100 150 200 250 300 350 0 20 40 60 (d) ● ● ● ● ● ● _● ●●●● ● ● ● ● ● ● ● ● ● 0 50 100 150 200 250 300 350 0 50 100 150 Wind direction [°] CO 7m − CO s e a s. [ppb] ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● _● ● ● 0 50 100 150 200 250 300 350 0 50 100 150 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 0 50 100 150 200 250 300 350 0 50 100 150 (e) ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● 0 50 100 150 200 250 300 350 0 50 100 150 Wind direction [°] CO 60m − CO s e a s . [ppb] ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● 0 50 100 150 200 250 300 350 0 50 100 150 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●_● ● 0 50 100 150 200 250 300 350 0 50 100 150 (f)

Figure 5.3 | Deviation of 7 m (left) and 60 m (right) mole fractions of COS (a,b), CO2(c,d) and CO (e,f) from their

seasonal cycle in Lutjewad. Data are separated between daytime (solar elevation angle > 0°; green) and nighttime (solar elevation angle < 0°), where nighttime data are divided over weakly (blue) and strongly (orange) stable nights, which are separated based on the temperature difference between 60 and 7 m being smaller or larger than 0.75 °C.

5.3.4. S

OURCES OF

COS

FROM AGRICULTURAL FIELD DURING PLOUGHING

We occasionally observed spikes of COS mole fractions up to ª3000 ppt (see Fig. 5.5) in the autumn and winter months. These spikes occurred in October 2014, January 2015 and February 2018, of which we were able to confirm that the spikes in 2014 and 2018 were coinciding with ploughing of nearby agricultural fields.

(11)

5

● ● ●● ● ●_●● ●● ● ● ●_● ● ● ● ● ● ● ● ● _● ● ● ● ● ● ● − 10 − 5 0 5 FCOS − Rn [pmol m − 2 s − 1 ] ● ● ●● ● ●_●● ●● ● ● ●_● ● ● ● ● ● ● ● ● _● ● ● ● ● ● ●

Aug'14 Oct'14 Dec'14 Feb'15 Apr'15

FCOS − Rn [pmol m − 2 s − 1 ] (a) ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● Feb'18 Mar'18 ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ●● ● ● ● ●_● ● ● ● ●● ● ● ●●● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 0 5 10 FCO2 − Rn [µ mol m − 2 s − 1 ] ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ●● ● ● ● ●_● ● ● ● ●● ● ● ●●● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

Aug'14 Oct'14 Dec'14 Feb'15 Apr'15

FCO2 − Rn [µ mol m − 2 s − 1 ] (b) ● ● ● ● ● ●●● ● ● ● ● ● ● ●●● ● Feb'18 Mar'18

Figure 5.4 | Nighttime fluxes of COS (a) and CO2(b) in Lutjewad based on the radon-tracer method. Note that

the x-axis jumps from April 2015 to February 2018.

5.3.4.1.SPIKES INOCTOBER2014

A large spike was observed in the COS mole fractions at 7 m in the afternoon of 28 October, which lasted for ª 5 hours before the mole fractions gradually decreased in the evening. The 40 and 60 m COS mole fractions also increased (to above 500 ppt) but not nearly as high as that at 7 m. The occurrence of the spike coincided with ploughing of the field directly adjacent to the measurement tower. The COS plume originating from this field may not have reached the upper levels of the tower. Unlike COS, no apparent elevations were observed in CO₂, CO or SF₆at the time of the COS spike. On this particular day, the diurnal cycle of CO₂follows a typical diurnal cycle with larger mole fractions during the night than during the day. Also the mole fractions of222Rn are higher during the night and are positively correlated with that of CO2, CO and SF6 (R2 = 0.86, 0.67 and 0. 65,

respectively), indicating that the variability of CO2, CO and SF6is driven by boundary layer

changes. A spike of SF6is observed on October 30, but no COS spike is detected then.

5.3.4.2.SPIKES INJANUARY2015

In January 2015, a double peak of COS mole fractions takes place at 7 m in the night and in the morning of 23 January. It coincides with a change of the wind direction between northeast and south, indicating the origin of the sources from the south. No increases occur at 40 and 60 m in the night when the 7 m mole fractions show a spike; however, COS does spike at all heights in the morning for about 2 hours at ª10:00 hr. The reason that the 40 and 60 m COS mole fractions do not spike in the night but in the morning could be due to very stable nighttime conditions with a shallow boundary layer, such that the 40 - 60 m

(12)

5.3.RESULTS

5

109

28 Oct 29 Oct 30 Oct 31 Oct

350 650 950 1250 1550 1850 COS [ppt] (a)

22 Jan 23 Jan 24 Jan 01 Feb 05 Feb 10 Feb 15 Feb 20 Feb

350 650 950 1250 1550 1850 COS [ppt]

380 420 460 CO 2 [ppm] (b)

380 420 460 CO 2 [ppm]

100

200

300

CO [ppb]

(c)

22 Jan 23 Jan 24 Jan 01 Feb 05 Feb 10 Feb 15 Feb 20 Feb 100

200

300

CO [ppb]

7 10 13 SF 6 [ppt] (d)

7 10 13 SF 6 [ppt]

0 2 4 6 222 Rn [Bq m − 3 ] (e)

0 2 4 6 222 Rn [Bq m − 3 ]

0 120 240 360 Wind dir . [ °] 2014 (f)

22 Jan 23 Jan 24 Jan

2015

01 Feb 05 Feb 10 Feb 15 Feb 20 Feb

0 120 240 360 Wind dir . [ °] 2018 ● _{60 m} ● _{40 m} ● _{7 m}

Figure 5.5 | Time series of COS (a), CO2(b), CO (c), SF6(d) and222Rn (e) mole fractions and wind direction (f)

in Lutjewad for the three periods with elevated COS mole fractions. COS, CO2and CO mole fraction data are

2-minute averages,222Rn data are 30-minute averages and wind direction data are hourly averages. The spike of COS mole fractions at 7 m on 23 January, 2015 reached up to ª 3000 ppt. The spike of SF6on that day reaches ª

17 ppt.

air layer is decoupled from the nocturnal boundary layer and is thus not seeing the sources from the surface. As soon as turbulence is enhanced after sunrise, the air with elevated COS mole fractions can reach the higher air layers, which is visible as a short spike of COS mole fractions at all heights. The air layer with high COS mole fractions mixes and dilutes quickly; therefore, the COS spike doesn’t last long. Mole fractions of CO2and CO at 7 m

also increase late in the night, but the timing is different and the increase is relatively small compared to that of COS. CO2and CO do spike at all heights in the morning at the same

(13)

5

sharp peak in the morning and is correlated with COS (R2= 0.58) over the period of two days, although the correlation is mainly driven by only two data points. Still, the correlation of COS with SF6is larger than that with CO2and CO (R2= 0.36 and 0.07 respectively).222Rn

also increases in the morning and shows a peak later in the day. For the period in January 2015 we have no record available of ploughing in the nearby fields; however, this peak also occurred in a cold period, just like in February 2018 (see next section), and therefore has a high chance of being associated with ploughing as well, perhaps at a somewhat larger distance.

5.3.4.3.SPIKES INFEBRUARY2018

The period of February 2018 is characterized by a number of elevations of COS mole fractions that sometimes last for a few hours, but also extend to a few days. This period was characterized by cold weather (air temperature < 0 °C) in which agricultural fields in the area were being ploughed. The ploughing happened specifically in this period because the cold weather made it possible to drive on the frozen clay soil with heavy tractors, which is otherwise not possible in winter. The fact that we observe COS elevations of ª100 ppt (at 60 m) above the background mole fraction over a period of a few days is likely because many agricultural fields in the region were being ploughed in that period. CO₂, CO and

222_{Rn are also elevated in the periods when COS is higher. CO}

2and CO mole fractions are

strongly correlated in this period (R2= 0.94) and the ratio of CO to CO₂elevations in this period is 5.3 ppb ppm°1. COS mole fractions are not as strongly correlated with CO2and

CO (R2= 0.50 and 0.48, respectively) and the correlation between COS and222Rn is weak (R2= 0.14).

The period of February 2018 is characterized by a number of elevations of COS mole fractions that sometimes last for a few hours, but also extend to a few days. This period was characterized by cold weather (air temperature < 0 °C) in which agricultural fields in the area were being ploughed. The ploughing happened specifically in this period because the cold weather made it possible to drive on the frozen clay soil with heavy tractors, which is otherwise not possible in winter. The fact that we observe COS elevations of ª100 ppt (at 60 m) above the background mole fraction over a period of a few days is likely because many agricultural fields in the region were being ploughed in that period. CO2, CO and222Rn are

also elevated in the periods when COS is higher. CO₂and CO mole fractions are strongly correlated in this period (R2= 0.94) and the ratio of CO to CO₂elevations in this period is 5.3 ppb ppm°1. COS mole fractions are not as strongly correlated with CO2and CO (R2=

0.50 and 0.48, respectively) and the correlation between COS and222Rn is weak (R2= 0.14). On February 9 there is a small spike of SF6. Unfortunately, measurements of SF6are not

continuous over the period, which makes it difficult to compare SF6observations with that

of COS.

5.4. D

ISCUSSION

5.4.1. U

NDERSTANDING THE DIURNAL CHANGE OF ATMOSPHERIC

COS

MOLE FRACTIONS

The nighttime drawdown of COS mole fractions can be explained by nighttime uptake of COS in a typically shallow nocturnal boundary layer. The nighttime uptake of COS was extensively studied for the Hyytiälä site byKooijmans et al.(2017), where they found that

(14)

5.4.DISCUSSION

5

111

during nighttime COS is taken up primarily by plants, secondly by the soil, due to the light-independence of the enzyme carbonic anhydrase that drives the COS uptake. The average nighttime flux in Hyytiälä was -6.8 ± 2.2 pmol m°2s°1for the months July–October (Kooijmans et al.,2017_{), which is higher than the average nighttime flux of -2.9 ± 2.2 pmol} m°2 _s°1 _{between August and November in Lutjewad, but is similar to the average flux}

of -7.2 ± 2.8 pmol m°2_s°1_{in April in Lutjewad. For the Lutjewad site we are not able to}

separate the plant and soil uptake with the measurements that we have.

During daytime, the vertical mixing of the boundary layer with air aloft—that has only been little affected by COS uptake at the surface—causes the increase in COS mole fractions in the early morning and decreases the vertical gradients. The vertical gradient that is still distinguishable in Hyytiälä during daytime would allow for calculation of fluxes using the flux-gradient method.

The diurnal cycle of COS mole fractions contrasts that of CO₂(Fig. 5.1), which is driven by ecosystem respiration during nighttime and restored mixing with the free troposphere during daytime. Since fossil fuel emissions increase from August to December (Van der Laan et al.,2010), the decrease of nighttime CO2fluxes from August to December (Fig. 5.4)

is most likely due to decreasing ecosystem respiration. Similarly, nighttime CO2fluxes

increase from January to April most likely due to increasing ecosystem respiration. The increasing nighttime uptake of COS that we observe in April at the Lutjewad site can be due to both the soil and vegetation that respond to higher temperature. We cannot separate the influence of the soil and vegetation with the data that we have; however, the nighttime uptake of COS in Hyytiälä was predominantly driven by vegetative uptake (Kooijmans et al.,2017), and the soil fluxes did not show a seasonal cycle (Sun et al.,2018a). If we assume that similar mechanisms control the fluxes in Lutjewad, then it is likely that the larger nighttime COS uptake in April is controlled by the vegetation and that the nighttime uptake in winter is driven by the soil.

The reason for the later increase of canopy COS mole fractions than the decrease of CO2mole fractions (Fig. 5.1b,d) in the morning is that COS does not increase until the

vertical mixing starts after the ground is heated up. Prior to that, both vegetation and soil continue to take up COS within the canopy. On the contrary, CO2mole fractions start to

decrease as soon as photosynthetic uptake of CO2dominates respired CO2, which is likely

earlier than the moment when sufficient vertical mixing is established.

5.4.2. COS

SPIKES IN THE ATMOSPHERE

The fact that we observed COS spikes when agricultural fields were ploughed (October 2014, February 2018) may be due to outgassing of air to the surface, which is otherwise trapped in the pore space of the deep soil. The spikes may be an indication that COS mole fractions are enriched in the soil and that COS is produced in the deep soil. On the other hand, a spike of COS that coincided with that of SF6(January 2015) could point to an

anthropogenic source. In the following two sections we will discuss both sources.

5.4.2.1.COSPRODUCTION IN AGRICULTURAL SOIL

Oxic soils are typically a sink of COS, but have also been observed to change from a sink to a source under dry and high-temperature conditions (Maseyk et al.,2014). This suggests that the net flux of COS is a balance of a concurrent sink and source (Sun et al.,2018a). Based on the nighttime ecosystem fluxes in Fig. 5.4 we learn that the soil in the Lutjewad

(15)

5

area is generally a sink of COS, which turns into a source when the deep soil is brought to the surface due to ploughing.

Soil COS production was found to increase with temperature (Maseyk et al.,2014) and light intensity (Kitz et al.,2017). However, light is not expected to drive COS production in the deep soil. The underlying mechanisms of soil COS production are poorly understood, but could potentially be the hydrolysis of thiocyanate (Katayama et al.,1992) or thermal degradation of organic matter (Maseyk et al.,2014;Kaisermann et al.,2018). Recently, Kaisermann et al.(2018) observed a strong correlation between COS production and soil nitrogen. The fact that many agricultural soils were found to emit COS (Whelan and Rhew,2015;Maseyk et al.,2014;Kitz et al.,2017) could be explained by high nutrient inputs in those soils that trigger degradation of organic matter (Kaisermann et al.,2018). Furthermore,Maseyk et al.(2014) observed COS emissions from roots, even when the above-ground part of the plant was harvested. They considered different processes for root emissions of COS, such as root metabolism, rhizosphere microbial activity and the decomposition of rhizosphere biota, but no conclusion on the exact process causing COS emissions could be drawn.

Also aerobic soils (e.g., wetlands and salt marshes) were found to act as a source of COS in soils with high levels of sulfate (from salt water). Emissions in those soils are triggered by the redox potential (Devai and Delaune,1995;Whelan et al.,2013;Ogée et al.,2016). It is unsure what the sulfate levels are in the agricultural soil at the Lutjewad measurement site, but given that the agricultural fields originally were wetlands (those close by until 1929), may be an indication that these soils still have high sulfate levels. On the other hand, we do not find indications that the current wetland area is a source of COS, as we found no elevated COS mole fractions from that direction (Fig. 5.3a,b).

Reicosky(1997) showed that ploughing can cause a significant loss of CO₂ to the atmosphere for several days after ploughing; however, we did not see that CO₂(or CO and

222_{Rn) changed consistently with COS during the ploughing in October 2014. This implies}

that the deep soil pore space concentrations of these gases do not differ from ambient concentrations as much as COS does.

5.4.2.2.ANTHROPOGENIC SOURCE OFCOS

It is difficult to determine whether the COS peak in January 2015 covaries with that of SF₆ due to a similar source. SF₆is used extensively as an insulator gas in equipment for the transmission and distribution of electricity where SF6emissions occur through leakage and

maintenance losses. In aluminum production foundries, an SF6/inert gas mixture serves

as a cleaning agent to remove impurities from the melt, releasing a small amount SF6to

the atmosphere (Schwarz and Leisewitz,1999). The aluminum industry is also a known source of industrial COS emissions (Harnisch et al.,1995;Zumkehr et al.,2018). Hence, coinciding SF₆and COS enhancement could point to similar industrial emission sources. An aluminum smelting factory is located 40 km southeast of the Lutjewad station. However, given that COS, CO2, CO and SF6peak at the moment that turbulence is enhanced in the

morning after the gases have built up in the night in a shallow nighttime boundary layer (given the nighttime peak at 7 m) suggests that COS and SF6originate from a more local

source than the aluminium factory 40 km away. Given the cold weather in this period the peaks may also be associated with ploughing, just like in October 2014 and February 2018. Whether COS and SF₆have the same source or are related to the same anthropogenic activity (e.g. machinery used for agriculture that emits SF6and soil emissions of COS) is

(16)

5.4.DISCUSSION

5

113

unknown.

Furthermore, we observed some similarities in the increase of COS, CO2, CO and222Rn

in February 2018. Again, these similarities are likely due to the change of atmospheric boundary layer heights. The ratio of elevations of CO and CO2in February 2018 of 5.3 ppb

ppm°1_{is typical for emission ratios of non-road transport (Lopez et al.,}_2013;_{Super et al.,}

2017) including i.a. machinery for agriculture; however, the ratio is more likely caused by a combination of various sources. Hence, the ratio does not provide insight into individual sources in the region.

5.4.3. S

PATIAL DISTRIBUTION OF

COS

AND

CO

2 SOURCES AND SINKS

The wind direction analysis in Fig. 5.3 aids in identifying the main sources and sinks of COS in the region of the Lutjewad measurement station, although we have to consider that sources and sinks can balance each other. In general, we find depletions of COS only coming from inland, which is likely driven by terrestrial vegetation and soil. In northern directions we did not observe a deviation from the seasonal cycle, indicating that the mud flats and salt marshes are not a strong net source or sink of COS. Still, a source of COS in the salt marshes could be balanced by uptake of COS by plants.

The fact that we observe COS depletion at 60 m during daytime is an indication that this is a regional signal. For the period between 1 and 8 September 2014, when COS mole fractions were lower, backward trajectories (NOAA HYSPLIT Trajectory Model) show that the air originates from Eastern Europe in the preceding 3 days. Terrestrial vegetation in these areas could have caused the depleted COS mole fractions in this period and could generally be the origin of the lower COS mole fractions with eastern wind.

The depletions of COS that we observe from the southwest are larger at 7 m under strongly stable nighttime conditions than during daytime and at 60 m, which implies that these depletions are caused by more local sinks of COS. On average we do not detect a sink from the south, even though this also covers continental air masses, including agricultural land nearby. Vegetative uptake of COS in these wind directions could be balanced by COS sources. Regionally, industrial activities in the Ruhr area in Germany could be an interesting area for further investigation of sources of COS, such as from aluminum production (Campbell et al.,2015;Zumkehr et al.,2018). Locally, we are not able to point to any specific activity or industry towards the south, apart from a small village (ª200 inhabitants). The data in Fig. 5.3 mainly represent the autumn and winter months with only the beginning and end of the growing season, larger COS depletions can be expected in the summer months if vegetation plays a dominant role on atmospheric COS.

The elevated CO₂mole fractions during daytime likely originate from anthropogenic activity from inland, which is substantiated by elevated CO mole fractions in the same range of wind directions. In the nighttime we find CO2mole fractions to be further elevated

than during the daytime, because the effects are amplified due to the shallow mixing layer. In both cases, we cannot attribute these elevations to anthropogenic sources alone because the net ecosystem exchange (NEE) of CO₂can contribute significantly to these elevations. Other tracers, e.g.,14CO₂, are needed to partition the CO₂elevations into anthropogenic emissions and NEE (Turnbull et al.,2009;Vogel et al.,2010;Van der Laan et al.,2010). The wind directions where CO2enhancements at 7 m shows a peak in the night (200º and 275º)

(17)

5

there is also a CO peak (200º). We are not aware of any anthropogenic activity that could lead to depletions of COS and at the same time emit CO2and CO, the sources and sinks of

these gases do therefore not have to be related. We also have to consider that a few nights with strongly stable conditions and a particular wind direction could have a large influence on the averages, which would affect all gases and would be detected as a peak.

5.5. C

ONCLUSION

We analyzed atmospheric measurements of COS and CO₂mole fractions at the Lutjewad station and compared them to those at a boreal forest site Hyytiälä to infer local to regional surface sources and sinks of COS. We detected lower COS mole fractions from inland, which is likely driven by vegetation and soil uptake, and found no indications that the mud flats and salt marshes at the coast are a net sink or a net source. The nighttime COS fluxes were determined to be -3.0 ± 2.6 pmol°2 _s°1 _{using the radon-tracer correlation}

approach. No evidence of anthropogenic sources of COS in the Lutjewad region was found. The diurnal cycle of COS mole fractions is driven by nighttime uptake together with a shallow nighttime boundary layer. This is in contrast to nighttime respiration and fossil fuel emissions of CO2that leads to elevated CO2mole fractions. Furthermore, we

observed a couple of spikes of COS mole fractions when nearby agricultural fields were being ploughed, which provides evidence that COS is produced within the agricultural soil. One COS peak coincided with a SF₆spike, suggesting an anthropogenic source; however, the exact origin could not be deduced.

We thank the technical staff from the Center for Isotope research in Groningen and from the SMEAR II station in Hyytiälä for the maintenance of the measure-ment stations. In particular we would like to thank Bert Kers, Marcel de Vries, Janne Levula and Juho Aalto for their efforts in organizing and maintaining the measurement campaigns. We also thank Kadmiel Maseyk, Ulli Seibt, Wu Sun, Kukka-Maaria Erkkilä and Timo Vesala for their contributions to the field work in Hyytiälä.

(18)

A5.APPENDIX

5

115

A5.

A

PPENDIX

● ●●●● ● ● ● ● ●● ● ● ●● ●● ● ●● ●● ●● ● ● ● ● ● ●●●●● ● ●● ● ● ● ●● ● ● ●● ● ● ●● ● ● ●● ● ● ● ●●● ● ● ● ● ●● ●_● ● ● ●● ●● ●_● ●● ●● ● ● ●● ● ● ● ● ●●● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●●● ●_● ● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ●_● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ●●● ● 350 400 450 500 550 600 DOY COS [ppt]

Jan Apr Jul Oct Jan

● _{Lutjewad, Netherlands − in situ}

Lutjewad, Netherlands − flasks

Figure A5.1 | Seasonal cycle of daytime average COS mole fractions at 60 m in Lutjewad. The data consist of in

situ measurements from August 2014–April 2015 and January–February 2018 (circles) and flask measurements between December 2013 and February 2016 (stars). The in situ measurements from August 2014–April 2015 are an update of the measurements presented inKooijmans et al.(2016). The seasonal cycle shows a peak-to-peak amplitude of 87 ppt, which was estimated to be 96 ppt byKooijmans et al.(2016) when no flask measurements were included.

(19)