The increasing use of intensive agricultural practices can lead to damaging
consequences for the atmosphere through enhanced emissions of air pollutants.
However, there are few direct measurements of the surface–atmosphere exchange
of trace gases and water-soluble aerosols over agricultural grassland,
particularly of reactive nitrogen compounds. In this study, we present
measurements of the concentrations, fluxes and deposition velocities of the
trace gases HCl, HONO,
As the demand for food production grows in line with an increasing global population, so too does the development of intensive agricultural practices. These can have deleterious impacts on the environment and human health (Godfray et al., 2010; Foley et al., 2011), particularly through the emission of trace gases and the formation of airborne particles generated by their reactive chemistry. It is therefore important that measurements be made of the surface–atmosphere exchange of trace gases and associated aerosol compounds to quantify the emissions from – and deposition to – land used for agriculture, in order to quantify the impact of agricultural activities on the atmosphere and environment. Such studies can inform the development of abatement strategies and legislation designed to control emissions from agricultural activities. This also provides important process understanding to represent better the dry deposition processes in chemistry and transport models used to predict air quality and climate change.
Of particular importance to the surface–atmosphere exchange over
agricultural land is the formation of atmospheric reactive nitrogen
(
The interaction of
HONO is similar to
The dry deposition of the acidic gases themselves can also induce soil
acidification, which on agricultural soils can limit the growth of crops
through perturbation of the uptake of nutrients. HCl, like
Measurements of trace gases and associated aerosols are, however, restricted by the availability of appropriate instrumentation, complications in their measurement due to their reactivity and water solubility, and the potential interference of gas–particle interactions.
Techniques to measure concentrations and fluxes of these trace gas and
associated aerosol components require multispecies quantification, low
detection limits and fast temporal resolution. Eddy covariance, the most
direct micrometeorological technique for the measurement of trace gas
fluxes, requires fast-response sensors that are not available for some
species (such as
The aerodynamic gradient method (AGM) derives fluxes of a tracer from its vertical
concentration gradient, which can be obtained from concentration
measurements at two or more heights, avoiding the requirement for fast
response measurement. Developments in automated wet-chemistry
instrumentation have in turn led to the development of the Gradient of
Aerosols and Gases Online Registrator (GRAEGOR), a two-point gradient system
that measures the concentrations of HCl, HONO,
Other wet-chemistry instruments have also been developed to measure individual species at one height, such as the Long Path Absorption Photometer (LOPAP), which measures concentrations of HONO with fewer artefacts than the GRAEGOR (Heland et al., 2001). A comparison study between LOPAP HONO measurements and the Gas and Aerosol Collector (GAC) – an instrument which uses similar measurement techniques to the GRAEGOR – was conducted by Dong et al. (2012), but there has not yet been a published comparison between the LOPAP and GRAEGOR in measurements of HONO. Similarly, measurements of trace gases and aerosols above agricultural grassland using the GRAEGOR are limited, and previous studies above these land systems have been restricted to measurements of a limited number of species within a limited particle size range.
The aim of this study was to use the GRAEGOR to measure concentrations and
fluxes of the trace gases HCl, HONO,
Location of the Bush Tower site (3
The campaign was conducted during the late spring–summer 2016 (21 May–24 June)
at the Easter Bush measurement site (3
Both fields are used for year-round (although not continuous) sheep grazing,
in rotation with adjacent fields, but the south field also typically has an
annual cutting for silage. Mineral fertilisation is carried out twice a year
on both fields. During this study, fertilisation of the two fields occurred
between 08:00 and 09:00 on 13 June, using urea mineral fertiliser
at a rate of 69.9
Over the years the Easter Bush field site has hosted several long-term
measurements of
The GRAEGOR (Energy Research Centre of the Netherlands) is a wet-chemistry
instrument that measures the concentrations of reactive trace gases (HCl,
HONO,
Each sample box contains a horizontal wet rotating annular denuder (WRD)
(Keuken et al., 1988) and a steam jet aerosol collector
(SJAC) (Khlystov et al.,
1995; Slanina et al., 2001) connected in series. Air is drawn through each
sample box simultaneously by an air pump at a rate of 16.7
A high-density polyethylene (HDPE) tube (0.3
Autonomous calibration of the FIA system was carried out 24
Measurements of the airflow into the sample boxes were conducted using an independent device (TSI Mass Flowmeter 4140) once every fortnight during the campaign. Additional checks of the field performance of the instrument included daily checks of the WRD tubes and sample box air inlets for signs of visible contamination.
The GRAEGOR sampling boxes have very short inlets with no size selection.
Consequently, the aerosol concentration reflects water-soluble total
suspended particles (TSP). It detects any compound that dissociates to
form the measured ions and therefore has a number of artefacts. These
include interferences in HONO measurements through
The GRAEGOR has been demonstrated to be capable of measuring fluxes in a number of studies both in identical form to the one used here (Wolff et al., 2010a, b; Twigg et al., 2011) and in related variants (Nemitz and Sutton, 2004; Rumsey and Walker, 2016). Ammonia-specific instruments based on the same technology (AMANDA, GRAHAM, ECN, Petten, the Netherlands; Wyers et al., 1993) represent the most commonly used instrument for the automated measurement of ammonia fluxes.
Vertical profiles of temperature were measured at the tower using
fine-thread, custom-made thermocouples set at the same heights as the
GRAEGOR sample boxes. Located 0.4
The aerodynamic gradient method (AGM), based upon flux–gradient similarity
theory, calculates the flux of a tracer (
A temperature gradient profile for the campaign was derived from
measurements of air temperature at the two heights at which concentrations
were measured (0.6 and 2.4
The dry deposition velocity (
The gradient technique is only applicable for inert species whose flux is
constant with height. Most studies of surface exchange fluxes of reactive
compounds do not have the information to assess whether chemical reactions
might interfere with the flux measurement, but in this study the behaviour
of
The concentration limit of detection (LOD) of the instrument for each of the
species measured was quantified from a field blank test. The field blank
test was carried out prior to the campaign on the 20 March over 24
Limit of detection (LOD, determined as 3 standard deviations
from average baseline signal), mean (
Mean (
Mean (
The minimum detectable flux for each aerosol and gas species measured by the
GRAEGOR is dependent upon atmospheric stability and the ambient
concentration of the given trace gas or aerosol species. Based on the method
described by Thomas et al. (2009), median minimum detectable fluxes
(
When calculating the flux of a species using the aerodynamic gradient method, it is apparent that errors in individual concentration measurements propagate into an error in the concentration differences and, subsequently, affect the accuracy of the calculated vertical concentration gradient. Some errors systematically affect both heights and therefore affect the gradient to a lesser extent than systemic errors in sampling efficiency at a single height, such as the difference in capture efficiency of the WRD tubes or slight differences in airflow caused by differences in the critical orifices, which may impact the accuracy of concentration measurements and resultantly affect the precision in the error of the concentration difference.
The overall random error in the measurements of the trace gas and
water-soluble aerosol concentrations (
Uncertainties for the trace gases and water-soluble aerosols measured
calculated by error propagation ranged from 8 % to 18 % (
The error in the concentration difference (
Errors in flux calculations can similarly be determined through the Gaussian
error propagation method applied to Eq. (1). Wolff et al. (2010b), using an
analogous form of this equation, showed that total error in the flux is
composed of (
Throughout this paper, stated errors for concentration measurements are derived from the measurement uncertainty as calculated by Eq. (10), while stated errors for flux calculations are derived from the flux uncertainty as calculated by Eq. (11). Calculated errors for the uncertainty in concentration measurements, the error in the concentration difference and the error in the calculated fluxes for all species measured are similar to values determined by previous studies which have used the GRAEGOR successfully to measure flux gradients (Thomas et al., 2009; Wolff et al., 2010b; Twigg et al., 2011).
Concentrations that were less than 5 times the limit of detection as
calculated before the campaign began (20 March) were discarded.
Calculated fluxes were filtered according to a standard protocol. Fluxes
were not calculated for periods of low wind speed (
Global radiation (orange line), rainfall (blue bars), relative humidity (green dots), air temperature (red line), wind direction (brown circles) and wind speed (grey line) recorded during the Easter Bush campaign, May to June 2016. The fertilisation period was 08:00–09:00 on 13 June and is highlighted in green.
Figure 2 shows time series of the rainfall, radiation, relative humidity,
air temperature, and wind speed and direction measured during the campaign.
The meteorology splits into two episodes. From 24 May to 5 June 2016, the dominant prevailing wind direction was north-easterly,
accompanied by dry and sunny conditions with air temperature displaying a
characteristic diel cycle that increased each day. Following a period of
cloudier conditions from 6 to 10 June, the prevailing wind
direction shifted to south-westerly for the remainder of the measurement
period. Conditions became wetter and the diel air temperature amplitude was
reduced. Relative humidity remained high throughout the campaign, with only
occasional periods
Time series of hourly concentrations of the water-soluble aerosol
species measured during the Easter Bush campaign. Results smoothed using a
5
Time series of hourly concentrations of the gaseous species
measured during the Easter Bush campaign. Results smoothed using a
5
Summary statistics for the concentrations of the trace gas and water-soluble
aerosol species measured at 2.4
Mean concentrations of
Median concentrations of particulate
In contrast,
Hourly median diel trace gas concentrations measured by the
GRAEGOR at 2.4
Hourly median diel water-soluble aerosol concentrations measured
by the GRAEGOR at 2.4
Maximum concentrations for
The time series of measurements presented in Figs. 3 and 4 show that both
aerosol and trace gas concentrations are affected by prevailing
meteorological conditions, with larger concentrations for each species
during the drier, warmer period of 28 May to 6 June, followed
by decreased concentrations from 6 to 10 June when
precipitation increased and temperature decreased. Concentrations were lower
– except for the peaks in
The concentrations of
Figure 5 shows the median diel concentrations of
Time series of hourly trace gas fluxes measured during the Easter Bush campaign. Results smoothed using a 5 h moving point average. The fertilisation period was 08:00–09:00 on 13 June and is highlighted in green. Flux uncertainties for each trace gas are included as error bars.
Figure 7 shows the time series of the fluxes for the traces gases measured during the campaign. Data gaps are due to either absent data points (unpaired concentrations) or periods where data were filtered (refer to Sect. 2.3.4).
Bidirectional fluxes were present for both
Summary statistics for the trace gas fluxes, deposition velocities,
theoretical maximum deposition velocities and canopy resistance values are
presented in Table 2. The maximum
Median diel cycles for deposition velocity (
Median diel cycles for the deposition velocity and calculated theoretical
maximum deposition velocity for the trace gases HCl, HONO,
Time series of hourly fluxes of water-soluble aerosol species measured during the Easter Bush campaign. Results smoothed using a 5 h moving point average. The fertilisation period was 10:00 on 13 June and is highlighted in green. Flux uncertainties for each aerosol are included as error bars.
The measured surface fluxes of the aerosol species
Pre-fertilisation, all aerosol species exhibited deposition fluxes. The
deposition fluxes were larger during the drier, warmer period from 31 May to 6 June
and close to zero during the wetter conditions at the
end of the campaign. An important exception was the emission of
Summary statistics for the fluxes and deposition velocities for the aerosol
species measured are shown in Table 3. As for the trace gases, the median
deposition velocities for the aerosol species exclude the period of flux
divergence which occurred during fertilisation. The maximum flux for
A comparison of HONO measurements from the GRAEGOR and two LOPAP instruments
was conducted from 26 May to 6 June to investigate the
potential artefacts in the WRD method used by the GRAEGOR. The LOPAPs were
part of a study to investigate the mechanisms controlling HONO fluxes over
managed grassland, including investigating the potential ground sources of
HONO, details of which are presented in Di Marco et al. (2018). A
series of simple linear regression analyses was conducted to determine the
level of agreement between the concentrations of HONO measured by each
sample box of the GRAEGOR and each of the LOPAPs. The two LOPAP instruments
were operated at the two heights of 0.6 and 2.0
The ion balance of measured selected anions (
On 7 June, a QCL with inlet at height 1.6
The ion balance for the hourly-measured cation (
The formation of
Fluxes of
Fluxes for the trace gases were bidirectional for
The normalised deposition velocity as a function of
Deposition velocities for
The dry deposition of particles can be modelled using a process-orientated
approach, which describes the deposition velocity as a function of particle
size based on removal mechanisms acting within the vegetation canopy, such
as Brownian diffusion, impaction, interception and sedimentation (Slinn and
Slinn, 1980; Davidson et al., 1982; Slinn, 1982). The models predict that
for particles
Secondary ammonium compounds are typically found in the accumulation mode
(0.1 to 1
While the dynamic range of
By contrast, the median deposition velocity of 0.37
Thus, the relatively high deposition velocities for
It should be noted that the increase in
Median diel deposition velocities for
Reductions in
It should be noted that during this period the aerodynamic gradient method does not derive accurate fluxes because the condition of flux conservation is not met (Wolff et al., 2010b), and this period has therefore not been included in the diel cycles and summary statistics presented above.
By contrast, fluxes of total ammonium (
Fluxes of tot-
The time series for tot-
A second potential pathway is the emission of HONO from the soil. As
described by Scharko et al. (2015), the oxidation of ammonium by microbes in
soils with high nitrification rates can lead to biogenic emissions of HONO.
The addition of urea to the agricultural soil at Easter Bush would lead to
an increase in soil
As shown in Fig. 5, the median diel concentrations for HONO recorded by
the GRAEGOR at 2.4
The comparison between the LOPAPs and the GRAEGOR revealed that both sample
boxes of the GRAEGOR measured higher HONO concentrations than the LOPAP,
principally due to the presence of a constant concentration offset of 0.01
to 0.02
The higher concentrations recorded by the GRAEGOR can be explained by the
presence of chemical interferences that occur on the inlet, at the
air–liquid interface and within the sampling solution. As the WRD uses a
liquid film to sample HONO, and as HONO can form heterogeneously on such
surfaces, overestimation of HONO can occur. Furthermore, interferences by
chemical reactions of
A comparison between daytime (06:00 to 18:00) GRAEGOR HONO concentrations
and LOPAP HONO concentrations found only a slightly greater difference than
the comparison between nighttime (19:00 to 05:00) concentrations recorded
by the GRAEGOR and LOPAP. While previous comparisons between the DOAS and
the WRD found that daytime concentrations measured by the WRD were higher
than the DOAS compared to nighttime measurements, these studies were
generally conducted in urban areas where both HONO and
Spindler et al. (2003) developed the following quantification of the
chemical artefact produced by the mixing of
Simple linear regression analyses between GRAEGOR (2.4
To determine whether the HONO concentration offset in the GRAEGOR measurements
impacted on the measurements of HONO flux, a comparison between the HONO
flux values derived from GRAEGOR and LOPAP measurements was conducted.
Concurrent fluxes of HONO derived from LOPAP and GRAEGOR measurements exist
for 72 hourly measurements, from 26 May to 6 June. Figure 15 shows (a) the full
time series of concurrent HONO flux values derived
from GRAEGOR and LOPAP measurements and (b) a scatter plot of GRAEGOR
against LOPAP HONO flux values. Overall, GRAEGOR HONO fluxes are biased
towards deposition, with greater deposition values and lesser emission
values compared to concurrent LOPAP values. This pattern would be consistent
with the concept of an artefact formation dependent upon
The comparison between the GRAEGOR and QCL found that, while there was
reasonable agreement between the instruments, the GRAEGOR measured somewhat
higher
A similar comparison between a WRD system (the Ammonia Measurement by
Annular Denuder with Online Analysis, AMANDA) and the QCL system was
conducted at the same site in 2004 and 2005 by Whitehead et al. (2008). This
comparison also found that the WRD system measured higher concentrations of
Any errors in the GRAEGOR's internal
While there remain significant differences in measured
In this paper, we have presented for the first time simultaneous
measurements of the trace gases HCl, HONO, Simultaneous measurements of the components of the
The deposition velocities measured for the aerosol compounds Evidence for a HONO daytime source at the site throughout the campaign adds
to the growing body of past measurements that has found evidence for HONO
daytime formation in rural, urban and agricultural areas. There is also
evidence for the emission of HONO postfertilisation at the site.
This also appears to be the first time a comparison between measurements of
HONO concentrations determined by the LOPAP and the GRAEGOR instruments has
been documented. While good linear agreement exists between HONO
measurements taken by GRAEGOR and LOPAP at both measurement heights, a
consistent offset in GRAEGOR HONO measurements suggests the presence of
chemically induced artefacts within the GRAEGOR system. This is potentially
linked to atmospheric
Furthermore, this paper presents a comparison between measurements of
Future measurements of aerosol deposition velocities should aim to
investigate the effect of particle size upon deposition velocity, using a
more robust measurement of particle size than used here. In addition, the
ability of urea pellets to act as a potential surface on which heterogeneous
formation of HONO and
Data from the MARGA were obtained from DEFRA (2018) and are subject to Crown copyright, DEFRA, licenced under the Open Government Licence (OGL).
EN, CFDM and WJB devised the study and secured some of the funding.
GRAEGOR measurements were taken by RR and CFDM. GRAEGOR data were processed by RR with
input from CFDM, MMT and EN. LOPAP HONO measurements were made by CFDM, MMT, LJK and
LC. The QCL
The authors declare that they have no conflict of interest.
This work was supported through a studentship funded jointly by the Max Planck Institute for Chemistry and the University of Edinburgh School of Chemistry, as well as by the UK Natural Environment Research Council (NERC) through the project Sources of Nitrous Acid in the Atmospheric Boundary Layer (SNAABL, NE/M013405/1), with additional field-site support from NERC National Capability funding. The QCL was operated within the framework of the UK-China Virtual Joint Centre for Improved Nitrogen Agronomy funded through the Newton programme and administered by the UK Biotechnology and Biological Sciences Research Council (BBSRC). Edited by: Leiming Zhang Reviewed by: four anonymous referees