Introduction
The availability of reactive nitrogen (or fixed nitrogen), including all
nitrogen species except N2, is often a determinant controlling the
primary production in terrestrial and marine ecosystems (Vitousek and
Howarth, 1991). However, since the mid-1800s the global nitrogen cycle has
been significantly perturbed by excessive production of reactive nitrogen
through human activities (at a similar rate to natural biological fixation)
(Dentener et al., 2006; Galloway et al., 2008; Fowler et al., 2013), such as
the Haber–Bosch process, fossil-fuel combustion, and legume cultivation
(Gruber and Galloway, 2008). A fraction of the reactive nitrogen is
transported from continents to oceans via rivers, groundwater, and the
atmosphere, exerting a significant influence on marine ecology and
biogeochemistry (e.g., acidification, eutrophication, and oxygen depletion)
(Seitzinger et al., 2005; Paytan et al., 2006; Doney et al., 2007;
Krishnamurthy et al., 2010). In particular, the supply of external or
new nitrogen to marine systems through atmospheric deposition can be
substantial in some marginal and coastal seas (Beddig et al., 1997; Castro
and Driscoll, 2002; de Leeuw et al., 2003; Uno et al., 2007; Zhang et al.,
2010), as well as the open ocean (Galloway et al., 2004; Duce et al., 2008).
The East Sea/Sea of Japan (EJS) is a semi-closed marginal sea (covering an area of
1.01 × 106 km2) surrounded by Russia, Korea, and Japan,
which has been suggested to be an ideal site to investigate the impacts of
atmospheric nutrient deposition on the northwestern Pacific Ocean (Kim and
Kim, 2013). It is a highly productive region where nitrogen is severely
limited (Talley et al., 2004; Jenkins, 2008; Kim and Kim, 2013). Therefore,
the supply of reactive nitrogen is particularly important in determining the
primary productivity of this marine ecosystem. In addition to the upwelling
of deep waters and N2 fixation by diazotrophs within the ocean,
land-derived reactive nitrogen (which is largely anthropogenic) can also
contribute to the nitrogen required for primary production in the EJS. Since
no major rivers flow into the EJS from the surrounding coasts, the fluvial
inputs of nitrogen can be ignored (Kang et al., 2009; Yoo and Park, 2009).
Nevertheless, terrigenous nitrogen transported via the atmosphere may reach
the open ocean, contributing to the nitrogen inventory across the EJS. In
fact, the mid- and long-range atmospheric transport of dust and pollutants
from land to the northwestern Pacific Ocean (including the EJS) has been
shown to be remarkable (Jo et al., 2007; Kang et al., 2009,
2011; T. W. Kim et al., 2011; Zhang et al., 2011), particularly since this
region is located downwind of East Asia, a densely populated area
characterized by intensive emissions of aerosols and gases by anthropogenic
and natural processes (Cooke et al., 1999; Richter et al., 2005; Kim, 2008).
Recently, T. W. Kim et al. (2011) suggested the significant atmospheric
nitrogen deposition has switched extensive parts of the northwestern Pacific
Ocean (including the EJS) from being nitrogen limited to phosphorous limited
during the last 3 decades. A follow-up study by Kim et al. (2013),
however, attributed this increasing trend in relative abundance of nitrogen
over phosphorous to nutrient transport by ocean currents. Nevertheless, Kim
and Kim (2013) argued that this oceanic region would remain nitrogen limited
for the next 100 years, based on observational data and model predictions.
The debate aroused here signifies the importance of an unambiguous budget of
nitrogen in the EJS, especially the fraction transported by the atmosphere,
which is poorly constrained due to a lack of data (Zhang et al., 2011).
The atmospheric deposition of nitrogen over the southern EJS was previously
evaluated by Kang et al. (2010) using dry deposition observations combined
with estimates for wet deposition. However, their approximation on nitrogen
deposition was an underestimate, since it did not take into account organic
nitrogen, which is a globally significant constituent of total reactive
nitrogen in the atmosphere (ca. 30 % on average) (Jickells, 2006; Cape et
al., 2011; Cornell, 2011). Indeed, the growing weight of evidence from global
observational studies suggests that a considerable fraction of deposited
nitrogen is in organic form (Cornell, 2011, and references therein), which is
able to stimulate the productivity of bacteria and phytoplankton in the
receiving marine ecosystems (Seitzinger and Sanders, 1999; Bronk et al.,
2007). In particular, atmospheric organic nitrogen deposition was shown to be
especially high over Asia (Cornell, 2011; Ito et al., 2014), implying it is
imperative to incorporate this fraction into the budget of atmospheric
nitrogen input to the EJS. Nevertheless, although quantitatively significant,
the sources, chemical compositions, and bioavailability of atmospheric
organic nitrogen are largely unknown (Mace et al., 2003a; Duce et al., 2008;
Jickells et al., 2013), hindering our understanding of its biogeochemical
role in the receiving ecosystems.
In order to provide better constraints on atmospheric nitrogen deposition to
the EJS, we analyzed reactive nitrogen species (inorganic and organic) and
major ions in precipitation samples collected at a coastal site upwind of
this marginal sea. The emphasis of the present work was placed on organic
nitrogen, which has been ignored in previous study and is generally poorly
characterized in comparison with its inorganic counterpart. With the aid of
air-mass back-trajectory (AMBT) and factor analysis, the geographical and emission
sources of organic nitrogen were explored. Moreover, using the calculated
total nitrogen depositional flux, the fraction of new primary production of
the EJS which is potentially supported by atmospheric nitrogen deposition
was estimated.
Materials and methods
Study site and sample collection
The sampling campaign was conducted at Uljin (37∘ N,
129.4∘ E), which is located on the eastern coastline of Korea
(Fig. 1). It is a rural area with a small population of around 52 000
inhabiting an area of 989.06 km2. Precipitation samples (N = 33)
were collected on an event basis on the rooftop of a four-story research
facility (ca. 16 m above the ground) at the Korean Institute of Ocean
Science and Technology (East Sea branch) from February 2011 to January 2012.
A home-made sampler was employed, which is composed of a polypropylene funnel
(diam. 250 mm) connected by Tygon FEP-lined tubing to a high-density
polyethylene (HDPE) bottle placed in a covered bucket. Prior to use, the
apparatus was thoroughly cleaned using dilute hydrochloric acid and rinsed
with deionized (Milli-Q) water (18.2 MΩ cm). The sampler was
manually deployed at the onset of the precipitation events and retrieved
after cessation. The impacts of dry deposition were minimized by restricting
the exposure time to dry conditions (i.e., <1 h for daytime events
or ≤5 h for overnight events). After collection, the samples were
transferred to a laminar flow clean room and filtered through pre-combusted
(5 h at 500 ∘C) Whatman 0.7 µm GF/F glass fiber filters.
Subsamples for dissolved organic carbon (DOC) analysis were placed in 20 mL
pre-muffled glass ampoules (5 h at 500 ∘C) and chemically
preserved with 6 M pure hydrochloric acid, followed by fire sealing.
Aliquots for dissolved nitrogen species and major ions were stored in
pre-cleaned HDPE bottles and kept frozen at -20 ∘C until
analysis.
Statistical summary of concentrations of dissolved chemical species
in precipitation analyzed in this study (unit: µmol L-1).
NO3-
NH4+
ON
Na+
K+
Mg2+
Ca2+
Cl-
SO42-
OC
Min.
2
1
BDa
8
2
2
7
13
4
9
1st quartile
12
5
7
55
8
7
18
82
23
61
Median
31
24
13
138
15
28
46
206
58
156
3rd quartile
74
34
40
900
25
108
68
1580
117
296
Max.
428
139
145
2930
69
337
259
3780
215
849
Mean
58
39
30
635
21
76
57
894
83
219
VWAb
20
14
13
430
13
47
25
578
43
97
a Below detection limit.
b Volume-weighted average.
Map showing the location of Uljin (sampling site;
37∘ N, 129.4∘ E), Donghae (37.5∘ N,
129.1∘ E), and Oki Island (36.3∘ N, 133.2∘ E).
![]()
Analytical methods
Total dissolved nitrogen (TDN) and DOC concentrations were measured
simultaneously by high-temperature catalytic oxidation (HTCO) using a
Shimadzu TOC / TN analyzer (Model TOC-VCPH / CPN) equipped with an ASI-V
autosampler. The acidified samples (pH <2 using HCl) were sparged
with carbon dioxide free carrier gas (UHP oxygen) at a flow rate of
150 ml min-1 for 2 min to remove inorganic carbon. Then the samples were injected
into a combustion column packed with Pt coated alumina beads heated to
720 ∘C. The carbon dioxide and nitrogen monoxide evolving from
combustion were detected by a non-dispersive infrared detector and a
chemiluminescence detector, respectively. Major ions including inorganic
nitrogen species were determined by high-performance liquid chromatography
using a Waters 2695 high-performance liquid chromatographic (HPLC) system equipped with a Waters 432 conductivity
detector. Cation analysis (i.e., Na+, K+, Ca2+, Mg2+, and
NH4+) was carried out using a mobile phase consisting of 3.0 mmol L-1 HNO3
and 0.1 mmol L-1 Ethylenediaminetetraacetic acid (EDTA), and a Waters IC-Pak C M/D
column (150 × 3.9 mm, 5 µm). Anions (i.e., Cl-,
NO3-, NO2-, and SO42-) were analyzed using a
borate/gluconate eluent containing 12 % acetonitrile and a Waters IC-Pak
A HR column (75 × 4.6 mm, 6 µm). The column was maintained
at 30 ∘C and the detector at 35 ∘C. Dissolved organic
nitrogen (DON) was quantified as the difference between TDN and dissolved
inorganic nitrogen (DIN), which is the sum of NO3- and NH4+
(NO2- in our samples was always below the detection limit and is
therefore not included). This approach could yield large uncertainties for
DON, especially when DIN dominates the nitrogen pool, as has been suggested by
previous studies (Cornell et al., 2003; Lesworth et al., 2010). The average
relative standard deviations associated with replicate measurements of
standards were 4, 2, and 3 % for TDN, NO3-, and
NH4+, respectively. Based on error propagation using uncertainties of
each analyte, the precision of DON was estimated to be less than 18 %
when DON % is greater than 20 %, 18–60 % when DON % varies
from 5 to 20 %, and greater than 100 % when DON % is less
than 5 % (e.g., 18 March 2011, 30 May 2011, and 16 August 2011). The limits of
detection calculated as 3 times the standard deviation of blanks are
0.5 µmol L-1 for NH4+, 0.2 µmol L-1 for
NO3-, 0.2 µmol L-1 for NO2-, 1.1 µmol L-1 for TDN, and 5.0 µmol L-1 for DOC. The
detection limit for DON was estimated to be 1 µmol L-1 (Cornell
et al., 1998; Zhang et al., 2001; Walker et al., 2012), by using detection
limits of relevant nitrogen species following error propagation rules for
standard deviations. Certified reference materials (from University of Miami
for DOC and TDN, and from National Research Council of Canada for DIN) were
employed during the sample analysis to confirm the data quality.
Concentrations (top) and percentages (bottom) of each
reactive nitrogen species in precipitation samples collected at Uljin from
2011 to 2012.
![]()
Results and discussion
Distributions of dissolved reactive nitrogen in precipitation
Concentrations and speciation
The concentrations of dissolved reactive nitrogen showed considerable
variations during the sampling year, ranging from 2 to
428 µmol L-1 for NO3-, from 1 to
139 µmol L-1 for NH4+, and from below the detection
limit to 145 µmol L-1 for DON (Table 1 and
Fig. 2). Rather similar temporal trends were found for all three nitrogen
species, with high abundances observed in spring and fall (Fig. 2). Aside
from the source strength, the distribution pattern displayed here is likely
to be influenced by several factors specific to this location. The
precipitation amount might be one of the most important factors, because the
variation patterns of these nitrogen species are in general accordance with
the precipitation regime of Korea (particularly high in summer), in a manner
that the concentrations decrease as the precipitation depth increases. In
addition, the wind systems (or air-mass origins) of Korea may also contribute
significantly to the observed temporal trend (see Sect. 3.1.2.). The
northwesterlies prevail most of the year, transporting large amounts
of terrestrially derived nitrogen from inland Korea and the Asian continent
(especially China) to Uljin and the EJS. In contrast, the prevailing wind
over the summer blows from the Pacific Ocean, carrying precipitation
associated with relatively pristine air masses.
Overall, NO3- is the most abundant nitrogen species with an annual
average concentration (AVG) of 58 µmol L-1 and a
volume-weighted average (VWA) (Topol et al., 1985) of
20 µmol L-1, followed by NH4+ (AVG = 39 µmol L-1; VWA = 14 µmol L-1) and DON
(AVG = 30 µmol L-1;
VWA = 13 µmol L-1) (Table 1). The reactive nitrogen
concentrations in precipitation observed in this study are generally at the
upper end of the range of literature values reported for coastal sites
globally over the past decade (Keene et al., 2002; Luo et al., 2002; Mace et
al., 2003b; Kieber et al., 2005; Calderon et al., 2007; Violaki et al., 2010;
Cape et al., 2011; Zamora et al., 2011). However, much higher concentrations
have been found in rainwater over coastal areas of China, which were
attributed to significant pollution associated with the rapid economic
development (Chen et al., 2011; Zhang et al., 2012). The relatively high
levels of reactive nitrogen in the atmosphere at Uljin probably can be linked
to the anthropogenic emissions of nitrogen in East Asia including Korea and
subsequent atmospheric transport.
In contrast to observations from regions characterized by intensive
agricultural practices (e.g., China and midwestern USA) (Fahey et al., 1999;
Zhang et al., 2008), NO3- showed higher abundances with respect to
NH4+, implying that combustion processes contributed more
significantly than agricultural activities to reactive inorganic nitrogen in
our samples (Galloway et al., 2004; Spokes and Jickells, 2005; Lee et al.,
2012). A few exceptions were observed in April, May, October, and November
(Fig. 2), when concentrations of NH4+ were similar to or higher than
those of NO3-. These exceptions can be attributed to enhanced
agricultural activities (i.e., application of N fertilizer and livestock
manure to the farmland) during these time periods in Korea (Lee at al.,
2012). Although being a minor component, DON made up a considerable fraction
(28 % on average) of the TDN in our samples, approximating to that
contributed by NH4+ (30 % on average). In comparison with
NO3- and NH4+, the relative proportion of DON showed
remarkable variations from sample to sample, which were found to be
relatively lower from April to August (Fig. 2). The seasonality and air-mass
origins might contribute to the pattern observed here to some extent.
However, the major factors that control the proportion of organic nitrogen in
atmospheric reactive nitrogen pool remain rather ambiguous (Cornell, 2011)
(see Sect. 3.3.1 for further discussion).
Potential source regions based on air-mass origins
Atmospheric deposition plays an important role in global biogeochemical
cycles, as it allows chemical substances to be transported to locations
remote from their source regions over relatively short timescales.
Therefore, the reactive nitrogen found in our samples was presumably subject
to influences of air masses originating from distant areas. In order to
explore the potential source regions and to assess the contributions by
long-range atmospheric transport, the provenances of the air masses
corresponding to each of the precipitation events were determined based on
AMBT analysis, using the Global Data Assimilation System (GDAS) data set and the
Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model
(version 4) developed at the Air Resources Laboratory of the National Oceanic and Atmospheric Administration
(NOAA) (Draxler and
Hess, 1998). Apart from the nearby continental regions (i.e., the Korean
peninsula), reactive nitrogen could be transported over long distance to
Uljin from mainland Asia as well as the surrounding oceans, as observed for
other airborne species in Korea (Kim et al., 2005; Kim, 2008). Accordingly,
based on calculated AMBTs, air-mass origins were sorted into three groups:
Korean Peninsula (type I), Asian continent (type II), and Pacific Ocean
(type III) (Fig. 3). A similar classification regime can be found in our
previous study (Yan and Kim, 2012), in which detailed descriptions for each
group were presented. It is noteworthy that the reactive nitrogen in each
group does not exclusively come from the corresponding source region
identified. For example, the precipitation events associated with long-range
transported air masses inevitably receive some contributions from local
sources.
Three-day air-mass backward trajectories at 500 m (above
ground level) for precipitation events at Uljin during the sampling period,
obtained using the HYSPLIT model. The air-mass provenances are sorted into
three groups: Korean Peninsula (type I), Asian continent (type II), and
Pacific Ocean (type III). The percentage values in parentheses indicate the
occurring frequencies of each type.
![]()
Volume-weighted average concentrations of reactive nitrogen
species for precipitation events associated with the three types of air
masses
at Uljin. Error bars represent standard deviations.
![]()
The distribution patterns of VWA concentrations for the three groups were
consistent for NO3- and NH4+, with the highest concentrations
observed in type II and the lowest in type III (Fig. 4). As these two
nitrogen species are known to be predominantly anthropogenic, this pattern is
consistent with the air-mass origins proposed for these groups. Type II air
masses also account for the largest fraction (47 %) of sampled
precipitation events, suggesting a significant contribution to inorganic
nitrogen at Uljin from the highly industrialized and urbanized areas of East
Asia, especially eastern and northeastern China (Richter et al., 2005; Jeong
and Park, 2008; Lee et al., 2014). The lowest VWA DON concentration was also
observed for air masses derived from surrounding marine areas (type III),
implying that marine biogenic emission is an insignificant source for DON at
Uljin. In contrast to the pattern seen for inorganic nitrogen, the highest
abundances of DON were associated with type I air masses. It is thereby
inferred that the emission sources within Korea play a dominant role in
supplying organic nitrogen to the atmosphere at our location. Despite the
fact that the largest concentrations of atmospheric organic nitrogen
worldwide have been observed in China (Cornell, 2011), type II air masses at
Uljin were characterized by relatively lower DON abundances than type I. This
could be partially due to the inclusion of clean marine air masses and/or the
loss of the labile portion of organic nitrogen during the long-range
transport. In addition, the highest proportion of DON (32 %) was also
observed in type I air masses, confirming the importance of organic nitrogen
from emissions in Korea.
Source identification for DON
While the major sources for inorganic reactive nitrogen in the atmosphere
are known to be anthropogenic (Galloway et al., 2004; Fowler et al., 2013),
the primary origins of the organic fraction remain poorly characterized (Cape
et al., 2011; Cornell, 2011). Previous studies have suggested that
atmospheric organic nitrogen may stem from a variety of natural and
anthropogenic processes, including resuspension of soil dust, marine
emissions, biomass burning, agricultural activities, industrial production,
and fossil-fuel combustion (Jickells et al., 2013 and references therein). In
this study, attempts were made to identify the major sources of DON in our
precipitation samples, on the basis of factor analysis in combination with
major ions as tracers. The results obtained from varimax-rotated factor
analysis suggest that most variations (91 %) in the nine variables
included can be accounted for by three extracted factors (eigenvalue
>1) (Table 2).
Varimax-rotated principal factor matrix*.
Factor loadings exceeding 0.7 are shown in bold.
Factor 1
Factor 2
Factor 3
NO3-
-0.12
0.94
0.25
NH4+
-0.12
0.59
0.46
ON
0.20
0.27
0.93
Na+
0.98
-0.05
0.10
K+
0.97
0.12
0.10
Mg2+
0.99
0.02
0.02
Ca2+
0.22
0.95
0.09
Cl-
0.99
-0.03
0.07
SO42-
0.75
0.45
0.26
Eigenvalue
4.53
2.44
1.21
% Variance
50.35
27.09
13.47
* The software package used for principal factor analysis is SPSS
16.0.
Enrichment factors of major ions in precipitation at Uljin
relative to seawater.
K+ / Na+
Mg2+ / Na+
Ca2+ / Na+
Cl- / Na+
SO42- / Na+
Ratios in seawater
0.022
0.227
0.044
1.160
0.121
Ratios in precipitation
0.124
0.298
0.844
1.384
0.847
EF*
5.64
1.31
19.17
1.19
7.00
* The enrichment factor for a given ionic species X is calculated using
the following equation:
EF = [X / Na+]precipitation / [X / Na+]seawater.
The first factor is characterized by high loadings of Na+, K+,
Mg2+, Cl-, and SO42-. The enrichment factors (EFs) for
these ions (except Na+) with respect to average seawater composition
were calculated using sodium as a reference element (Keene et al., 1986). The
EFs fall in between 1 and 10 (Table 3), suggesting these chemical species are
primarily derived from marine emissions (Poissant et al., 1994). This
interpretation is confirmed by the significant positive correlations between
the abundances of these ions and Na+ (R2 = 0.86 for K+,
0.93 for Mg2+, 0.98 for Cl-, and 0.51 for SO42-).
Therefore, this factor probably can be attributed to the contributions from
marine sources. The largest fraction of variance (50 %) (Table 2) in the
nine variables explained by this factor is consistent with the marine feature
of Uljin. The second factor has high loadings of NO3- and Ca2+,
which are generally considered to be of different origins. In addition,
moderate loadings are shown for NH4+ and SO42- on this
factor. In the atmosphere, NO3- is mainly derived from NO2
emitted by fossil-fuel combustion (Dentener et al., 2006), whereas Ca2+
comes from both crustal and marine sources (Gabriel et al., 2002). The
relatively high-EF value (19.17) indicates that crustal contribution is more
important for Ca2+ in our samples. This factor is therefore likely to
represent a mixed source involving inputs from combustion processes as well
as soil resuspension. The association of these ions (i.e., Ca2+,
NH4+, NO3-, and SO42-) has frequently been observed
in factor analysis conducted in previous studies (Hu et al., 2003; Wai et
al., 2008; Song and Gao, 2009), and is attributed to the neutralization of
acidic anions by alkaline cations in precipitation (Wai et al., 2008).
Since the loadings of DON on these two factors are rather low, the
contributions from marine emissions, soil dust, and combustion-related
processes to DON are expected to be less significant.
The third factor is associated with a moderate loading of NH4+ and a
high loading of DON. NH4+ in precipitation is derived from its
gaseous precursor NH3, which is primarily released during agricultural
activities such as animal husbandry and the application of synthetic
fertilizer (Galloway et al., 2004). These activities are also known to be the
origins of several important atmospheric organic nitrogen species (e.g.,
urea) (Cornell et al., 1998). According to statistical data from
International Fertilizer Industry Association (2013), a considerable fraction
of the nitrogen fertilizer consumed in Korea each year is in the form of
urea. In addition, Lee et al. (2012) found NH4+ and DON in
precipitation collected in a southern Korean city were mainly derived from
fertilizer use and livestock excretion in surrounding rural areas using
nitrogen isotopic analysis. This suggests that agricultural practices
conducted within Korea could be a crucial source of NH3 and organic
nitrogen in the atmosphere, which is in line with the conclusions drawn from
the AMBT analysis that the Korean Peninsula contributes significantly to
NH4+ and DON in our samples (Fig. 4). Taken together, it is concluded
that the primary fraction of DON in our samples originated from agricultural
activities in Korea, albeit contributions from other anthropogenic and
biogenic sources cannot be ruled out.
Seasonal wet depositional fluxes of NO3-,
NH4+, DON, TDN, and DOC at Uljin. The precipitation amounts for
each season are also shown.
![]()
Biogeochemical impacts of atmospheric nitrogen deposition on the EJS
Seasonal variations in nitrogen fluxes and DON proportions
The wet depositional fluxes (defined as the product of concentration and
precipitation depth) of the three reactive nitrogen species showed remarkable
seasonal variations, with relatively high values observed in spring and fall
for NO3- and NH4+, and in fall and winter for DON,
(Fig. 5). While being a minor component in spring and summer,
DON made up the largest fraction of the TDN fluxes in fall and winter.
Precipitation amount is discounted as the cause of this temporal trend as it
shows no correlation with the variations in nitrogen fluxes. The seasonal
variations in nitrogen fluxes are mainly attributed to the strength of
emission sources located in Korea as well as other distant regions (see the
discussion in Sect. 3.1.2).
In addition, the relative distributions of DON, TDN, and DOC fluxes were
found to be in good agreement over all seasons (Fig. 5), implying there might
exist an inherent link among these species. A statistically significant
positive correlation (R2 = 0.64) was observed between the proportion
of organic nitrogen (ON) in total reactive nitrogen (TN) and ON to organic
carbon (OC) atomic ratio in dissolved organic matter in precipitation
(Fig. 6). This study is the first to point out this correlation, but analysis
of a larger data set consisting of the available literature values and data
obtained at our site and another location in Korea (Seoul) shows an even
stronger correlation (R2 = 0.72) (Fig. 6). The correlation found in
this data set is significant, since the data incorporate rainwater and aerosol
samples collected worldwide from geographically varied locations (e.g., urban
vs. rural, terrestrial vs. marine, and anthropogenic vs. pristine) (Jordan et
al., 1995; Eklund and McDowell, 1997; Campbell et al., 2000; Kieber et al.,
2005; Miyazaki et al., 2010; Gioda et al., 2011). Therefore, the trend
observed herein is likely not simply coincidental, but universal.
Plot of ON / TN versus ON / OC in precipitation samples
collected at Uljin (N = 33) and Seoul (N = 33) in Korea and other
locations worldwide, including the western North Pacific (aerosols,
N = 4), Chesapeake Bay (N = 14), Puerto Rico (N = 5), Costa Rica
(N = 3), New England (N = 15), and Wilmington (N = 4). The solid
lines denote the best-fit correlations for Uljin data (red) and the entire
data set (green).
![]()
Annual atmospheric deposition fluxes of reactive nitrogen
(mmol N m-2 yr-1) at coastal and marine locations estimated based
on long-term measurements during the last decade.
Location
Period
Deposition
NO3-
NH4+
DIN
DON
%DON
Reference
mode
Uljin, Korea
2011–2012
wet
28
20
48
19
28
This study
Donghae, Korea
2002–2003
dry
27
6
33
–
–
Kang et al. (2010)
Taiwan
2006
dry
–
–
39
22
36
Chen et al. (2010)
Shenzhen, China
1986–2006
wet
37
58
95
–
–
Huang et al. (2013)
Singapore
2007–2008
wet
51
26
77
34
31
He et al. (2011)
dry
19
3
21
21
50
Crete, Greece
2003–2006
wet
10
7
17
5
23
Violaki et al. (2010)
dry
25
2
27
17
39
Baltic Sea
2001–2002
wet plus dry
21
19
40
4
9
Rolff et al. (2008)
Tampa Bay, USA
2005
wet
23
10
33
2
6
Calderon et al. (2007)
Barnegat Bay, USA
1999–2001
wet
29
19
48
–
–
Gao (2002)
dry
4
2
6
–
–
Miami, USA
2007–2009
wet
15
12
27
2
7
Zamora et al. (2011)
Puerto Rico
2004–2007
wet
6
4
10
7
41
Gioda et al. (2011)
Variations in the ON / OC ratio of airborne organic matter could reflect
source variations, as being suggested by previous studies (Neff et al., 2002;
Cape et al., 2011; Kanakidou et al., 2012). Specifically, the highest
ON / OC ratios were observed for aerosols (total concentrations for
aerosol sizes between 0.39 and 10.0 µm) over the remote North
Pacific Ocean by Miyazaki et al. (2010), who also proposed that aerosols
subject to increased biological influences were associated with higher
ON / OC ratios. In contrast, the lowest ON / OC ratios were mostly
found for dissolved organic matter in precipitation samples collected in
Seoul, where atmospheric organic matter is mainly derived from anthropogenic
processes, especially fossil-fuel combustion (Yan and Kim, 2012). Atmospheric
processing may also affect ON / OC ratios, as being seen in different
types of aerosols that undergo various physical and photochemical atmospheric
processes (Sun et al., 2011). In addition, it has been suggested that organic
nitrogen tends to be less effectively removed from the atmosphere than
inorganic nitrogen (Cornell, 2011 and references therein). Therefore, higher
proportions of organic nitrogen might be linked to higher contributions from
aged aerosols characterized by elevated ON / OC ratios (Sun et al.,
2011).
The organic fraction of total reactive nitrogen in the atmosphere is highly
variable in time and space, lacking a consistent trend on a global scale
(Cornell, 2011; Jickells et al., 2013). In addition, the bioavailability of
atmospheric organic nitrogen is poorly constrained (Duce et al., 2008;
Kanakidou et al., 2012). Therefore, despite the fact that organic nitrogen is
increasingly recognized as a significant factor in the atmospheric deposition
of reactive nitrogen, its biogeochemical impact on receiving ecosystems
(especially the ocean) remains unclear. The empirical relationship found in
the present study between ON / TN and ON / OC (Fig. 6) suggests that the
utilization potential (evaluated using ON / OC ratios) of atmospheric
deposited organic nitrogen by marine biota is positively correlated with its
proportion in total atmospheric reactive nitrogen. Thus, atmospheric
deposition of organic nitrogen may play a vital role in supplying
bioavailable nitrogen to the surface layer of marine systems, especially when
inorganic nitrogen is less abundant, as is the case for the northwestern
Pacific Ocean.
Annual depositional fluxes of TDN at Uljin and over the EJS
The annual wet depositional flux is calculated to be
28 mmol N m-2 yr-1 for NO3-,
20 mmol N m-2 yr-1 for NH4+, and
19 mmol N m-2 yr-1 for DON (Table 4). Our
estimates of inorganic nitrogen fluxes are higher than those obtained for dry
deposition by Kang et al. (2010) at Donghae (Fig. 1), a coastal site adjacent
to our sampling location. This is mainly due to the higher scavenging
efficiency of airborne species by wet deposition comparing with dry
deposition. In addition, the increasing anthropogenic emissions in East Asia
may also contribute to this discrepancy (Galloway et al., 2004), as the
sampling campaign by Kang et al. (2010) was conducted a decade ago. However,
the fluxes of NO3- for both deposition modes were similar, which can
be attributed to the effective dry scavenging of NO3- associated with
coarse mode aerosols (Nakamura et al., 2005; Matsumoto et al., 2009). In
general, atmospheric depositional fluxes of reactive nitrogen are higher in
the coastal areas of East Asia (including the western coast of the EJS) than
at other coastal locations around the world (Table 4). This is in agreement
with the distribution of oceanic regions characterized by the most intensive
atmospheric nitrogen deposition on a global basis (Dentener et al., 2006;
Krishnamurthy et al., 2007). The considerable contributions by DON to
atmospheric TDN fluxes (7–50 %) at these coastal/marine sites suggest
the significant role of DON in supplying nitrogen to surface waters of the
ocean.
Fractions of new primary production supported by
atmospheric nitrogen input in world oceans on an annual basis.
Ocean
Deposition
Nitrogen
Contribution to
Method
Reference
regions
mode
species
new production
Southern East Sea/Sea of Japan
wet plus dry
TDN
12–14 %
field observation
This study
Southern East Sea/Sea of Japan
wet plus dry
DIN
∼10 %
field observation
Kang et al. (2010)
East Sea/Sea of Japan
wet plus dry
DIN
2–12 %
modeling
Onitsuka et al. (2009)
Southern Yellow Sea
wet plus dry
DIN
10.5 %
field observation
Lv et al. (2005)
Yellow Sea
wet plus dry
TDN
0.3–6.7 %
field observation
Qi et al. (2013)
Southern East China Sea
dry
TDN
8.3 %
field observation
Chen et al. (2010)
East China Sea
wet plus dry
DIN
1.1–3.9 %
modeling
Zhang et al. (2010)
East China Sea
dry
DIN
0.1–9 %
field observation
Nakamura et al. (2005)
South China Sea
wet plus dry
TDN
20 %
modeling
Kim et al. (2014a)
Bay of Bengal
dry
TDN
up to 25 %
field observation
Srinivas and Sarin (2013)
Arabian Sea
dry
TDN
<1 %
field observation
Srinivas and Sarin (2013)
Southeast Mediterranean Sea
wet plus dry
DIN
8–20 %
field observation
Herut et al. (1999)
Western Mediterranean Sea
wet
DIN
10–20 %
field observation
Loye-Pilot et al. (1993)
Eastern Atlantic
wet plus dry
DIN
4–6 %
field observation
Neuer et al. (2004)
Bermuda
wet
TDN
7–15 %
field observation
Kim et al. (2014b)
World oceans
wet plus dry
TDN
1.5–6.9 %
modeling
Duce et al. (2008)
World oceans
wet plus dry
DIN
5.1 %
modeling
Krishnamurthy et al. (2010)
Since our sampling site is located on the eastern coast of Korea, potential
losses during transport must be considered when using the results of this
study as a proxy for atmospheric depositional fluxes of reactive nitrogen
over the remote EJS. Oki Island (Japan), an island located downwind of Uljin
in the eastern section of the southern EJS (Fig. 1), is mainly influenced by
air masses originating from China and Korea, and shows a similar wet
depositional flux of inorganic nitrogen to the values found in this study
(45 mmol N m-2 yr-1; Japanese Acid Deposition Survey) (Kitayama
et al., 2012). This implies that losses between the coast of Korea and
offshore areas are small for wet depositional fluxes, as being shown for wet
depositions of nitrogen over the eastern China Sea (Zhang et al., 2010).
Therefore, our estimated atmospheric wet deposition flux can be extrapolated
to the offshore region in the southern EJS. By taking the dry depositional
flux of inorganic nitrogen observed at Oki Island
(35 mmol N m-2 yr-1 for gases plus particles in 2003–2008)
(Endo et al., 2011), the total (wet plus dry) inorganic nitrogen flux was
calculated to be 83 mmol N m-2 yr-1. Further, if assuming the
same DON / TDN ratio in dry deposition as in wet deposition, the total
atmospheric depositional flux of reactive nitrogen (organic plus inorganic)
amounts to 115 mmol N m-2 yr-1 over the southern EJS.
Biogeochemical implications
Considering the prevailing westerly winds and the anthropogenic origins of
these atmospheric reactive nitrogen species, their atmospheric deposition
represents a source of new rather than recycled oceanic nitrogen to the EJS.
Therefore, the atmospheric deposition of this nitrogen could presumably fuel
the new primary production of this marine ecosystem. While inorganic nitrogen
is readily utilized by marine biota, organic nitrogen is only partially
available. The bioavailability of atmospheric organic nitrogen is largely
determined by its chemical composition. Specifically, most reduced organic
nitrogen species (e.g., amino acids, urea, and amines) can be taken up by
marine microorganisms on very short timescales, whereas the organic nitrogen
associated with humic-like substances is probably refractory (Bronk et al.,
2007). Incubation experiments estimate the bioavailability of atmospheric
organic nitrogen to be 20–80 % (Peierls and Paerl, 1997; Seitzinger and
Sanders, 1999; Wedyan et al., 2007). Although not explicitly measured for the
DON in our samples, a significant fraction is expected to be bioavailable,
since it is mostly derived from agricultural activities that are known to
produce large amounts of atmospheric reduced organic nitrogen, especially
urea. Moreover, the severely N-limited conditions of the EJS are likely to
promote the utilization of atmospheric organic nitrogen. Assuming a
bioavailability of 20–80 % for DON and 100 % for inorganic nitrogen,
atmospheric deposition can supply approximately 89–109 mmol m-2 yr-1
of reactive nitrogen to support primary production in the southern
EJS.
Using a Redfield C / N ratio of 6.625, this nitrogen flux can be
converted to 7.1–8.7 g C m-2 yr-1 fixed by marine
phytoplankton. This would account for 12–14 % of new primary production
in the southern EJS annually, which is taken to be
62 g C m-2 yr-1 based on the particulate organic carbon export
fluxes reported by Hahm and Kim (2001) (using tritium and helium isotopes)
and Kim et al. (2011) (using 234Th / 238U disequilibrium). Our
estimate is higher than those obtained by Onitsuka et al. (2009) using a
coupled physical–ecosystem model (6–12 % for the southern EJS from 1996
to 2003) and by Kang et al. (2010) based on field observations (ca. 10 %)
(Table 5). These discrepancies can be ascribed to the inclusion of organic
nitrogen in our study as well as a potential increase of nitrogen deposition
resulting from growing anthropogenic emissions over the past decade. On a
global scale, the value obtained for the southern EJS is generally at the
higher level, comparing with other oceanic regions (Table 5). Taking into
account that the southern EJS (especially the Ulleung Basin) is highly
productive (Hyun et al., 2009), the estimate obtained in this study is rather
remarkable. Such high estimates can be explained by the exceptionally high
levels of atmospheric anthropogenic nitrogen deposition over the coastal seas
downwind of East Asia (Krishnamurthy et al., 2007; Duce et al., 2008).
Moreover, the contribution of atmospheric reactive nitrogen to biological
productivity in the southern EJS is expected to be more pronounced during
summer to fall, when the upward flux of nitrogen from the deep layer is
suppressed by water column stratification (Kang et al., 2010; D. Kim et al.,
2011).
The results presented here might be subject to large uncertainties, such as
those associated with spatial variations of depositional fluxes across the
EJS, the proportions of organic nitrogen in dry deposition, and the
bioavailability of organic nitrogen in the atmosphere. Nevertheless, our
estimation evidently suggests atmospheric transport represents a significant
source of external nitrogen input to the southern EJS. Previous studies
indicate that N2 fixation is insignificant to the nitrogen budget of
this marine system (Yanagi, 2002; Kang et al., 2010). Therefore, in addition
to nitrogen upwelled from below the euphotic zone and transported by the
Tsushima Warm Current through the Korea/Tsushima Strait (Onitsuka et al.,
2007; Yoo and Park, 2009), atmospheric deposition could sustain a
considerable fraction of the nitrogen demand by primary productivity in the
southern EJS, especially the offshore region. Moreover, the elevated
nitrogen deposition may modify the nutrient limitation regime (Duce et al.,
2008) of the EJS and thus impact the phytoplankton community structure
therein (Krishnamurthy et al., 2007; Galloway et al., 2008). Due to climate
warming, El Niño is likely to become more prevalent in the future, which
could result in suppression of the nutricline and thus restriction of
nutrient upwelling from deep waters (Mackey et al., 2010). Conversely, the
anthropogenic nitrogen deposition over the northwestern Pacific Ocean is
predicted to increase in the near future (Duce et al., 2008). Consequently,
atmospheric deposition of nitrogen is expected to play an increasingly
important role in the biogeochemistry of the northwestern Pacific Ocean and
the EJS over the coming decades. Special attentions should be given to the
organic fraction, since little is known with respect to its sources and
distributions, rendering the regulation of emission as well as the
prediction on its future trend rather difficult.