Journal of Atmospheric Chemistry (2005) 52: 81–99

DOI: 10.1007/s10874-005-8354-6

C Springer 2005

Factors Influencing Nitrogen Speciation

in Coastal Rainwater

R. J. KIEBER, M. S. LONG and J. D. WILLEY

Department of Chemistry and Biochemistry, University of North Carolina at Wilmington,
Wilmington, North Carolina 28403-5932, USA, e-mail:Kieberr@uncw.edu

(Received: 3 June 2005; accepted: 2 June 2005)

Abstract. Rainwater was collected from 129 rain events between February 2002 and August 2003
and analyzed for ammonium (NH+ −
4 ), nitrate (NO3 ), organic nitrogen (ON) and free amino acids
− +
(AA). Inorganic nitrogen (NO3 + NH4 ) was the dominant form of N representing 85% of total
nitrogen based on volume-weighted averages. The remainder of the N occurred as organic nitrogen
species of which free amino acids contribute approximately 17%. A significant, and in some cases the
majority (>75%), of the remaining ON could be accounted for by macromolecular uncharacterized
humic like substances. This has important ramifications with respect to the long range transport of
atmospheric ON because humic materials are recalcitrant and therefore may travel long distances from
their source. There was a distinct seasonality to the N speciation data with maximum concentrations
of NH+ 4 , ON and AA occurring in the spring. Air-mass back trajectory analysis indicates there is a
strong anthropogenic component to the NO− +
3 , NH4 and AA signal but not ON. There was a strong
positive correlation between amino acid concentrations and ammonium which suggests they have
similar sources and sinks in rainwater. Finally, large episodic additions of NH+ 4 and AA during
tropical events could significantly impact short term bioavailable N budgets in estuaries impacted
by these storms. Approximately three times as much NH+ 4 and AA were deposited during Hurricane
Isabel (317 µmoles · m−2 and 84 µmoles · m−2 respectively) compared to the mean impact of average
summertime rain events at this location.

Key words: amino acids, ammonium, organic nitrogen, rainwater

1. Introduction
Inorganic nitrogen species including nitrate (NO− 3 ) and ammonium (NH4 ) are
+

among the most widely measured analytes in precipitation. The majority of atmo-
spheric nitrate is formed by oxidation of NO and NO2 , emitted primarily from fossil
fuel combustion forming nitric acid HNO3 (g) which is subsequently scavenged into
the aqueous phase. Ammonia (NH3 ) is emitted from numerous sources including
both natural and anthropogenic. The main natural sources include emission from
soils and the oceans and wild animal excreta while the main anthropogenic sources
are agriculturally based including domestic animal excreta, fertilizers and biomass
burning (Russell et al., 1998). Atmospheric ammonia is lost via deposition and con-
version to ammonium very close to its source with an atmospheric residence time
of 1 to 5 days (Warneck, 1988). Recent studies demonstrate that organic nitrogen
82 R. J. KIEBER ET AL.

(ON) is also a significant component of the nitrogen pool in rainwater (Cornell

et al., 2003). It occurs in precipitation by incorporation of gas phase materials or
scavenging of aerosols. Significant effort has gone into quantifying ON concen-
trations along with inorganic N concentrations in atmospheric waters because of
the potential to impact aquatic ecosystems where N is often the primary nutrient
limiting plant, algal and microbial production. Current estimates indicate 35–60%
of the nitrogen entering Atlantic coastal and offshore waters is via atmospheric
deposition (Paerl et al., 1994; Paerl and Whitehall, 1999).
One of the most important uncertainties with regards to ON in atmospheric
waters is its chemical composition. Current estimates indicate that only half the ON
in precipitation can be accounted for by specific N compounds while the remaining
50% is still uncharacterized (Cornell et al., 2003). One potentially important class of
N containing organic compounds which have been overlooked are uncharacterized
macromolecular humic like substances. Earlier studies have suggested that they
comprise as much as 50% of the DOC in rainwater (Facchini et al., 1999; Likens
et al., 1983; Willey et al., 2000). One of the goals of the present manuscript is
to elucidate the relative importance of these humic compounds to the dissolved
organic N composition of rainwater. In concert with this, we will also address the
quantitative relationship between the DOC and ON content of rainwater.
In addition to humic substances another important component of the organic
nitrogen pool are amino acids (AA). Their significance to atmospheric waters is
enhanced by their involvement in a variety of processes occurring in the condensed
phase such as complexation with trace metals such as Fe and Cu (Zhang and
Anastasio, 2003). They are also photochemically labile and react readily with ozone
to form products such as ammonium, organic acids and hydroxylamines (Berger
et al., 1999; McGregor and Anastasio, 2001; Milne and Zika, 1993).
Despite the significance of amino acids to the chemistry of atmospheric wa-
ters there is a paucity of data on concentration ranges and patterns of variability.
Gorzelska et al. (1992) measured free amino compounds over a one year period
in central Virginia precipitation. The authors found significant quantities of amino
compounds in rainwater and suggested that these compounds could comprise a sig-
nificant fraction of reduced nitrogen in precipitation, especially in remote regions.
Mopper and Zika (1987) also found significant concentrations of amino acids in
marine rain while Zhang and Anastasio (2003) quantified the relative importance
of amino compounds in fog waters and atmospheric fine particles in northern Cali-
fornia. A second goal of this study is to generate a detailed concentration profile of
amino acids in precipitation and determine what factors control their concentrations
in rainwater. We will also present the first data on the relationship between amino
acids and other N species in precipitation such as ON and ammonium.
Rainwater from tropical events can have significant impacts on the biogeochem-
istry of natural waters near where they make landfall. Avery et al. (2004) reported
the four hurricanes that came ashore near Wilmington, North Carolina during the
late 1990’s deposited between 3 and 9% of the entire annual flux of bioavailable
FACTORS INFLUENCING NITROGEN SPECIATION IN COASTAL RAINWATER 83

DOC to Long Bay over a 1 or 2 day period. The authors suggested this high deposi-
tion of bioavailable DOC to the nearby coastal waters had significant ramifications
with respect to the carbon budget because it likely spurred short term secondary
productivity following the hurricanes. Because of the potential influence on N bud-
gets as well as carbon, a final opportunistic goal of the present manuscript is to
determine what impact extreme weather events such as hurricanes have on N spe-
ciation in rainwater and how important they are in deposition of nitrogen to the
receiving watershed.

2. Methods
2.1. SAMPLE COLLECTION

Rainwater samples were collected on an event basis at the University of North

Carolina at Wilmington in a large open area of approximately one hectare located
within a turkey oak, long leaf pine and wire grass community, typical of inland
coastal areas in southeastern North Carolina. The site (34◦ 13.9 N, 77◦ 52.7 W) is
approximately 8.5 km from the Atlantic Ocean. Due to the close proximity of
the collection site to the laboratory, analyses could be initiated within minutes of
collection, which reduced the possibility of compositional changes between the
time of collection and analysis.
Four Aerochem-Metrics (ACM) Model 301 Automatic Sensing Wet/Dry Pre-
cipitation Collectors were used to collect event rain samples. One of these ACM
collectors contained a 4 L muffled Pyrex glass beaker from which samples for dis-
solved organic carbon were collected and analyzed. The remaining three ACM
collectors consisted of a high density polyethylene (HDPE) funnel connected by
Tygon
R
FEP-lined tubing to a 2.2 L Teflon bottle and these were used to collect the
samples for the remaining data. Real time precipitation maps were used to define
the end of specific rain events, which initiated the sample collection process. Rain
samples were generally collected within 1 h of the end of an event. In the case of
overnight rain events, samples were collected before 8:00 AM the following morn-
ing. Upon collection, rainwater was immediately analyzed for pH and hydrogen
peroxide. Samples for inorganic anions, and N analysis were transferred to clean
125 mL HDPE bottles on site and stored frozen until analysis. Nearly all samples
were analyzed for inorganic nitrogen and amino acids (NO− +
3 , NH4 , AA) within one
week of collection.

2.2. NH+
4

NH+ 4 in all rainwater samples was determined using a modified version of a flu-
orometric method (Holmes et al., 1999). The method is based on the reaction
of o-phthaldialdehyde (OPA) with ammonia forming a fluorescent chromaphore.
OPA in the presence of sulfite blocks the fluorescent signal of free amino acids.
84 R. J. KIEBER ET AL.

Fluorescence was measured on a Turner Designs Model 450 Fluorometer (Turner

Designs, Sunnyvale CA, USA) with 360 nm excitation and 440 nm emission filters.
1.5 mL sample aliquots were pipetted into six 8 mL Nalgene HDPE sample vials
to which 5 mL of a working reagent containing the OPA was added and allowed to
incubate for 3 to 8 h in the dark after which the fluorescence was recorded. NH+4
was quantified in samples by the method of standard additions to eliminate matrix
effects. The detection limit defined as three times the standard deviation of the
blank is 0.17 µM with a precision of 8% RSD.

2.3. FREE AMINO ACIDS

Free amino acid concentrations were determined by a modified fluorometric tech-

nique where OPA, in the presence of 2-mercaptoethanol, reacts with primary amines
to form a fluorescent chromaphore with free amino acids rather than ammonia
(Parsons et al., 1984). Fluorescence of the derivatized amino acids was measured
on a Turner Designs Model 450 Fluorometer (Turner Designs, Sunnyvale CA, USA)
with 360 nm excitation and 440 nm emission filters. For AA analysis, 2 mL sample
aliquots and 2 mL of working reagent are pipetted into 8 mL Nalgene HDPE sample
vials. The fluorescence was recorded after a minimum 4 min incubation period for
the sample and standard addition vials. Free amino acids were again quantified in
samples by the method of standard additions to eliminate matrix effects. The de-
tection limit defined as three times the standard deviation of the blank is 0.13 µM
with a precision of 3% RSD.

2.4. INORGANIC ANIONS AND pH

Anions (chloride, nitrate, nitrite and sulfate) were measured using suppressed ion
chromatography using synthetic rain as a standard. The detection limit for ni-
trate was 0.1 µM with a relative standard deviation of 3%. Nitrite concentrations
were always less than the detection limit of 0.5 µM and so were a small part of
the total nitrogen in this study. Rainwater pH was determined using a Ross electrode
calibrated with low ionic strength 4.10 and 6.97 buffers. Ionic strength adjuster
(pHix from Orion Research Incorporated, Boston, Massachusetts) was added to
each sample to match the ionic strength of samples to that of the buffers.

2.5. TOTAL NITROGEN AND ORGANIC NITROGEN

Total nitrogen (TN) was analyzed by high temperature catalytic oxidation (HTCO)
(Alvarez-Salgado and Miller, 1998). Measurements were performed using a com-
mercial Shimadzu TOC-5050A coupled to an Antek 9000N nitrogen-specific
chemiluminescence detector (Antek Instruments, TX). Samples were injected
(50 µL injection volume) from pre-treated and capped glass cuvettes into the
Shimadzu TOC-5050A furnace, filled with a preconditioned Shimadzu catalyst
FACTORS INFLUENCING NITROGEN SPECIATION IN COASTAL RAINWATER 85

(Al2 O3 impregnated with 0.5% platinum), at 680 ◦ C. The combustion gases from
the TOC-5050A furnace were routed to the Antek 9000N detector by pulling with
a Vacuubrand MZ 2D diaphragm vacuum pump (ABM, Germany) at the exit of
the Antek permeation-tube drier to lower the pressure within the NOx /O3 reaction
chamber. A high precision controller (Orme Scientific, Manchester, UK) was used
to keep a constant vacuum which is critical in maintaining the constant flow through
the reaction cell necessary to perform precise measurements. The flow through the
Antek 9000N was set to ∼75% of the total flow (=150 mL min−1 ) by means of
an extra-fine Nupro (swage-lock) metering valve (Bristol Valve and Fitting, Bristol
UK). The detection limit defined as three times the standard deviation of the blank
was 0.51 µM with a precision of 2% RSD. A Hansel Laboratory Deep Seawater
Reference (Lot # 06–00, Bermuda Biological Station for Research Inc.) was mea-
sured to confirm the accuracy of the analysis. Total nitrogen in the reference was
21.3 ± 0.24 µ M N (n = 10) which was within the accepted values ranging from
20.5–21.5 µM N.
Organic nitrogen (ON) was determined as the difference between total nitro-
gen and inorganic nitrogen where inorganic nitrogen is the sum NO− −
3 , NO2 and
+ +
NH4 · NH4 was reanalyzed in all samples prior to determination of ON to correct
for any losses due to frozen sample storage. Loss of ammonium during storage
causes ON values to be low or negative which was observed for samples with
high initial ammonium concentrations. The importance of measuring NH+ 4 imme-
diately prior to TN analysis was demonstrated by following changes in ammonium
concentration in six frozen rain samples as a function of storage time. The rain
events were divided into two identical samples and stored in clean 125 mL HDPE
bottles. Initial, unfrozen NH+4 concentrations were measured from samples in one
bottle while the other was frozen at −20 ◦ C. Initial NH+ 4 concentrations ranged
from 0.64–57.6 µM. Subsequent measurements were taken from the frozen sam-
ples after periods ranging from 221–297 days. There was a positive correlation
between NH+ 4 loss rate and initial sample concentration, with losses being largest
for samples with highest initial concentrations. According to the data in Figure 1,
a rain sample with 50 µM NH+ 4 initially would decrease by approximately 9 µM
after six months of frozen storage. The detection limit for ON was 0.6 µM based on
three times the standard deviation of the blank, and the relative standard deviation
calculated by propagation of errors in the variances for TN, NH+ −
4 and NO3 was
18% for a sample with average rainwater concentrations of the N analytes.

2.6. H2 O 2 AND DOC

Hydrogen peroxide was measured on rain within four hours of collection using
a scopoletin induced fluorescence decay technique (Kieber et al., 2001; Kieber
and Helz, 1986). Dissolved organic carbon was determined by high temperature
combustion (HTC) using a Shimadzu TOC 5000 total organic carbon analyzer
equipped with an ASI 5000 autosampler (Shimadzu, Kyoto, Japan). Standards
86 R. J. KIEBER ET AL.

0.08
NH4+ Loss (µM day-1)

y = 0.0011x - 0.0079
0.06 R = 0.989
n = 6, p < 0.001

0.04

0.02

0
0 20 40 60
-0.02
Initial NH4+ (µM)

Figure 1. Plot of NH+ −1

4 loss (µM day ) versus initial sample concentration (µM) in samples
stored frozen in separate bottles between 221 and 297 days. Error bars represent the standard
deviation based on triplicate analyses.

were prepared from reagent grade potassium hydrogen phthalate (KHP) in Milli-Q
Plus Ultra Pure Water. Samples and standards were acidified to pH 2 with 2 M HCl
and sparged with carbon dioxide free carrier gas for 5 min at a flow rate of 125 ml
min−1 to remove inorganic carbon prior to injection onto a heated catalyst bed
(0.5% Pt on alumina support, 680 ◦ C, regular sensitivity). A nondispersive infrared
detector measured carbon dioxide gas from the combusted carbon. Each sample
was injected 4 times. The relative standard deviation was ≤3%. A previous study
demonstrated that there was no difference between DOC in filtered (0.7 µm glass
fiber or 0.4 µm Teflon pore size ) versus unfiltered rain samples at this location
(Willey et al., 2000) therefore it is also very likely that the ON measured in this
study is primarily dissolved.

2.7. AIR MASS BACK - TRAJECTORY CLASSIFICATION

Precipitation events were categorized using air-mass back-trajectories generated

using version 4 of the Hybrid Single Particle Lagrangian Integrated Trajec-
tory Model (HYSPLIT) developed at the National Oceanic Atmospheric Ad-
ministration – Air Resources Laboratory (NOAA/ARL - http://www.arl.noaa.gov/
ready/hysplit4.html). Trajectories were generated using a stand-alone PC version of
the model; and calculated using pre-processed gridded horizontal and vertical wind
fields generated at 6-hour intervals from the National Center for Environmental
Prediction’s Global Data Assimilation System (GDAS) using the Medium Range
Forecast model (MRF) to produce the forecast wind fields.
Single back-trajectories were run for each measured precipitation event col-
lected at UNCW starting at the recorded onset of precipitation. Trajectories were
generated for a 72 h hind-cast since this is a mid-range value for the residence time
of NH+ 4 in the atmosphere (Walker et al., 2000). Trajectories were run starting
FACTORS INFLUENCING NITROGEN SPECIATION IN COASTAL RAINWATER 87

at the 500 m level to represent the air-mass near the well mixed boundary layer
likely to contribute more heavily to in-cloud processes contributing to wet depo-
sition (Walker et al., 2000). They were then visually categorized based on origin
(compass direction) and path (terrestrial, oceanic, or mixed): (1) N-Mixed, (2)
W/SW-Terrestrial, (3) N/NW-Mixed, (4) SW-Coastal, (5) E-Oceanic. Terrestrial
air masses are those whose pathway for the 72 h period was predominantly over a
landmass, and like-wise over the ocean for oceanic types. Mixed trajectories were
determined to have the same potential for oceanic as terrestrial influence based on
a visual analysis of their pathway. Trajectories unclassifiable by either origin or
pathway were not included in the dataset used for this analysis.

3. Results
Rainwater was collected from 129 rain events from February 2002 through Au-
gust 2003 and analyzed for various nitrogen species and pH (Table I). A subset
of samples from September 1, 2002 to August 31, 2003 (78 rain events) was used
to represent an average calendar year for comparison purposes. The sample year
had 163.4 cm of rain (21.0 mm per event) which is near the 147 cm per year av-
erage at the Wilmington International Airport (ILM) based on data from the last
10 years collected by the National Weather Service. This indicates that the average
year presented in Table I was not abnormally wet or dry. Table II contains the vol-
ume weighted average concentrations of other relevant analytes (SO2− 4 , non-seasalt
sulfate, hydrogen peroxide and DOC) in the rainwater samples analyzed. These
parameters are all at or near concentrations reported earlier at this location (Kieber
et al., 2002; Willey et al., 2000) suggesting the compositional characteristics of the
rainwater were near normal for the sampling period.
Inorganic nitrogen (NO− +
3 plus NH4 ) was the dominant form of N representing
85% of total nitrogen based on the yearly volume-weighted averages (Table I).
Rainwater ammonium concentrations ranged from below detection (<0.17 µM)
to 77.2 µM (median: 4.42 µM) while NO− 3 ranged from 0.3–55.8 µM (median:
11.2 µM). The yearly average volume weighted concentrations of NO− 3 and NH4
+
−
were very similar (11.1 µM, and 10.1 µM respectively) corresponding to a NO3 to
NH+ 4 ratio of 1.1:1 which is lower than an earlier study at this location where the
ratio was 1.4:1.0 (Willey and Kiefer, 1993). The decrease in the ratio of NO− 3 to
NH+ 4 is most likely the result of a 12% increase in ammonium concentrations and
13% decrease in nitrate concentrations since the earlier study at this location.
The volume weighted average organic nitrogen (ON) concentration during this
study was 4.6 µM with a range from below detection (<0.6 µM) to 83.3 µM (me-
dian: 2.54) (Table I). ON concentrations were approximately half the NO− 3 or NH4
+

values and constituted approximately 20% of the total dissolved N signal observed
during this study. Both ON and total nitrogen (TN) concentrations are comparable
to values at a similar site in Morehead City on the North Carolina coast during a
1994–1996 study where concentrations were 3.6 and 17.4 µM respectively (Peierls
 88
Table I. Volume-weighted average concentration and standard deviations of N containing species in rainwater collected at Wilmington, NC between
February 2002 and August 2003

n Rain (mm) pH NH+

4 (µM) NO−
3 (µM) AA (µM) ON (µM) TN (µM) % NH+
4 % ON

129 All data 2603 4.62 9.8 ± 1.3 11.5 ± 1.1 0.86 ± 0.10 4.6 ± 1.3 22.0 ± 0.4 45 21
78 Avg year 1634 4.72 10.1 ± 1.6 11.1 ± 1.0 0.93 ± 0.10 5.5 ± 1.1 24.8 ± 0.5 41 22
13 Winter 234 4.85 6.8 ± 1.7 15.5 ± 3.7 0.84 ± 0.22 2.4 ± 0.6 24.2 ± 1.2 28 10
21 Spring 567 4.69 18.6 ± 4.5 12.4 ± 2.0 1.37 ± 0.25 7.5 ± 2.3 32.7 ± 0.9 57 23
31 Summer 618 4.65 6.0 ± 0.7 8.9 ± 1.2 0.70 ± 0.09 5.5 ± 0.4 20.2 ± 0.2 30 27
13 Fall 215 4.87 3.0 ± 0.6 7.7 ± 1.1 0.55 ± 0.10 3.3 ± 1.6 14.6 ± 0.4 21 23

The average year includes data from September 1, 2002 through August 31, 2003. The %NH+ 4 and %ON are relative to TN concentration. n is the
number of samples included in their respective calculations. Volume weighted hydrogen ion concentrations converted to pH values for the various
time periods are also given.

R. J. KIEBER ET AL.
FACTORS INFLUENCING NITROGEN SPECIATION IN COASTAL RAINWATER 89

Table II. Volume-weighted average concentrations and standard deviations of SO2− 4 , non-
seasalt sulfate (NSS), H2 O2 , and dissolved organic carbon (DOC) in rainwater collected at
Wilmington NC between September 1, 2002 and August 31, 2003

n SO2−
4 (µM) NSS (µM) H2 O2 (µM) DOC (µM)

78 Avgyear 13.2 ± 1.1 11.1 ± 1.0 13.6 ± 1.1 70.7 ± 10.4

13 Winter 15.2 ± 2.9 12.4 ± 2.8 5.9 ± 1.4 59.8 ± 17.1
21 Spring 15.5 ± 2.1 12.9 ± 1.9 15.9 ± 1.4 84.9 ± 28.8
31 Summer 10.3 ± 1.3 9.0 ± 1.4 15.8 ± 2.0 70.7 ± 11.6
13 Fall 13.0 ± 0.8 10.5 ± 1.6 9.1 ± 2.3 45.5 ± 11.4

n is the number of samples. Non-seasalt sulfate is the sulfate that does not come from seaspray
and is calculated as follows: NSS = total sulfate – 0.052 ∗ chloride concentration for each
individual rain sample.

and Paerl, 1997). The volume weighted average ON concentration reported here is
also similar to the 6.2 µM average (n = 60) reported for a coastal site on the Chesa-
peake Bay (Russell et al., 1998) and the 3.1 µM (n = 83) and 4.2 µM (n = 50)
observed in Charlottesville, Virginia and Newark, Delaware respectively (Keene
et al., 2002). The rainwater ON concentration at the Wilmington location is also
very similar to the 7 µM average reported for marine rains in a recent review of N
deposition by Cornell et al. (2003) taking into account all published modern ON
data.
The concentration of free amino acids (AA) ranged from below detection to
4.56 µM (median: 0.64). They constituted approximately 17% of the rainwater ON
and are approximately 1% the magnitude of the NH+ 4 concentration. The volume
weighted yearly average concentration of 0.93 µM falls within the reported range
of <0.002–6.4 µM for free amino compounds collected over a one year period
in central Virginia precipitation where the total amine nitrogen contributed from
less than 1 to approximately 10% of ammonium nitrogen level (Gorzelska et al.,
1992). The volume weighted average concentration observed during this study is
also similar to concentrations observed in rain collected near the coast of Florida
of 0.61 µM (Mopper and Zika, 1987).

3.1. CORRELATION ANALYSIS

Correlation analysis was performed using data for N species and other rainwater
analytes in order to evaluate patterns of variation among rainwater components
(Table III). A weak negative correlation ( p < 0.05) was observed between rain
amount, NO− 2−
3 , and SO4 indicating these species, which tend to be associated with
aerosols, are washed out of the atmosphere by rainfall. In contrast, NH+
4 , AA and ON
concentrations did not correlate with rain amount suggesting these nitrogen species
90 R. J. KIEBER ET AL.

Table III. Correlation matrix of Pearson correlation coefficients for rainwater components

H+ NO−
3 NH+
4 AA ON DOC SO2−
4 Cl− H2 O2

Amount −0.173 −0.342∗ −0.039 −0.128 −0.151 −0.352∗ −0.254∗ −0.026 −0.084
H+ 0.615 0.517 0.617 0.086 0.794 0.626 −0.287∗ 0.350∗
−
NO3 0.651 0.702 0.097 0.718 0.844 0.157 0.206
+
NH4 0.786 −0.081 0.470 0.673 −0.035 0.214
AA 0.156 0.608 0.720 −0.060 0.176
ON 0.302∗ 0.009 −0.062 0.028
DOC 0.724 −0.102 0.244
2−
SO4 0.153 0.198
Cl− −0.125

Bold values indicate significance at p < 0.001; and asterisk ∗ indicates significance at p < 0.05.
The number of samples varies from 48–78; all samples came from the average year in Tables I
and II.

may be generated in situ during rain events. Free amino acids and ammonium
could, for example, be generated by photochemical decomposition of dissolved
humic material present in rainwater (Tarr et al., 2001). Some fraction of the ON
may be produced in situ from reactions of NOx and/or NH+ 4 with hydrocarbons
present in atmospheric waters (Scudlark et al., 1998).
Concentrations of NO− 2− +
3 , SO4 and H were highly correlated in the current
study as in previous studies at this location (Kieber et al., 2004; Kieber et al., 2002;
Willey and Kiefer, 1993) and at other locations in the eastern United States (Gorham
et al., 1984; Hooper and Peters, 1989). A strong correlation between NO− 3 , SO4
2−
+
and H is generally interpreted as reflecting their common anthropogenic sources;
hence they are often used as pollution indicators in rain. There was also a strong
correlation between NH+ − 2− +
4 , amino acids and NO3 , SO4 and H in rainwater while
ON did not correlate with these pollutant indicators.
There was a strong positive correlation between amino acid concentrations ob-
served in rainwater and ammonium (Figure 2). Gorzelska and Galloway (1990)
found a similar positive correlation between total amino acid N and NH+ 4 in aerosol
samples collected from the North Atlantic Ocean (Gorzelska and Galloway, 1990).
The strong positive correlation between free amino acids and NH+ 4 observed here
and in the oceanic aerosol samples measured by Gorzelska and Galloway (1990)
suggests that these N species have similar source and sink functions in rainwater.
This interpretation is consistent with the positive correlation patterns to pollutant
indicators and lack of washout with increasing rain amount (Table III) for both
NH+ 4 and amino acids species.
In contrast to free amino acids and NH+ 4 , ON did not correlate with any rainwater
parameters suggesting that organic nitrogen behaves as an independent variable in
precipitation. In a similar study of ON in precipitation at a coastal site in the mid
FACTORS INFLUENCING NITROGEN SPECIATION IN COASTAL RAINWATER 91

5
y = 0.0509x + 0.4668
R = 0.821
n = 88, p < 0.001

4
Amino acids (µM)

0
0 10 20 30 40 50 60 70 80 90
+
NH4 (µM)
Figure 2. Concentration of amino acids (µM) and ammonium (µM) in rainwater samples
collected between February 2002 and August 2003 near Wilmington, North Carolina.

Atlantic United States, Scudlark et al. (1998) also found that ON concentrations did
not exhibit a significant correlation with anthropogenic constituents (NO− 3 , SO4
2−
+ +
and H ) or NH4 . This suggests that ON concentrations are controlled by a more
complex and probably different set of sources and sinks relative to NH+4 and amino
acids and most likely ON levels are not dominated by anthropogenic inputs at this
location.

3.2. SEASONALITY

All rain events were separated based on sampling date in order to assess potential
seasonal differences in N speciation (Figure 3). Seasons were defined as winter:
January 1–March 31, spring: April 1–June 30, summer: July 1–September 30 and
fall: October 1–December 31. Seasonal averages of NH+ 4 and AA were maximum
( p < 0.05) in the spring while NO− 3 was maximum during the winter season.
+
The springtime maximum of NH4 most likely results from increased agricultural
activity in the surrounding region (Aneja et al., 2001; Walker et al., 2000). These
agricultural sources in springtime could be directly linked to a jump in emissions
associated with the liming of livestock lagoons and crop soils and the application of
fertilizers. In the spring, livestock waste is applied as fertilizer to crop fields while
soil and waste liquid pH is controlled by the addition of basic compounds. These
92 R. J. KIEBER ET AL.

40

Winter
Concentration (µM)

30 Spring
Summer
20
Fall
10

H+ NH4+ NO3- AA ON TN

Figure 3. Volume weighted average concentrations (µM) and standard deviations of the various
N species and H+ determined during different seasons defined as winter: January 1–March 31,
spring: April 1–June 30, summer: July 1–September 30 and fall: October 1–December 31.

common practices increase NH3 volatility from soils, waste slurries and fertilizers
which results in increased gas phase concentrations which are ultimately seen as
higher rainwater concentrations once the NH3 (g) is scavenged into the aqueous
phase.
The AA concentrations were also at a maximum during the spring season similar
to NH+ 4 . In a similar study of water soluble amine compounds in rural precipitation
in central Virginia, the monthly volume weighted average concentration of amino
acids were also highest in samples collected in March, April and May (Gorzelska
et al., 1992). A maximum in free amino compounds in fog waters and atmospheric
fine particles in northern California was also observed during March and April
(Zhang and Anastasio, 2003). The reasons behind this springtime maximum of free
amino acids are unclear primarily because the mechanisms responsible for their
introduction and transformation in the atmosphere are unknown. One possibility
is that amino acid concentrations, like ammonium, are largest in springtime at
this location because of increased agricultural activity in the surrounding region.
Gorzelska et al. (1992) hypothesized that the maximum in AA concentrations in
precipitation may occur because of increased biogenic emissions from plants and
increased degradation of organic litter during the springtime.
The volume weighted average concentration of ON was at a maximum in the
springtime. This springtime peak in ON is similar to what was observed by Keene
et al. (2002) in a study of organic nitrogen in precipitation in eastern North Amer-
ica. In a similar study, Scudlark et al. (1998) also found a springtime maximum in
organic nitrogen at a coastal location (Lewes, DE). The authors suggest this spring-
time maximum is the result of the scavenging and wet deposition of ON associated
with elevated biogenic particulate material such as pollen, spores and small insects
FACTORS INFLUENCING NITROGEN SPECIATION IN COASTAL RAINWATER 93

during springtime (Keene et al., 2002). The peak in ON in the springtime most
likely does not result from the increased AA signal during the same time period
because free amino acids comprise a relatively small percentage of the ON signal.

3.3. AIR MASS BACK - TRAJECTORIES

Precipitation events were also categorized using air-mass back-trajectories in order

to evaluate the influence of storm origin on N speciation in rainwater (Table IV).
The two end members were trajectory 2 storms which were terrestrial air masses
whose pathway for the 72 h period preceding the rain event was predominantly over
a landmass and trajectory 5 where the air mass traveled almost exclusively over the
ocean. Trajectories 1, 3, and 4 were mixed events with both oceanic and terrestrial
influences based on a visual analysis of their pathway.
An analysis of concentrations as a function of air-mass origin shows a maximum
from trajectory 2 for all N analytes except ON. Hydrogen ion concentration also
has a maximum from trajectory 2. In contrast to trajectory 2, the event air masses
of marine origin in trajectory 5 had the lowest concentration of NH+ 4 and were near
the minimum concentration of the remaining analytes with the exception of ON.
The high NH+ 4 concentration observed in trajectory 2 is a direct result of in-
creased continental and/or anthropogenic sources consistent with the correlation
of NH+ 4 with the pollutant indicators in rain presented in Table III. The increased
anthropogenic signal observed in trajectory 2 storms most likely results from the
large number of livestock farms and population centers to the west of the sampling
site. Mixed events from trajectories 1, 3 and 4 had intermediate NH+ 4 concentrations
between the oceanic and continental end members.
Free AA followed a pattern similar to NH+ 4 with highest concentrations in
trajectory 2 storms and near minimum concentrations from trajectory 5 storms.

Table IV. Total rain, rainfall per event, and VWA concentrations of H+ and N analytes in
rainwater at UNCW grouped by trajectory category (Traj)

Total Rain per

rain event NH+
4 NO−
3 AA TN ON H+
Traj Dir Class n (mm) (mm) (µM) (µM) (µM) (µM) (µM) (µM)

1 N M 8 232 29 5.5 4.7∗ 0.49 16.1 4.3 9.9

2 W/ SW T 15 257 17 22.8∗ 16.2∗ 1.7∗ 39.3∗ 5.3 35.6∗
3 N/ NW M 15 305 20 16.0 13.2 1.5 29.5 4.7 15.4
4 SW C 15 294 20 7.0 12.6 0.84 24.6 6.6 27.6
5 E O 23 391 17 4.0∗ 8.2 0.50 18.4 6.1 13.6

The direction of air mass origin is given (Dir) along with the classification (Class) as terrestrial
(T) mixed (M), coastal (C) or oceanic (O). Statistically significant values ( p < 0.05) based on
a Student’s t-test are marked with an asterisk ∗ .
94 R. J. KIEBER ET AL.

Amino acids are not typically thought to be components of urban emissions, though
some compounds in the urban and coastal atmosphere are known to be involved in
the breakdown of organic aerosols (Blando and Turpin, 2000; Mopper and Zika,
1987). This mechanism could increase the AA signal as more complex ON forms
are reduced to free amino acids in the region. Air masses of purely oceanic origin
contribute the least to AA in rainfall (trajectory 5).
Interestingly organic nitrogen concentrations did not display any significant
storm origin effect in contrast to NO− +
3 , free AA and NH4 at this location (Table IV).
This is similar to the conclusion drawn by Cornell et al. (2003) who presented a
summary of ON concentrations in rainwater collected from recent literature values
at a variety of locations. The authors reported that the average mean concentrations
of ON at continentally dominated locations were very similar to coastal/island
locations defined as <100 km from the sea. This would suggest that ON has a
different source function than the other N species studied and that there could be
long range transport of some relatively recalcitrant portion of ON between the
continents and the ocean.

3.4. HURRICANE ISABEL

In order to evaluate the potential impacts of tropical events on N speciation, rain-

water from Hurricane Isabel, which made landfall just north of Cape Lookout,
North Carolina on September 18, 2003 as a category-2 hurricane, was analyzed
sequentially for AA, NH+ +
4 and H (Figure 4). The National Weather Service at
Wilmington International Airport (ILM) 16 km from the coast reported maximum
sustained winds of 20 m s−1 (45 mph) during the course of the hurricane. Rainfall

AA H+
NH4+
10 100 4 4

8
75 3 3
Rain (mm)

6
µM

µM

µM

50 2 2
4

25 1 1
2

0 0 0 0
7:00:00 AM 7:00:00 PM 7:00:00 AM 7:00:00 AM 7:00:00 PM 7:00:00 AM 7:00:00 AM 7:00:00 PM 7:00:00 AM

Figure 4. Concentration of NH+ +

4 , AA and H as a function of sampling time during Hurricane
Isabel, which made landfall just north of Cape Lookout, North Carolina on 18 September, 2003
as a category-2 hurricane. Each data point represents a different sequential sample collected
during the five back trajectories listed in Figure 5 with the amount of rain in mm in each sample
as indicated.
FACTORS INFLUENCING NITROGEN SPECIATION IN COASTAL RAINWATER 95

began on the UNCW campus at approximately 1:30AM EST on the 18th of Septem-
ber and lasted until approximately 9:00 PM the same day netting 6.0 cm. Sequen-
tial rainfall samples were taken on the UNCW campus: (A) 8:30AM (17.5 mm),
(B) 11:20AM (11.4 mm), (C) 2:30PM (21.8 mm), (D) 5:40PM (7.1 mm), and (E)
8:30AM (2.3 mm) the following day.
Five NOAA/ARL HYSPLIT back-trajectories (A-E) were calculated with the
end-point time set to that when each sequential sample was collected (Figure 5).
The majority of a hurricane’s moisture comes from seawater reflected in the low
initial H+ concentrations in the first three data points in Figure 4 representing
trajectories A, B, and C. The increase in H+ observed later in the storm results
from a transition from marine to continental air masses where trajectories D and E
are essentially all terrestrial.
The peak in NH+ 4 concentration and deposition corresponding to trajectory
C is consistent with this air mass traversing major NH+ 4 agricultural sources in

Figure 5. Plots of back trajectories calculated using the HYSPLIT model and interpolated
rainfall intensity field for hurricane Isabel sequential samples. The end point was set to the
time when each sequential sample was collected. Rainfall intensity (mm/h) is calculated from
values given at each trajectory point.
96 R. J. KIEBER ET AL.

conjunction with high relative wind intensities which aid NH+ 4 transport (Figures 4
and 5). The concentration in trajectory C remains high in spite of the heavy rain
volume that would tend to dilute the NH+ 4 concentration. The sequential profile
suggests a complete washout of NH+ 4 from the atmosphere by the end of the storm
even though trajectory E traversed over primarily continental locations. The reasons
for the AA maximum in trajectory C are not as clear as for NH+ 4 since sources of AA
to the atmosphere are not as well defined. The coincidence of the AA maximum
near the NH+ 4 maximum during Hurricane Isabel suggests similar sources. This
is consistent with the earlier conclusion based on the strong positive correlation
between NH+ 4 and AA for rainwater samples collected at the Wilmington site pre-
sented in Figure 2 and Table III.
The impact of Hurricane Isabel on nitrogen loading in the nearby Cape Fear
River system was also evaluated. Approximately three times as much total NH+ 4
and AA (317 µmoles · m−2 and 84 µmoles · m−2 respectively) were deposited to the
watershed compared to the mean impact of average summertime rain events. This
large episodic addition of NH+ 4 and AA could significantly impact the short term
bioavailable N budget in the estuary similar to what was suggested for labile carbon
loading in receiving watersheds resulting from tropical systems (Avery et al., 2004).

4. Discussion
One of the most important uncertainties regarding N speciation in atmospheric
waters is the chemical composition of the organic nitrogen pool. The results pre-
sented here and in other studies suggest that amino acids comprise a relatively small
fraction of the nitrogen in ON (Sidle, 1967; Willey et al., 2000). Other potentially
significant ON components which have been measured in precipitation include or-
ganic nitrates as well as other soluble reduced-N forms (urea and amines) although
there are relatively few measurements of individual N compounds and ON in the
same precipitation samples (Cornell et al., 2003 and references therein.)
One potentially important class of N containing organic compounds which have
been overlooked are uncharacterized macromolecular humic like substances in
rainwater. Preliminary data indicates they may comprise as much as 50% of the
DOC in rainwater at this location (Willey et al., 2000). This estimate is in relatively
close agreement with other studies of the organic composition of rainwater. Facchini
et al. (1999) suggested that humics comprise 40% of the DOC in fog in the Po
valley while Likens et al. (1983) estimated that 50–60% of the DOC in rainwater
is uncharacterized and associated with macromolecules such as humics.
There is a significant positive correlation (n = 51, p < 0.02) between the ON
and DOC concentrations in rainwater at this location, although there is scatter in
the data resulting from the somewhat variable input of DOC which does not contain
N (i.e. formic and acetic acids) (Figure 6). When the DOC/ON ratio is calculated
for individual rain events and averaged, the resulting C:N ratio is approximately
FACTORS INFLUENCING NITROGEN SPECIATION IN COASTAL RAINWATER 97

40

30
ON (µΜ)

20

10

0
0 100 200 300 400 500
DOC (µM)
Figure 6. Concentration (µM) of organic nitrogen (ON) and dissolved organic carbon (DOC)
in rainwater samples collected between February 2002 and August 2003 near Wilmington,
North Carolina.

20:1 with a range of 1:1 to 45:1. The average agrees well with the 25:1 C:N ratio
determined for C18 extracted humic material in a summer rainwater sample at the
same location (Reid, 2003). The C:N ratio determined for rainwater samples at
this location is also close to the value reported for humic acid in soils of about
20:1 (Schwarzenbach et al., 1993) which suggests that a possible source of the
atmospheric humic substances are wind blown soil material.
The C:N ratio can be used, in conjunction with the relative abundance of humic
material, to estimate the contribution of humic N to the ON pool in rainwater in
the following manner. The summer rainwater sample analyzed by Reid (2003) had
a DOC concentration of 86 µM and an ON concentration of 1.9 µM. Assuming
humics comprise 40% of the DOC and utilizing the C:N ratio of the humic material
determined by Reid of 25:1 yields a humic contribution 1.4 µM N to the ON pool in
this rain sample. The total free amino acids measured in this rain event contributed
0.7 µM N. Addition of the amino acid and humic N contributions therefore accounts
for essentially all the organic nitrogen present in this rain event.
This calculation should be viewed with caution as the exact humic contribution
to rainwater DOC and its C:N ratio is not well constrained and is currently an
area of active research in our laboratory. This calculation is, however, very useful
because it demonstrates that macromolecular uncharacterized humic like substances
comprise a significant, and in some cases the majority, of nitrogen species in the
organic nitrogen pool in atmospheric waters. This has important ramifications with
respect to the long range transport of atmospheric ON because humic materials
are recalcitrant and therefore may travel long distances from their source. The
98 R. J. KIEBER ET AL.

long range transport of ON may help explain why there are no major differences
between ON concentrations observed in marine and continental dominated rain
events, and why the ON concentrations do not correlate with concentrations of any
other analytes.

Acknowledgments
This work was supported by NSF Grants ATM 0342420 and OCE 0326685 and
the North Carolina Sea Grant program. The Marine and Atmospheric Chemistry
Research Laboratory group at UNC-Wilmington and Rob Whitehead assisted with
sampling and analyses.

References
Alvarez-Salgado, X. A. and Miller, A. E. J., 1998: Simultaneous determination of dissolved organic
carbon and total dissolved nitrogen in seawater by high temperature catalytic oxidation: Conditions
for precise shipboard measurements, Mar. Chem. 62, 325–333.
Aneja, V. P., Bunton, B., Walker, J. T., and Malik, B. P., 2001: Measurement and analysis of atmospheric
ammonia emissions from anaerobic lagoons, Atmos. Environ. 35, 1949–1958.
Avery, G. B., Kieber, R. J., Willey, J. D., Shank, G. C., and Whitehead, R. F., 2004: Impact of
hurricanes on the flux of rainwater and Cape Fear river water dissolved organic carbon to Long
Bay, southeastern United States, Global Biogeochem. Cycles 18, 3015–3020.
Berger, P., Leitner, N. K., Dore, M., and Legube, B., 1999: Ozone and hydroxyl radicals induced
oxidation of glycine, Water Res. 33, 433–441.
Blando, J. D. and Turpin, B. J., 2000: Secondary organic aerosol formation in cloud and fog droplets:
A literature evaluation of plausibility, Atmos. Environ. 34, 1623–1632.
Cornell, S. E., Jickells, T. D., Cape, J. N., Rowland, A. P., and Duce, R. A., 2003: Organic nitrogen
deposition on land and coastal environments: A review of methods and data, Atmos. Environ. 37,
2173–2191.
Facchini, M. C., Fuzzi, S., Zappoli, S., Andracchio, A., Gelenscer, A., Kiss, G., Krivancsy, Z.,
Meszaros, E., Hannsson, H. C., Alsberg, T., and Zebuhr, Y., 1999: Partitioning of the organic
aerosol component between fog droplets and interstitial air, J. Geophys. Res. 104, 26821–
26832.
Gorham, E., Martin, F. B., and Litzau, J. T., 1984: Acid rain: Ionic correlations in the eastern United
States 1980–1981, Sci. 225, 407–409.
Gorzelska, K. and Galloway, J. N., 1990: Amine nitrogen in the atmospheric environment over the
north Atlantic Ocean, Global Biogeochem. Cycles 4, 309–333.
Gorzelska, K., Galloway, J. N., Watterson, K., and Keene, W. C., 1992: Water soluble primary amine
compounds in rural continental precipitation, Atmos. Environ. 26, 1005–1018.
Holmes, R. M., Aminot, A., Kerouel, R., Hooker, B. A., and Peterson, B. J., 1999: A simple and
precise method for measuring ammonium in marine and freshwater ecosystems, Can. J. Fish. Aq.
Sci. 56, 1801–1808.
Hooper, R. P. and Peters, N. E., 1989: Use of multivariate analysis for determining sources of so-
lutes found in wet atmospheric deposition in the United States, Environ. Sci. Technol. 23, 1263–
1268.
Keene, W. C., Montag, J. A., Maben, J. R., Southwell, M., Leonard, J., Church, T. M., Moody, J. L.,
and Galloway, J. N., 2002: Organic nitrogen in precipitation over eastern North America, Atmos.
Environ. 36, 4529–4540.
FACTORS INFLUENCING NITROGEN SPECIATION IN COASTAL RAINWATER 99

Kieber, R. J., Cooper, W. J., Willey, J. D., and Avery, G. B., 2001: Hydrogen peroxide at the Bermuda
Atlantic Time Series Station. Part 1: Temporal variability of atmospheric hydrogen peroxide and
its influence on seawater concentrations, J. Atmos. Chem. 39, 1–13.
Kieber, R. J. and Helz, R. G., 1986: Two method verification of hydrogen peroxide determinations in
natural waters, Anal. Chem. 58, 2312–2315.
Kieber, R. J., Skrabal, S. A., Smith, C., and Willey, J. D., 2004: Redox speciation of copper in rainwater:
Temporal variability and atmospheric deposition, Environ. Sci. Technol. 38, 3587–3594.
Kieber, R. J., Willey, J. D., and Zvalaren, S. D., 2002: Chromium speciation in rainwater: Temporal
variability and atmospheric deposition, Environ. Sci. Technol. 36, 5321–5327.
Likens, G. E., Edgerton, R. S. and Galloway, J. N., 1983: The composition and deposition of organic
carbon in precipitation, Tellus 35B, 16–24.
McGregor, K. G. and Anastasio, C., 2001: Chemistry of fog waters in California’s Central valley: 2
Photochemical transformations of amino acids and alkyl amines, Atmos. Environ. 35, 1091–1104.
Milne, P. J. and Zika, R. G., 1993: Amino acid nitrogen in atmospheric aerosols-occurence, sources
and photochemical modification, J. Atmos. Chem. 16, 361–398.
Mopper, K. and Zika, R. G., 1987: Free amino acids in marine rains: Evidence for oxidation and
potential role in nitrogen cycling, Nature 325, 246–249.
Paerl, H. W., Fogel, M. L., Bates, P. W., and O’Donnell, P. M., 1994: In K. R. Dyer and R. J. Orth
(eds.), Changes in Fluxes to Estuaries: Implications from Science Management, Olsen and Olsen,
Paerl, H. W. and Whitehall, D. R., 1999: Anthropogenically derived atmospheric nitrogen deposition,
marine eutrophication and harmful algal bloom expansion: Is there a link?, Ambio 28, 307–311.
Parsons, T., Maita, Y., and Lalli, C., 1984: In A manual of Chemical and Biological Methods for
Seawater Analysis, Pergamon Press, Oxford.
Peierls, B. L. and Paerl, H. W., 1997: Bioavailability of atmospheric organic nitrogen deposition to
coastal phytoplankton, Limnol. Oceanogr. 42, 1819–1823.
Reid, S., 2003: Master’s Thesis, Chromophoric dissolved organic matter in coastal rainwater. Univer-
sity of North Carolina at Wilmington.
Russell, K. M., Galloway, J. N., Macko, S. A., Moody, J. L., and Scudlark, J. R., 1998: Sources of
nitrogen in wet deposition to the Chesapeake Bay region, Atmos. Environ. 32, 2453–2465.
Schwarzenbach, R. P., Gschwend, P. M., and Imboden, D. M., 1993: In Environmental Organic
Chemistry, John Wiley & Sons, New York.
Scudlark, J. R., Russell, K. M., Galloway, J. N., Church, T. M., and Keene, W. C., 1998: Organic
nitrogen in precipitation at the mid Atlantic U.S. coast – methods evaluation and preliminary
measurements, Atmos. Environ. 10, 1719–1728.
Sidle, A. B., 1967: Amino acid content of atmospheric precipitation, Tellus 19, 128–135.
Tarr, M. A., Wang, W. W., Bianchi, T. S., and Engelhaupt, E., 2001: Mechanisms of ammonia and
amino acid photoproduction from aquatic humic and colloidal matter, Water Res. 35, 3688–3696.
Walker, J. T., Aneja, V. P. and Dickey, D. A., 2000: Atmospheric transport and deposition of ammonium
in North Carolina, Atmos. Environ. 34, 3407–3418.
Warneck, P., 1988: In Chemistry of the Natural Atmosphere, Academic Press, New York.
Willey, J. D., Kieber, R. J., Eyman, M. S., and Avery, G. B., 2000: Rainwater dissolved organic carbon:
Concentrations and global flux, Global Biogeochem. Cycles 14, 139–148.
Willey, J. D. and Kiefer, R. H., 1993: Atmospheric deposition in southeastern North Carolina: Com-
position and quantity, J. Elisha Mitchell Sci. Soc. 109, 1–19.
Zhang, Q. and Anastasio, C., 2003: Free and combined amino compounds in atmospheric fine particles
(PM2.5) and fog waters from Northern California, Atmos. Environ. 37, 2247–2258.