Article
Evidence for the importance of atmospheric nitrogen deposition to eutrophic Lake Dianchi, China Xiaoying Zhan, Yan Bo, Feng Zhou, Xuejun Liu, Hans W. Paerl, Jianlin Shen, Rong Wang, Farong Li, Shu Tao, Yanjun Dong, and Xiaoyan Tang Environ. Sci. Technol., Just Accepted Manuscript • Publication Date (Web): 01 Jun 2017 Downloaded from http://pubs.acs.org on June 2, 2017
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 37
Environmental Science & Technology
1
Evidence for the importance of atmospheric nitrogen
2
deposition to eutrophic Lake Dianchi, China
3 4
Xiaoying Zhan#†, Yan Bo#†, Feng Zhou*†, Xuejun Liu‡, Hans W. Paerl§|, Jianlin Shenǁ,
5
Rong Wang┴, Farong Li , Shu Tao†, Yanjun Dong≡, Xiaoyan Tang+ ∥
6 7
†
8
Processes, College of Urban and Environmental Sciences, Peking University, Beijing,
9
100871, P.R. China
Sino-France Institute of Earth Systems Science, Laboratory for Earth Surface
10
‡
11
Beijing, 100193, P.R. China
12
§
13
City, North Carolina, 28557, USA
14
ǁ
15
Subtropical Agriculture, Chinese Academy of Sciences, Changsha, Hunan, 410125,
16
P.R. China
17
┴
18
94305 USA.
19
College of Resources and Environmental Sciences, China Agricultural University,
Institute of Marine Sciences, University of North Carolina at Chapel Hill, Morehead
Key Laboratory of Agro-Ecological Processes in Subtropical Regions, Institute of
Department of Global Ecology, Carnegie Institution for Science, Stanford, California,
∥
Kunming Environmental Monitoring Center, Kunming, Yunnan, 650028, P.R. China
20
+
21
100871, P.R. China
22
≡
23
Pearl River Hydraulic Research Institute, Guangzhou 510611, P.R. China
24
|
College of Environmental Sciences and Engineering, Peking University, Beijing,
Laboratory of the Pear River Estuarine Dynamics and Associated Process Regulation,
College of Environment, Hohai University, Nanjing, 210098, P.R. China
25 26
* Corresponding author: Phone / fax: +86 10 62756511; Email:
[email protected] 27
#
X.Y.Z. and H.Y. contributed equally to this work.
28
ACS Paragon Plus Environment
Environmental Science & Technology
29
TOC Art
30
ACS Paragon Plus Environment
Page 2 of 37
Page 3 of 37
Environmental Science & Technology
31
Abstract
32
Elevated atmospheric nitrogen (N) deposition has significantly influenced aquatic
33
ecosystems, especially with regard to their N budgets and phytoplankton growth
34
potentials. Compared to a considerable number of studies on oligotrophic lakes and
35
oceanic waters, little evidence for the importance of N deposition has been generated
36
for eutrophic lakes, even though emphasis has been placed on reducing external N
37
inputs to control eutrophication in these lakes. Our high-resolution observations of
38
atmospheric depositions and riverine inputs of biologically-reactive N species into
39
eutrophic Lake Dianchi (the 6th largest freshwater lake in China) shed new light onto
40
the contribution of N deposition to total N loads. Annual N deposition accounted for
41
15.7% to 16.6% of total N loads under variable precipitation conditions, twofold
42
higher than previous estimates (7.6%) for the Lake Dianchi. The proportion of N
43
deposition to total N loads further increased to 27~48% in May and June when toxic
44
blooms of the ubiquitous non-N2 fixing cyanobacteria Microcystis spp. are initiated
45
and proliferate. Our observations reveal that reduced N (59%) contributes a greater
46
amount than oxidized N to total N deposition, reaching 56-83% from late spring to
47
summer. Progress toward mitigating eutrophication in Lake Dianchi and other
48
bloom-impacted eutrophic lakes will be difficult without reductions in ammonia
49
emissions and subsequent N deposition.
ACS Paragon Plus Environment
Environmental Science & Technology
50
INTRODUCTION
51
Human activities have significantly increased emissions of reactive nitrogen (N,
52
including reduced and oxidized forms) to the atmosphere and their resultant
53
deposition1, 2. Results of global atmospheric chemical transport models indicated that
54
atmospheric deposition of N has increased by approximately three-fold since
55
pre-industrial times, especially over the East and South Asia3. The ecological effects
56
of elevated atmospheric N deposition have been observed or modeled for oligotrophic
57
lakes and oceanic waters, showing that such N inputs can stimulate phytoplankton
58
growth4-7. This also leads to an increase in the stoichiometric N/phosphorus (P) ratio,
59
which can induce a shift from N-limited to P-limited phytoplankton3, 4. In contrast, the
60
contribution made by atmospherically deposited N to the total N input for eutrophic
61
lakes has rarely received attention8, 9, because it was commonly considered to be
62
much less than N inputs from watersheds to these lakes10.
63 64
However, recent evidence from the few field observations demonstrated that the
65
contribution of N deposition cannot be neglected compared to riverine inputs for
66
eutrophic lakes. For instance, Luo et al9. measured wet N deposition in mesotrophic
67
Lake Taihu, the third largest freshwater lake in China, indicating that atmospheric
68
inputs of N accounted for approximately 14% of total N inputs. A similar percentage
69
was reported for a mesotrophic lake in Venezuela11 (Lake Maracaibo, wet deposition,
70
19%). Furthermore, the percentage increased considerably when both dry and wet
71
deposition (i.e., gaseous, particulate and dissolved N) were included12. With long-term
72
controls of point and non-point sources from watersheds taking place in developed
73
countries, atmospheric N deposition is becoming an increasingly dominant N source
74
to these waters (e.g., 15-42% of the total N inputs13 in the northeastern and
ACS Paragon Plus Environment
Page 4 of 37
Page 5 of 37
Environmental Science & Technology
75
mid-Atlantic regions of the United States).
76 77
However, the contribution made by atmospheric N deposition to total inputs remains
78
challenging to accurately assess, especially at a high-resolution temporal scale.
79
Multiple factors are associated with the difficulty in assessment, especially when
80
quantifying dry deposition and N inputs from watersheds to receiving waters. First,
81
gaseous N (i.e., ammonia [NH3], nitrogen dioxide [NO2], nitric acid and nitrous acid
82
[HNO2/HNO3]) deposition have generally been ignored or are only partially
83
considered by current observation networks globally (e.g., EANET14, CASTNET15,
84
AMoN15, IMPROVE NHx15, EMEP16) and in previous regional assessments12.
85
Likewise, wet or bulk deposition of inorganic N (NH4+-N+NO3−-N) has been
86
measured systematically in China, but organic N was seldom analyzed12. Those
87
omissions may result in a considerable underestimation of direct atmospheric N
88
deposition to these waters. Second, most measurements applying passive samplers
89
yield ambient exposures on a monthly interval, mainly due to low concentrations of
90
gaseous or particulate N12. Although active samplers (e.g., denuder systems) can be
91
deployed to determine hourly or daily concentrations of gaseous constituents and
92
fine-fraction particulates17, they are unsuitable for long-term field observations
93
without a power supply12. In addition, riverine N inputs (i.e., N concentration × river
94
discharge) were generally determined by using process-based models that simulate the
95
N available for transport to the waters from different watershed N sources (runoff
96
from agriculture, urban areas and upland forests, point sources)13,
97
limitation in such models is the large uncertainties arising from model structure and
98
parameter choices20. Such uncertainties can be reduced by performing observations of
99
river discharge and N concentrations at high spatial and temporal resolutions.
ACS Paragon Plus Environment
18, 19
. A key
Environmental Science & Technology
100 101
The primary objective of this study is to test the hypothesis that atmospheric N
102
deposition makes an important contribution to total N loads for eutrophic Lake
103
Dianchi, the sixth largest freshwater lake in China, through high-resolution (daily or
104
biweekly) systematic observations of N deposition and inputs. Dry and wet deposition
105
fluxes of all N species and riverine N inputs were quantified and their annual and
106
biweekly contributions to total N loads were determined. In addition, we considered
107
the reliability of the fluxes as well as the importance of N deposition on lake
108
eutrophication, and implications for lake ecosystem management.
109
MATERIALS AND METHODS
110
Study area. Lake Dianchi is located in Southwest China (24º29′ to 25º28′N,
111
102º29′to 103º01′E). It has a surface area of 310 km2 at the elevation of 1887.5 m, a
112
mean depth of 5.3 m, a shoreline of 150 km, and water retention time of 3.5 years21
113
(Fig. 1). Watershed area for Lake Dianchi is 2920 km2, of which 1088.6 km2 is
114
controlled by 10 large to medium-size reservoirs that have no outflows (Fig. 1). The
115
lake is artificially divided into two segments by an embankment and a dam, where the
116
area of the northern part (Caohai) is ~12 km2 surrounded by urban area of Kunming
117
city and the southern part (Waihai, 298 km2) is bordered by an intensively managed
118
farmland. Land use is primarily forest (47%), cropland (20%), and urban areas (16%;
119
Fig. 1). The total population in the Lake Dianchi Watershed has increased steadily
120
from 1.8 million in 1992 to 4.1 million in 2015. Consequently, the lake has suffered
121
from severe eutrophication, despite the fact that approximately 52 billion RMB (equal
122
to 7.7 billion in USD) has been invested to control lake eutrophication by the Chinese
123
Central Government during 1996-201522.
ACS Paragon Plus Environment
Page 6 of 37
Page 7 of 37
Environmental Science & Technology
124 125
Figure 1. Locations of Lake Dianchi watershed and sampling sites. Nineteen
126
rivers, accounting for 95% of lake’s inflow, were chosen for measuring N
127
concentration and discharge. The rivers numbered 1 to 6 flowed into Caohai, while
128
the rest flowed into Waihai. Hatch marks indicate the area controlled by 10 large to
129
medium-size reservoirs.
130 131
Dry N deposition samples around the lake. Sampling of dry N deposition, including
132
gaseous and particulate N, was conducted from April 1, 2010 to March 31, 2011. Five
133
deposition monitoring stations were evenly located around Lake Dianchi (Fig.1),
134
including Kunming (KM), Baofeng (BF), Chenggong (CG), Jinning (JN), and Haikou
135
(HK). For each station, Ogawa passive samplers (Ogawa & Co., FL, USA) were used
136
to collect ammonia [NH3] and nitrogen dioxide [NO2] samples, which showed no
ACS Paragon Plus Environment
Environmental Science & Technology
137
significant differences from the continuous active samplers used in previous studies23.
138
Another passive sampler (USDA Forest Service, CA, USA) was used for collecting
139
nitrous acid/nitric acid [HNO2/HNO3] samples. Previous comparative experiments
140
showed that this sampler was reliable for determining wide ranges of ambient
141
HNO2/HNO324. Detailed descriptions of these passive samplers can be found in
142
Roadman et al.25 and Bytnerowicz et al.26. High volume aerosol samplers (Laoshan
143
Elec. Inc., Qingdao, China) were applied to collect total suspended particulates (TSP)
144
at a flow rate of 1.05 m3 min−1. Three replicate samplers for dry N deposition were
145
installed at a height of 3.5 m from the ground and at a distance of 2 mm. All collected samples were frozen at −18°C at each site until delivery to
165
the laboratory for analysis. Collectors were cleaned for each event with 1% HCl and
166
Nanopure water.
167 168
River discharge and water samples. Although river discharges data were not
169
available except for Panlongjiang (No.8) and Baoxiang (No.10) Rivers, water levels
170
in the other 17 rivers were continuously measured in this study (Fig. 1). These 19
171
rivers contribute more than 95% of lake’s streamflow27. The rating curves were
172
developed by simultaneously measuring water levels and the current velocities for a
173
cross-section of each river following the USGS guideline for the current-meter
174
method28, where the sizes of cross-section (i.e., width, depth, area) were measured for
175
each river (Table S1). To ensure rating curves over a wide range of flow rates,
176
enhanced observations were conducted for three stormwater events (precipitation > 30
177
mm day−1). Each sampling procedure lasted more than 48 h per event. The 10-minute
178
water levels were continuously observed for each tributary by HOBOTM data logger
179
(U20-001-01, Onset Computer Corporation, MA, USA), and the current velocities
180
were measured by the Global WaterTM flow probe (FP201) in each subsection of a
181
channel cross-section. Water level was further converted into daily discharge based on
182
the derived rating curves for each river. In addition, vertical water samplers were used
183
to collect water samples of 19 rivers on biweekly interval. The procedure was
184
referenced to the Technical Specifications Requirements for Monitoring of Surface
185
Water and Waste Water in China (HJ/T91-2002). As all rivers have width of ≤ 50m
186
and depth of ≤ 5m (Table S1), one sample was collected in the river centerline at 0.5
ACS Paragon Plus Environment
Environmental Science & Technology
Page 10 of 37
187
m below or half depth of the river. All water samples were rapidly transported to the
188
KEMC for analysis.
189 190
Chemical analysis. After the exposures, all filters for gaseous and particulate N were
191
placed in polyethylene tubes into which 8 ml and 20 ml distilled/deionized water was
192
added, respectively. The tubes were shaken on a shaker for 15 min at mid-range speed.
193
The extracted solutions of dry deposition, wet deposition samples, and water samples
194
were then filtered with a 0.45 µm syringe filter. N quantities in the filtered solutions
195
were analyzed following the Standard Methods for the Examination of Water and
196
Wastewater29 in China. Ammonium nitrogen [NH4+-N], nitrate nitrogen [NO3−-N],
197
total nitrogen (TN), and Kjeldahl nitrogen (KTN) were measured by Nessler's reagent
198
colorimetric
199
oxidation-ultraviolet spectrophotometric method, as well as Semi-Micro-Kjeldahl
200
method, respectively. Organic N was determined by the difference between KTN and
201
NH4+-N. The detection limits for NH4+-N, NO3−-N, TN, and KTN are 0.025, 0.02,
202
0.05, and 0.025 mg N L−1, respectively. For gaseous N deposition, ambient
203
concentrations of NH3, NO2, and HNO2/HNO3 were calculated as product of their
204
quantities collected in passive samplers and conversion coefficients divided by
205
exposure time (see details at http://ogawausa.com/wp-content/uploads/2014/04/
206
ConcentrationInAmbientAir.pdf and Bytnerowicz et al.26). Additionally, for each
207
batch of field exposed samples for gaseous and particulate N, three blank filters were
208
also prepared to represent extracts from unexposed samplers that are prepared in the
209
laboratory and transported between laboratory and monitoring sites. The blank
210
samples were extracted and analyzed using the same methods as the exposed samples.
211
The details of blank samples can be found in Text S1.
method,
ultraviolet
spectrophotometric
ACS Paragon Plus Environment
method,
K2S2O8
Page 11 of 37
Environmental Science & Technology
212 213
Estimation of deposition fluxes and riverine inputs. Biweekly flux of dry
214
deposition (mg km−2 s−1) was calculated as a product of N concentration and
215
deposition velocity (Vd, m s−1) at each site. The concentrations of gaseous and
216
particulate N were determined as described above. Vd of gaseous N and particulate N
217
over lake surface was estimated by a well-tested deposition velocity model30
218
combined with the on-site hourly meteorological data. One of the national
219
meteorological stations was located at the north end of Lake Dianchi (Fig. 1), where
220
meteorological variables used in the model included air temperature, relative humidity,
221
air pressure, solar radiation, wind speed at 10-m height (v_10), precipitation, and dew
222
point on hourly intervals. Prior to the estimate, v_10 was converted into wind speed at
223
3.5-m height, v_3.5, according to a power law (i.e., v_3.5= v_10×(3.5/10)α, where α
224
was set as 0.1)31. Gas Vd over Lake Dianchi was calculated using the big-leaf
225
resistance analogy model32. In contrast to vegetation, roughness length was set as
226
0.0003 m based on the results for the East China Sea33, and surface resistance was set
227
as 0 s m−1 for NH3 and HNO3 but 20000 s m−1 for NO234. Particulate Vd was
228
parameterized according to Slinn35, where parameters related to surface resistance and
229
gravitational settling velocity were determined according to Zhang et al36. Details of
230
these models were described in Text S2. In addition, the flux of wet deposition was
231
calculated as a product of N concentration and rainfall amount for each rainfall event
232
and monitoring site.
233 234
Daily riverine N inputs were calculated as a function of river discharge, by means of
235
Load Estimator (LOADEST)37:
ACS Paragon Plus Environment
Environmental Science & Technology
ln ( Li ) = a0 + a1 ln Q + a2 ln Q2 + a3 sin ( 2π dtime ) +
236
a4 cos ( 2π dtime ) + a5dtime + a6 dtime2 + ε
Page 12 of 37
(1)
237
where Q is daily river discharge, dtime is decimal time, a0~a6 are the fitted
238
coefficients in the multiple regression model, and ε is estimate error. Note that the
239
biweekly TN concentration was multiplied by the corresponding daily discharge so
240
that equation 1 is a function of load (Li) instead of concentration. Within LOADEST,
241
the model to estimate N loads was set to be automatically selected from predefined
242
regression models37. To select the best one, LOADEST calculated model coefficients
243
using each calibration data set (i.e., observed TN loads in few of days), and models
244
with the lowest Akaike information criterion (AIC) values were selected for load
245
estimations.
246
RESULTS
247
Dry N deposition. The N flux of dry deposition over the Lake Dianchi was 42.1±10.3
248
mg km−2 s−1 in 2010-2011 (1σ as the standard deviation of fluxes occurring in 5 sites;
249
Fig. 2), with 37.1±10.5 mg km−2 s−1 for gaseous N and 5.0±0.7 mg km−2 s−1 for
250
particulate N. The major species, accounting for 99% of the dry deposition, was NH3
251
(32.8±10.1 mg km−2 s−1, 77.9%), HNO2/HNO3 (4.0±0.6 mg km−2 s−1, 9.5%),
252
particulate ON (2.9±0.6 mg km−2 s−1, 6.8%), and particulate NO3−-N (2.0±0.1 mg
253
km−2 s−1, 4.7%), while the rest of 1% were NO2 (0.7%) and particulate NH4+-N (0.3%).
254
The corresponding concentrations and Vd of gases and TSP at 5 sites can be found in
255
Fig. S1 and S2. Regardless of three missing samples in early April-June, the dry
256
deposition N flux showed strong seasonality (Fig. 2b). High fluxes were observed
257
from May to August, which were 1.7-13.3 times higher than the remaining periods.
ACS Paragon Plus Environment
Page 13 of 37
Environmental Science & Technology
258
Dry deposition of NH3 was the major source (>86.0%) of total dry deposition N flux
259
from May to August, with biweekly coefficient of variation [CV] of 115%. The NO2
260
flux was highest in autumn and early winter (0.4±0.2 mg km-2 s-1) and remained low
261
in early spring and summer (0.2±0.1 mg km-2 s-1). The remaining N fluxes were
262
negligible with a seasonal variability with CV of ~27%.
263 264
Wet N deposition. Wet deposition over the Lake Dianchi yielded an N flux of
265
46.5±13.2 mg km−2 s−1, 10.5% higher than dry deposition (Fig. 2c). NO3−-N
266
dominated the total flux of wet deposition (24.3±7.5 mg km−2 s−1, 52.2%), followed
267
by ON (11.9±5.4 mg km−2 s−1, 25.6%) and NH4+-N (10.3±3.9 mg km−2 s−1, 22.2%).
268
The corresponding concentrations of diverse N species at all 5 sites can be found in
269
Fig. S1. Peak N fluxes of wet deposition occurred in the wet season (i.e., from later
270
July to early October; Fig. 2d), following the dry deposition peaks. The N flux during
271
this period were up to 172.9±72.2 mg km−2 s−1, six-fold greater than mean value of the
272
rest of the year. N fluxes from January to March were only 4.4±7.7 mg km−2 s−1, due
273
to this being the period of
274
deposition fluxes was highly consistent with precipitation patterns over all 5 sites
275
(r=0.83, P