Interactive comment on “ Atmospheric reactive nitrogen concentrations at ten sites with contrasting land use in an arid region of Central Asia

Dear Editor, First of all, we thank you and two referees’ comments/suggestions on our manuscript “Atmospheric reactive nitrogen concentrations at ten sites with contrasting land use in an arid region of Central Asia”. The manuscript has been made substantial revisions based on two referees’ comments. The revisions include almost all Sections of the manuscript from Introduction, Materials and Methods, Results (including Tables and Figures), Discussion, Conclusion and References. We are feeling the quality of the manuscript has been largely improved and can be accepted for publication in the Special Issue of Biogeosciences. Detailed responses to two referees’ comments are


Introduction
Anthropogenic activities have greatly accelerated the emissions of reactive nitrogen (N r ) species worldwide (Galloway et al., 2008), with a consequent increase in atmospheric deposition of N r species.Elevated N deposition/addition may lead to eutrophication (Vitousek et al., 1997;Liu et al., 2011), declining biodiversity (Phoenix et al., 2006;Clark and Tilman, 2008;Song et al., 2011), soil acidification (Guo et al., 2010) and increased N 2 O emission (Li et al., 2012a, b) in aquatic and terrestrial ecosystems.N r species emissions have resulted in profound deterioration of local air quality (Chan and Yao, 2008;Kulshrestha et al., 2009;Shen et al., 2011a).N r deposition has therefore become an important public concern (Liu et al., 2011;Sutton et al., 2011).
Arid and semi-arid regions account for 47 % of the global land area and play a significant role in global N cycling.Xinjiang Uyghur Autonomous Region is located at the center of the arid and semi-arid regions of central Asia and is one of the main source areas of Asian dust, which can be transported from central Asia to the Pacific Ocean, and occasionally to the west coast of the United States (Li et al., 2008).Rapid economic development has led to significant increases in the combustion of fossil fuels by industry, power plants and vehicles in Xinjiang since 2000, with rates of consumption of diesel oil, crude oil, raw coal and coke increasing by 12, 27, 68 and 92 %, respectively (Mamtimin and Meixner, 2011).In addition, N fertilizer application has also increased steadily since the 1980s due to the expansion of cotton and cereal/fruit/vegetable production in Xinjiang.Increasing atmospheric N r pollution would therefore be expected in this region.On the other hand, N limitation of plant growth often occurs in arid and semi-arid regions and atmospheric N r deposition may act as an important source of plant nutrients (Liu et al., 2010).Thus, N inputs from N r deposition can have an important N fertilizer effect in arid and semi-arid regions (Bai et al., 2010;Song et al., 2012).However, very little is known about the influence of air pollutants and the dynamics of N r concentrations in the arid regions of China.The current study attempted to determine the spatial and seasonal characteristics of atmospheric N r pollution in different ecosystems within the arid Xinjiang region of northwest China.

Monitoring sites
The study was carried out at ten sites in the Xinjiang arid region in northwestern China.Site BYB is at Bayinbuluk Grassland Ecosystem Research Station, Chinese Academy of Sciences (CAS), which is located in the southern Tian Shan Mountains (3000-4500 m a.s.l.) of central Asia, covering a total area of approximately 23 000 km 2 .Site TZZ is at the Taklimakan Desert Research Station, CAS, in the center of the Taklimakan Desert.The Taklimakan Desert is in the arid center of Eurasia and covers a total area of 337 600 km 2 .Site CLZ is located at Cele Desert Research Station, CAS, near the southern fringe of the Taklimakan Desert, one of the three most extreme arid zones in the world.Site TLF denotes Turpan Eremophyte Botanic Garden, CAS; its main landform is the improved shifting sand land in the hinterland of the Turpan Basin.Site FKZ is at Fukang Research Station for Desert Ecology, CAS, which is situated at the southern fringe of the Gurbantunggut Desert in central Asia.Cotton is the most important cash crop in the region, and fertilizer is usually applied as (NH 4 ) 2 HPO 4 at 450 kg ha −1 and mixed in the top 20 cm of the soil profile before sowing (Tang et al., 2010).Site AKS is located at Aksu Research Station for Farmland Ecosystems, CAS, which is located at the northern fringe of the Tarim Basin and near downstream Aksu River, a tributary of the Tarim River.This is a large and nationally important cotton production area that is characterized by dramatic watercourse changes and high agricultural water consumption.Site YPH is located at an experimental farm of Xinjiang Academy of Agricultural Sciences and is a typical farmland monitoring site where wheat and maize are the main crops.Site BTH is an experiment of Xinjiang Academy of Agricultural Sciences and is a typical farmland monitoring site adjacent to the west of Kongque River where cotton is also an important regional crop.Site TFS is located at an experimental farm of Xinjiang Academy of Agricultural Sciences and is a typical suburban monitoring site.Site SDS is located in Urumqi city, the most inland city in the world about 2500 km from the Arctic, Indian and Pacific Oceans.It is surrounded by the Tian Shan Mountains to the north, east and west.The city covers a total area of approximately 140 km 2 with a population of 2.2 million and is a typical urban monitoring site.Sites SDS and TFS are only about 11 km apart.Additional characteristics of the sites are shown in Table 1 and Fig. 1.

Sampling procedure and sample analysis
NH 3 and NO 2 samples were collected using Radiello passive samplers (Aquaria Italy, Trident Equipments Pvt. Ltd., Mumbai, India).At each site three NH 3 and three NO 2 samplers were exposed during each measurement period inside a PVC shelter (2 m above the ground) to protect the samplers from precipitation and direct sunlight.NH 3 and NO 2 concentrations were measured monthly by exposing the samplers for two weeks in the middle of the month.After sampling, absorption cartridges of the passive samplers were placed in airtight plastic tubes and stored in a refrigerator at 4 • C until analysis with an automated, segmented continuous flow analyzer (Seal AA3, Norderstedt, Germany) within two months.Detailed information about the passive samplers has been provided by Shen et al. (2009).Airborne PM 10 particles (particulate matter whose aerodynamic equivalent diameter is < 10 µm) were sampled using a particulate sampler (BGI, Omni, Waltham, MA) with a flow rate of 5 l min −1 , and 7-10 daily samples of PM 10 were collected at each site during each month.The sampler was placed about 2 m above the ground and ran for 24 h to obtain a particulate matter sample on 47 mm quartz filters (Whatman, Maidstone, UK).Before and after sampling, each filter was conditioned for at least 24 h inside a chamber at a relative humidity of 50 % and a temperature of 25 • C, and then weighed (Sartorius, Göttingen, Germany; precision 10 µg).PM 10 mass concentrations were determined from the mass difference and the sampled air volume.Each sampling filter was extracted with 10 ml deionized water by ultrasonication for 30 min, and the extract solution was filtered through a syringe filter (0.45 mm, Tengda Inc., Tianjin, China) and stored in a refrigerator.Chemical analysis of PM 10 was conducted within two months.Ammonium and nitrate in PM 10 (pNH + 4 and pNO − 3 ) were measured by continuous flow analyzer (Seal AA3).Monthly mean air temperature, precipitation, wind speed and relative humidity from August 2009 to November 2011 at all sites except TFS and YPH are shown in Fig. 2.

Statistical analysis
Values of NH 3 , NO 2 , PM 10 , pNH + 4 and pNO − 3 concentrations each month at all sites are means ± standard errors (s.e.).All statistical analyses (Pearson correlation analysis and one-way analysis of variance) were performed using the SPSS 13.0 statistical package (SPSS Inc., Chicago, IL).

Seasonal variation in N r concentrations
NH 3 exhibited distinct and significant temporal variation, with lowest concentrations in winter compared to in spring, summer or autumn (Fig. 4a).Monthly mean concentrations measured for NH 3 and NO 2 are given in Fig. 4a and c, and the values ranged from 0.58 to 21.7 and 0.29 to 57.2 µg N m −3 across all sites.In contrast to NH 3 , high monthly mean NO 2 concentrations were found in winter (except site BYB).Monthly mean PM 10 concentrations ranged from 3.06 to 876.2 µg m −3 , and PM 10 concentrations reached their maximum values in April at TZZ (876.2 µg m −3 ) and CLZ (852.7 µg m −3 ) (Fig. 4e).PM 10 concentrations were higher in winter and spring than in summer or autumn (except site BYB).Monthly mean NH + 4 concentrations in PM 10 (pNH + 4 ) ranged from 0.02 to 25.8 µg N m −3 , and pNH + 4 concentrations peaked in winter with lower values in summer across almost all sites (Fig. 4b).Monthly mean particulate concentrations of NO − 3 in PM 10 (pNO − 3 ) ranged from 0.02 to 11.1 µg N m −3 , and were higher in winter than in spring, summer or autumn (Fig. 4d).

Mass concentrations of inorganic N ions in PM 10
Mass concentrations of PM 10 and particulate NH + 3 ) in PM 10 accounted for 0.35-13.6% of PM 10 across the ten sampling sites.

Correlations between various N r species and effects of environmental factors on atmospheric N r concentrations
Correlations between various atmospheric N r species concentrations across the ten sites are shown in Fig. 5. NO 2 concentrations were significantly and positively correlated with pNH + 4 (Fig. 5a), pNO − 3 (Fig. 5b) and NH 3 (Fig. 5c).Although no significant correlation was found between NH 3 and pNO − 3 concentrations (Fig. 5d), pNH + 4 was significantly and negatively correlated with NH 3 (Fig. 5e) and positively correlated with pNO − 3 (Fig. 5f).These positive or negative correlations reflected the complicated interactions between these N r species, in particular the transformations between gaseous N r species (NH 3 and NO 2 ) and secondary particles (e.g.pNH + 4 and pNO − 3 ).The effects of environmental factors on atmospheric N r concentrations and PM 10 are shown in Fig. 6.Both pNH + 4 and pNO − 3 concentrations showed significant negative correlations with air temperature and positive correlations with relative humidity (Fig. 6a-d).PM 10 concentrations showed significant negative correlations with relative humidity (Fig. 6e).NH 3 concentration showed significant positive correlations with air temperature (Fig. 6f) and wind speed (Fig. 6g) and negative correlations with relative humidity (Fig. 6h).No significant correlations were found between NO 2 concentrations and environmental factors (data not shown).

Concentrations of N r compounds in different ecosystems
The annual NH 3 concentration showed significant spatial variation from alpine grassland to farmland in the Xinjiang arid region of central Asia.The highest NH 3 concentration found was in farmland (YPH), which is affected by agricultural pollution sources.The lowest NH 3 concentration was found in alpine grassland (BYB) where the absence of fertilizer applications and the lowest air temperatures may have contributed to lower NH 3 volatilization and concentrations.Although agriculture is the main source of NH 3 , motor vehicles can be important sources of ammonia in urban environments (Ianniello et al., 2010).This may explain the moderate NH 3 concentrations at the urban site (TFS).
In the current study, annual NO 2 concentrations also exhibited significant spatial variation.The highest NO 2 concentration was found at an urban site (SDS) which is subject to anthropogenic pollution sources, and the lowest NO 2 www.biogeosciences.net/9/4013/2012/Biogeosciences, 9, 4013-4021, 2012 concentration was found in alpine grassland which is relatively undisturbed and free from anthropogenic activities.
The highest PM 10 concentrations were found in the desertoasis ecotone (CLZ) and desert (TZZ) in spring where sandstorm events often occur due to low precipitation and almost no vegetation cover, and the values were much higher than reported from numerous other sites (Kumar and Joseph, 2006;Li et al., 2006;Namdeo and Bell, 2005).Higher annual pNH + 4 concentrations were found at sites SDS and TFS and were much higher than reported at many other urban and suburban sites (Hayashi et al., 2007), possibly due to anthropogenic influences.Higher annual pNO − 3 concentrations were also found at SDS and TFS.We assume that Bayinbuluk alpine grassland (BYB, 3.09 µg N m −3 ) is representative of background conditions, because it is remote from urban areas and relatively undisturbed by anthropogenic activities.Extremely high N r concentrations at an urban location (site SDS, 47.0 µg N m −3 ) may reflect substantial air pollution due to increased emissions by traffic, industry and domestic heating, with consequent effects on the eutrophication of water bodies (Liu et al., 2011) and soil acidification in neighboring natural and semi-natural ecosystems (Yang et al., 2012).Yang et al. (2011) have pointed out that the arid desert region of northwest China is one of the main global sources of atmospheric dust, and the Taklimakan Desert in the Tarim Basin is an important source of sand/dust storms.Our results further demonstrate three types of air pollution by N r species: from natural sources (e.g.sand/dust storms due to dry weather conditions in the Taklimakan Desert region at CLZ, AKS, TZZ and YPH), from anthropogenic N r emissions (e.g. at arid suburban/urban locations such as TFS and SDS) and from combined natural and anthropogenic sources (e.g. at oasis/farmland locations such as TLF, FKZ and BTH).These differences require policy makers to formulate different strategies to control air pollution in the arid region of central Asia.Another implication is that the extent of air pollution of N r species at suburban and urban sites in Xinjiang is similar to or even greater than at intensively managed agricultural and industrial regions in China such as the North China Plain (Shen et al., 2011a, b).This may be due to the concentration of the human population and anthropogenic activities on the very limited oasis areas, which comprise only 5-6 % of the total land area in Xinjiang.

Seasonal variation in concentrations of N r compounds
Higher NH 3 concentrations were found in summer and autumn in farmland, most likely due to high NH 3 emissions from N fertilizer applications, rapidly increasing livestock production, and motor vehicles.In contrast to other sites, summer NH 3 concentrations were low in alpine grassland.This is most likely due to the transfer of almost all livestock to higher mountain areas (over 3000 m a.s.l.) and lower air temperatures.In north and northwest China, NH 3 concentrations are relatively high in summer and autumn, mainly due to N fertilization and higher air temperatures, both of which promote NH 3 volatilization from arable soils.For example, the N fertilizer application rate can reach 240 kg N ha −1 per crop to achieve high yields in spring and summer at site TFS.However, less than 30 % of applied N fertilizer may be taken up by the crops (Zhang et al., 2008) and much of the unaccounted-for fertilizer N may have been lost by NH 3 emission.Source strength and removal efficiency can explain seasonal variation in NH 3 concentrations (Hong et al., 2002).Moreover, air temperature and wind speed determine the NH 3 concentrations.Shen et al. (2011a) also reported the lowest seasonal NH 3 concentration in winter at six sites on the North China Plain.NO 2 concentrations exhibited distinct and significant seasonal variation, with higher concentrations in winter at almost all sites (except BYB), and reached their highest value at site SDS.These NO 2 values were very similar to the high concentrations observed in other regions in China (Aas et al., 2007;Shen et al., 2009).Vehicle emissions and coal consumption for heating may be the main explanation for the higher NO 2 concentrations.For example, Urumqi is only 140 km 2 in area and had a population of 2.2 million and a total number of motor vehicles of 140 000 units in 2003.However, motor vehicles reached 240 000 units and were increasing at a rate of 260 units per day in 2009.Therefore, the higher NO 2 concentrations are likely related to anthropogenic sources such as vehicle emissions at the urban site, and the use of coal for heating in north and northwest China in winter may be the main factor responsible for the increased NO x concentrations in the atmosphere.In addition, due to the paucity of rainfall or snowfall in winter in these arid areas, scavenging of atmospheric NO 2 by wet deposition will be greatly limited.Higher PM 10 concentrations in this study were found in spring and winter and were much higher than reported from numerous other sites (Kumar and Joseph, 2006;Li et al., 2006).We observed PM 10 concentrations reaching maximum values due to dust storm events in spring, notably 852.7 and 876.2 µg m −3 at sites CLZ and TZZ, and they were even higher than in Lanzhou, a city with some of the highest air pollution worldwide with a peak PM 10 concentration (541.9 µg m −3 ) occurring in April due to dust events (Wang et al., 2009).Higher PM 10 concentrations occurred in winter due to domestic heating.Moreover, increasing numbers of vehicles and low precipitation and relative humidity in the arid regions resulted in considerable PM 10 pollution.The lowest pNH + 4 concentrations were found in summer at almost all sites, while the highest pNO − 3 concentrations were found in winter, especially at sites SDS, TFS and FKZ where the values were much higher than reported for numerous other sites (Hayashi et al., 2007).This was most likely generated from coal combustion in winter and from other anthropogenic sources such as vehicles, manufacturing and industrial processes at other times of year.Meteorological parameters such as low temperature and high relative humidity point towards poor dilution of pollutants during the winter period.Regarding the presence of nitrate associated with ammonium, it was more likely to form secondary particulate ammonium nitrate during periods of low temperatures and relatively high humidity in the arid areas.

Ratios of concentrations of secondary particles to PM 10
Secondary particles (pNH + 4 and pNO − 3 ) have been regarded as the main contributor to fine particulate matter (PM), which is harmful to human health and reduces visibility (Erisman and Schaap, 2004;Pinder and Adams, 2007) and is also implicated in climate change due to its effects on direct and indirect radiative forcing.In North China the ratios of secondary inorganic N particles to PM 10 were relatively high.For example, concentrations of secondary particles (pNH + 4 and pNO − 3 ) accounted for 14.6-22.3% of the PM 10 concentrations (Shen et al., 2011a).Heavy secondary particle pollution is a reflection of the high emission intensities of NH 3 and NO x in north China.In our study, lower ratios were found in desert (TZZ, 0.35 %) and desert-oasis ecotone (CLZ, 0.63 %) sites, while farmland near a desert (FKZ, 13.6 %) and urban (SDS, 11.7 %) and suburban (TFS, 8.3 %) areas had higher ratios.The use of coal for heating may be a major factor because other sites (except BYB) have relatively high air temperatures with almost no coal consumption in winter.This indicates that anthropogenic activities have greatly increased the emissions of N r species and influenced local air pollution.Therefore, a decline in the concentration of secondary inorganic particles would be a worthwhile target to improve air quality in these arid regions.Further studies should focus on the fluxes of atmospheric N deposition and their ecological impacts along anthropogenic N r concentration gradients in the arid oasis Xinjiang region.

Conclusions
This study provides original monitoring data on spatial and temporal variations in atmospheric N r species and PM 10 from an arid region of central Asia.Overall annual mean concentrations of NH 3 , NO 2 , pNH + 4 and pNO − 3 were 7.71, 9.68, 1.81 and 1.13 µg N m −3 , and the annual average PM 10 concentration was 249.2 µg m −3 in the Xinjiang arid region of northwest China.The increasing order of total concentrations of N r species was, respectively, alpine grassland; desert, desert-oasis ecotone; desert in an oasis; farmland; suburban and urban ecosystems.NO 2 concentrations exhibited distinct and significant seasonal variation with higher concentrations in winter.Lower pNH + 4 concentrations were found in summer at almost all sites, while the highest pNO − 3 concentrations were found in winter.Lower ratios of secondary parti-cles (NH + 4 and NO − 3 ) were found in desert and the desertoasis ecotone, while urban and suburban sites had higher ratios, indicating that anthropogenic activities such as home heating in winter have greatly influenced local air quality and pollution and requiring effective controlling measures.

Fig. 1 .
Fig. 1.Distribution of the ten monitoring sites in the Xinjiang arid region of northwest China.

Fig. 2 .
Fig. 2. Air temperature, precipitation, wind speed and relative humidity from August 2009 to November 2011 at the monitoring sites.

Fig. 3 .
Fig. 3. Total annual average concentrations of reactive nitrogen species at the ten sites.
pNH + 4 and pNO −3 ) at the ten sites are shown in Table2.Mean concentrations of pNH + 4 and pNO − 3 in PM 10 ranged from 0.18 to 8.23 µg m −3 and from 0.96 to 11.2 µg m −3 , and the concentrations of pNH + 4 and pNO − 3 in PM 10 accounted for 0.8-5.2 and 0.3-8.4% of PM 10 across all sites.Concentrations of total inorganic N ions (pNH + 4 and pNO −

Fig. 4 .
Fig. 4. Seasonal concentrations of (A) NH 3 , (B) pNH + 4 , (C) NO 2 , (D) pNO − 3 and (E) PM 10 at the ten sites.Spring, summer, autumn and winter denote the periods from December to February, March to May, June to August, and September to November, respectively.

Table 1 .
Characteristics of the ten monitoring sites in the Xinjiang arid region of northwest China.

Table 2 .
Concentrations of PM 10 , pNH + 4 and pNO − 3 in PM 10 , NH 3 , NO 2 and ratios of concentrations of secondary particles (pNH + Values in a column without the same letters are significantly different at p < 0.05.Ratio denotes (pNO − 3 + pNH + 4 ) concentrations as a percentage of PM 10 .