lable at ScienceDirect

Atmospheric Environment 43 (2009) 928–939

Contents lists avai

Atmospheric Environment

journal homepage: www.elsevier .com/locate/a tmosenv

Past and future trends in concentrations of sulphur and nitrogen compoundsin the Arctic

Lars R. Hole a,b,*, Jesper H. Christensen c, Tuija Ruoho-Airola d, Kjetil Tørseth a,Veronica Ginzburg e, Piotr Glowacki f

a Norwegian Institute for Air Research (NILU), P.O. Box 100, 2027 Kjeller, Norwayb Norwegian Meteorological Institute (met.no), Forecasting Division for Western Norway, Norwayc National Environmental Research Institute, Roskilde, Denmarkd Finnish Meteorological Institute, Helsinki, Finlande Russian Federal Service for Hyrometeorology and Environmental Monitoring and Russian Academy of Sciences, Moscow, Russiaf Polish Academy of Sciences, Warsawa, Poland

a r t i c l e i n f o

Article history:Received 30 May 2008Received in revised form17 October 2008Accepted 20 October 2008

Keywords:ArcticNOx

NHx, SOx

MonitoringDepositionPrecipitationAMAPTrends

* Corresponding author. Norwegian Institute for A100, 2027 Kjeller, Norway. Tel.: þ47 91769141; fax: þ

E-mail address: lrh@nilu.no (L.R. Hole).

1352-2310/$ – see front matter � 2008 Elsevier Ltd.doi:10.1016/j.atmosenv.2008.10.043

a b s t r a c t

Recent trends in nitrogen and sulphur compounds in air and precipitation from a range of Arcticmonitoring stations are presented, with seasonal data from the late 70s to 2004 or 2005. Earlier findingsof declining sulphur concentrations are confirmed for most stations, while the pattern is less clear forreduced and oxidized nitrogen. In fact there are positive trends for nitrogen compounds in air at severalstations. Acidity is generally reduced at many stations while the precipitation amount is either increasingor stable. Variability of sulphate concentrations in air for the period 1991–2000 is reasonably wellreproduced at most stations using an Eulerian, hemispherical model. Results for nitrogen compounds areweaker. Scenario studies show that even if large sulphur emission reductions take place in importantsource regions in South-East Asia in the coming decades, only small changes in Arctic deposition can beexpected. This is because South-East Asian emissions have small influence north of the Arctic circle.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction level of wintertime air pollution remained roughly constant,

Arctic acidification occurs in areas with both sensitive ecologyand levels of acid deposition elevated to a point that exceeds thesystem’s acid neutralizing capacity. Sulphur is the most importantacidifying substance in the Arctic, with nitrogen of secondaryimportance (Kamari and Joki-Heiskala, 1998). Significant anthro-pogenic sources of sulphur emissions, and to a lesser extentnitrogen emissions, exist within the arctic region. In addition, long-range transported air pollutants contribute to acidification andarctic haze. Emissions from natural sources within the Arctic(volcanoes, marine algae, and forest fires) are difficult to quantifyand project (Kamari and Joki-Heiskala, 1998), but comprise a rela-tively small contribution compared to the anthropogenic sources.

Based on firn core analysis from the Canadian high Arctic, Barrieet al. (1985) suggested that in the first-half of the 20th century the

ir Research (NILU), P.O. Box47 55236703.

All rights reserved.

consistent with a pattern of little change in European sulphurdioxide (SO2) emissions. However, between 1956 and 1977 therewas 75% increase of Arctic air pollution which seems to be associ-ated with a marked increase in SO2 and total oxidized nitrogen(NOx) emissions in the industrialized world. Weiler et al. (2005)analysed an ice core from a North Siberian ice cap and found thatmaximum sulphate and nitrate concentrations in the ice could berelated to maximum SO2 and nitrogen oxide (NOx) anthropogenicemissions in the 1970s, probably caused by the nickel- and copper-producing industries in Norilsk and on the Kola peninsula or byindustrial combustion processes occurring in the Siberian Arctic. Inaddition, they found that during recent decades, sulphate (SO4

2�)and nitrate (NO3�) concentrations declined by 80% and 60%,respectively, reflecting a decrease in anthropogenic pollution of theArctic basin. However, this NO3� reduction is much stronger thanthe emission reductions reported by EMEP1 for the Northern

1 Co-operative Programme for Monitoring and Evaluation of the Long-rangeTransmission of Air pollutants in Europe, www.emep.int.

Table 1Monitoring stations referred to in the text.

AMAP code Station name Parameters Latitude Longitude Operation period Heigh above sea level

CA0420G Alert Air 82 28 62 30 W 1980- 1DK0010G Nord, Greenland Both 81 36 16 40 W 1990–2002

2006-20

FI22 Oulanka Both 66 19 29 24 E 1990- 310IS02 Irafoss Both 64 5 21 1 W 1980- 66NO30 Jergul (with NO55) Both 69 27 24 36 E 1977–1997 255NO15 Tustervann Both 65 50 13 55 E 1977 439NO42 Zeppelinfjell Air 78 54 11 53 E 1989- 474NO47 Svanvik Both 69 27 30 2 E 1986- 30NO55 Karasjok (with NO30) Both 69 28 25 13 E 1997- 333NO57 Ny-Ålesund Precip 78 55 11 55 E 1982- 8RU01 Janiskoski Both 68 56 28 51 E 1990- 118RU13 Pinega Both 64 42 43 24 E 1979- 28SE05 Bredkal Both 63 51 15 20 E 1979- 404SE93 Abisko Precip. 68 21 18 49 E 1992- 0

Snare Rapids Precip. 63 31 116 00 W 1989-Hornsund Precip. 77 00 15 33 E 1988-

Russian precipitation stations Zarechensk Precip. 66 7 31 4 E 1990-Padun Precip. 68 6 31 8 E 1991-Krasnoshelie Precip. 67 4 37 1 E 1990-Nariyan-Mar Precip. 67 7 53 0 E 1962-Urengoy Precip. 66 0 78 4 E 1989-Turukhansk Precip. 65 8 87 9 E 1962-Deputatskiy Precip. 69 3 139 7 E 1990-Ust-Moma Precip. 66 5 143 2 E 1990-Palatka Precip. 60 1 150 9 E 1962-

2 Arctic Monitoring and Assessment Programme, www.amap.no.

L.R. Hole et al. / Atmospheric Environment 43 (2009) 928–939 929

Eurasian continent, which is about 20% (assessment report avail-able at www.emep.int).

Kamari and Joki-Heiskala (1998) concluded that there were notrends in atmospheric concentrations of acidifying compounds inCanada and Alaska during the 1980s, but that there weredecreasing trends on Svalbard. Background data from Russia werenot available. It was considered that about 75% of the depositioncould be as dry deposition, but that there was a lack of observationsand knowledge at this point. Model output for SO2 and SO4

2�

compared well with time series observations series at one station(Nord, Greenland) and for long-term averages at a number of EMEPstations.

Although atmospheric lifetimes of SO2, NOx and their oxidationproducts are of the order of a some days at temperate latitudes(Schwarz, 1979; Levine and Schwarz, 1982; Logan, 1983), theatmospheric half-life of SO4

2� has been reported to reach even twoweeks or more in the high Arctic during winter (Barrie, 1986). Thetransport distances range from hundreds to thousands of kilometres(Seinfeld and Pandis, 1998). Thus, many factors, besides the primaryemissions, affect the observed concentrations and trends of thecompounds involved in the acid deposition process, including theirrelative concentrations in the atmosphere, the reversible nature ofsome of the reactions and the meteorological situation.

Trend analysis for several indicators of Arctic haze has beenperformed for the spring months by Quinn et al. (2007). Themonthly average SO4

2� concentration in air in March and April hasdecreased in the Canadian, Norwegian and Finnish Arctic by 30–70% from the early 1990s to early 2000. NO3

� concentration in airhas increased by 50% in Alert, Canada during the same period. ForZeppelin Norway (see Table 1), there is no trend, in accordance withAas et al. (2008).

Here we present a comprehensive overview of trends in Arcticacidification from the late 70s to recent years, including Russianprecipitation stations. Summer and winter mean values have beencalculated separately in order to avoid the influence of the possiblechanges in the seasonal variation. New numerical modeling resultsusing a Eulerian hemispheric model are compared to monitoringdata for the period 1991–2004. In addition, we present total

oxidized sulphur (SOx) and NOx concentrations in air and deposi-tion from the late 1800s to 2020 for the whole Arctic, based on thesame Eulerian model.

2. Description of monitoring stations

There is a lack of long-term time series of background concen-trations of main atmospheric compounds in the high Arctic. Alsothere are few stations with co-located air and precipitationsampling. Table 1 lists the stations included here. The AMAP2

atmospheric monitoring network consists of a number of stationsspread across the Arctic. Most of these are EMEP stations that alsoreport to the AMAP database. In addition, a few national stationsreport data. Some stations have been reporting data since the mid1970s. As of 2002, 24 stations reported data to AMAP relevant foracidification and eutrophication (Fig.1). Most stations are located inthe European sector. The nitrogen compounds in air are at theEMEP stations measured as the sum of particulate nitrate and gasphase nitric acid and, respectively, the sum of particulate ammo-nium and gas phase ammonia. They are referred later in the text astotal nitrate and total ammonium in air. Alert measures nitrate andammonium in the particulate phase only.

The Russian national network for monitoring of precipitationchemical composition and acidity consists of 110 monitoringstations. Precipitation samples collected at these stations are ana-lysed by regional laboratories for the main atmospheric compounds.The coordinating and analytical centre for the precipitation chem-istry monitoring network is the Voieykov Main Geophysical Obser-vatory, Roshydromet whose data are mainly used for this article. Inaddition to these stations, there are 105 monitoring sites where onlypH value is analysed. Stations are unevenly distributed over theterritory of Russia. Less than 40% of the stations are situated in thevast Siberian region. The period of observations reaches up to 40years for some stations. For analysis of the acid precipitation andacidity we have used nine background monitoring stations situated

60° N

70° N

80° N

90° N

Alert

Nord

Irafoss TustervannBredkäl

Snare Rapids

Nariyan−Ma

Ureng

Turuk

DeputaskiyUst−Moma

Palatka

Oulanka

Jergul

Tustervann

Zeppelin/Ny−Ålesund

Svanvik

Janiskoski

PinegaBredkäl

Abisko

Hornsund

Zarachensk

PadunKrasnoshelie

10 ° E 15 ° E 20° E 25° E 30° E 35° E 40° E 45

° E 50° E

65 ° N

70 ° N

75 ° N

80 ° N

Fig. 1. Location of monitoring stations for air and precipitation chemistry.

L.R. Hole et al. / Atmospheric Environment 43 (2009) 928–939930

in the Russian Arctic. For these stations, average summer (June–August) and winter (December–February) values were reported.Except for the two EMEP stations reported here (Janiskoski andPinega), there are no background air concentrations monitoringsites situated in the Arctic region of Russia.

3. Statistical method

As pointed out by Macdonald et al. (2005), detection of recenttrends in the Arctic is difficult due to the combination of short or

a b

Fig. 2. Estimated SOx–S and N

incomplete data records at some sites and interference fromnatural variations on seasonal, annual and decadal timescales(Quinn et al., 2007). In order to remove seasonal variability from thetrend analyses, we focus here on monthly concentrations for winter(December–February) and summer (June–August) separately.

Existence of a monotonic increasing or decreasing trend in thetime series 1980–2005 and 1990–2005 was tested with thenonparametric Mann–Kendall test at the 10% significance level asa two-tailed test (Gilbert, 1987). Some of the stations opened in the1970s, but we choose to test for the same periods at all stations to

Ox–N emissions in 2000.

L.R. Hole et al. / Atmospheric Environment 43 (2009) 928–939 931

be able to compare trends. An estimate for the slope of a lineartrend was calculated with the nonparametric Sen’s method (Sen,1968). The Sen’s slope is the median of the slopes calculated fromall pairs of values in the data series. The method is not greatlyaffected by data outliers, and it can be used when data are missing(Salmi et al., 2002).

It is likely that significant trends in deposition are partly a resultof changes in emissions. However, it is not obvious which emissionareas contribute to deposition. The relative contributions ofdifferent regions are suspected vary from year to year depending onatmospheric transport paths.

4. Description of Danish Eulerian Hemispherical Model

The Danish Eulerian Hemispheric Model (DEHM) systemconsists of a weather forecast model, the PSU/NCAR MesoscaleModel version 5 (MM5) modelling subsystem (see Grell et al.,1994),which is driven by meteorological data from ECMWF, and a 3-Datmospheric transport model, the DEHM model. The model hasa horizontal resolution of 150 km� 150 km and 20 irregularly

78 92 060

0.9Oulanka

mg

S

l−

1

78 92 060

1.8Irafoss

78 92 060

4.2Janiskoski

mg

S

l−

1

78 92 060

0.9Pinega

78 92 060

1.3Tustervann

mg

S

l−

1

78 92 060

2.2Karasjok

78 92 060

2.3Bredkal

mg

S

l−

1

78 92 060

0.5Snare−Rapids

Year

78 92 060

2.6Ny−Ålesund

mg

S

l−

1

Year

Fig. 3. Non-seasalt sulphate concentrations (nss SO42�) in precipitation (Simpson et al.,

2003) for summer (June–August (red)) and winter (December–February (blue)).Monthly average values in mg S l�1. Significant Sen slopes within 10% are shown assolid lines (whole period) and dashed lines (after 1990). See also Table 2. [For inter-pretation of the references to colour in this figure legend, the reader is referred to theweb version of this article.]

spaced vertical layers up to 16 km. The coverage is close to hemi-spheric from nearly 10� N at the corners and 25� N at midpoints ofthe model domain boundaries.

The original version of the DEHM model was developed forstudying the long-range transport of SO2, SO4

2� and Pb to the Arctic(Christensen, 1997) and has been used since 1991. The sulphurversion has been used in the first and the second phase of the AMAPprogram (see Kamari and Joki-Heiskala, 1998; Hole et al., 2006a,b)and the Pb version was used in the last AMAP heavy metalassessment. It has been further developed to study transport,transformation and deposition of reactive and elemental mercury,and this version was also used in the heavy metal assessment, seealso Christensen et al. (2004) and Heidam et al. (2004). Otherversions calculate the concentrations and depositions of variouspollutants (Frohn et al., 2002a,b) through the inclusion of theextensive chemistry scheme, and transport and exchange ofatmospheric carbon dioxide (Geels et al., 2004) and PersistentOrganic Pollutants (Hansen et al., 2004).

In this work we are using the extensive chemical version whichincludes 63 species of which 4 groups relate to primary particulates(PM2.5, PM10, TSP and seasalt), other variables are SOx, NOx, totalreduced nitrogen (NHy),VOC and secondary inorganic particulates(Frohn et al., 2003). The chemical scheme was based on a schemewith 51 species presented in Flatøy and Hov (1996), which were an

78 92 060

0.3Oulanka

mg

N

l−

1

78 92 060

0.7Janiskoski

78 92 060

0.8Pinega

mg

N

l−

1

78 92 060

0.4Tustervann

78 92 060

0.7Karasjok

mg

N

l−

1

78 92 060

0.5Bredkal

78 92 060

0.2Abisko

mg

N

l−

1

78 92 060

0.6Snare−Rapids

Year

78 92 060

1.1Ny−Ålesund

mg

N

l−

1

Year

Fig. 4. Same as Fig. 3, but for NO3�, mg N l�1.

L.R. Hole et al. / Atmospheric Environment 43 (2009) 928–939932

ozone chemistry scheme with most of the important inorganicspecies and as well the most abundant hydrocarbons (explicittreatment of alkanes up to C4, longer alkanes lumped, explicittreatment alkenes up to C3, longer alkenes lumped, xylene, tolueneand isoprene). There were added reactions to extend the chemistryto eutrophication issues by using ammonium chemistry based onthe old EMEP acidification model and adding reactions in orderextend to acidification issues by using aqueous chemistry based onJonson et al. (2000). The scheme contains 120 chemical reactionswhere 17 are photolysis reactions calculated by the Phodis routine(Kylling et al., 1998) depending on sun-angle, altitude, Dobson unitand 3-D cloud cover. The used chemical scheme is quite similar tothe EMEP scheme described in Simpson et al. (2003).

The dry deposition module used in the DEHM model is based onthe resistance method and is very similar to the dry depositionmodule of the EMEP model (for details and documentation seeSimpson et al., 2003). This module calculates deposition of bothgaseous species and particulates to 16 different land-use categoriesbased on Olson World Ecosystem Classes, version 1.4D. The drydepositions of gaseous species to water surfaces are depending onthe wind speed (surface roughness) and on solubility of thechemical species (see Hertel et al., 1995).

Wet deposition is parameterized by a scavenging ratio formu-lation, where the scavenging is divided into two contributions. Thefirst contribution is the in-cloud scavenging, which represents theuptake in droplets inside a cloud. The second contribution origi-nates from precipitation events and is uptake in droplets below thecloud base. The scavenging coefficients are also very similar to theEMEP model.

78 92 064

8Oulanka

pH

78 92 064

8Irafoss

78 92 064

8Janiskoski

pH

78 92 064

8Pinega

78 92 064

8Tustervann

pH

78 92 064

8Karasjok

78 92 064

8Bredkal

pH

78 92 064

8Snare−Rapids

Year

78 92 064

8Ny−Ålesund

pH

Year

Fig. 5. Same as Fig. 3, but for acidity (pH).

5. Emission data

Many emission areas in the Northern Hemisphere contribute tothe air pollution levels in the Arctic, but Russian sources are mostimportant with more than 50% for the black carbon contribution(Stohl, 2006). Nonferrous metal production is the greatest source ofenvironmental pollution in the Russian Arctic (Hole et al., 2006a).Increased shipping activity as a result of reduced ice cover in thecoming decades will probably be compensated by improved tech-nology, so that significant increases in nitrogen and sulphur emis-sions are not expected (Hole et al., 2006a).

For the model studies in this work, several emission scenarioshave been used, depending on the model runs are for present time(1990–2004) historical run (preindustrial, and 1890–1980), orfuture scenarios (2010 and 2020). The emission inventories for themodel run for the present time are based on the global anthropo-genic emissions on a 1�1 grid from the EDGAR database of SO2,NOx, VOC (Volatile Organic Compounds) CO and CH4 for the years1990, 1995 and 2000. Furthermore emissions of anthropogenic andnatural NH3 and NOx from lightning and soil emissions from theglobal GEIA are used. From the RETRO emissions database (seehttp://www.retro.enes.org) the wildfire emissions of NOx, SO2, CO,VOC, PM on monthly basis are obtained for the period 1987–2000.These global emissions scenarios are combined by replacing similaranthropogenic emissions with the EMEP expert emissions forEurope of SO2, NOx, NH3, VOC and CO for the years, 1985, 1990,1995–2003. Emissions of SOx–S and NOx–N for year 2000 areshown in Fig. 2.

The model runs for the historical runs are based on the globalEDGAR–HYDE anthropogenic emissions inventories of SO2, NOx,VOC, CO, CH4 and NH3 for the years 1890, 1920, 1940, 1960 and1980. Furthermore the emissions of natural NH3 and NOx fromlightning and soil emissions and the wildfire emissions for the year1987 are used. The preindustrial emissions are only natural NH3,lightning, soil and wildfire emissions.

The model runs for the future scenarios are based on IIASAanthropogenic emissions scenarios of SO2, NOx, VOC, CO and CH4

(Frank Dentener, Joint Research Centre, ISPRA, Italy, personalcommunication). For the years 2010, 2020 there are two types ofemissions scenarios: The ‘‘Business As Usual’’ (BAU) scenario whichreflects the current perspectives of individual countries on futureeconomic development and takes the effects of presently decidedemission control legislation in the individual countries intoaccount. The other scenario is the ‘‘Maximum technically FeasibleReduction’’ (MFR) scenario that outlines the scope for emissionreductions offered by full implementation of the presently availableemission control technologies, while maintaining the projectedlevels of anthropogenic activities. Furthermore the emissions ofanthropogenic and natural NH3 and NOx from lightning and soilemissions (from GEIA database) and the wildfire emissions for theyear 2000 are used. These global emissions scenarios are combinedby replacing similar anthropogenic emissions with the futurescenarios for the EMEP area in Europe (Expert emissions estimatein CAFE (clean air for Europe)) for 2010 and 2020 (see http://webdab.emep.int).

For all the model runs the available information about emissionsfrom the large smelters in Russia are used. Furthermore for all themodel runs natural emissions of isoprene are based on the GEIAnatural VOC emission model, see Guenther et al. (1995). Emissionsare split into 14 different emissions categories (12 SNAP, SelectedNomenclature for Air Pollution in addition to lightning and wild-fires). Each emission category has different vertical distribution andtemporal variation (monthly, weekly and daily).

L.R. Hole et al. / Atmospheric Environment 43 (2009) 928–939 933

6. Trends in concentrations in air and precipitation,1980–2005 and 1990–2005

Figs. 3 and 4 show summer and wintertime series with signifi-cant trends in non-seasalt SO4

2� (nss SO42�, see Simpson et al., 2003)

and NO3� in precipitation from 1980 and from 1990 until 2003 or

2004. Time series for acidity (pH) are shown in Fig. 5. Trends after1990 are also shown on the map in Fig. 6 for nss SO4

2� and NH4þ in

precipitation. For the SO42� concentration, the values are usually

higher during summer months than during winter months. Lowconcentrations are measured at the Oulanka, Pinega and SnareRapids stations. The level of the monthly SO4

2� concentration in thebeginning of the monitoring period is higher than at the end of theperiod but there is not a significant trend at all of the stations. Forthe NO3

� concentration, values are on the contrary higher duringthe winter months than during the summer months. The inter-annual variation in the NO3

� concentration is larger than in thesulphate concentration. The level of the nitrate concentration at theend of the monitoring period is lower than in the beginning at onlythe Pinega station. At the Janiskoski station, the concentration hasincreased during the winter months.

SO42− winter

60° N

70° N

80° N

90° N

NH4+ winter

60° N

70° N

80° N

90° N

−0.05 0

Fig. 6. Significant trends (within 10%) in SO42� and NH4

þ in precipitation after 1990 for winternot shown because of few significant trends. Units are the same as in Table 2.

Table 2 summarises the trend statistics for the precipitationdatasets and also main compounds in air for the period since 1990.There are increasing trends in sulphate in precipitation at Ust-Moma in winter but at this station background concentrations arevery low. This could be due to changes in emissions from Norilsk(NE Siberia, 69�21’ N 88�12’ E) or variability in transport pattern(Hole et al., 2006b). However, Norilsk emissions are not wellquantified, so no clear conclusions can be drawn.

SO42� concentrations measured in air at monitoring stations in

the High Arctic (Alert, Canada; and Zeppelin, Svalbard) and atseveral monitoring stations in subarctic areas of Fennoscandia andnorthwestern Russia show decreasing trends since the 1990s,which corresponds well with Quinn et al. (2007). At many stationsthere are significant downward trends for SO4

2� and SO2 in air, bothsummer and winter. There are significant reductions of SO2 inSvanvik probably because emissions in the area are stronglyreduced. For the concentrations of nitrogen compounds in air thereis no clear pattern, but it is interesting to see a positive trend inthe total NO3

� concentration during summer at 3 stations. Totalammonium in air also has both positive and negative trends insummer.

SO42− summer

60° N

70° N

80° N

90° N

NH4+ summer

60° N

70° N

80° N

90° N

0.05

(December–February) and summer (June–August). No trend is shown as green. NO3� is

Table 2Statistically significant trends at the 10% level after 1990. For precipitation the unit is mm per year, for pH the unit is 10�7 Hþ ions l�1, for the chemical compounds the units are mg S l�1 and mg N l�1 for precipitation and mg S m�3 andmg N m�3 for air. n.a.¼ data not available. Empty cell means no significant non-zero trend at the 10% level.

Precipitationamount

Concentration in precipitation Concentrations in air

Hþ SO4 NO3 NH4 SO2 SO4 NO3 NH4

winter summer winter summer winter summer winter summer winter summer winter summer winter summer winter summer winter summer

Alerta n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. �0.02 �0.004Oulanka �2.8 �0.01 �0.01 �0.004 �0.07 �0.02 �0.02 �0.01 �0.003Irafloss n.a. n.a. �1.1 �1.4 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.Janiskoski 0.71 �0.01 �0.02 0.01 �0.05 0.003 0.01Pinega 0.65 �8 �0.03 �0.01 �0.06 0.002 0.03Tustervann �1.89 �2.8 �5.7 �0.01 �0.01 0.002 0.002Zeppelin n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. �0.02 �0.01Jergul/karasjok �5 �8.3 0.01 0.01 �0.02 �0.01Bredkal �3.6 �6.5 �0.01 �0.01 �0.007 �0.02 �0.01 �0.01 �0.02 �0.01Abisko �2.8 �6.2 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.Snare Rapids 2.47 �0.01 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.Homsund n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.Ny-Alesund �2 �0.05 0.01 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.Svanvik n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. �0.50 n.a. n.a. n.a. n.a. n.a. n.a.Nord n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. �0.02 �0.001 �0.004Zarechensk �1.68 �1.1 �0.01 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.Padun �0.02 �0.01 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.Krasnoshelie n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.Nariyan-Mar �1.1 �1.1 �0.04 0.04 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.Urengoy �2.58 �4 �1.2 1.1 �0.03 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.Turukhansk �1.96 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.Zhigansk n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.Deputatskiy n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.Ust-Moma 0.07 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.Palatka n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.

a Alert measures particulate nitrate and ammonium. The other stations measure total nitrate and ammonium.

L.R.H

oleet

al./A

tmospheric

Environment

43(2009)

928–939934

L.R. Hole et al. / Atmospheric Environment 43 (2009) 928–939 935

7. Air concentrations: comparison of observations withmodel results

The model has been run for the period 1991–2004 usingactual emission data as derived from the emission inventories.Figs. 7 and 8 show modelled monthly time series for SO4

2� andNOx together with observed concentrations in air for the stationswith longest data series in the Arctic. Total sulphur concentration(SOx – the sum of sulphur measured as SO2 and SO4

2�) showsa strong seasonal variation with high values in the winter/springand very low values in the summer. But the partition of thetwo components changes during the winter/spring season froma high content of sulphur dioxide during the dark winterperiod to a very low content after the polar sunrise in earlyspring.

Stations like Alert (Canada), Nord (Greenland) and Zeppelin(Svalbard-Norway) are not influenced by local or regional airpollution (e.g. Quinn et al., 2007), but measured concentrationsdepend strongly upon conditions along the pathway to the Arcticand on local meteorological conditions. In summer, atmospheric

1990 1995 2000 20050

0.5

1

g S

m

−3

Alert

R2 =0.65

1990 1995 2000 20050

0.2

0.4

0.6

0.8

g S

m

−3

Zeppelin

R2 =0.46

1990 1995 2000 20050

0.5

1

1.5

2

g S

m

−3

Oulanka

R2 =0.73

1990 1995 2000 20050

1

2

3

g S

m

−3

Pinega

R2 =0.51

Fig. 7. Comparison of measurements (solid lines) and DEHM model predictions (dashed lineare shown for each station.

substances from temperate latitudes hardly reach the Arctic, andduring winter, transport time can change significantly from week toweek (Stohl, 2006). Therefore, single measurements at a limitednumber of stations reflect only poorly the actual emissions in thesource areas, and trend analysis on the raw concentration datamight give misleading results. These should therefore be accom-panied by transport model analysis.

Variations of modelled concentrations from year to year at thestations are consequently mainly due to changes in the meteoro-logical conditions while the long-term trends to a larger extentreflect changes in emissions. In Fig. 7, SO4

2� concentrations in air areshown for stations with good long time series. Significant correla-tion at the 10% level, R2, is shown for each plot. Given the coarseresolution of the model, it is not expected to have very highcorrelation at these surface stations, where local meteorology canhave strong influence. Vertical concentration profiles would havebeen better suited for comparison. The poor correlation at Jan-iskoski is surprising, given the reasonably good fit to the measureddata after 1999. From the plot it appears like there have beenproblems with the measurements in the early 1990s, as the curve

1990 1995 2000 20050

0.2

0.4

0.6

0.8Nord

R2 =0.38

1990 1995 2000 20050

0.5

1

1.5Jergul

R2 =0.51

1990 1995 2000 20050

1

2

3Janiskoski

MeasurementsDEHM

s) of SO42� in air (mg S m�3) 1991–2004. Significant correlation coefficients (within 10%)

1990 1995 2000 20050

0.02

0.04

0.06

0.08Alert

1990 1995 2000 20050

0.05

0.1

0.15

0.2Nord

1990 1995 2000 20050

0.05

0.1

0.15

0.2

g N

m

−3

Zeppelin

1990 1995 2000 20050

0.1

0.2

0.3

0.4Jergul

1990 1995 2000 20050

0.1

0.2

0.3

0.4

g N

m

−3

Oulanka

R2 =0.36

1990 1995 2000 20050

0.1

0.2

0.3

0.4

Year

Janiskoski

R2 =0.19

1990 1995 2000 20050

0.1

0.2

0.3

0.4

Year

Pinega

R2 =0.11

MeasurementsDEHM

g N

m

−3

g N

m

−3

Fig. 8. Same as Fig. 6, but for total NO3� (mg N m�3). For Alert the measured data are particulate NO3

�.

L.R. Hole et al. / Atmospheric Environment 43 (2009) 928–939936

deviates from the generally decreasing trend seen at the otherstations.

For NOx in air (Fig. 8), correlations are lower than for SO42�. This

is to be expected since emission data are more uncertain. However,except for Pinega, the general concentration level is well inagreement with observations at most stations.

8. Historical and expected trends 2000–2030 with ‘‘constant’’climate

The DEHM model with extensive chemistry has been run withtwo different emissions scenarios: The ‘‘Business As Usual’’ (BAU)and the ‘‘Maximum technically Feasible Reduction’’ (MFR), asdescribed in Section 5 and in Hole et al. (2006b). For each emissionscenario the DEHM model has been run for the same meteorolog-ical input for the period 1991–1993 in order to reduce the meteo-rological variations of the model results.

Fig. 9a, b shows the mean concentrations and depositions ofSOx and NO3

� for the area north of Polar Circle for year 2000. Themost interesting feature is found in contribution of the Norilsksmelter in central Siberia to SOx deposition. It also appears that thepollution penetrates further north in the eastern Arctic comparedto the western Arctic. This is in accordance with the polar domeconcept (see e.g. Stohl, 2006 and Iversen and Jordanger, 1985) andis a result of differences in circulation patterns and highertemperatures in the Barents Sea region which allows air massesfrom temperate regions to move to higher latitudes withoutisentropic lifting. Also, in Siberia, wintertime emissions take placeinside the polar dome, and pollution can thus penetrate north atlow altitudes in winter.

In Fig. 10 we present the overall variation of concentrationand deposition of SOx and total NOx

þ and NHx in the Arctic since1860, based on DEHM model runs and emission data asdescribed in Sections 4 and 5. The patterns for NHx and NOx are

a b

Fig. 9. SO42� (a) and NO3

� (b) deposition in 2000 as estimated by the Danish Eulerian Hemispherical Model.

L.R. Hole et al. / Atmospheric Environment 43 (2009) 928–939 937

very similar to each other. It is not clear why concentrations anddeposition do not have exactly the same trend, but changes intemperature and precipitation patterns will influence thehistorical deposition record. The general trend with an acceler-ating deposition during the 19th century and a decline afterabout 1980 corresponds well with ice core observations such asWeiler et al. (2005).

9. Discussion

The data presented here leave the impression that there aresignificantly decreasing concentrations of acidifying compounds inair and precipitation in the Arctic. This is particularly true forsulphur compounds and in Fennoscandia and Northwestern Russia(Fig. 3 and Table 2). For other areas, the data coverage is limited. InEastern Siberia trends are basically non-existent, but backgroundconcentrations are also very low here. The acidity of precipitation isalso reduced (increasing pH) at many stations (Fig. 5 and Table 2),but this development does not follow exactly the same pattern asSOx and NOx trends partly due to its influence by basecations.Interestingly, the acidity in precipitation has increased in the Ure-ngoy station (Western Siberia) in summer. In addition, in theeastern Siberia an increase in the SO4

2� concentration in precipita-tion was detected. Even if the changes are relatively small theycould be of concern with regard to the Arctic environment. Thereare only few stations reporting the level of the compounds in air. Allthe significant trends in NO3

� concentration in summer werepositive, whereas in winter only one station reported increasingNO3� concentration. For NH4

þ there are some stations with reduc-tions in Fennoscandia, but also a few with increasing trends in the

same area. This unclear pattern is in accordance with observationsin Hole et al. (2008a). It has earlier been suspected that decreasingNHy-deposition in continental Europe due to lower surface NHy

affinity could lead to higher NHy-concentrations in remote areas(e.g. Sutton et al., 2003), but this is not confirmed by our findings.

For the period 1990–2001, the DEHM model is able toreproduce 60–85% of the variability for SOx in air but only 30–60% for NOx at a range of Arctic stations where there is significantcorrelation with observations (Figs. 7 and 8). For SOx there issignificant correlation at 6 out of 7 stations, whereas for NOx thecorresponding number is 3 out of 7. Emissions of sulphur aremore well defined than nitrogen emissions and dry deposition ofnitrogen in northern regions is a big uncertainty Hole et al.(2008b), so this difference is not surprising. It is shown elsewhere(i.e. Hole and Engardt, 2008) that changes in precipitationpatterns due to climate change will have a significant influenceon the deposition of sulphur and nitrogen compounds innorthern regions.

For the BAU emission scenarios, northern hemisphere sulphuremissions will only decline from 52.3 megatonnes (mt) to 51.3 mtfrom 2000 to 2020. For the MFR scenario 2020 emissions will beonly 20.2 mt. However, the two different scenarios show muchsmaller differences in concentration and deposition of sulphur inFig. 10. This is because the largest potential for improvement in SO2

emissions is in China and SE Asia. These regions have little influenceon the pollution of the Arctic boundary layer according to Stohl(2006) and others. For oxidized and reduced nitrogen compoundsthere are larger reductions in the emissions in Russia and Europe inthe MFR scenario, and hence the potential for improvements in theArctic is larger.

1850 1900 1950 2000 20500

0.2

0.4

0.6

0.8

µg S

m

−3

µg N

m

−3

µg N

m

−3

SOx Concentration

1850 1900 1950 2000 20500

500

1000

1500

2000

mg

S

m

−2 y

r−

1

mg

N

m

−2 y

r−

1

mg

N

m

−2 y

r−

1

SOx Deposition

1850 1900 1950 2000 20500

0.01

0.02

0.03NOx Concentration

1850 1900 1950 2000 20500

200

400

600NOx Deposition

1850 1900 1950 2000 20500

0.05

0.1NHy Concentration

Year

1850 1900 1950 2000 20500

200

400

600NHy Deposition

Year

Fig. 10. Development of SOx, NOx and NHx concentration in air and dryþwet deposition north of the Arctic circle from industrialization to 2020. Solid lines: Business As Usualscenario (BAU). Dashed lines: Most Feasible Reduction scenario (MFR).

L.R. Hole et al. / Atmospheric Environment 43 (2009) 928–939938

10. Conclusions

SO42� concentrations are decreasing significantly at many Arctic

stations. Also SO2� concentrations are reduced, reflecting the

significant reductions in sulphur emissions in Europe and NorthernAmerica. In particular the reduced emissions in the Eurasian regionis the major explanation due to this regions influence on the Arcticenvironment. For NO3

� and NH4þ in air and precipitation the pattern

is unclear (some positive and some negative trends), and there arealso few signs of significant trends in precipitation amount for theperiod studied.

There is relatively good monitoring data coverage in Fenno-scandia and on the Kola peninsula in Russia, but there are other-wise few stations for background air and precipitationconcentration measurements in the Arctic. Our analysis reveals fewcharacteristic differences between trends during summercompared to winter. An exception is NO3

� wet deposition which ishigher in winter in some stations in NW Russia and Fennoscandia(Pinega, Oulanka, Bredkal and Karasjok). The explanation for this isnot clear, but in Hole et al. (2006b) seasonal directional exposuredifferences for SO2 at Oulanka are revealed which can indicate thatdifferences in the synoptic transport may partly explain theseasonal patterns.

Because of new emission control technologies, a changedclimate and an increase in human activities, future emissions anddeposition patterns are particularly uncertain and difficult toproject in the arctic region. Increased extraction of naturalresources and increased sea traffic can be expected. Climate changeis also likely to influence transport and deposition patterns (ACIA,2004). There is a need for a deeper insight in plans and conse-quences with respect to the Arctic. Modelling results presented

here seem to rule out SE Asia as an important contributor topollution close to the surface in the Arctic atmosphere. This is inaccordance with earlier studies (e.g. Iversen and Jordanger, 1985;Stohl, 2006) giving thermodynamic arguments why SE Asianemissions will have minor influence in the Arctic.

Future increased occurrence of rain events in both summer andwinter can result in increasing wet deposition in the Arctic (ACIA,2004), but to study the effects of global warming on Arctic airpollution is beyond the scope of this paper.

Acknowledgements

Authors acknowledge the Voieykov Main Geophysical Obser-vatory, Roshydromet for providing data on Russian Arctic precipi-tation chemistry to the AMAP database. Data from Snare Rapids arekindly provided by the National Air Pollution Surveillance (NAPS)Network, a co-operative program of the federal, provincial, terri-torial and municipal government monitoring agencies in Canada.Other data were obtained from EMEP, www.emep.int. Thanks toAndrew Ruddock and two anonymous reviewers for their help withthe manuscript. The Danish Environmental Protection Agencyfinancially supported (part of) this work with means from theMIKA/DANCEA funds for Environmental Support to the ArcticRegion. The findings and conclusions presented here do notnecessarily reflect the views of the Agency.

References

Aas, W., Solberg, S., Manø, S., Yttri, K.E., 2008. Monitoring of Long Range Trans-ported Pollution in Norway. Atmospheric Transport, 2005 (in Norwegian).Norwegian Pollution Control Authority. TA-2423/2008. NILU OR 29/2008.Available from: www.nilu.no.

L.R. Hole et al. / Atmospheric Environment 43 (2009) 928–939 939

Arctic Climate Impact Assessment (ACIA), 2004. Impacts of a Warming Arctic.Cambridge University Press, 133 pp.

Barrie, L.A., Fisher, D., Koerner, R.M., 1985. Twentieth century trends in Arctic airpollution revealed by conductivity and acidity observations in snow and ice inthe Canadian high Arctic. Atmos. Environ. 19 (12), 2055–2063.

Barrie, L.A., 1986. Arctic air pollution: an overview of current knowledge. Atmos.Environ. 20, 643–663.

Christensen, J., Brandt, J., Frohn, L.M., Skov, H., 2004. Modelling of mercury in theArctic with the Danish Eulerian Hemispheric Model. Atmos. Chem. Phys. 4,2251–2257.

Christensen, J., 1997. The Danish Eulerian Hemispheric Model – a three dimensionalair pollution model used for the Arctic. Atmos. Environ. 31, 4169–4191.

Frohn, L.M., Christensen, J.H., Brandt, J., Geels, C., Hansen, K., 2003. Validation of a3-D hemispheric nested air pollution model. Atmos. Chem. Phys. 3, 3543–3588.

Flatøy, F., Hov, Ø., 1996. Three-dimensional model studies of the effect of NOxemissions from aircrafts on ozone in the upper troposphere over Europe andthe North Atlantic. J. Geophys. Res. 101, 1401–1422.

Frohn, L.M., Christensen, J.H., Brandt, J., 2002a. Development of a high-resolutionnested air pollution model: the numerical approach. J. Comput. Phys. 179 (1),68–94.

Frohn, L.M., Christensen, J.H., Brandt, J., 2002b. Development and testing ofnumerical methods for two-way nested air pollution modelling. Phys. Chem.Earth, Parts A/B/C 27 (35), 1487–1494.

Geels, C., Doney, S.C., Dargaville, R.J., Brandt, J., Christensen, J.H., 2004. Investigatingthe sources of synoptic variability in atmospheric CO2 measurements over thenorthern hemisphere continents: a regional model study. Tellus B 56 (1), 35–50.doi:10.1111/j.1600-0889.2004.00084.x.

Gilbert, R.O., 1987. Statistical Methods for Environmental Pollution Monitoring. VanNostrand Reinhold, New York.

Grell, G., Dudhia, J., Stauffer, D., 1994. A Description of the Fifth-generation PennState/NCAR Mesoscale Model (MM5). NCAR Tech. Note TN-398. Natl. Cent. forAtmos. Res., Boulder, Colo.

Guenther, A., Hewitt, C.N., Erickson, D., Fall, R., Geron, C., Graedel, T., Harley, P.,Klinger, L., Lerdau, M., McKay, W.A., Pierce, T., Scholes, B., Steinbrecher, R.,Tallamraju, R., Taylor, J., Zimmerman, P., 1995. A global-model of natural volatileorganic-compound emissions. J. Geophys. Res. 100 (D5), 8873–8892.

Hansen, K.M., Christensen, J.H., Brandt, J., Frohn, L.M., Geels, C., 2004. Modellingatmospheric transport of a-hexachlorocyclohexane in the northern hemispherewith a 3-D dynamical model: DEHM-POP. Atmos. Chem. Phys. 4, 1125–1137.

Heidam, N.Z., Christensen, J., Wåhlin, P., Skov, H., 2004. Arctic atmosphericcontaminants in NE Greenland: levels, variations, origins, transport, trans-formations and trends 1990–2001. Sci. Total Environ. 331 (1–3), 5–28.

Hertel, O., Christensen, J., Runge, E.H., Asman, W.A.H., Berkowicz, R., Hovmand, M.F.,1995. Development and testing of a new variable scale air pollution model –ACDEP. Atmos. Environ. 29, 1267–1290.

Hole, L.R., Brunner, S.H., Hansen, J.E., Zhang, L., 2008a. Low cost measurements ofnitrogen and sulphur dry deposition velocities at a semi-alpine site: gradientmeasurements and a comparison with deposition model estimates. Environ.Pollut. 154, 473–481 (Special issue on biosphere-atmosphere fluxes).

Hole, L.R., de Wit, H., Aas, W., 2008b. Trends in N deposition in Norway: a regionalperspective. Hydrol. Earth Syst. Sci. 12, 405–414.

Hole, L.R., Christensen, J., Ginzburg, V.A., Makarov, V., Polishok, A.I., Ruoho-Airola, T.,Vasilenko, V.N., 2006a. Concentrations and Deposition of Acidifying AirPollutants. AMAP assessment report, chapter 3.

Hole, L.R., Christensen, J., Forsius, M., Nyman, M., Stohl, A., Wilson, S., 2006b.Sources of Acidifying Pollutants and Arctic Haze Precursors. AMAP assessmentreport, chapter 2.

Hole, L.R., Engardt, M., 2008. Climate change impact on atmospheric nitrogendeposition in northwestern Europe – a model study. AMBIO 37 (1), 9–17.

Iversen, T., Jordanger, E., 1985. Arctic air pollution and large scale atmosphericflows. Atmos. Environ. 19, 2099–2108.

Jonson, J.E., Kylling, A., Berntsen, T., Isaksen, I.S.A., Zerefos, C.S., Kourtidis, K., 2000.Chemical effects of UV fluctuations inferred from total ozone and troposphericaerosol variations. J. Geophys. Res. 105, 14561–14574.

Kamari, J., Joki-Heiskala, P. (Eds.), 1998. Acidifying Pollutants, Arctic haze, andAcidification in the Arctic. Arctic Monitoring and Assessment Programme.AMAP assessment report ch. 9, pp. 621–658. Available from: www.amap.no.

Kylling, A., Bais, A.F., Blumthaler, M., Schreder, J., Zerefos, C.S., Kosmidis, E., 1998. Theeffect of aerosols on solar UV irradiances during the photochemical activity andsolar radiation campaign. J. Geophys. Res. 103, 21051–26060.

Levine, S.Z., Schwarz, S.E., 1982. In-cloud and below-cloud scavenging of nitric acidvapor. Atmos. Environ. 16, 1725–1734.

Logan, J.A., 1983. Nitrogen oxides in the troposphere; global and regional budgets.J. Geophys. Res. 88, 10785–10807.

Macdonald, R.W., Harner, T., Fyfe, J., 2005. Recent climate change in the Arctic andits impact on contaminant pathways and interpretation of temporal trend data.Sci. Total Environ. 342, 5–86.

Quinn, P.K., Shaw, G., Andrews, E., Dutton, E.G., Ruoho-Airola, T., Gong, S.L., Feb2007. Arctic haze: current trends and knowledge gaps. Tellus B 59 (1), 99–114.

Salmi, T., Maatta, A., Anttila, P., Ruoho-Airola, T., Amnell, T., 2002. Detecting Trendsof Annual Values of Atmospheric Pollutants by the Mann–Kendall test and Sen’sSlope Estimates – the Excel Template Application MAKESENS. FMI-AQ-31. In:Publications on Air Quality, 31. FMI, Helsinki, Finland.

Schwarz, S.E., 1979. Residence times in reservoirs under non-steady-state condi-tions: application to atmospheric SO2 and aerosol sulphate. Tellus 31, 520–547.

Sen, P.K., 1968. Estimates of the regression coefficient based on Kendall’s tau. J. Am.Stat. Assoc. 63, 1379–1389.

Seinfeld, J.H., Pandis, S.N., 1998. Atmospheric Chemistry and Physics: from AirPollution to Climate Change. John Wiley & Sons, Inc., New York.

Simpson, D., Fagerli, H., Jonson, J., Tsyro, S., Wind, P., 2003. Transboundary Acidi-fication, Eutrophication and Ground Level Ozone in Europe. Part 1. Technicalreport, Norwegian Meteorological Institute. Available from: www.emep.int.

Stohl, A., 2006. Characteristics of atmospheric transport into the Arctic troposphere.J. Geophys. Res. 111. doi:10.1029/2005JD006888 D11306.

Sutton, M.A., Asman, W.A.H., Ellermann, T., van Jaarsveld, J.A., Acker, K., Aneja, V.,Duyzer, J., Horvath, L., Paramonov, S., Mitosinkova, M., Tang, Y.S.,Achtermann, B., Gauger, T., Bartniki, J., Neftel, A., Erisma, J.W., 2003. Establishingthe link between ammonia emission control and measurements of reducednitrogen concentrations and deposition. Environ. Monit. Assess. 82, 149–185.

Weiler, K., Fischer, H., FritzscheRuth, U., WilhelmsFand Miller, H., 2005. Gla-ciochemical reconnaissance of a new ice core from Severnaya Zemlya, EurasianArctic. J. Glaciology 51 (172), 64–74.