16year simulation of Arctic black carbon ... - Wiley Online Library

JOURNAL OF GEOPHYSICAL RESEARCH: ATMOSPHERES, VOL. 118, 943–964, doi:10.1029/2012JD017774, 2013

16-year simulation of Arctic black carbon: Transport, source contribution, and sensitivity analysis on deposition S. Sharma,1 M. Ishizawa,2 D. Chan,1 D. Lavoué,3 E. Andrews,4 K. Eleftheriadis,5 and S. Maksyutov6 Received 11 March 2012; revised 28 October 2012; accepted 14 November 2012; published 19 January 2013.

[1] Arctic regional climate is influenced by the radiative impact of aerosol black carbon (BC) both in the atmosphere and deposited on the snow and ice covered surfaces. The NIES (National Institute for Environmental Studies) global atmospheric transport model was used, with BC emissions from mid-latitude fossil fuel and biomass burning source regions, to simulate BC concentrations with 16 year period. The model-simulated BC agreed well with the BC observations, including the trends and seasonality, at three Arctic sites: Alert (Nunavut, Canada), Barrow (Alaska, USA), and Zepplin, Ny-Ålesund (Svalbard, Norway). The equivalent black carbon (EBC, absorption inferred BC) observations at the three Arctic locations showed an overall decline of 40% from 1990 to 2009; with most change occurring during early 1990s. Model simulations confirmed declining influence on near surface BC contribution by 70% , and atmospheric BC burden by one half from the Former Soviet Union (FSU) BC source region over 16 years. In contrast, the BC contribution from the East Asia (EA) region has little influence at the surface but atmospheric Arctic BC burden increased by 3 folds. Modelled dry deposition is dominant in the Arctic during wintertime, while wet deposition prevails at all latitudes during summer. Sensitivity analyses on the dry and wet deposition schemes indicate that parameterizations need to be refined to improve on the model performance. There are limitations in the model due to simplified parameterizations and remaining model uncertainties, which requires further exploration of source region contributions, especially from growing EA source region to Arctic BC levels in the future is warranted. Citation: Sharma, S., M. Ishizawa, D. Chan, D. Lavoue´, E. Andrews, K. Eleftheriadis, and S. Maksyutov (2013), 16-year simulation of Arctic black carbon: Transport, source contribution, and sensitivity analysis on deposition, J. Geophys. Res. Atmos., 118, 943–964, doi:10.1029/2012JD017774.

1. Introduction [2] Light-absorbing, short-lived (e.g., less than a few days to weeks) black carbon (BC) aerosols, have a positive radiative feedback on the climate and contribute to climate warming, adding to the effect of greenhouse gases [Forster et al., 2007]. However, in contrast to longer lived greenhouse gases, mitigation of atmospheric BC emissions could have an “immediate climate impact” as atmospheric levels of 1 Climate Research Division, S&T Branch, Environment Canada, Toronto, Ontario, Canada. 2 Toronto, Ontario, Canada. 3 DL Modelling and Research, Brampton, Ontario, Canada. 4 CIRES, University of Colorado, Boulder, Colorado, USA. 5 ERL, Demokritos National Center of Scientific Research, Institute of Nuclear Technology and Radiation Protection, Attiki, Greece. 6 Center for Global Environmental Research National Institute for Environmental Studies 16-2 Onogawa, Tsukuba, Ibaraki, Japan.

Corresponding author: S. Sharma, Environment Canada, 4905 Dufferin Street, Toronto, Ontario, M3H5T4, Canada. ([email protected]) ©2012. American Geophysical Union. All Rights Reserved. 2169-897X/13/2012JD017774

short-lived forcers deplete faster in response to reduced emissions [Baron et al., 2009; UNEP 2011]. Also in contrast with greenhouse gases, BC aerosols have an additional effect on climate through surface deposition. This effect is strongest when BC is deposited onto the snow and ice surfaces, where it can reduce the snow albedo and thereby perturb the regional Arctic radiative balance [Clarke and Noone, 1985; Flanner et al., 2007; Wang et al., 2011; Warren and Wiscombe, 1985]. The darkening of Arctic snow and ice by BC deposition may be one of the processes contributing to the rapid melting of sea-ice in the Arctic in recent decades [Hansen and Nazarenko, 2004]. [3] The Arctic atmosphere is influenced by emissions from a variety of BC sources, in particular fossil fuel and biofuel combustion, and both natural and anthropogenic biomass burning (e.g., wildfires, deforestation fires and agricultural fires). The vertical distribution in the atmosphere of BC aerosols from such sources is highly variable. Transport of aerosols from various sources occurs along the isentropic surfaces and vertical distribution thus reflects the winter/ summer stratified structure of the Arctic atmosphere. Consequently, haze from fossil fuel emissions that are dominant

943

SHARMA ET AL.: 16 YR TRANSPORT OF BC TO THE ARCTIC

during winter remain near the surface [e.g., Sharma et al., 2004; 2006; Hirdman et al., 2010] while smoke from biomass burning, which is largely a spring/summer phenomenon tends to be highly stratified but mostly present at higher altitudes. [Lavoué, 2000; Hsu et al., 1999; Stohl et al., 2006; 2007; Warnecke et al., 2009; Brock et al., 2011]. Atmospheric transport of smoke typically occurs at high altitude because forest fire emissions can be injected well above the boundary layer during the flaming phase [Lavoué et al., 2000, 2009] and even into the lower stratosphere by pyrocumulus clouds [Fromm et al., 2005]. Due to the latitudinal locations of the boreal fires, BC emissions may be transported quasi-isentropically to the Arctic surface resulting in higher BC concentrations above the polar dome, and lower concentrations at the surface. However, due to occurrence of some synoptic events during winter and spring, high BC concentrations from fires can reach at the surface [Stohl et al., 2006, 2007]. [4] Particulate BC concentrations in the Arctic atmosphere have also been shown to be highly seasonal. During the winter, BC from anthropogenic combustion processes at midlatitudes is transported to the Arctic surface contributing to the phenomenon referred to as Arctic haze [Rahn 1981; Barrie, 1986; Law and Stohl, 2007; Quinn et al., 2007; Sharma et al., 2006]. In contrast, summertime Arctic atmospheric BC concentrations at the surface are generally much lower [Sharma et al., 2004; Garrett et al., 2011]. Many factors influence this strong intra-annual variability in concentrations. During the summer, the advection of warmer air masses from lower latitudes over the Arctic cold surfaces favors the formation of stratus type cloud. These low clouds and associated drizzle enable the removal of atmospheric pollutants, including smoke, through in-cloud and belowcloud scavenging [Raatz, 1991]. In contrast, several factors contribute to the higher concentrations of pollutants in the lower layers of the atmosphere during polar nights. First, colder air masses are drier, resulting in lower precipitation amounts. This means that wet deposition of particles in the winter/spring period may not be as important as in the summer. Second, poleward atmospheric transport of pollution from the mid-latitudes to the Arctic in the winter and spring is more frequent than in summer [Klonecki et al., 2003]. Strong emission areas in the northern part of the temperate zone are found within the polar vortex (or polar anticyclone) due to its expansion southwards during the winter, while during summer the polar vortex retreats to within the higher latitude region. Finally, the northern region boundary layer is shallow in the winter, enhancing BC concentrations at low altitudes, which may then be mixed to the surface. To simulate the seasonality of particulate concentrations in chemical transport models, it is necessary to include the above mentioned processes in the numerical models. [5] State-of-the-art global transport models have been recently used to simulate BC concentrations in both source and in remote regions [Croft et al., 2005; Koch and Hansen, 2005; Koch et al., 2009; Shindell et al., 2008; Huang et al., 2010a, 2010b; Bourgeois and Bey, 2011; Liu et al., 2011]. Overall, most of these models underpredict BC near-surface concentrations compared to observations, especially in the Arctic region. This could be due to uncertainties in the emissions data [Bond et al., 2004] and/or poorly modeled chemical, transport and deposition processes [Koch et al., 2009].

A recent model intercomparison project [Textor et al., 2006, 2007] showed that the use of the same BC emission inventory did not improve the discrepancies between simulations by different models. This result suggests that the poorly modeled processes and parameterizations are largely responsible for the underestimation of BC concentrations. [6] In general, it has been found that most advanced models underestimate BC at northern latitudes by overestimating BC scavenging (especially wet deposition) [Koch et al., 2009; Bourgeois and Bey, 2011], and vertical mixing during poleward transport [Koch et al., 2009]. These models include various aerosol schemes, such as black carbon transformation from fresh to aged particles, microphysical mixing schemes [Vignati et al., 2010; Liu et al., 2011], wet and dry deposition algorithms [e.g., Petroff and Zhang, 2010], and BC removal by in-cloud processes for different types of clouds and phases [Koch et al., 2009]. These aerosol transformation and deposition schemes are characterized by different levels of uncertainties. Because these schemes are complex and interact with one another, it is difficult to determine which process is contributing to the underestimation of surface BC concentrations in any given transport model. These complex model simulations are typically conducted for short time periods (one to two years) and have not addressed interannual variations in BC source contributions over longer time frames [Koch and Hansen, 2005; Shindell et al., 2008; Bourgeois and Bey, 2011]. [7] As anthropogenic BC concentration levels have bearing on the development of BC mitigation policies [UNEP, 2011], the purpose of this study is to examine (using an atmospheric transport model) the impact of anthropogenic BC emissions from different Northern Hemispheric regions on the Arctic, on a decadal time scale. This study will mainly focus on the wintertime period when the BC concentrations are higher in the surface measurement records. Several studies have applied various numerical and statistical methods to determine the BC source regions and their contributions to BC levels at measurement locations in the Arctic: (1) by using proportions of clusters of 10-d back-trajectories to determine different source regions (Potential Source Contribution Function or PSCF method) at Alert and Barrow [Sharma et al., 2004, 2006], and Ny-Ålesund [Eleftheriadis et al., 2009]; (2) by calculating transport functions, from Eurasia and North America, determined from 700 hPa geopotential height anomalies [Gong et al., 2010] - BC source contributions at Alert were deduced from these transport functions; (3) and by applying the Lagrangian particle dispersion model FLEXPART, in combination with BC measurements, to determine emission source regions and sensitivities of these regions to equivalent BC measurements at Alert, Zeppelin, Barrow, and Summit [Hirdman et al., 2010]. [8] In contrast, this study adopts a different approach by using the global atmospheric transport model of the National Institute for Environmental Studies (NIES) [Maksyutov and Inoue, 2000] to simulate BC in the Arctic over a 16-year period. The model includes various parameterizations for dry and wet aerosol deposition, to simulate BC in the Arctic over a 16-year period. Wintertime BC modelling scenarios were run to determine (a) which mid-latitude fossil fuel emission source regions have a predominant impact in the Arctic; and (b) how the respective source region contributions have changed over the 16-year period. Summertime simulations

944


were also included for the purpose of comparing model output to measurements, and to determine the impact of biomass burning emissions on the Arctic BC surface concentrations, atmospheric burden and budget. These issues are addressed by testing model sensitivities to BC wet/dry aerosol deposition parameterizations. The following four questions are addressed: (1) Are anthropogenic BC emissions and/or atmospheric transport of BC the governing factors controlling BC concentrations in the Arctic? (2) How have the contributions to Arctic BC concentration from the main anthropogenic BC source regions changed over the 16-year period? (3) What is the relative importance and sensitivity of the Arctic BC column burden to emissions from different source regions, including biomass burning source regions? (4) What role does deposition (wet and dry) play in the accumulation of BC on Arctic snow and ice surfaces?

2. Method 2.1. Transport Model [9] The NIES model has a horizontal resolution of 2.5 2.5 with 15 vertical layers defined as s = p/p0, corresponding to the ratio of layer pressure (p) to surface pressure (p0). Values of s range from 0.985 to 0.065 and correspond to approximately 0.15 km to 20 km in altitude, respectively. Simulations were performed with a time step of two hours. The advection scheme, implemented in the NIES model, is semi-Lagrangian with mass adjustment to conserve tracer mass. The global wind fields used in our modeling exercise were obtained from the NCEP (National Center for Environmental Prediction) reanalysis [Kalnay et al., 1996] with a 2.5 2.5 horizontal resolution and a 12-hour temporal resolution. NCEP winds were linearly interpolated to the two-hour model time-step. The seasonal variation of the planetary boundary layer (PBL) height at each grid point in the model is specified with the climatological monthly mean heights obtained from the Data Assimilation Office at NASA’s Goddard Space Flight Center. The well-mixed state in the PBL is accomplished with a constant turbulent diffusion coefficient set to 40 m2 s1. The NASA climatological PBL heights do not capture the synoptic variability of the PBL near the mid-latitude source region. But the impact of this effect is reduced with the diffusive transport of BC to the Arctic observation sites (typically 7–10 days) as the transport from the mid-latitude source region to the Arctic is strongly dependent on the horizontal (diffusive) transport. Previous studies using the NIES model (Worthy et al., 2009, Chan et al., 2008; Ishizawa et al., 2006; Patra et al., 2005; Fujita et al., 2003) were able to simulate well the observed CO2 and CH4 seasonally and interannually. Thus the NIES model is used in this study to simulate the seasonal and interannual variations in BC. The important new additions to the model are the wet and dry BC removal processes described in the next section. Other vertical mixing processes included in the model are turbulent diffusion, as a function of the vertical temperature stability, and cumulus convection derived from the humidity, temperature, and wind fields obtained from the NCEP reanalysis.

[10] The NIES model has been extensively used for forward simulation and inversion modeling of greenhouse gases in previous studies (e.g., Chan et al., 2008; Fujita et al., 2003; Patra et al., 2005). The model performance has been validated for the transport of gases such as CO2 and CH4 to the Arctic [Ishizawa et al., 2006; Worthy et al., 2009]. 2.2. Dry and Wet Removal Schemes [11] The NIES model employs simplified wet and dry aerosol removal parameterizations, applied to a static particle size, for the wet and dry removal processes. No other atmospheric microphysical processes (e.g., particle growth, condensation, and coagulation) were considered in the simulations and all particles were assumed to be hygroscopic. Selection of the dry and wet deposition parameters was based on considerations described below. Long-range transport of BC from mid-latitude source regions to the Arctic takes several days to weeks, as determined by 10-d back-trajectory cluster analysis [Sharma et al., 2004, 2006; Hirdman et al., 2010]. Croft et al. [2005] suggested that in the aging of fresh BC released from sources, nonhygroscopic (i.e., hydrophobic) BC is transformed into hygroscopic (i.e., hydrophilic) aerosol in less than a day. Therefore, most BC en route to Arctic receptor sites becomes hygroscopic during transport, and was assumed to favour water uptake in the present modeling study. Simulation of the BC aging process is more complex in chemical transport models as they take into effect sulfuric acid production [e.g. Koch, 2001] and aerosol number density [Riemer et al., 2004 and Croft et al., 2005] on the BC aging rate. The simplified approach used in this model might result in higher BC deposition rates due to the lack of BC aging that depends on presence of sulphuric acid, especially in clean remote regions. [12] The dry deposition flux was assumed to be proportional to BC concentration in the lowest model layer by using a constant dry deposition velocity: Fd ¼ C z Vd

(1)

where Fd is the deposition flux (ng m2 s1), Cz is the BC concentration (ng m3) in the lowest layer, and Vd is the dry deposition velocity (m s1). Vd depends on the surface characteristics and meteorological conditions, and therefore is highly variable. Numerical schemes of Vd currently implemented in aerosol transport models have large uncertainties [Petroff and Zhang, 2010] and estimation of these uncertainties to the present study is beyond the scope of this paper. The dry deposition Vd = 0.05 cm s1 was used for the control case simulations. A more in-depth discussion on the choice of control case dry deposition is given in section 3.4.1. [13] Some studies [e.g., Bourgeois and Bey, 2011] suggested wet deposition to be the dominant removal process for BC in the Arctic during winter. Wet removal occurs in and below clouds, namely with in-cloud and below-cloud scavenging. The wet removal parameterization for the stratiform, below-cloud scavenging in Croft et al. [2005] was implemented into the NIES model. The wet deposition occurring in the model levels below the clouds was assumed to be proportional to the precipitation rate at the surface and applied equally to snow/rain/drizzle. Clear sky ice crystal

945


C

S3

S2

S4

S4 S3 C

S2 S1

Figure 1. Wet scavenging coefficients considered in this study are taken from Croft et al. [2009] (modified Figure 2). The curves indicate the wet scavenging coefficients as a function of aerosol radius for various types of rain and snow. For the control run for this study, static scavenging coefficients used are 0.1 (2.8 105s1) and 0.005 m2kg1(1.4 106s1) for the rain and snow as indicated by the red star and circle, respectively. These values are chosen with the assumption that the bulk of the aerosol is in the accumulation mode and there is no distinction between in- and below-cloud scavenging. The sensitivity parameters used for scavenging coefficients are also shown for control case (C) and changes to control case labeled S1 to S4. precipitation is not included in the model. The change in BC concentration over a time step is calculated with: @C ¼ C R Pr @t

(2)

where C = BC concentration (kg m3), R = scavenging coefficient (m2 kg1), and Pr = precipitation rate (kg m2 s1), and @t = time step (s). [14] Figure 1 has been modified from Croft et al., 2009, which shows increasing scavenging as a function of increasing particle size distribution. Fresh emissions from BC

sources release small, black carbon particles in large numbers with a modal size range of less than 40 nm (i.e., vehicular emissions; personal communication, John Liggio, Environment Canada, 2012). These particles age and grow to larger sizes during transport to the Arctic. The ranges of particle sizes chosen for the control and sensitivity test cases lie within the accumulation size range of 50 to 500 nm as shown in Figure 1 (shaded light blue region). The scavenging coefficients (marked as C in Figure 1), used in the control case, correspond to a static radius of 200 nm and was

946


arbitrarily chosen as median radius of the accumulation size range between 50 to 500 nm range selected. The values of 0.1 and 0.005 m2 kg1 (2.8 105 and 1.4 106 s1) were used for rain (red star) and snow (red circle). [15] The choice of particle size is critical since both dry and wet removal schemes depend on it. The static particle radius of 200 nm, used for all simulations, is a factor of 2 higher than modal radius measured in the particle number distribution in the Arctic, but is chosen to qualitatively account for particles with radii >80 nm that deposit preferentially en route to the Arctic. Measurements made during Arctic haze events at Alert (Leaitch et al., Dimethyl Sulphide Control of the Clean Summertime Arctic Aerosol and Cloud, for Nature-Geoscience, 2012), Ny-Ålesund, Zeppelin Station [Ström et al., 2003] and in the background Arctic air [Brock et al., 2011] show particle modal radius around 90 nm. The NIES model, with horizontal resolution of 2.5 2.5 cannot properly simulate cloud scale processes to give cloud base and cloud top heights. The in-cloud scavenging is approximated by applying the same process in the cloud up to the height of s = 0.5 (approximately 500 mb). The deep cloud layer and larger scavenging coefficients could partly compensate for the stronger in-cloud scavenging process [Croft et al. 2010]. Snow and rain were distinguished using the surface skin temperature. Daily mean surface temperatures and daily precipitation rates were provided from the NCEP reanalysis surface dataset. Distinguishing rain and snow scavenging processes by surface temperature is not always correct. Particularly for rain events near and above 0 C, the scavenging above the surface to 500mb could be due to ice processes. This error is more likely in the spring and fall seasons and less likely in the winter season that this study is focused on. 2.3. Black Carbon Emissions [16] BC emissions from fossil fuel combustion and biomass burning (i.e., vegetation fires and agricultural burning) were included in the present study as these sources are dominant over other sources at the mid- and high-latitudes [see Table 1]. Emissions from biofuel combustion were not included in the BC emission inventories, although there is evidence that residential wood combustion in northern Europe during winter may be a significant source of BC [UNEP, 2011]. Emissions inventories used in earlier modeling studies are also listed in the Table 1 for comparison. [17] BC emissions in the present study and in Huang et al. [2010a, 2010b] use the same methodology but cover different years, and therefore give different totals. Moreover, emission inventories used in the modeling studies in Table 1 are not from the same year as the simulation period. It is crucial to use the same year for both meteorology and emissions to get the correct transport and atmospheric loading, especially when highly variable biomass burning emissions are included. 2.3.1. Fossil Fuel Combustion [18] Global emission inventories of anthropogenic BC were generated annually for 1990–2005 at a horizontal resolution of 1 1 . The same emission inventories through 1998 were previously used by Sharma et al. [2004] to interpret the trend of atmospheric BC concentrations at Alert. These emissions were subsequently updated through 2005 in Gong et al. [2010]. [19] Annual emissions were calculated by country, utilizing the methodology of Cooke et al., [1999] applied to the

United Nations global fuel consumption dataset for 19902005 [United Nations, 2007]. Since it has been previously shown that northern mid-latitude emissions influence BC concentrations in the Arctic [Sharma et al., 2004, 2006; Hirdman et al., 2010; Huang et al., 2010c], the emitting countries in the northern mid-latitudes were grouped into four regions: North America (NA), the former Soviet Union (FSU), Europe (EU), and East Asia (EA) (Figure 2a). BC emissions from these four regions are the focus of this study (Figure 2b). Other BC emission inventories available in the literature do not provide the interannual variation in BC emissions required for our modelling needs. For example, Bond et al. [2007] developed global 10-year averaged emissions from the beginning of the industrial era until the year 2000 (http://www.hiwater.org, accessed 6 December 2010). These emissions were updated by Lamarque et al. [2010] for IPCC decadal historical scenarios. A third inventory from the IIASA GAINS model (http://gains.iiasa.ac.at; Cofala et al., 2007) was used in a recent United Nations Environmental Program report [UNEP, 2011]. [20] In comparison, the average decadal BC fossil fuel (hereafter FF) emissions of 6.6 Tg for 1990 to 2000 in the present study is a factor of 2 higher than Bond et al.’s [2007] estimate for fossil fuel alone (excluding biofuels as these were not included in our inventory), although our estimate lies within their stated uncertainty range between 1 and 14 Tg BC per year estimated by Bond et al. [2007]. For the same time period, anthropogenic BC emissions from IIASA GAINS [Cofala et al., 2007] and Lamarque et al. [2010] agree well with the Bond et al. [2007] estimates. However, in a recent modeling study, Wang et al. [2011] used 7 Tg yr1 BC which is closer to the emissions values used in this study. To improve on the agreement of model output to the BC measurements, Wang et al. increased FF emissions estimates from Russia and Asia by a factor of two above the original Bond et al. [2007] emissions used in their simulations. The differences between Cooke et al. [1999] and others are likely due to the different emission factors used for various fuel types in each of these emission inventories [e.g., Junker and Liousse, 2008]. 2.3.2. Biomass Burning [21] For the present study, 7-year BC emissions from the Global Fire Emissions Database GFEDv3.1 [van der Werf et al., 2010] with the spatial resolution of 1 1 were included in the simulations (hereafter GFED). Monthly fire BC emissions were estimated from a combination of biogeochemical modeling and satellite derived area burned, fire activity, and plant productivity with an original spatial resolution of 0.5 0.5 . In this study, the biomass burning (hereafter BB) term is used and refers to the GFED fire database that includes vegetation fires, land clearing fires, and agricultural crop burning. [22] Emission fields in the GFED are available from 1997 onwards. Figure 3 shows the interannual variability in total BC emissions from global BB for 1997 to 2005. For the model runs, regions with fire emissions were divided in a similar manner to fossil fuel emissions (Table 1). Unlike fossil fuel BC emissions, there are large year-to-year changes in total and regional fire BC emissions, as illustrated in Figure 3. Additionally, BC aerosols from fire emissions are injected at different altitudes, depending on type of vegetation and thus the nature of the fire [Lavoué et al., 2000].

947

948

decadal average

a

BB UNEP 2011 FF BB

FF

2008 transport FF BB

2001 transport FF BB - Domestic and agricultural fires and Tropical and savannah - Boreal and temperate fires

FF 2001 transport

FF 1990-2005 transport BB 1997-2005 transport

Emission sources

1980 1981

Liousse et al. [1996]

2000 2005

Cofala et al. [2007]

5.5 3-3.7

2.61

-------

0.11

0.48

Lamarque et al. [2010] 2000 5.02a a

-------

0.27 -

0.67

0.5-0.9

0.7 (0.6–0.9) 0.09 (0.01–0.2)

NA

5.14 3.05

2008 2008

Bond et al. [2007] Junker and Liousse [2008] GFEDv2 van der Werf et al. [2006]

Jacob et al. [2010] FLAMBE (Reid et al. [2009]) GFEDv3 (van der Werf et al. [2010])

5.55 0.33

4.96

Bourgeois and Bey [2011]

1984

Lavoué et al. [2000]

Total

7.3 (6.3–9.3) 1997–2005 2.2 (1.5–2.9) Shindell et al. [2008] Variable 20% less BC scenario Huang et al. [2010a, 2010b, 2010c]

This study 1990–2005

Emission years used

Cooke et al. [1999]

Various methods

GFEDv3 van der Werf et al. [2010]

Cooke et al. [1999]

References

BC Emissions (Tg yr-1)

-------

0.02

1.64

-------

0.36 ----

1.83

1.5-2.1

2.6 (2.2–3.9) 0.02 (0.01–0.03)

EA

-------

0.01

0.51

-------

0.07 ----

1.81

0.7-2.1 (incl. FSU)

0.9 (0.8–1.0) 0.004 (0.001–0.003)

EU

-------

0.21

0.31

-------

0.07 ----

0.19

-

0.6 (0.4–1.3) 0.17 (0.06–0.4)

FSU

-------

2.26

2.09

-------

4.78 ----

----

-

2.5 (1.6–3.0) 1.9 (1.3–2.7)

Rest

Table 1. Average and Range of Annual Black Carbon Emissions from Fossil Fuel (FF) Combustion and Biomass Burning (BB) in Various World Regions. Emissions Used in the Present Study are Compared to Emissions Published in Previous Studies. It is Worth Noting that Similar Emission Methods Used in Different Studies Give Different Total BC Emissions Due to Different Emission Years are Used in the Simulation. The Boundaries of Political Regions Also Differ Among the Studies and Will Also Influence the Regional Emissions. SHARMA ET AL.: 16 YR TRANSPORT OF BC TO THE ARCTIC


(a)

(b)

Figure 2. (a) Locations of the mid-latitude BC source regions with geopolitical boundaries (EA = East Asia, FSU = Former Soviet Union, EU = Europe, and NA = North America. The three measurement sites are indicated with red dots; (b) annual anthropogenic BC emissions in source regions from 1990 through 2005; emissions have been increasing since 2000 especially in EA. In this study, the injection heights were set to constant values; 4–5 km in boreal NA and 2–3 km for other regions [Dentener et al., 2006]. Considering the seasonality of injection heights would mean taking into account the weather conditions in the calculation of the plume heights over 16 years, which is beyond the scope of the present study. The interannual variation of total and regional fire emissions is such that climatological average emissions from fires should not be used in model runs, as these averages cannot capture the large variability that occurs on an annual basis. 2.4. Equivalent Black Carbon (EBC) Observations [23] The EBC concentrations at the three Arctic sites were measured using an aethalometer model AE-6 (1-wavelength; Magee Scientific; Hansen et al. [1984]) at Alert, model

AE-8 (1-wavelength) at Barrow, and model AE-31 (7-wavelength) at Ny-Ålesund. The instruments measure the attenuation of light transmitted through particles that accumulate on a quartz fiber filter at 880 nm. A detailed description of the method is given for Alert by Sharma et al. [2004], for Ny-Ålesund by Eleftheriadis et al. [2009], and for Barrow by Bodhaine [1989]. Because there were no corrections applied for filter loading or scattering effects to the aethalometer measurements [Arnott et al., 2005], these measurements will be referred to as the “equivalent” black carbon (EBC). Some of the uncertainty stems from the use of a recommended value for the specific attenuation (k), of 19 m2 g1. This parameter, which is typically used to convert the light absorption to the black carbon mass, is based upon calibrations during instrument development and theoretical

949


Figure 3. Black carbon emissions from Global Fire Emissions Database Version 3.1. The regions were divided in a similar manner to those of fossil fuel BC emissions. The insert shows BC emissions excluding the rest of the world (others). Although magnitude of Savannah and tropical BB emission are dominant, their influence in the Arctic is less than 5% as shown in Table 4. calculations, and was employed in the present study. The accuracy of this factor for the Arctic has been explored in a comprehensive study at Alert [Sharma et al., 2004] and at Ny-Ålesund [Eleftheriadis et al., 2009], but not at Barrow. The specific attenuation k was found to be close to 19 m2g1 and 16 m2g1 for the winter/spring at Alert and Ny-Ålesund, respectively. The uncertainties in the value of k in the literature could be as much as a factor of 2, because k varies anywhere between 7 and 35 m2 g1 for various locations [Liousse et al., 1996; Sharma et al., 2002]. [24] All archived filters were examined from aethalometers at Alert and Barrow. Filters from Barrow indicated a possible instrument leak. The filter spots had dispersed edges for the first two years of the sampling period corresponding to the time period where the measured BC levels were low. The BC concentrations were adjusted for the dispersed spot size resulting in a 27% increase in BC but that was not enough to resolve the discrepancy. These two years of the Barrow observations were excluded from the analysis in this paper. 2.5. Simulation Cases [25] For the purpose of evaluating model performance with respect to seasonal and annual observations, both FF and BB emissions (available after 1997) were used in the simulations from 1997–2005. Simulations covering the longer 16-year period of 1990–2005 were mainly based on fossil fuel emissions because there is interest in understanding the contribution of this dominant source in winter to Arctic BC levels in order to help guide mitigation decisions. To investigate whether changes in transport alone, or in emissions and transport combined, affected the inter-annual variation

and wintertime trends of BC concentrations at the three Arctic sites, several model runs were conducted for the years 1990–2005. These influences were investigated in the following manner: 1) Case 1: fossil fuel combustion emissions were kept constant by using emissions for a single year (1990) to determine the effect of meteorology alone on the inter-annual variability of wintertime (January to March) BC concentrations for 16 years. 2) Case 2: yearly variable fossil fuel combustion emissions were used to determine which geographical regions influenced the trends and inter-annual variability in BC during the wintertime in the Arctic over the 16 years. 3) Under both cases 1, and 2, the NIES model output was compared to the winter time-series of EBC measurements at Alert, Barrow, and Ny-Ålesund (sections 3.2.1 and 3.2.2). In both cases, contributions from individual source regions (FSU, NA, EU and EA) were estimated by masking emissions from other regions. Near surface BC concentrations, atmospheric burden and deposition amounts were estimated. [26] In addition, various sensitivity tests were performed to assess the impact of dry and wet depositions on surface BC concentrations at the three Arctic sites shown in section 3.4.1 for winter (greater anthropogenic influence and drier atmospheric period) and summer (minimum anthropogenic period, predominant circumpolar fire influence and wetter period). Three different cases for the dry deposition velocity were used. First, a constant dry deposition rate of 0.05 cm s1 was used in the control run. In a second case, the control run

950


concentrations was shown to be related to declining emissions in the FSU during the 1990s. As an update from the previous analysis, Figure 4 shows a decline of ~40% (by using linear regression) between 1990 and 2009 in combined EBC measurements from the three Arctic stations (Alert, Barrow and Ny-Ålesund). Ny-Ålesund measurements have influence on this trend analysis after 2002 since the beginning of aethalometer sampling at this location. Despite the overall increase in fossil fuel BC emissions since 2000 in the source regions, especially in the EA region, the observed concentrations at the three Arctic locations did not increase. In the next sections, various simulation scenarios as described in the section 2.5 will address this feature.

Figure 4. Surface BC (EBC) measurements at the three Arctic sites, Alert (82 N, 62.3 W), Barrow (71 N, 156.6 W) and Ny-Ålesund (79 N, 12 E). There is overall decline of 40% in BC measurements in the Arctic from 1990 to 2009. In spite of increase in BC FF emissions in the source regions since 2000, there is no increase in the observed surface BC measurements at all three locations. dry deposition value was doubled to 0.1 cm s1 (enhanced case), a rate which has been commonly used as a bulk dry deposition velocity in various models [Chung and Seinfeld, 2002; Kim et al., 2008; Liousse et al., 1996; Wang, 2004; Wang et al., 2011]. A third, variable case used dry deposition values of 0.025 cm s1 over land (dry surface) and 0.2 cm s1 over the ocean (wetted surface) for the soluble/mixed BC as in Croft et al. [2005]. [27] For the wet deposition scheme, sensitivity tests were also performed by using four different sets of the scavenging coefficients besides the control case, as shown in Figure 3. The values used for the control case were multiplied by 0.5 (S2) and 2 (S3). For the extreme cases, the control scavenging parameters were modified as S1 with rc 0.01 and sc 0.2 for the low case and as S4 with rc 3 and sc 6 for the high case. The scavenging coefficients in the extreme cases were chosen to include the lowest and highest particle sizes corresponding to 50 and 500 nm particle radius of the accumulation mode.

3. Arctic BC Distribution, Burden and Deposition 3.1. Is the Arctic Surface EBC Increasing with Increasing BC Emissions? [28] The motivation of this modeling exercise emerged from the need to investigate the impact of increasing BC emissions from some source regions, specifically those from EA, on the Arctic (Figure 2b). The changes in BC fossil fuel emissions, between 1990 and 2005, show that emissions from the FSU declined in the early 1990s, following the collapse of the Soviet Union. Overall, BC emissions tended to stabilize in the mid-1990s, but have been on the rise since 2000 in all regions with the largest increase in EA [Peters et al., 2012]. A decline, from the early 1990s until 2003, was identified in the long term surface measurements of EBC during the springtime Arctic haze period at Alert in the Canadian high Arctic and at Barrow in the western Arctic [Sharma et al., 2006]. This decline in surface EBC

3.2. Spatial Distribution of BC Concentrations [29] Sixteen years of simulations were performed with time-dependent meteorological fields, while BC emissions were kept constant using the 1990 values (Figures 5a to 5f, left panel), and with variable emissions (Figures 5g to 5l, right panel). These simulations were conducted to investigate the impact of transport alone (with constant emissions) versus transport and variable emissions (yearly) together, for conditions at both the surface and 700 mb. [30] Two periods were considered in the following simulations; 1990-1994 and 2000-2005. These periods were selected for analysis on the basis of the changes in observed EBC (Figure 4) and BC emissions in the source regions (Figure 2b). During the first period, the largest change occurred in observed EBC concentrations and BC emissions in the FSU region. During the second period, BC emissions in the FSU did not change much and were a factor of 2.5 lower than in 1990–1991, but they increased in EA approximately by a factor of 2 from 1990 values (Figure 2b). A similar change to BC emissions in FSU was also observed in EBC concentrations at all measurement sites during the latter period as compared to 1990 period (Figure 4). 3.2.1. Constant BC Emissions and Variable Transport [31] The average wintertime BC spatial distribution (Figure 5 (a,d) (upper left panel)) is characterized by a large concentration gradient between the eastern (0–180oE) and western (0–180oW) Arctic regions at the surface (1) and at 700 mb (2). Higher BC source regions in the mid-latitudes are located in EA, western FSU, and EU. BC distribution anomalies are defined as differences of the mean concentrations for 1990–1994 (Figures 5b, 5e, 5h, and 5k; middle panel) and 2001–2005 (Figures 5c, 5f, 5i, and 5l; bottom panel) from the mean of the entire 1990–2005 time period. With constant emissions at 1990 levels, the 1990–1994 period (Figures 5b and 5e) shows a negative anomaly over northeastern Eurasia and positive anomaly over the Arctic at both altitudes. However, the 2001–2005 period shows positive anomalies over Eurasia and negative anomalies over the Arctic. A positive anomaly indicates higher BC in the atmosphere and a negative anomaly shows lower BC for that period with respect to the average over the 16 years. This indicates that transport from Eurasia to the Arctic was much more frequent during 1990 to 1994 than 2001 to 2005 periods. 3.2.2. Variable BC emissions during 1990–2005 [32] Figures 5g and 5j (right panel) show the average spatial distribution of BC concentrations calculated by taking into account the annual variability of fossil fuel emission sources and the changing transport that has occurred over

951


(B) Variable emissions

(A) Constant emissions

(a)

(d)

(g)

(j)

(b)

(e)

(h)

(k)

(c)

(f)

(i)

(l)

Figure 5. The four plots in the top panel show the spatial distribution of wintertime (January–March) mean BC concentrations at the (1) surface and at (2) 700 mb during 1990–2005 for both (A) constant and (B) variable emission scenarios. The middle and bottom panels present the anomalies in BC spatial distribution during 1990–1994 and 2001–2005, respectively, from the 1990–2005 period. The scale is in ng m3. 16 years. The anomaly plots (Figure 5h and 5k), show a strong positive anomaly over the western FSU source region that extends out over the eastern Arctic. More positive BC anomalies were determined over Eurasian source region that extends farther into the Arctic during 1990–1994 period under the variable emissions scenario compared to the constant 1990 BC emissions scenario. On the other hand, the surface anomaly simulated for 2001–2005 is strongly negative over the European region of FSU and extends to Siberia and the western Arctic (Figure 5i and 5l). This simulated anomaly is a result of decreased average emissions in FSU during this period. Moreover, the EA emissions (Figure 2b) during 2001–2005 increased compared to 1990–1994 values. The influence is evident in the positive anomaly in the BC concentrations over EA under the variable emission scenario. The simulation results suggest that the increase in emissions from EA has not influenced the BC concentrations at the Arctic near the surface but the impact is more evident aloft. 3.2. Model Comparison to Observations 3.2.1. Seasonal Cycle [33] The observed seasonal cycle of BC at the three sites is characterized by high BC concentrations during the winter/

spring Arctic haze, and lower concentrations during summer. Figure 6 shows average monthly BC concentrations and standard deviations for the model results under the variable emissions scenario compared to the observed concentrations at the three Arctic locations. At all the sites, the observed seasonality from in-situ measurements is reproduced by the model simulations reasonably well (regression coefficients between r = 0.8 and r = 0.94). There is more variation in BC during the winter/spring as indicated by large standard deviations compared with summertime values. The 16 years of model results agree with the observations within the 95% confidence levels, at both Alert and Barrow. For the Ny-Ålesund site, the model results for 2002-2005 were compared to the observations for the same time-period. The seasonal variability and magnitude of the modeled BC concentrations are larger than observations at Ny-Ålesund. [34] Observations were also compared to model simulations that included BB emissions. By including the BC emissions from BB in this model, the BC concentrations at all three locations (green curve) increase during the summer, compared to the FF scenario (red curve). However, observed EBC concentrations (black curve) during the summer show lower BC concentrations than modeled results (green curve).

952


Figure 6. Comparison of averaged seasonal variations between model simulations and observations at (a) Alert (1997–2005), (b) Barrow (1997–2005), and (c) Ny-Ålesund (2002–2005). Variable BC emissions were used in this modeling exercise. Monthly standard deviations are indicated by vertical bars on the plots. Monthly values were derived from daily values for the observations and from 6-hourly outputs for the simulations. FF is curve with the fossil fuel alone and FF + BB is curve with fossil fuel and biomass burning emissions. Using monthly biomass burning and fixed injection height in the model may not yield realistic simulated BC concentrations as both are characterized by a strong diurnal variability that cannot be reproduced in transport models at coarse resolution. The over prediction of BC concentrations during the summer by this model might also be due to the inability of this model to produce rain by convective clouds/fog that readily form during the summer in the Arctic. [35] Vertically, the model performance was within the measurement variability when compared to observations from three recent Arctic campaigns as shown in Figure 7. The altitudinal dependence of BC was determined by averaging model output vertically, with BB emissions included, for the region 70–90 N and 0–180 W longitude (pink region shown on the flight map in Figure 7), for April between 1997 and 2005. Flight tracks for all campaigns are shown

in Figure 7. Two coordinated aircraft campaigns with BC measurements were conducted in April 2008 over Alaska: (1) the Arctic Research of the Composition of the Troposphere from Aircraft and Satellites (ARCTAS) [Jacob et al., 2010] campaign and (2) the Aerosol, Radiation and Cloud Processes affecting Arctic Climate (ARCPAC) [Brock et al., 2011; Spackman et al., 2010; Koch et al., 2009]. For comparison purposes, a larger domain was selected for BC concentrations from model output over 1997 to 2005 period. The model output along the flight path could not be retrieved for 2008 and 2009 as the simulations were only until 2005 and average meteorology from 1997 to 2005 may not support the transport for 2008 and 2009. The ARCPAC data (blue curve) were higher than ARCTAS measurements (green curve) due to the large frequency of BB plumes encountered during the campaign. The grey curve shows measurement data with the influence of BB plumes removed. Finally, a third data-set from PAM-ARCMIP [Stone et al., 2010] conducted in April 2009 measured much lower BC concentrations due to the cleaner air-masses encountered along the flight-path. 3.2.2. Trends in Wintertime BC Concentrations [36] The time series of the simulated averaged wintertime BC concentrations are shown for the 1990–2005 period in Figure 8a–8c for the constant 1990 emissions scenario and in Figure 8d,e,f for the variable emissions scenario (which uses monthly emissions estimates for 1990–2005). The impact of each source region (FSU, EU, EA and NA) on concentrations at the three Arctic sites was obtained by turning off the emissions in the other regions. The inter-annual variability in the modeled BC concentrations under the constant emissions scenario is then related to changes in meteorology alone. Overall, change in the observations (black curve, Figure 8a–8c) at Alert and Barrow are not reproduced by the model, when BC emissions are held at 1990 levels. Transported BC from the FSU region (pink curve) appears to be the dominant factor that determines the overall, modeled, inter-annual variability at the three sites, including the sharp decline in simulated BC (red curve, Figure 8a and 8b) at Alert and Barrow from 1990–1996. The model also shows larger inter-annual variability in the BC concentrations at Barrow and Ny-Ålesund than at Alert. [37] Under the variable emissions scenario (Figure 8d–8f) the model agrees well with the observed BC concentrations at Alert and Barrow except at the beginning of the study period at Barrow. Whereas the model shows a decrease in BC concentrations at both Alert and Barrow during the first few years, a similar decrease is only evident in the measurements at Alert. At Ny-Ålesund (Figure 8f), the BC measurements only cover years from 2002–2005. The model over-predicts observed BC concentrations estimated at the site altitude of 474 m by a factor of three. This site is in the free troposphere [Beine et al., 2001] and above the surface based inversion during all winter/spring. This feature might not be reproduced well by this model. Additionally this over-prediction may be the effect of the simple wet scavenging scheme not capturing the strong removal of BC by the precipitation near this site. Significant cloud amount is frequently observed over the Norwegian/Barents Seas due to many cyclonic systems in these regions [Raatz, 1991]. Moreover, higher washout ratios and scavenging efficiencies than previously determined were found at Ny-Ålesund for the aged aerosol

953


Figure 7. Altitudinal dependence of BC in the Arctic from NIES model output (red) for April, 1997– 2005 is compared to BC observations from several Arctic campaigns such as ARCTAS (April 2008; green curve), ARCPAC (April 2008;blue curve) with std dev and without (grey curve) biomass burning plumes and PAMARCMIP (April 2009;black curve).

Figure 8. Simulated wintertime (January–March) BC concentrations at the three Arctic sites with plots on left showing results assuming constant 1990 emissions and plots on right showing results assuming variable emissions between 1990 and 2005. Each plot shows the comparison of BC concentrations between the observations (black curve with dots) and the simulations with global emissions (red curve with open circle) and regional emissions from FSU (pink), EU (blue), NA (green) and EA (amber). The grey dots in (b) and (d) are Barrow measurements omitted from the analysis as these data are questionable. 954

SHARMA ET AL.: 16 YR TRANSPORT OF BC TO THE ARCTIC Table 2. Averaged Simulated Wintertime (Jan-Mar) BC from 1990–2005 at Alert, Barrow and Ny-Ålesund Indicating that EU and FSU are the Dominant BC Contributors at the Surface. Largest Decline in Surface and Burden BC from the FSU are Significant to 95% Confidence Limit. Increase in BC in the 16 years of Simulation in NA, EU and EA are not Significant. Increases in the Arctic BC Burden from EU, EA and NA are Significant. Model Output

Total

NA

EA

EU

FSU

Rest

33 (35%)

32 (36%)

9 (10.5%)

45 33 (-26%) 11.8

105 32 (-68%)d 28.2

11 9 (-19%) 1.8

32 (29%)

58 (46%)

11(10.4%)

48 38.9 (-19%) 9.5

174 51 (-70%)d 43.1

14 11 (-21%) 2.5

77.8 (39%)

103(47.3%)

13 (6.6%)

298 90 (-70%)d 69.1

15 11 (27%) 2.8

Surface BC concentrations (winter) Alert -Avg. BC (ng m-3) -Measured BC 1990 BC (ng m-3) 2005 BC (ng m-3) Δ (%) SD (1990-2005 )

98.5 112 176 98.5 (-43%)d 32.3

6.4 (11.3%)

-Avg. BC (ng m-3) -Measured BC 1990 BC (ng m-3) 2005 BC (ng m-3) Δ (%) SD (1990-2005 )

116.7 82 251 120 (-52%)d 49.7

6.4 (6.1%)

-Avg. BC (ng m-3) -Measured BC 1990 (ng m-3) 2005 (ng m-3) Δ ( %) SD (1990-2005)

209 50 427 211 (-57%)d 83.8

7.1 (3.6%)

-Change in BC 1990 (%) 2005 (%) Other studies Huang et al., 2010b a Bourgeois and Bey, 2011b

8.6 16.4 (+96%) 4.7

5.9 6.0 (-1.7%) 2.4

7.9 (7.2%) 6.6 7.9 (+20%) 2.6 Barrow 9 (8.6%) 8 12 (+50%) 3.9 Ny-Ålesund 7.2 (3.8%)

6 6 102 9 8 91 (+50%) (+33%) (-10.5%) 2.5 2.8 31.3 Atmospheric BC Burden (for winter and North of 70 N)

Total (Tg) 0.0196 0.0130

4.6% 8.5%

6.4% 18.5%

23% 31%

57% 28%

9% 14%

N/A N/A

7% 17%

16% 27%

44% 25%

33% 29%

N/A N/A

Surface – simulation for a single year; N/A = not available BC burden – simulation for a single year c contribution from Eurasia (EU + FSU) d statistically significant with 95% confidence a

b

[Hegg et al., 2011] suggesting higher removal from the atmosphere. [38] Figures 8g–8i show anomalies, as calculated from differences between the yearly wintertime BC and the BC concentrations averaged over 16 years. Results show that: 1) With respect to EBC observations (grey curve) at Alert and Barrow, anomalies were similar during early 1990s and 2000s, except for year 2000 at Barrow. The anomalies were negative from the mid-1990s until 2005; 2) the interannual variability in the simulated anomalies show that decreasing BC concentrations were similar to those observed in early 1990s (grey). By keeping the emissions constant (pink curve), the magnitude of the transport from FSU was emphasized during mid-1990s at Alert and in 2000 at Barrow. These differences may be attributed to several reasons that also include difference in transport of BC mass to the Arctic. Higher BC was measured by 40% at Alert and Barrow during the positive phase of the North Atlantic Oscillation (NAO) than the negative phase [Sharma et al., 2006]. The agreement between the anomalies in the observations and those in the simulated BC improved when variable emissions were used (red curve). [39] During 1990-2005, the decrease in simulated BC concentrations at all sites (red curve, Figures 8d,e,f) can be attributed to the decreasing FSU emissions, because the

results under the constant emissions scenario (Figures 8a–8c) show that large interannual variability in transport alone did not cause this decrease in BC concentrations. Under both the constant and variable scenarios, the FSU and EU regions are the primary contributors to BC concentrations observed at the surface in the Arctic. There were minimum impacts from NA and EA source regions. Moreover, the overall change in simulated BC is sensitive to any changes in the FSU source region. These results are in agreement with the earlier papers of Huang et al. [2010b, 2010c], Hirdman et al. [2010], and Shindell et al. [2008]. 3.3. Relative Contributions of the Source Regions at the Three Arctic Sites [40] As discussed in the previous section, the relative contributions of BC from the different source regions to concentrations at the three sites appears to be a function of both the variable transport patterns and changing BC emissions. Table 2 shows these contributions and their changes for the different time periods. The results in Table 2 were obtained by modeling each source region separately. From 1990 to 2005, on average, FSU and EU contributed the most to surface BC, up to 85%, at all three measurement locations. The influence from FSU is higher at Barrow and Ny Ålesund than at Alert. BC contribution from EA is higher at Barrow than at the other two sites, although

955


Figure 9. Winter-time altitudinal BC gradient (a) and relative source contributions (b) averaged for 1990–2005 over north of 70 N region from various source regions. FSU and EU are dominant in the lower troposphere but EA contribution increases aloft; (c) interannual variations in the source contributions from FSU and EU in the lower troposphere (solid) and aloft (dotted). it is still quite small (70 N) among the NIES model and the other studies in Table 4 are mainly due to differences in deposition processes. Table 4 does not differentiate between dry and wet deposition processes, but we explore that further below. The total amount of BC deposited is 0.065 Tg yr1 with 0.043 Tg yr1 (65%) from fossil fuel (FF) combustion sources and 0.022 Tg yr1 (35%) from BB sources. Most of the BC deposited in the Arctic comes from EU (50%) and FSU (25%). Less than 10% of the deposited amount originates from each of EA and NA. With respect to BB BC, FSU and NA contribute the most (92%), 65% and 27% respectively. In support to our modeling results large proportions of BB fractions were also found in the snow measurements from Russian BB sector [Hegg et al., 2009]. [44] BC particles near source regions, en route to the Arctic and in the Arctic are subject to removal processes such as dry and wet deposition that occur at different rates. The scavenging rate of BC depends on the transport pathway

of the air mass and the atmospheric conditions encountered along the way. The seasonal variability of BC deposition is one of the key parameters in the simulation of BC during long-range transport from mid-latitude source regions to the Arctic. The spatial distribution of the precipitation rate is critical, as this determines how much BC emitted in the various source regions is scavenged (i.e., by wet deposition) before it can reach the Arctic. In this section we discuss (a) the relative contribution of dry vs wet deposition as a function of season and region; (b) the sensitivity to wet and dry deposition parameter assumptions; and (c) how the results from the work described relate to other studies analyzing the importance of deposition in influence BC in the Arctic. 3.4.2. Wet Versus Dry Deposition as a Function of Season and Region [45] Figure 10a shows that the Arctic region (north of 70oN) is very dry during the winter with a daily average precipitation rate between 0 and 1 mm day1. Precipitation rates over some of the BC source regions such as Eurasia are the same order of magnitude as Arctic precipitation rates. However, precipitation rates are much higher (5–10 mm day1) over the oceanic regions that air masses cross from NA and EA en route to the Arctic, which likely results in higher BC scavenging along these transport routes. In the summertime (Figure 10b), the Arctic region experiences 2 to 3 times higher precipitation rate relative to wintertime. Higher summertime precipitation also affects the BC source regions. The variability in precipitation will influence the ratio of dry to total deposition as is discussed next. [46] Figures 10c and 10d show the spatial distribution of dry to total deposition ratios averaged over 9 years for the winter (January to March) and the summer (June to August), respectively, based on control run model outputs. The time period from 1997–2005 selected for this simulation was based on the availability of BB emissions in the GFED emissions inventory. The figures show that, in the winter, dry deposition is the dominant removal mechanism (>70%) over much of the Arctic region (Figure 10c). In the summer, the contribution of dry to total deposition is significantly lower over the mid- and high- latitudes (often less than 40%) (Figure 10d). During the winter, less wet deposition and a shallower relative PBL result in higher BC concentrations near the surface. This results in a relative increase of dry deposition during wintertime poleward transport of air masses. [47] Figures 10e and 10f show the zonal mean (SD) of dry to total deposition ratios for both the winter and summer Table 4. Arctic BC Deposition Contribution from NIES Model Output from Various Source Region Total BC Deposition by region Model Output Fossil Fuel Fires Total Huang et al. [2010b] Bourgeois and Bey [2011] Shindell et al. [2008]

957

Total (Tg)

NA (%)

EA (%)

NIES 0.04Tg 7% 10% 0.02Tg 27% 2% 0.06Tg 9.1% 5.8% Other studies 0.11Tg 7% 10% 0.23Tg 14% 12% 10%

19%

EU (%)

FSU (%)

Rest (%)

48% 1% 42%

27% 65% 29%

8% 5% 14.1%

47% 32%

30% 42%

N/A N/A

69%

N/A

N/A


Figure 10. Seasonally averaged precipitation rate during 1997–2005 were calculated for the winter (a) and summer (b) in the Northern Hemisphere. Spatial distributions of the relative contributions of dry deposition to total deposition are also shown for the winter (c) and summer (d) zonally averaged dry deposition contributions for both seasons, as shown by the shaded area for the Arctic (e, f), were determined in the control case scenario (in red), and for high (in blue and black) and low (in green and amber) wet scavenging coefficients. Winter corresponds to the months of January–March and summer to June–August. for the control case (red curve) as well as for various sensitivity tests (high scavenging scenarios in blue and black and low scavenging scenarios in green and amber) discussed in the next section. A variance of only 10% in the zonal mean over nine years represents very consistent dry deposition north of 30 N. Wintertime dry deposition represents about 70% of the total deposition in the Arctic region (shaded area in Figure 10e). In contrast, wet deposition is predominant at all latitudes during the summer (shaded area Figure 10f) and south of 70 N during the winter. This suggests, in general, that a large portion of BC is removed in

the regions south of 70 N prior to air mass transport into the Arctic. In particular, the BC emissions in the EA region (30–40 N) are subject to more efficient scavenging (90% wet deposition) than other temperate source regions, meaning the EA influence on the Arctic is significantly inhibited. Similar conclusions were also drawn from results of ARCTAS conducted in April, June and July 2008 in the Alaskan Arctic [Matsui et al., 2011]. In this study, which spanned over 3 months, Matsui et al. [2011] showed that due to stronger scavenging in EA, this source region had less influence on the Arctic compared to biomass burning sources in

958

SHARMA ET AL.: 16 YR TRANSPORT OF BC TO THE ARCTIC Table 5. Annual Global BC Budget: Results from Other Modeling Studies (GEM-AQ, Aerocom, ECHAM5) and from the NIES Model Used in this Study Model

Emission (Tg yr-1)

GEM-AQc 10.9 11.9 Aerocomd e ECHAM5 8.5 This study NIES, control 9.00.8 This study – wet deposition sensitivity 9.00.8 S1 (rc0.01,sc0.2) 9.00.8 S2 (rc0.5, sc0.5) 9.00.8 S3 (rc2.0, sc2.0) 9.00.8 S4 (rc3.0, sc6.0) This study – dry deposition sensitivity Enhanced 9.00.8 Variable 9.00.8

Wet Deposition (Tg yr-1)

Dry Deposition (Tg yr-1)

Burden (Tg)

Life Time (day)a

Particle radius b

8.0 8.8 8.0

3.0 2.5 0.6

0.28 0.24 0.13

9.4 7.3 5.6

8.10.78

0.90.08

0.260.03

10.5

180 nm

2.00.29 7.80.63 8.30.67 8.40.78

6.50.99 1.20.11 0.70.06 0.60.06

4.220.88 0.430.06 0.160.02 0.120.02

170.3 17.4 6.5 4.9

60 nm 150 nm 250 nm 500 nm

7.80.65 8.20.68

1.20.15 0.80.07

0.20.03 0.30.04

8.1 12.2

a

Life Time = (Burden/Emission) x 365day/year; particle sizes corresponding to the S parameters Huang, L., S. L. Gong, C. Q. Jia, and D. Lavoué (2010a), Importance of deposition processes in simulating the seasonality of the Arctic black carbon aerosol, J. Geophys. Res., 115(D17207). doi: 10.1029/2009JD013478; d Textor, C., M. Schulz, S. Guibert, S. Kinne, Y. Balkanski, S. Bauer, T. Berntsen, T. Berglen, O. Boucher, M. Chin, F. Dentener, T. Diehl, J. Feichter, D. Fillmore, P. Ginoux, S. Gong, A. Grini, J. Hendricks, L. Horowitz, P. Huang, I. Isaksen, T. Iversen, S. Kloster, D. Koch, A. Kirkevåg, J. E. Kristjansson, M. Krol, A. Lauer, J. F. Lamarque, X. Liu, V. Montanaro, G. Myhre, J. E. Penner, G. Pitari, M. S. Reddy, . Seland, P. Stier, T. Takemura, and X. Tie (2007), The effect of harmonized emissions on aerosol properties in global models-an AeroCom experiment, Atmos. Chem. and Phys.s, 7, 4489-4501. doi: 10.5194/ acp-7-4489-2007. e Bourgeois, Q., and I. Bey (2011), Pollution transport efficiency toward the Arctic: Sensitivity to aerosol scavenging and source regions, J. Geophys. Res., 116(D8), D08213. doi: 10.1029/2010jd015096. b c

Russia, where precipitation rates were lower. Their analyses were based on adding precipitation amounts as proxy for deposition, along the trajectory cluster pathway from EA source and Russian BB source regions to the Alaskan Arctic in April and July. Several other field campaigns in the Arctic, which also concluded that there is a strong seasonality in the dry and wet deposition contributions to the aerosols, with dry deposition being dominant in the winter [Davidson et al., 1985, 1987; Koerner et al., 1999; Garrett et al., 2011]. [48] It is also worth noting that the results of the present study are based on a simple wet scavenging scheme that does not include in-cloud processing, different parameterizations for the different cloud phases, and frontal effects. The wet deposition sensitivity study (Section 3.4.2) will discuss how different values of high and low wet scavenging schemes influence the relative contribution of dry deposition as shown in Figure 10e and 10f. 3.4.3. Sensitivity Tests for Wet Deposition Parameters [49] In this section the sensitivity of the relative contribution of dry to total deposition as a function of the wet deposition parameters rc (for rain scavenging) and sc (for snow scavenging) is assessed. Results from the control case (red line, Figures 10e,f) are compared to four different wet scavenging scenarios. S1 represents an extreme case of minimal wet scavenging – the control scavenging parameters were modified by reducing the control case coefficients by a factor of 100 for rain and a factor of 5 for snow (i.e., rcx0.01 and scx0.2). S2 and S3 use wet deposition values spanning a more realistic range: for S2 the control case rc and sc values were multiplied by 0.5 and while for S3 the rc and sc values were multiplied by 2 (see Figure 4). S4 was designed to test sensitivity to higher wet deposition rates – the control case scavenging was multiplied by 3 for rain and by 6 for snow. The scavenging coefficients in the S1 and S4 cases were chosen to include the smallest and largest particle sizes

corresponding to 50 and 500 nm particle radius of the accumulation mode. [50] Figures 10e,f show that under the extremely low wet scavenging conditions of S1 (blue curve), the relative contribution of dry to total deposition is dominant, as expected, and is relatively stable at 80–90% from mid to high latitudes in both summer and winter. Scenario S1 corresponds to the scavenging coefficients for particle sizes of radius 60 nm according to Figure 4. This is close to the 80 nm modal size measured at Alert, but simulations using particles of this size extended the lifetime of BC in the atmosphere to over 4 months (Table 5), which is very unrealistic. If the more realistic dry S2 (black curve) and wet S3 (amber curve) conditions prevail, dry deposition will still be dominant in the Arctic (70–90 ). However, under extremely wet conditions, S4 (green curve), wet deposition is dominant at all latitudes in both seasons, with dry deposition contributing less than 10% in summer (regardless of latitude) and from 10% (at low latitudes) to 40% (at higher latitudes) in winter. The dry deposition contribution decreased by up to 50% in the Arctic, for the S4 scenario. Scenarios S3 and S4, correspond to particle sizes of 250 and 500 nm (Figure 4); these particle sizes are unrealistic in terms of modal size distribution measured in the Arctic and also in the source regions. Not surprisingly, these scenarios resulted in atmospheric lifetimes that are shorter than the control case (Table 5) as scavenging is more efficient for larger particles. These results suggest that all model outputs are sensitive to the magnitude of the dry and wet scavenging parameters that are used in the model to remove BC from the atmosphere. The choice of the right deposition scheme will affect the aerosol BC burden and the surface BC concentrations [Liu et al., 2011]. [51] Table 5 shows how changes in the wet scavenging parameters influence the budget and lifetimes of BC in the NIES model. When the wet scavenging parameters were decreased (i.e., S1 and S2) compared to the control case, the

959


Figure 11. Comparison of simulated BC at the Arctic sites with three different dry deposition velocities: (a) 0.05 cm s1 for all surface (control case), (b) 0.1 cm s1 for all surface (enhanced case), (c) 0.2 cm s1 over oceans and 0.025 cm s1 over lands (variable case). All simulations were performed with the year 1990 emissions and the wet deposition scheme for the period of 1990 to 1996. The ratios of the model simulation to observation (R) and linear correlation coefficients (r) for Alert and Barrow are also shown. The suffix, win, indicates winter (Jan–Mar). No BC observation data were available for Ny Ålesund during 1990 to 1996 for comparison to model output. lifetime of BC was longer than that projected by other models. A factor of 2 increase in the wet scavenging coefficients (i.e., scenario S3) from the control case will result in a change of the total BC wet deposition by 2.5% globally and by 1% in the Arctic. 3.4.4. Sensitivity Tests on Dry Deposition Velocities [52] Huang et al. [2010a] and Liu et al. [2011] showed that simulated BC concentrations are very sensitive to dry deposition velocities, especially in the Arctic region. During transport from the mid-latitude source regions to the receptor sites, air masses encounter various surface types and boundary layer conditions, so a constant dry deposition velocity may not account for the changes in resistance over different surfaces or to varying atmospheric stability. In the winter and spring seasons, air transported from the source regions (mainly the FSU and EU) tends to follow potential

temperature surfaces, [Sharma et al., 2006] and crosses over frozen land and sea ice before reaching our observation sites. In contrast, air transported from EA and NA generally cross the ice-free North Pacific and North Atlantic Ocean, respectively. In the present study, we used Vd = 0.05 cm s1 for the control case to represent the mixture of surface types over which air is transported, although values between 0.01 and 0.2 have been used in the literature depending on the assumptions about the surface type [e.g., Wang et al., 2011; Fisher et al., 2011, Liu et al., 2011; Croft et al., 2005]. Here, three cases of dry deposition velocities were used in the sensitivity test: (1) control case (0.05 cm s1), (2) enhanced case (0.1cm s1) and (3) variable case to represent the different surfaces encountered by air-masses during pole-ward transport, Vd = 0.2 cm s1 for wet surface (ice surfaces and open ocean) and Vd = 0.025 cm s1 for land, as used by Croft et al. [2005]. The simulations were performed for the time period 1990-1996 with constant emissions representative of 1990 and variable meteorology. Wet deposition and meteorology were variable as a function of time, however they were the same for all three cases so the effects seen here are only due to changing dry deposition velocities. 3.4.4.1. Control (0.05 cm s1) vs. Enhanced (0.1 cm s1) Dry Deposition Cases [53] Figure 11a shows the seasonal variation in the model output for the control case with Vd = 0.05 cm s1. Since BC emissions were kept constant for the 7-year simulation period, the inter-annual variability, as shown by the standard deviation, is due to meteorology alone. The magnitude and inter-annual variability in simulated BC concentrations at Ny-Ålesund are larger than those at Alert and Barrow due to the proximity of Ny-Ålesund to the major source regions. In addition, Ny-Ålesund is also not always in the “polar dome” stable airmasses, since it is near open water and the North Atlantic storm track. The control case agreement with the EBC measurements, as represented by R (ratio of simulated BC to measurements) and Rwin (ratio of simulated BC to measurements for winter data) is better at Alert than Barrow during winter and a factor of 2 higher for all data. [54] The first sensitivity test on Vd is the ‘enhanced’ case where the 0.05 cm s1 control case dry deposition velocity value was increased to 0.1 cm s1. As a result, simulated BC concentrations at the sites were always lower than those observed for the control case (Figure 11b). The seasonal amplitude at Alert, Barrow and Ny-Ålesund with Vd = 0.1 cm s1 is half the seasonal amplitude obtained with the control case. The agreement with the measurements improved for Barrow, as indicated by the R-values, especially during winter. In contrast, at Alert the wintertime R decreased from 0.97 in the control case to 0.58 in the enhanced case suggesting the increase in Vd was too high. The ‘enhanced’ Vd results suggest that increased dry deposition velocities strongly affect wintertime BC concentrations and, accordingly the seasonal amplitude. Temporal (seasonal/inter-annual) variability and the low summertime BC concentrations are minimally impacted. 3.4.4.2. Control (0.05 cm s1) vs. Variable Case (0.2 cm s1 for Ocean and 0.025 cm s1 for Land) [55] Model simulations for the control and variable cases are shown in Figures 11a and 11c. Relative to the control case, the variable case includes lower dry deposition velocity over land (by factor of 2) and higher deposition (by factor of 4) over the oceans. The amplitude of the seasonal variation

960


in BC is similar for both the control and variable case scenarios for all three Arctic sites (Figure 11c). A comparison of Figure 11a,c indicates that enhanced dry deposition over the ocean and less dry deposition over the land does not produce a noticeable change in the BC concentrations during the summer, although wintertime BC concentrations are slightly higher for the variable case conditions. 3.4.5. Comparison with Previous Studies [56] The global BC budget as shown in Table 5, suggests that the NIES model is comparable to other modeling studies and model results, e.g., Aerocom [Textor et al., 2006, 2007], enhanced GEM/AQ [Huang et al., 2010a, 2010b] and ECHAM5-Hammoz [Bourgeious and Bey, 2011]. Global BC annual emissions of 9.0 Tg in NIES (this study) includes fossil fuel and biomass burning emissions but does not include biofuels. Enhanced GEM/AQ and Aerocom studies use higher BC emissions than NIES model with 10.9 and 11.9 Tg yr1 respectively. The lowest BC emissions of 8.5 Tg yr1 was used in ECHAM5-Hammoz model. [57] Global annual BC wet deposition derived from this study using the NIES model is comparable to the other three studies in Table 5. The global annual dry deposition by NIES and ECHAM5-Hammoz models are a factor of ~3 lower than the Aerocom and GEM-AQ results. The NIES global annual aerosol BC burden of 0.26 0.03 Tg is also comparable to other studies except for ECHAM-Hammoz model which is a factor of 2 lower than other studies and reflects the lower emissions and higher wet deposition used in that model. The NIES global average lifetime of BC is estimated to be 10.5 days which is within the range of 5 to 11 estimated by other studies as listed in Table 5 and the ensemble model intercomparison by Koch et al. [2009]. [58] Table 5 also allows us to compare our results on the relative importance of dry deposition to some other recent studies. In this study, the annual average global dry deposition contribution was found to be 10% using the NIES model which is lower than the ~25% dry deposition contribution found by the GEM-AQ and Aerocom studies. In contrast, Bourgeois and Bey [2011] estimated global and Arctic annual BC loss by wet deposition to be 98%, i.e., only a 2% contribution from dry deposition. The large amount of BC loss by wet scavenging calculated by Bourgeois and Bey [2011] for latitudes north of 70oN seems unrealistic due to the dry atmospheric conditions that prevail during the winter/spring Arctic haze period.

4. Conclusions [59] The global transport model NIES was used to examine the inter-annual variation and trends in black carbon for 16 years (1990-2005) of atmospheric transport, from the mid-latitude source regions to the Arctic. Simulations included scenarios with constant and annual variable emissions to understand whether it was meteorology and/or emissions that influenced the BC transport to the three Arctic sites. In spite of increases in BC emissions by 10–15% in the NA, FSU and EU source regions, and a factor of 2 BC emissions increase in the EA source region since 2000, corresponding increases in BC were not observed at the three measurement locations. To address this issue, transport and contributions from individual source regions were estimated by masking the emissions from other regions. Various

parameterizations of dry and wet deposition parameters were implemented in the model to perform sensitivity tests. The following are the findings from this study: 1. Observations of surface EBC, at Alert, Barrow and NyÅlesund showed a decline of 40% during wintertime between 1990 and 2009. 2a. Using reanalysis fields and highly parameterized wet and dry scavenging schemes, a simple aerosol model closely simulate surface BC concentrations atBarrow and Alert. NyÅlesund results were not well simulated because unlike the other two sites, this site is above the inversion layer and is also not always in the “polar dome” stable airmasses, since it is near open water and the North Atlantic storm track. 2b. Variability in boreal BB emissions were found to be important in explaining temporal variability in BC concentrations at Arctic surface sites. 3. Changes in meteorology alone could not explain the observed trends, but were reflected in the interannual variations in wintertime EBC measurements at the three Arctic sites (Alert, Barrow and Ny-Ålesund) over the 16 years. This is consistent with results of previous studies [Sharma et al., 2004, 2006] that used different methods. 4. The trends in wintertime BC from 1990 to 2005 at the three sites are governed by changes in emissions in the northern mid-latitude regions, mainly FSU [also previously determined by other statistical methods based on trajectory analyses and FLEXPART in Sharma et al., 2004, 2006; Hirdman et al., 2010; Huang et al., 2010c; Gong et al., 2010; other models with one to two year simulations in Koch and Hansen, 2005; Huang et al., 2010b; Bourgeious and Bey, 2011]. The maximum contribution to the surface BC concentrations at Alert, Barrow, and Ny-Ålesund are from the combined FSU and EU regions which contributed a total of up to 71%, 75% and 86% of the modeled BC concentrations at the three sites respectively. EA and NA were found to have little influence on surface BC concentrations at any of the sites. The EA influence in the upper atmosphere was found to be significant in a relative sense (40% of the BC aloft is from EA) [also shown in Koch and Hansen, 2005], however, high altitude BC concentrations are quite low (less than 50 ngm-3) due to greater scavenging losses along the transport pathway before the air-masses reach the Arctic. 5. The relative contribution from the FSU to the column burden decreased by one half from 1990 to 2005. However, BC integrated north of 70 N in 100–300 mb level, increased by a factor of 3 as a result of increasing EA emissions. 6. In the present study, dry deposition was found to be an important mechanism for BC deposition on snow and ice surfaces north of 70 N during winter. Use of a single representative value of dry deposition velocity (Vd = 0.05 cm s1) gave consistent results with other studies that used variable Vd [Croft et al., 2005]. 7. The zonal average for the contribution of dry to total deposition indicated that wet deposition is the dominant process for atmospheric BC removal south of 70 N regardless of season. Emissions from EA source regions are subjected to efficient wet deposition along the transport pathways during the winter/spring before it reaches the Arctic. This is evidenced by the small BC concentrations from EA in the layers of lower (700–1000 mb) and upper (100–300 mb) Arctic atmosphere relative to large BC emissions from the

961


EA region. During the summer, the wet deposition was found to be the dominant removal process at all latitudes. 8. Sensitivity tests on both dry and wet deposition parameterizations revealed that choices made for these parameters are extremely important in controlling the modeled atmospheric BC concentrations, burdens and depositions amounts in the Arctic. However, there are limitations of this model with assumptions made for the wet and dry parameterizations and remaining model uncertainties. Thus, more experimental and modeling work is required on this subject. [60] There is currently a global effort to decrease ShortLived Climate Forcers (SLCFs), including BC, under the United Nations Environmental Program [UNEP, 2011] and the Arctic Contaminants Action Program (ACAP) of the Arctic Council (http://www.arctic-council.org, access 29 January 2012). The present study shows that the collapse of FSU led to a decrease in black carbon concentrations at the three surface measurement sites and in the atmospheric column since the 1990s. More recently, increases in BC emissions, specifically in EA, has little bearing on the surface BC concentrations while the BC levels in the upper atmosphere have increased from 25% to 40% from 1990 to 2005. The BC mitigation policies applied in both FSU and EU regions, as major contributors to Arctic BC levels, would further decrease the BC burden in the Arctic atmosphere. On the other hand, BC mitigation efforts in EA are less likely to have an impact according to our modeling results, which shows that this region has a limited but slowly growing influence on Arctic BC concentrations at the near surface. The impact may be more prominent aloft. Near surface BC has greater potential in the warming due to deposition on snow/ice surfaces whereas increasing BC levels in the upper atmosphere might not be as prominent due to increase in co-mixed aerosols (defined as BC co-emitted with non-absorbing species and gas-phase precursors). It is possible that cooling via further mitigation of BC in the Arctic may be partly offset by a collateral decrease in co-mixed aerosol that normally cool the surface radiatively (reduction in first indirect effect), a factor that should be considered when making future assessments of Arctic climate. [61] Given the sensitivity of model simulations of BC concentrations to parameterizations of wet and dry deposition, and the relatively large uncertainties about future changes in precipitation, these results may be helpful in determining the relative effectiveness of BC mitigation in different regions. Certainly, further study is warranted to explore the impact of various mitigation scenarios explicitly. [62] Acknowledgments. The authors would like to thank Drs. Leonard Barrie and Fred Hopper for starting the black carbon measurement program at Alert; technical support by Joe Kovalick, Armand Gaudenzi and Dave Halpin over the years; Alert operators over the years for maintaining the program and members of CFS Alert for site maintenance and flights to Alert. Special thanks to Elton Chan for trend analysis of black carbon concentrations at the three Arctic sites; Dr. Knut Von Salzen, Climate Research Division/Environment Canada for internal review of this manuscript; and Dr. Robert Stone of NOAA/CIRES for useful discussions and Elizabeth Bush of Climate Research Division, Environment Canada for editing of this manuscript.

References Arnott, W.P., K Hamasha, H Moosmüller, P J. Sheridan & JA. Ogren, (2005), Towards aerosol light-absorption measurements with a 7-wavelength aethalometer: Evaluation with a photoacoustic instrument and

3-wavelength nephelometer, Aero. Sci. and Tech., 39, Issue 1, doi: 10.1080/027868290901972 Baron, R. E., W. D. Montgomery, and S. D. Tuladhar (2009), An Analysis of Black Carbon Mitigation as a Response to Climate Change, 31 pp., Copenhagen Consensus Center. Barrie, L. A. (1986), Arctic air pollution: an overview of current knowledge, Atmos. Environ., 20(4), 643‐663. doi: 10.1016/0004-6981(86)90180-0. Beine, H. J., S. Argentini, A. Maurizi, G. Mastrantonio, and A. Viola (2001), The local wind field at Ny-Alesund and the Zeppelin mountain at Svalbard, Meteorol. Atmos. Phys., 78, 107 – 113. Bodhaine, B., (1989), Barrow surface aerosol: 1976–1986, Atmos. Env., 23, 11, 2357–2369 Bond, T. C., E. Bhardwaj, R. Dong, R. Jogani, S. Jung, C. Roden, D. G. Streets, and N. M. Trautmann (2007), Historical emissions of black and organic carbon aerosol from energy-related combustion, 1850-2000, Global Biogeochem. Cycles, 21(GB2018). doi: 10.1029/ 2006GB002840. Bond, T. C., D. G. Streets, K. F. Yarber, S. M. Nelson, J.-H. Woo, and Z. Klimont (2004), A technology-based global inventory of black and organic carbon emissions from combustion, J. Geophys. Res., 109 (D14203). doi: 10.1029/2003JD003697. Bourgeois, Q., and I. Bey (2011), Pollution transport efficiency toward the Arctic: Sensitivity to aerosol scavenging and source regions, J. Geophys. Res., 116(D8), D08213. doi: 10.1029/2010jd015096. Brock, C.A., et al. (2011), Characteristics, sources, and transport of aerosols measured in spring 2008 during the aerosol, radiation, and cloud processes affecting Arctic Climate (ARCPAC) Project, Atmos. Chem. Phys., 11, 2423–2453, 2011,doi:10.5194/acp-11-2423-2011. Chan, D., M. Ishizawa, K. Higuchi, S. Maksyutov, and J. Chen (2008), Seasonal CO2 rectifier effect and large-scale extratropical atmospheric transport, J. Geophys. Res., 113(D17309). doi: 10.1029/2007JD009443. Chung, S. H., and J. H. Seinfeld, (2002), Global distribution and climateforcing of carbonaceous aerosols, J. Geophys. Res., 107(D19), 4407, doi:10.1029/2001JD001397. Clarke A.D. and K.J. Noone, (1985), Soot in the Arctic snowpack: a cause for perturbations in radiative transfer, Atmos. Environ., 19, pp. 2045‐2053. Cofala, J., M. Amann, Z. Klimont, K. Kupiainen, L. Höglund-Isaksson, (2007), Scenarios of global anthropogenic emissions of air pollutants and methane until 2030, Atmos. Env., 41, 38, 8486–8499. Cooke, W. F., C. Liousse, H. Cachier, and J. Feichter (1999), Construction of a 1 x1 fossil fuel emission data set for carbonaceous aerosol and implementation and radiative impact in the ECHAM4 model, J. Geophys. Res., 104(D18), 22137‐22162. doi: 10.1029/1999JD900187. Croft, B., U. Lohmann, R.V. Martin, P. Stier, S. Wurzler, J. Feichter, C. Hoose, U. Heikkilla, A. van Donkelaar, and S. Ferrachat, (2010), Influences of in-cloud aerosol scavenging parameterizations on aerosol concentrations and wet deposition in ECHAM5-HAM, Atmos. Chem. Phys., 10, pp. 1511–1543 Croft, B., U. Lohmann, R. V. Martin, P. Stier, S. Wurzler, J. Feichter, R. Posselt, and S. Ferrachat (2009), Aerosol size-dependent below-cloud scavenging by rain and snow in the ECHAM5-HAM, Atmos. Chem. and Phys.s, 9, 4653‐4675. doi: 10.5194/acp-9-4653-2009. Croft, B., U. Lohmann, and K. von Salzen (2005), Black carbon ageing in the Canadian Centre for Climate Modeling and Analysis atmospheric general circulation model, Atmospheric Chemistry and Physics, 5(7), 1931‐1949. doi: 10.5194/acp-5-1931-2005. Davidson C.I., R.E. Honrath, J.B. Kadane, R.S. Tsay, P.A. Mayewski, B.W. Lyons and N.Z. Heidam. (1987), The scavenging of atmospheric sulfate by Arctic snow, Atmos. Environ., 21(4), 871‐882. Davidson, C. I., S. Santhanam, R. C. Fortmann, and M. P. Olson (1985), Atmospheric transport and deposition of trace elements onto the Greenland Ice Sheet, Atmos. Environ., 19(12), 2065‐2081. Dentener, F., et al. (2006), Emissions of primary aerosol and precursor gases in the years 2000 and 1750 -prescribed data-sets for AeroCom, Atmos. Chem. Phys., 6, 4321‐4344. Eleftheriadis, K., S. Vratolis, S. Nyeki (2009), Aerosol black carbon in the European Arctic: Measurements at Zeppelin station, Ny-Ålesund, Svalbard from 1998–2007, Geophys. Res Lett., 36, L02809, 5, doi:10.1029/ 2008GL035741 Fisher, J. A., et al. (2011), Sources, distribution, and acidity of sulfate ammonium aerosol in the Arctic in winter-spring, Atmos. Environ., 45, 7301–7318, doi:10.1016/j.atmosenv.2011.08.030. Flanner, M. G., C. S. Zender, J. T. Randerson, and P. J. Rasch (2007), Present-day climate forcing and response from black carbon in snow, J. Geophys. Res., 112(D11202). doi: 10.1029/2006JD008003. Forster, P. M., et al. (2007), Changes in Atmospheric Constituents and in Radiative Forcing, In: Solomon; S; Qin D; Manning M; Chen Z; Marquis M; Averyt KB; Tignor M; Miller HL (Ed) Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change,

962

SHARMA ET AL.: 16 YR TRANSPORT OF BC TO THE ARCTIC Cambridge University Press. Cambridge, U. K. and New York, N. Y., USA. Fromm, M., R. Bevilacqua, R. Servranckx, J. Rosen, J. Thayer, J. Herman, and D. Larko (2005), Pyro-cumulonimbus injection of smoke to the stratosphere: observations and impact of a super blowup in northwestern Canada on 3-4 August 1998, J. Geophys. Res., 110, D08205, doi:10.1029/2004JD005350. Fujita, D., M. Ishizawa, S. Maksyutov, P. Thornton, T. Saeki, and T. Nakazawa (2003), Interannual variability of the atmospheric carbon dioxide concentrations as simulated with global terrestrial biosphere models and atmospheric transport model, Tellus, 55B, 530‐546. doi: 10.1034/ j.1600-0889.2003.00044. Garrett, T. J., S. Brattström, S. Sharma, D. E. J. Worthy, and P. Novelli (2011), The role of scavenging in the seasonal transport of black carbon and sulfate to the Arctic, Geophys. Res. Lett., 38, L16805, doi:10.1029/ 2011GL048221. Gong, S. L., T. L. Zhao, S. Sharma, D. Toom-Sauntry, D. Lavoué, X. B. Zhang, W. R. Leaitch, and L. A. Barrie (2010), Identification of trends and interannual variability of sulfate and black carbon in the Canadian High Arctic: 1981-2007, J. Geophys. Res., 115(D07305). doi: 10.1029/2009JD012943. Hansen, J., and L. Nazarenko (2004), Soot climate forcing via snow and ice albedos, Proc. Natl. Acad. Sci. USA, 101(2), 423‐428. doi: 10.1073/ pnas.2237157100. Hansen, A., H. Rosen and T. Novakov (1984), The Aethalometer: an instrument for the real-time measurement of optical absorption by aerosol particles, The Science of the Total Environment, 36, 191‐196. Hegg, D.A., A. D. Clarke, S. J. Doherty, J. Ström (2011), Measurements of black carbon aerosol washout ratio on Svalbard, Tellus B, 63, Issue 5, pp 891–900, 2011. Hegg, D., S. G. Warren, T.C. Grenfell, S. J. Doherty, T. V. Larson, A. D. Anthony (2009), Source Attribution of Black Carbon in Arctic Snow, Env. Sci. & Tech., 43, 11, 4016‐4021. Hirdman D., H. Sodemann, S. Eckhardt, J.F. Burkhart, A. Jefferson, T. Mefford, P.K. Quinn, S. Sharma, J. Ström and A. Stohl (2010), Source identification of short-lived air pollutants in the Arctic using statistical analysis of measurement data and particle dispersion model output. Atmos. Chem. Phys., 10, pp. 669‐693. Hsu, N. S., J. R. Herman, J. F. Gleason, O. Torres, and C. J. Seftor (1999), Satellite detection of smoke aerosols over a snow/ice surface by TOMS, Geophys. Res. Lett., 26(8), 1165‐1168. doi: 10.1029/1999GL900155. Huang, L., S. L. Gong, C. Q. Jia, and D. Lavoué (2010a), Importance of deposition processes in simulating the seasonality of the Arctic black carbon aerosol, J. Geophys. Res., 115(D17207). doi: 10.1029/2009JD013478. Huang, L., S. L. Gong, C. Q. Jia, and D. Lavoué (2010b), Relative contributions of anthropogenic emissions to black carbon aerosol in the Arctic, J. Geophys. Res., 115(D19208). doi: 10.1029/2009JD013592. Huang, L., S. L. Gong, S. Sharma, D. Lavoué, and C. Q. Jia (2010c), A trajectory analysis of atmospheric transport of black carbon aerosols to Canadian High Arctic in winter and spring (1990-2005), Atmos. Chem. Phys., 10, 5065‐5073. doi: 10.5194/acp-10-5065-2010. Ishizawa, M., D. Chan, K. Higuchi, S. Maksyutov, C.-W. Yuen, and J. Chen (2006), Rectifier effect in an atmospheric model with daily biospheric fluxes: impact on inversion calculation, Tellus, 58B(5), 447‐462. doi: 10.1111/j.1600-0889.2006.00219. Jacob, D.J. et al. (2010), The Arctic Research of the Composition of the Troposphere from Airctaft and Satellites (ARCTAS) mission: Design, execution, and first results, Atmos. Chem. Phys., 10(11), 51915212, doi:10.5194/acp-10-5191-2010. Junker C. and C. Liousse (2008), A global emission inventory of carbonaceous aerosol from historic records of fossil fuel and biofuel consumption for the period 1860-1997, Atmos. Chem. Phys., 8, 1‐13. Kalnay, E., et al. (1996), The NCEP/NCAR 40-year reanalysis project, Bull. Am. Meteorol. Soc., 77, 437‐471. doi: 10.1175/1520-0477(1996) 0772.0.CO;2. Kim, D., C. Wang, A. M. L. Ekman, M. C. Barth, and P. J. Rasch (2008), Distribution and direct radiative forcing of carbonaceous and sulfate aerosols in an interactive size-resolving aerosol-climate model, J. Geophys. Res., 113, D16309, doi:10.1029/2007JD009756. Klonecki, A., P. Hess, L. Emmons, L. Smith, J. Orlando, and D. Blake (2003), Seasonal changes in the transport of pollutants into the Arctic tropospheremodel study, J. Geophys. Res., 108(D4, 8367). doi: 10.1029/2002JD002199. Koch, D., and J. Hansen (2005), Distant origins of Arctic black carbon: A Goddard Institute for Space Studies ModelE experiment, J. Geophys. Res., 110(D04204). doi: 10.1029/2004JD005296. Koch, D., et al. (2009), Evaluation of black carbon estimations in global aerosol models, Atmos. Chem. and Phys., 9, 9001‐9026. doi: 10.5194/ acp-9-9001-2009. Koch, D. (2001), Transport and direct radiative forcing of carbonaceous and sulfate aerosols in the GISS GCM, J. Geophys. Res., 106(D17), 20,311–20,332, doi:10.1029/2001JD900038.

Koerner, Roy M.,, David A. Fisher, Kumiko Goto-Azuma (1999), 100 year record of ion chemistry from Agassiz Ice Cap Northern Ellesmere Island NWT, Canada, Atmos. Environ ,33, 347‐357. Lamarque, J.-F., et al. (2010), Historical (1850–2000) gridded anthropogenic and biomass burning emissions of reactive gases and aerosols: methodology and application, Atmos. Chem. Phys., 10, 4963‐5019, 2010 doi:10.5194/acpd-10-4963-2010. Lavoué, D., T.L. Zhao, S.L. Gong, P. Huang, S. Sharma, and B.J. Stocks (2009), Interannual variability of carbonaceous particles from boreal wildland fires and their contributions to aerosol loading and deposition in the Arctic, American Geophysical Union (AGU) Fall Meeting, San Francisco, CA, December 14-18. Lavoué, D. (2000), Transport vers la région arctique de l’aérosol carboné par les feux de biomasse des régions boréales et tempérées, Ph.D. thesis (in French), Laboratoire des Sciences du Climat et de l’Environnement, CNRS/CEA, Gif-sur-Yvette, France, p 270. Lavoué, D., C. Liousse, H. Cachier, B.J. Stocks, and J. G. Goldhammer (2000), Modeling of carbonaceous particles emitted br Boreal and temperate wildfires at northern latitudes, J. Geophys. Res., 105, 26,871‐ 26,890. Law, K., and A. Stohl (2007), Arctic air pollution: origins and impacts, Science, 315, 1537‐1540. doi: 10.1126/science.1137695. Liousse, C., J. E. Penner, C. Chuang, J. J. Walton, H. Eddleman and H. Cachier (1996), A global three-dimensional model study of carbonaceous aerosols, J. Geophys. Res., 101, 19411‐19432. Liu, J.F., S. M. Fan, L.W. Horowitz and H. Levy (2011), Evaluation of factors controlling long range transport of black carbon to the Arctic, J. Geophys. Res., 116, D04307, doi:10.1029/2010jd015145. Maksyutov, S., and G. Inoue (2000), Vertical profiles of radon and CO2 simulated by the global atmospheric transport model, in CGER supercomputer activity report CGER I039-2000, edited by CGER NIES, pp. 39‐41, Tsukuba, Japan. Matsui, H., et al. (2011), Seasonal variation of the transport of black carbon aerosol from the Asian continent to the Arctic during the ARCTAS aircraft campaign, J. Geophys. Res., 116(D05202), doi:10.1029/2010JD015067. Patra, P., M. Ishizawa, S. Maksyutov, T. Nakazawa, and G. Inoue (2005), Effects of biomass burning and meteorological conditions on landatmosphere CO2 flux from atmospheric CO2 inverse modeling, Global Geochem. Cycles, 19(GB3005). doi: 10.1029/2004GB002259. Peters, G.P., G. Marland, C. Le Quere, T. Boden, J. G. Canadell, M.R. Raupach (2012), Rapid growth in CO2 emissions after the 2008–2009 global financial crisis, Nature Climate Change, 2, 2–4. Petroff A. and L. Zhang (2010), Development and validation of a size-resolved particle dry deposition scheme for applications in aerosol transport models, Geosci. Model Dev. Discuss. 3, 1317‐1357. Quinn, P. K., G. Shaw, E. Andrews, E. G. Dutton, T. Ruoho-Airola, and S. L. Gong (2007), Arctic haze: current trends and knowledge gaps, Tellus, 59B, 99‐114. doi: 10.1111/j.1600-0889.2006.00238. Raatz W.E. (1991), The climatology and meteorology of Arctic air pollution, in: W.T. Sturges, Editor, Pollution of the Arctic atmosphere, 13‐42. Rahn, K. A. (guest editor)(1981), Arctic air chemistry. Atmos. Environ. 1.5, 1345‐1516. Riemer, N., H. Vogel, and B. Vogel (2004), Soot aging time scales in polluted regions during day and night, Atmos. Chem. Phys., 4, 1885–1893, doi:10.5194/acp-4-1885-2004. Sharma, S., E. Andrews, L. A. Barrie, J. A. Ogren, and D. Lavoué (2006), Variations and sources of the equivalent black carbon in the high Arctic revealed by long-term observations at Alert and Barrow: 1989-2003, J. Geophys. Res., 111(D14208). doi: 10.1029/2005JD006581. Sharma, S., D. Lavoué, H. Cachier, L. A. Barrie, and S. L. Gong (2004), Long-term trends of the black carbon concentrations in the Canadian Arctic, J. Geophys. Res., 109(D15203). doi: 10.1029/2003JD004331. S. Sharma, J. R. Brook, H. Cachier, J. Chow, A. Gaudenzi, and G. Lu (2002), Light absorption and thermal measurements of black carbon in different regions of Canada, J. Geophys. Res., 107, D24, 4771, doi:10.1029/2002JD002496. Shindell, D. T., et al. (2008), A multi-model assessment of pollution transport to the Arctic, Atmos. Chem. and Phys.s, 8, 5353‐5372. doi: 10.5194/ acp-8-5353-2008. Spackman, J. R., R. S. Gao, W. D. Neff, J. P. Schwarz, L. A. Watts, D. W. Fahey, J. S. Holloway, T. B. Ryerson, J. Peischl, and C. A. Brock (2010), Aircraft observations of enhancement and depletion of black carbon mass in the springtime Arctic, Atmos. Chem. Phys., 10, 9667–9680, doi:10.5194/acp-10-9667-2010 Stohl, A., et al. (2007), Arctic smoke - record high air pollution levels in the European Arctic due to agricultural fires in Eastern Europe, Atmos. Chem. Phys. 7, 511‐534. Stohl, A., et al. (2006), Pan-Arctic enhancements of light absorbing aerosol concentrations due to North American boreal forest fire during summer 2004, J. Geophys. Res., 111(D22214). doi: 10.1029/2006JD007216.

963

SHARMA ET AL.: 16 YR TRANSPORT OF BC TO THE ARCTIC Stone, R.S., et al. (2010), A Three-dimensional Characterization of Arctic Aerosols from Airborne Sun Photometer Observations: PAM-ARCMIP, April 2009, J. Geophys. Res., 115, D13203, doi:10.1029/2009JD013605. Ström, J., J. Umegard, K. Torseth, P. Tunved, H. –C. Hansson, K. Holmen, V. Wismann, A. Herber and G. Konig-Langlo (2003), One year of particle size distribution and aerosol chemical composition measurements at the Zeppelin Station, Svalbard, Phys. Chem. Earth, 28, 1181‐1190. Textor, C., et al. (2006), Analysis and quantification of the diversities of aerosol life cycles within AeroCom, Atmos. Chem. and Phys.s, 6, 1777‐1813. doi: 10.5194/acp-6-1777-2006. Textor, C., et al. (2007), The effect of harmonized emissions on aerosol properties in global models-an AeroCom experiment, Atmos. Chem. and Phys.s, 7, 4489‐4501. doi: 10.5194/acp-7-4489-2007. United Nations (2007), The United Nations energy statistics database (2005), 5 pp, United Nations Statistics Division, New York. UNEP 2011 – Integrated Assessment of Black Carbon and Tropospheric Ozone: Summary for Decision Makers, Nairobi, Kenya UNEP/GC/26/ INF/20ISBN: 978-92-807-3142-2. Van der Werf, G. R., J. T. Randerson, L. Giglio, G. J. Collatz, and P. S. Kasibhatla (2006), Interannual variability in global biomass burning emission from 1997 to 2004, Atmos. Chem. Phys., 6, 3423‐3441. SRefID: 1680-7324/acp/2006-6-3423.

Van der Werf et al. (2010), Global fire emissions and the contribution of deforestation, savanna, forest, agricultural, and peat fires (1997–2009), Atmos. Chem. Phys., 10. Vignati, E., M. Karl, M. Krol, J. Wilson, P. Stier, and F. Cavalli (2010), Sources of uncertainties in modelling black carbon at the global scale, Atmos. Chem. Phys., 10, 2595‐2611, doi:10.5194/acp-10-2595-2010. Wang, C. (2004), A modeling study on the climate impacts of black carbon aerosols, J. Geophys. Res., 109, D03106. doi:10.1029/ 2003JD004084. Wang, Z. L., H. Zhang, and X. S. Shen (2011), Radiative Forcing and Climate Response Due to Black Carbon in Snow and Ice, Adv. Atmos. Sci., 28 (6), 1336‐1344, doi: 10.1007/s00376-011-0117-5, Warneke C., et al. (2009), Biomass burning in Siberia and Kazakhstan as an important source for haze over the Alaskan Arctic in April 2008, Geophys. Res. Lett., 36, pp. L02813, doi: 10.1029/2008gl036194. Warren S. and W. Wiscombe (1985), Dirty snow after nuclear war, Nature 313, 467‐470. doi: 10.1038/313467a0. Worthy, D. E. J., E. Chan, M. Ishizawa, D. Chan, C. Poss, E. J. Dlugokencky, S. Maksyutov, and I. Levin (2009), A decrease in anthropogenic methane emissions in Europe and Siberia inferred from continuous carbon dioxide and methane observations at Alert, Canada, J. Geophys. Res., 114 (D10301). doi: 10.1029/2008JD011239.

964