High temporal resolution measurements of black carbon (BC) and organic carbon
(OC) covering the time period of 1956–2006 in an ice core over the
southeastern Tibetan Plateau show a distinct seasonal dependence of BC and OC
with higher respective concentrations but a lower OC
Carbonaceous aerosol, released from fossil fuel, biofuel and/or biomass combustion, contains both black carbon (BC, a.k.a. elemental carbon, EC), a strong light absorber, and organic carbon (OC), which also absorbs the near infrared, but more weakly than BC (Kirchstetter et al., 2004; Bond and Bergstrom, 2006). Often mixed with other aerosol species, BC impacts human health, crop yields and regional climate (Auffhammer et al., 2006; Tie et al., 2009), and is believed to be the second strongest climate warming forcing agent after carbon dioxide (Jacobson, 2001; IPCC, 2013).
Because of their high population density and relatively low combustion efficiency, developing countries in South and East Asia such as India and China are hotspots of carbonaceous aerosol emissions (Ramanathan and Carmichael, 2008). During the cold and dry winter season, haze (heavily loaded with carbonaceous aerosols) builds up over South Asia, and exerts profound influences on regional radiative forcing (Ramanathan et al., 2007; Ramanathan and Carmichael, 2008), hydrologic cycles (Menon et al., 2002; Ramanathan et al., 2005), and likely Himalaya–Tibetan glacier melting that could be accelerated by the absorption of sunlight induced by BC in the air and deposited on the ice and snow surfaces (Ramanathan et al., 2007; Hansen and Nazarenko, 2004), although BC deposited in snow and glaciers at some locations may not significantly affect the energy balance (Ming et al., 2013; Kaspari et al., 2014).
Due to the lack of long-term observations of emissions and concentrations of atmospheric carbonaceous aerosols, it is difficult to evaluate the effects of BC and OC on historical regional climate and environment before the satellite era. Some studies have evaluated historical anthropogenic emissions based on the consumption of fossil fuels and biofuels (Novakov et al., 2003; Ito and Penner, 2005; Bond et al., 2007; Fernandes et al., 2007). While fossil fuel is the major energy source in the urban areas of South Asia and East Asia, biomass combustion, such as fuel wood, agricultural residue and dung cake, is prevalent in rural areas (Revelle, 1976; Venkataraman et al., 2010; Streets and Waldhoff, 1998). Biomass burning has been considered as the major source of black carbon emissions (Reddy and Venkataraman, 2002; Venkataraman et al., 2005). However, as reliable biomass consumption data are hard to obtain, estimates of BC and OC emissions from biomass burning are ambiguous and incomplete.
Measurements of carbonaceous aerosol concentrations in glacier ice are an
ideal means of reconstructing historical emissions and revealing long-term
trends of anthropogenic aerosol impacts on local climate. Greenland ice-core
measurements were previously used to reconstruct the North American BC
emission history and its effects on surface radiative forcing back to the
1880s (McConnell et al., 2007). Himalayan ice cores retrieved from the
Tibetan Plateau have revealed the mixed historical emissions from South Asia,
Central Asia and the Middle East, and have also been used to evaluate
radiative forcing from BC in snow (Ming et al., 2008; Kaspari et al., 2011).
Using the Snow, Ice, and Aerosol Radiative (SNICAR) model, Flanner et
al. (2007) estimated an instantaneous regional forcing exceeding
20 W m
By using five ice-core records, Xu et al. (2009a) elucidated an important contribution of BC to the retreat of Tibetan glaciers in addition to greenhouse gases. Due to the short atmospheric lifetime of carbonaceous aerosols compared to greenhouse gases, emission reductions may be an effective way to mitigate their warming effects. Thus, it is particularly important to identify the source regions and the source types of carbonaceous aerosols observed in Tibetan glaciers. Xu et al. (2009a) suggested that BC deposited on the Tibetan Plateau was broadly from Europe and Asia. However, they did not perform in-depth analysis on emissions from more specific source regions and the source types. In this study, we use the ice core retrieved from the southeastern Tibetan Plateau, also known as the Zuoqiupu ice core in Xu et al. (2009a), to reconstruct the history of atmospheric deposition of carbonaceous aerosols in this glacier, and to characterize emissions and source–receptor relationships with the help of a global climate model in which BC emitted from different source regions can be explicitly tracked. We also estimate the respective contributions from BC and OC to radiative forcing in the Zuoqiupu glacier using the ice-core measurements and the SNICAR model.
Zuoqiupu glacier is in the southeastern Kangri Karpo Mountains, located on
the southeastern margin of the Tibetan Plateau (Fig. 1). In 2007, an ice core
of 97 m in depth (9.5 cm in diameter) was retrieved within the accumulation
zone of Zuoqiupu glacier at 96.92
Site location of Zuoqiupu glacier (top): black circle represents the location of Zuoqiupu glacier, and warm colors indicate high elevations over the Tibetan Plateau. Detailed elevation contours of the Zuoqiupu glacier are shown in the bottom panel. Red circle marks the ice-core drilling site.
We use the Community Atmosphere Model version 5 (CAM5; Neale et al., 2012) to help understand the emissions, transport and dry/wet deposition of carbonaceous aerosols in the atmosphere. In the default three-mode modal aerosol scheme of CAM5 used for this study, BC and primary OC are emitted into an accumulation size mode, where they immediately mix with co-existing hygroscopic species such as sulfate and sea salt (Liu et al., 2012). Hygroscopic aerosol particles in the accumulation mode are subject to wet removal by precipitation. Recent model improvements to the representation of aerosol transport and wet removal in CAM5 by H. Wang et al. (2013) have substantially improved the model prediction of global distribution of aerosols, particularly over remote regions away from major sources. To minimize the model biases in simulating meteorological conditions and, particularly, circulations that are critical to aerosol transport, we configure the CAM5 model to run in an offline mode (Ma et al., 2013) with wind, temperature, surface fluxes and pressure fields constrained by observations. However, cloud/precipitation fields and interactions between aerosol and clouds are allowed to evolve freely. A source tagging technique has recently been implemented in the CAM5 model to allow for explicit tracking of aerosols emitted from individual source regions and, therefore, to assist in quantitatively characterizing source–receptor relationships (Wang et al., 2014). This tagging technique along with the CAM5 model is used in the present study to do source attribution for carbonaceous aerosols deposited onto the Zuoqiupu glacier.
We conducted an 11-year (1995–2005) CAM5 simulation on a horizontal grid
spacing of 1.9
As the ice-core drilling site was located in a remote and elevated area over the southeastern Tibetan Plateau, local emissions are minimal. Deposition of carbonaceous aerosols is most likely contributed by the non-local major emission sources (e.g., distributions of mean BC emissions during non-monsoon and monsoon seasons shown in Fig. 2) in South Asia and East Asia. These two regions, along with Southeast Asia and Central Asia, are identified as the potential source contributors. Thus, BC emissions from the four regions and the rest of the world are explicitly tracked in the CAM5 simulation.
BC and OC concentrations in the Zuoqiupu ice core both exhibit statistically significant seasonal variations corresponding to the stable oxygen isotope variability, which shows high values during the winter and low values during the summer (Xu et al., 2009a). As shown in Fig. 3, concentrations of BC and OC have distinct differences between the summer monsoon and non-monsoon seasons. Seasonally varying emissions and meteorological conditions that determine the transport pathways of BC and OC emitted from major sources, removal during the transport, and local precipitation rate, can cause the seasonal variations of BC and OC in ice at the sampling site. The seasonal dependence of BC and OC in ice cores is consistent with available observations of atmospheric aerosols on the southern slope of the Himalayas and the southeastern Tibetan Plateau, where the high concentration of carbonaceous aerosols during the cold and dry season suggested an association with the South Asian haze (Cong et al., 2009; Marinoni et al., 2010; Kaspari et al., 2011; S. Zhao et al., 2013; Z. Zhao et al., 2013). The consistency between the seasonal dependence of airborne BC and OC concentrations and the seasonal variation of ice-core measurements indicates that seasonal differences in precipitation rates at the sampling location are less likely to be the determining factor. Our model results (details discussed in Sect. 3.2) suggest that the seasonal dependence of BC deposition flux in the target region could be mainly due to meteorological conditions that determine the transport pathways (and associated wet removal processes during the transport). The small seasonal contrasts in BC emissions from the major source regions (see Table 1) that are used in the model simulation do not seem to be able to explain the large seasonal difference in BC deposition, although the BC emissions are known to have large uncertainties.
10-year (1996–2005) mean wind vectors (denoted by arrows) at
500 hPa
Scatterplots for yearly monsoon and non-monsoon mean OC and BC concentrations during 1956–2006, obtained from the ice-core measurements, and corresponding linear regressions.
Source regions (South Asia, East Asia, Southeast Asia, and Central
Asia) and corresponding monthly mean BC emissions (Tg a
Our further analysis shows that the ratio of OC to BC also has clear seasonal
dependence. In Fig. 3, the slope of the fitted line to measured OC vs. BC
concentrations during monsoon season is
To attribute the source of BC at the drilling site (as a receptor region) quantitatively, we use the CAM5 model with the BC source tagging capability to conduct an 11-year simulation, with the last 10 years (1996–2005) used for analysis. The surrounding area is divided into four source regions (see Table 1 and Fig. 4): South Asia, East Asia, Southeast Asia and Central Asia. BC emissions from each of the four regions and the rest of the world are explicitly tracked, so that the fractional contributions by emissions from the individual source regions to BC deposition in the receptor region can be explicitly calculated. Figure 4 shows the spatial distribution of fractional contribution from the four source regions. BC deposition at the drilling site (indicated by the black box in Fig. 4), which has a consistent seasonal dependence (i.e., more during the non-monsoon season; Fig. 5) with ice-core measurements, is predominately (over 95 %) from South Asia and East Asia. The seasonal dependence of BC deposition is also consistent with a recent regional climate modeling study on BC deposition on the Himalayan snow cover from 1998 to 2008 (Ménégoz et al., 2014).
Spatial distributions of fractional contribution from the four source regions (South Asia, East Asia, Southeast Asia, and Central Asia) to monsoon, non-monsoon, and annual mean BC deposition fluxes during 1996–2005. The large black boxes indicate the boundary of source regions, and the small black box marks the model grid cell where the Zuoqiupu drilling site is located. Color in the small black box in each panel corresponds to the fraction contribution to BC deposition at the sampling site. Exact percentage contributions are provided in Table 1.
Seasonal dependence (“NM” for non-monsoon season and “M” for monsoon season) of BC deposition flux at the Zuoqiupu site from 1995 to 2005 simulated in CAM5. The dashed line represents a linear regression of all data points.
Time series of annual (dotted line with circles) and 5-year averaged
(solid line) OC
The 10-year (1996–2005) average wind fields (at the surface and 500 hPa from MERRA reanalysis data sets), as shown in Fig. 2, indicate distinct circulation patterns during the summer monsoon (June–September) and non-monsoon (October–May) seasons, which in part determine the seasonal dependence of transport of aerosols emitted from the different major sources. During the non-monsoon season, a strong westerly dominates the transport from west to east at all levels. Emissions from northern India and Central Asia can have an influence on BC in the direct downwind receptor region over the southeastern Tibetan Plateau. During the summer monsoon season, the westerly moves northward, and the monsoon flow from the Bay of Bengal at the surface and middle levels (e.g., 500 hPa), coupled with the monsoon from the Indochina peninsula and the South China Sea, exert an influence on BC in the receptor area. The strong monsoon precipitation removes BC from the atmosphere during the transport. The high Himalayas can partly block the further transport of emissions from South Asia to the Tibetan Plateau, although small local topographical features such as the Yarlung Tsangpo River valley can provide a gate for the pollution to enter the inner Tibetan Plateau (Cao et al., 2010). Elevated emissions from the west (or the northern part of South Asia) can take the pathways at middle and upper levels, but they have minimal contribution to deposition. Therefore, BC emissions from East Asia play a relatively more important role in affecting deposition at the Zuoqiupu site during the monsoon season.
The fractional contributions to 10-year mean BC deposition at the drilling
site from the four tagged regions are summarized in Table 1. Results show
that South Asia is the dominant contributor (
For comparison, seasonal and annual mean BC emissions from the individual tagged source regions are also included in Table 1. Apparently, the contrast in strengths of regional emissions alone cannot explain their relative contributions to BC deposition at the sampling site, and the small seasonal variations in emissions are unlikely to be the cause of seasonal dependence of source attribution. Note that the BC emission inventory (Lamarque et al., 2010) used in CAM5 does not consider seasonal variations in anthropogenic emissions, which is likely to have introduced biases in the quantitative model estimates of seasonal dependence of contributions, but the relative importance of source regions should be robust.
Based on annual snow accumulation and BC and OC concentrations derived from
the ice-core record, the annual BC and OC deposition fluxes can be estimated,
which are then used to examine the interannual variations and long-term trend
in the fluxes and the ratio of OC
As shown in the CAM5 model simulation, the annual mean atmospheric deposition
of BC over the southeastern Tibetan Plateau is mostly contributed by
emissions from South Asia, particularly in the non-monsoon season. The BC and
OC deposition fluxes derived from the ice-core measurements may reflect
changes in South Asian emissions to some extent. The temporal variations of
BC and OC deposition fluxes (see Fig. 6) are compared with the primary BC and
OC emissions from fossil fuel and biofuel combustion in South Asia during
1955–2000 (Bond et al., 2007). BC and OC emissions during 1996–2010 from Lu
et al. (2011) are also illustrated in Fig. 6 to extend the emission data to
cover the entire time period that the ice-core data span. Note that the emission data from
Lu et al. (2011) are only for India, which is the largest energy consumer and
carbonaceous aerosol-emitting country in South Asia. There are differences
between the emissions of Bond et al. (2007) and Lu et al. (2011) during the
overlap time period (1996–2000). However, good agreements on the increasing
trend can be found in the respective deposition fluxes and emissions of BC
and OC (Fig. 6). The OC
BC and OC in the atmosphere are co-emitted from a variety of natural and
anthropogenic sources, including combustion of fossil fuel, biofuel and/or
biomass burning. In general, open biomass burning typically produces more
abundant OC (i.e., a larger OC
The temporal variations of BC and OC in the Zuoqiupu ice core, along with the
source attribution analysis of the CAM5 model results, suggest an increasing
trend in emissions and altered emission sources in South Asia during the late
twentieth century. Coal has been the primary energy source in South Asia. For
example, in India, coal accounted for 41 % of the total primary energy
demand in 2007, followed by biomass (27 %) and oil (24 %) (IEA,
2009). The consumption data of coal and crude oil in South Asia (BP Group,
2009) is compared with the BC and OC fluxes in Fig. 6 (bottom right). Coal
consumption had an increasing trend from 1965 to 2008, particularly in the
two time periods 1980–1995 and 2003–2008 after a leveling off during
1996–2002. This trend is consistent with the variations of BC and OC
deposition fluxes in the Zuoqiupu ice core. The correlations between coal
consumption and BC (
Biomass is the second largest energy resource in South Asia, and it is
essential in rural areas. In India, 70 % of the population lives in rural
areas, and depends substantially on solid fuels (i.e., firewood, animal dung,
and agriculture residues) for cooking and heating (Heltberg et al., 2000).
Even in urban areas, biomass contributes 27 % of the household cooking
fuel (Venkataraman et al., 2010). Although the consumption of biomass is
lower than coal, the OC
BC is often the most important light-absorbing impurity in surface snow because of its strong absorption of solar radiation. The effect of BC in snow on surface albedo reduction and the resultant positive radiative forcing have been widely addressed and reported (e.g., Warren and Wiscombe, 1980; Clarke and Noone, 1985; Hansen and Nazarenko, 2004; Hadley and Kirchstetter, 2012; Flanner et al., 2007; 2009; McConnell et al., 2007; Ming et al., 2008; Kaspari et al., 2011; Qian et al., 2011, 2014, 2015). In contrast, the impact of OC in snow has not been widely assessed because of its relatively weak light absorption over the entire spectrum compared to BC, and because of large uncertainties associated with OC light-absorbing properties and measurements of OC in snow. However, there has been increasing interest in light-absorbing OC (a.k.a. brown carbon) and its radiative effect in both the atmosphere and snow. A growing number of studies (e.g., Kirchstetter et al., 2004; Andreae and Gelencsér, 2006; Hoffer et al., 2006; Yang et al., 2009; Kirchstetter and Thatcher, 2012) have reported that airborne brown carbon can contribute significantly to aerosol light absorption in the atmosphere, although there are still substantial uncertainties in quantifying optical properties of brown carbon, which makes the model estimation of OC radiative forcing difficult. Similarly, the importance of OC absorption in snow has been recognized and suggested for inclusion in modeling aerosol snow-albedo effects (e.g., Flanner et al., 2009; Aoki et al., 2011). Observational analysis of light-absorbing particles in Arctic snow reported that the main non-BC component is brown carbon, which accounted for 20–50 % of the visible and ultraviolet absorption (Hegg et al., 2009, 2010; Doherty et al., 2010). In the rural area of central northern China, brown carbon in winter snow also played an important role in visible light absorption, which contributed about 60 % to light absorption at 450 nm and about 40 % at 600 nm (X. Wang et al., 2013). A more recent observational study by Dang and Hegg (2014) quantified the light absorption by different light-absorbing particulates in snow, and suggested that humic-like substances and polar OC contributed 9 and 4 % to the total light absorption, respectively. Despite the substantial uncertainties in brown carbon optical properties, a recent global modeling study (Lin et al., 2014), in which a range of optical properties of brown carbon taken from the literature were applied to OC-in-snow concentrations simulated in a global chemical transport model, showed that the global OC forcing in land snow and sea ice is up to 24 % of that caused by BC. Thus, the contribution of OC in snow to the surface albedo reduction is likely to be important, which has also been considered in recent climate modeling studies (Qian et al., 2015).
In this study, we use the SNICAR online model (available at
The SNICAR model currently does not support the calculation of OC-in-snow
forcing in the same way as that for BC due to a lack of reliable OC optical
properties that span the dimensions of snow grain size and OC particle size
(M. Flanner, personal communication, 2014). We take a MAC value of
0.6 m
Two main assumptions could have caused our first-order estimate of OC forcing
to have large biases. First, the MAC value of 0.6 m
It is also important to note that we did not consider variations in chemical compounds of OC, the changes in OC during sample filtration, and the different spectral dependence of OC and BC absorption. Although such uncertainties can also cause bias in the estimation of OC radiative forcing herein, the increasing trend should be robust.
BC and OC concentrations in the ice core increased rapidly after 1980, and the induced radiative forcing rose as a consequence. According to the estimates using the SNICAR model, the average BC radiative forcing had increased 43 % after 1980, and OC radiative forcing had an increase of 70 %. These numbers are by no means accurate, but the stronger increasing trend in the ice-core recorded OC than in BC during 1990–2006 (Fig. 6) suggests that the contribution of OC to the total radiative forcing in the glacier induced by snow/ice impurities deserves more attention.
Light-absorbing carbonaceous aerosols can induce significant warming in the
atmosphere and in snow and glaciers, which likely accelerates the melting of
glaciers over the Himalayas and the Tibetan Plateau. Ice-core measurement of
carbonaceous aerosols is a useful mechanism for evaluating historical
emission inventories and revealing long-term changes in anthropogenic
aerosols and their impacts on regional climate. In this study, we analyze
carbonaceous aerosols recorded in an ice core (97 m in depth and 9.5 cm in
diameter) retrieved from the Zuoqiupu glacier (96.92
BC and OC concentrations in small segments of the Zuoqiupu ice core were
measured using a thermal-optical method. Ice-core dating based on significant
seasonal variations of oxygen isotope ratios (
The MERRA reanalysis products used to drive the CAM5 model simulation show distinct circulation patterns during the summer monsoon (June–September) and non-monsoon (October–May) seasons. Both the circulation patterns (and the associated aerosol transport and wet removal) and seasonal variation of emissions in major source regions influence the seasonal deposition of aerosol at the Zuoqiupu site. The CAM5 simulation with tagged BC regional sources shows that South Asia is the dominant contributor (81 %) to the 10-year mean BC deposition at the Zuoqiupu site during the non-monsoon season, with 14 % from East Asia, while the contribution of East Asia (56 %) is larger than that of South Asia (39 %) during the monsoon season. For the annual mean BC deposition, South Asia (75 %) is the biggest contributor, followed by East Asia (21 %).
The annual mean BC and OC deposition fluxes into the ice core are also estimated to explore the interannual variations and long-term trends. Results show stable and relatively low BC and OC fluxes from the late 1950s to 1979, followed by a steady increase through the mid-1990s. A more rapid increase occurred after the minimum in 2002. The BC and OC deposition fluxes in 2006 were 2 and 3 times the respective average before 1980.
The overall increasing trend in deposition fluxes since 1980 is consistent
with the BC and OC emissions in South Asia as the major contributor.
Moreover, the increasing trend of the OC
Our offline calculation using the SNICAR model shows a significant increase in radiative forcing induced by the observed BC and OC in snow after 1980, which has implications for the Tibetan glacier melting and availability of water resources in the surrounding regions. More attention to OC is merited because of its non-negligible light absorption and the recent rapid increases evident in the ice-core record.
This work was supported by the China National Funds for Distinguished Young Scientists and the National Natural Science Foundation of China, including 41125003, 41101063, and 2009CB723901. H. Wang, Y. Qian and P. J. Rasch were supported by the US Department of Energy (DOE), Office of Science, Biological and Environmental Research as part of the Earth System Modeling program. R. Zhang acknowledges support from the China Scholarship Fund. PNNL is operated for DOE by Battelle Memorial Institute under contract DE-AC05-76RLO1830. The National Center for Atmospheric Research is sponsored by the National Science Foundation. We thank Z. Guo and S. Yang for providing the observations of snow. Edited by: X. Xu