As with other USAP facilities, Palmer Station operates on petroleum and station activities produce hydrocarbon emissions. For Fiscal Years 2006–2010, fuel consumption at Palmer Station averaged 104,792 gallons annually. Of this, 80% was consumed by the station’s power plant, 19% for heating and hot water and 1% for vehicles (RSA Engineering, 2011). Two 250 kW Caterpillar engine generators (circa 1998) supply power to Palmer Station and heat to the building in which they are located. Oil boilers provide hot water for the other buildings. Machinery and small boats, at the time of this study, primarily inflatable Zodiacs, are also operated on station. A small incinerator is operated sporadically to incinerate poultry waste and a small wood-burning stove operates in one of Palmer’s two main buildings. Palmer Station has ∼20% of the population of the South Pole at its maximum.

The Research/Vessel (R/V) Laurence M. Gould (LMG), operated by Lockheed Martin, supports Science and operations at Palmer Station. The LMG is a 230 foot long ice-classed ABS-A1 research and resupply vessel on a long-term charter from Offshore Service Vessels LLC. The LMG has two main engines (Caterpillar 3606) and three generators (Caterpillar 3508). USAP’s larger vessel, the Nathaniel B. Palmer, services Palmer Station occasionally and approximately a dozen cruise vessels visit the station each year.

The LMG typically averages 8–9 visits to Palmer Station each year (RSA Engineering, 2011). During the period of snow accumulation in this study, the LMG made three cruises to Palmer Station LMG 15-02 (2/13-3/16), 15-03 (3/20-4/4), and 15-04 (4/7-5/11). Based on analysis of the LMGs navigational data (Aronson, 2015; O’Brien et al., 2015) and cruise history (United States Antarctic Program[USAP], 2018), the LMG spent approximately 13 days docked at Palmer Station from the first day of snow accumulation March 9th through sample collection on May 5th.

On 5 May 2015, during a period without recent snowfalls, surface snow samples were collected for rBC analysis at eight sites along a 1.7 km transect extending eastward from Palmer Station. These sites were augmented by samples collected from two more remote sites located approximately 14 km from Palmer Station on Biscoe Point (now an island) situated within the Biscoe Point Antarctic Specially Protected Area (ASPA) No. 139 on 1 May 2015. Entry to an ASPA requires special permitting to sample these pristine sites (Figure 1).

At each site three replicate sample duplicates were collected. rBC concentrations reported in Table 1 are the average of the three sample duplicates. Snow was collected in 500 ml glass jars which were pre-cleaned to meet Environmental Protection Agency (EPA) Protocol A Level 1 standards for contaminants. Samples were obtained by placing the opening of the glass jar perpendicular to the horizontal surface of the snowpack and pressing the jar through the snowpack until the snow completely filled the jar. Care was taken so the surface was only sampled once per replicate.

After collection, snow samples were kept frozen at Palmer Station, during their transport to Punta Arenas, Chile and onward to Texas A&M University until they were shipped to the Institute of Arctic and Alpine Research (INSTAAR) in Boulder, CO and then transferred to the Earth System Research Laboratory at the National Oceanic and Atmospheric Administration for rBC analysis. Snow samples were then melted for the first time immediately before analysis. The thirty surface snow samples were analyzed for rBC mass mixing ratios (MMR). Briefly, snow samples were melted, then immediately aerosolized with a carefully calibrated nebulizer and sampled with an SP2 following the approach of Schwarz et al. (2012). The SP2 has excellent sensitivity and selectivity for rBC with minimal interference from non-BC materials and was calibrated with fullerene soot (Alfa Aesar, Woodbridge, MA, United States, Lot F12S011). Here, a CETAC Marin-5 nebulizer interfaced with the SP2, following the same protocol as Katich et al. (2017). The size-dependent nebulization efficiency was characterized with concentration standards of polystyrene latex spheres (PSLs) in the size range 220–3000 nm diameter, confirming low sensitivity to particle size over this range, consistent with recent results with concentric pneumatic nebulizers (Lim et al., 2014; Wendl et al., 2014; Mori et al., 2016; Katich et al., 2017). Size distributions were measured for samples with higher concentrations, and these indicated no appreciable contribution to rBC mass above 0.7 μm. Hence no size dependent corrections were applied for nebulizer size dependence. Background rBC contamination levels were bounded by measuring blank samples. Deionized water was used to flush the nebulization system between snow samples to reduce line contamination. Further, measurements of rBC concentrations in blank water samples were interspersed amongst snow samples to bound background contamination levels in the system. This averaged 0.43 μg/L over 15 blank water measurements, indicating a low enough background to avoid significant interference with the rBC MMR measurements. The blanks provide an upper bound on possible contaminants associated with the estimated 60% uncertainty for rBC MMR in the snow, which conservatively accounts for uncertainties dominated by calibration issues (Schwarz et al., 2012).

Information on local meteorological conditions during the period of early snow accumulation in 2015 was obtained from two sources. Local Daily Summaries provided daily weather observations including snow depth and accumulation measurements recorded by Palmer Station personnel and were obtained from the Antarctic Meteorological Research Center (AMRC) at the University of Wisconsin (Antarctic Meteorological Research Center [AMRC], 2018) 2-min observations from the Palmer Automated Weather System (PALMOS) available from the Palmer Station Antarctica LTER (Long Term Ecological Research Network-Network Data Portal [LTER-NDP], 2017) were used to investigate wind speed and directions during the period of snow accumulation. Analysis of the weather observations was accomplished using the Anaconda Scientific Python Distribution. The wind roses are produced using IDL software (Fanning, 2019). All mapping was accomplished using ArcMap or ArcGIS Pro.

Smoke aerosol optical depth (AOD) and concentrations from biomass burning comes from the Navy Aerosol Analysis and Prediction System (NAAPS) reanalysis (Lynch et al., 2016), which was developed and validated at the Naval Research Laboratory. The NAAPS model was used to explore regional and long-range biomass burning smoke transport events during the study period. It is a global 1°×1° decade-long 6-hourly aerosol reanalysis product. It utilizes a modified version of the NAAPS as its core and assimilates quality controlled retrievals of AOD from Moderate Resolution Imaging Spectroradiometer (MODIS) on Aqua and Terra, as well as the Multi-angle Imaging Spectroradiometer (MISR) on Terra (Zhang and Reid, 2006; Hyer et al., 2011; Shi et al., 2014). NAAPS characterizes anthropogenic and biogenic fine aerosol (including primary and secondary organic aerosols and sulfate), biomass burning smoke, dust, and marine sea salt aerosols. Biomass burning smoke is derived from near-real time thermal anomaly satellite based data used to construct smoke source functions (Reid et al., 2009), with orbital corrections on MODIS-based regional tunings and emissions. The fine and coarse mode reanalysis AOD at 550 nm is shown to be in agreement with the ground-based global scale sun photometer network Aerosol Robotic Network AOD (Holben et al., 1998).

rBC concentrations in surface snow decrease with increasing distance from Palmer Station (Figure 1 and Table 1). Elevated rBC concentrations were found near the station, 16.5–1.5 μg-rBC/L-H₂O, especially as compared to the low concentrations found on Biscoe Island, 0.4 and 1.2 μg-rBC/L-H₂O. This suggests that the station’s influence on rBC extends at least 1 km from the station. rBC concentrations in this study near Palmer station were higher than other studies of rBC in snow on the Antarctic continent, such as in the MDV, 0.3–1.2 μg-rBC/L-H₂O, except on the more remote Biscoe Island, 0.4 and 1.2 μg-rBC/L-H₂O, which like the MDV, has restrictions to entry, and had similar background concentrations indicative of long-range atmospheric transport. However, concentrations in this study were less than previous findings downwind of the generator and along the runway at the South Pole Station (6.6–7000 μg-rBC/L-H₂O) (Casey et al., 2017). It should be noted that Palmer Station has ∼20% of the population of the South Pole at its maximum. Additionally, samples from the clean air sector at the South Pole (0.14–0.80 μg-rBC/L-H₂O) (Casey et al., 2017) contained similar background concentrations of rBC as found in this study and in the MDV. Although, significant increases in rBC concentrations are found within a km of Palmer Station, like the South Pole Station, rBC is limited to a few kms from station.

During 2015, Palmer Station recorded some of its highest snow depths over the period of record (Figure 2). The period preceding snow sampling on May 1st and 5th saw intermittent snow on the ground from March 9 to April 5 after which snow was continuously present at the station. It reached a depth of 40 cm in late April. However, April 28–30 saw air temperatures remaining above freezing and maximum air temperatures exceeding 4° causing snow to disappear at the Palmer snow stake on May 1st. Therefore, the snow sampled on May 1st and 5th represent the snow that remained in patches or on the glacier from the early period of well-above average snow accumulation.

The wind roses illustrated in Figure 3 show that surface winds during the snow accumulation period of the sampling differ from long-term averages for the entire year and the March to May (MAM) period. The period of snow accumulation captured by the sampling was characterized by a much higher occurrence of SSW surface winds than is typical.

According to the NAAPS analysis, a small amount of biomass burning smoke was transported southward from South America to Palmer Station on April 9, 2015 (Figure 4a) and on April, 28, 29, 2015 (Figure 4b). As evidenced by the NAAPS model plots, when there is southward wind blowing from South America, there is possible smoke transport to the Antarctic Peninsula (Figure 4a,b). Thus, it is possible that biomass burning smoke was deposited on the sampled snow layer. However, AOD time series of the study site for the whole year of 2015 (Figure 4c) show that the amount of smoke transported over the study period is smaller compared to the peak biomass season (September–October) transport. Palmer Station is located in the roaring circumpolar strong westerly wind belt; therefore sea salt is excluded in the time series plot, as it dominates the AOD signal.