Data Package Metadata   View Summary

Dust samples were collected at four sites along an urban-rural transect extending nearly 100 km from Columbus to Gambier, OH, during June and July 2021. We collected additional samples from March – December 2021 at the rural Kenyon College site to capture the seasonal variability in dust geochemistry. We used passive samplers to collect weekly bulk deposition samples at each site, which were then filtered to isolate the particulate dust. Three samplers were installed at Scioto Audubon, Blendon Woods, and Rocky Fork; five samplers were installed at Kenyon College. Each sampler consisted of a 25.4 cm diameter plastic funnel that drained into a 1 L amber Nalgene bottle through flexible tubing. Funnels were filled with marbles to retain dust particles once they were deposited. Although the samplers were installed 0.5 m above the ground surface, the collectors were surrounded by restored prairie or mown grass at each site, limiting the potential for local soil to enter the collectors. Each week, the 1 L amber Nalgene bottles containing precipitation were removed. The funnels and marbles were then thoroughly rinsed with deionized water, and the rinse was collected in a clean Nalgene bottle. Finally, a clean Nalgene bottle was attached to the collector for the following weekly sample. To isolate the particulate dust from the liquid precipitation and deionized water rinse, we filtered samples through pre-weighed 0.45 micron polyethersulfone (PES) filters. We reweighed the filters and dust after drying to calculate the total mass of dust in each sample. For dust geochemistry and Pb isotope analyses, we selected 18 weeks over the course of our 2021 sampling effort to capture as much seasonal variability as possible, and for each week randomly selected one collector from each site that was in operation, resulting in a total of 36 samples. The majority of samples consisted of two filters with dust particles: one from the filtration of the precipitation and one from the filtration of the collector rinse. Eight samples only contained the rinse filter, due to a lack of precipitation during the collection week. These samples included 1 at Scioto Audubon, 1 at Blendon Woods, and 6 at Kenyon College. Dried filters were stored at Kenyon College until they were sent to the Trace Elements Clean Lab at the Wisconsin State Laboratory of Hygiene (WSLH) for geochemical and isotopic analysis.

At WSLH, samples (PES filters plus dust) were digested by hot block at 95 degrees C for two hours in polypropylene tubes, with 3.5 mL 16M nitric acid, 0.5 mL 12M hydrochloric acid, and 0.2 mL 28M hydrofluoric acid. Digested samples were then subsampled for elemental analysis and Pb isotopic analysis. For elemental analysis, 55 elements were analyzed on a magnetic sector ICP-MS (Thermo-Finnigan Element2). Samples were run in duplicate, with average values used for data analysis. Reference materials, method blanks, and filter blanks were included in each run. Reference materials were used to calculate recovery for quality assurance. Method blanks were subtracted from each sample value. The filter blanks each included two PES filters to mimic the majority of the samples. Values for the filter blanks were then subtracted from each sample value; for the eight samples with only one filter, one half of the blank value was subtracted from the sample value. Due to high and/or low recoveries throughout the runs, data for As and Se are not reported, and data for Ce, Cu, La, Pt, and U are incomplete. For these elements, values are included in the enrichment factor calculations but are not included in the principal component analyses that require complete datasets.

Lead isotopes were measured at the Trace Elements Clean Lab at WSLH using a Neptune Plus (Thermo-Fisher) multi-collector inductively coupled plasma mass spectrometer (MC-ICP-MS). Prior to the analyses, aliquots of the samples were dried and redissolved in 1 M hydrobromic acid (HBr) for Pb separation using anion exchange chromatography. Samples were loaded onto columns containing previously cleaned AG1-X8 anion exchange resin conditioned in 1 M HBr. The bulk of the sample matrix was eluted with 1 M HBr, while Pb remained adhered to the resin. Lead was then removed from the resin using 1 M nitric acid (HNO3). The collected Pb fraction was dried and then redissolved in 2% HNO3 for isotopic analysis. Samples and standards were diluted to ~5 ppb Pb and introduced into the plasma through an Aridus 3 desolvating nebulizer system coupled to an ESI SC-DX2 autosampler. The Jet sampler and “X” skimmer cones were used to enhance sensitivity. Thallium (Tl) was added to each sample at a Pb:Tl ratio of ~3:1 to correct for mass bias using the exponential law and assuming 205Tl/203Tl = 2.387. Lead and Tl isotopes were measured on six Faraday cups in static mode using 1011 Ohm resistors. An additional Faraday cup was used to monitor 202Hg, which was not present in any sample. Isotope ratios were measured in static mode for 60 cycles of 8 second integrations. Certified Pb isotope reference material NBS981 was measured periodically, and final Pb isotopic data are reported relative to NBS981 values of (Thirlwall, 2002). The NIST reference materials urban particulate matter (NIST 1648a), auto catalyst (NIST 2556), San Juan soil (NIST 2709), and marine sediment (NIST 2702) were run for quality assurance and for comparison purposes. We obtained good agreement with literature values for NIST 2709 and report what may be the first published Pb isotope values for NIST 1648a and NIST 2556.

Thirlwall, M.F., 2002. Multicollector ICP-MS analysis of Pb isotopes using a 207pb-204pb double spike demonstrates up to 400 ppm/amu systematic errors in Tl-normalization. Chemical Geology 184, 255–279. https://doi.org/10.1016/S0009-2541(01)00365-5

To identify elements with anthropogenic sources, we calculated crustal enrichment factors (EFs) normalized to the reference element aluminum (Al):

EFUCC = [Cx-sample/CAl-sample]/[Cx-UCC/CAl-UC]

where Cx-sample is the concentration of a particular element in the sample and Cx-UCC is the average concentration of that element in the upper continental crust. Crustal enrichment factors have been used in dust and aerosol studies to differentiate elements with mostly crustal sources (EF < 10) from elements with anthropogenic sources (EF >> 10).