Hi-Vol Filter Sampling:
In early 2014 a high-volume aerosol filter sampling system was installed at a site situated on the top of a 67-meter wooded hill (4.948936N, ‑52.309692W) located directly on the coast that lies northeast of the city of Cayenne (population 57,000) (see Fig. 1). The site is a fenced and gated government radar facility with restricted access. Cayenne lies in the Trade Wind belt and winds have a very strong NE component all year long. Figure S2 shows selected seasonal wind roses from a site 1.3 km SE of the sampling site. Sampling is carried out on a cooperative basis by personnel of ATMO. Samples are changed on a nominal 24-hour cycle starting around local noon although the start and end times are variable, depending on the workload of the personnel. Exposed filters are replaced with a new filter the following morning except over weekends and holidays when multi-day samples are taken. Most 3-day filter samples in the archive are weekend samples that were begun on a Friday morning and terminated on a Monday morning. Also, because of work schedule limitations and holidays, some filter changes may be missed during the week or extend over longer times. Note that the filter change cycle starts approximately at local noon and, thus, the schedule is not precisely coordinated with the daily PM measurement cycle which starts at midnight.
Samples are collected by drawing air through 20cm x 25cm Whatman 41 filters (W41) at a flow rate of about 0.75 m-3 min‑1. In the marine boundary layer, the collection efficiency of W-41 filters is greater than 95% for dust (Kitto & Anderson, 1988; Arimoto et al., 1990). W-41 filters yield consistently low blank values for a wide range of elements and they are routinely used for aerosol trace element sampling in ocean environments (Morton et al., 2013). Exposed filters are periodically returned to Miami where all analyses are carried out.
For mineral dust analyses, a quarter section of each filter is extracted with three aliquots of water and then placed in a muffle furnace for about 14 hours (i.e., overnight) at 500°C. Earlier studies on Barbados show that the soluble components consist primarily of sea salt aerosol (Savoie et al., 2002; Savoie, Prospero, & Saltzman, 1989). The ash residue weight (less filter blank) is assumed to be mineral dust. The ash blanks yield a standard error in the mineral dust concentration that is essentially constant at ±0.1 µg m-3 for concentrations less than 1 µg m-3; at higher levels, the standard error is about ±10%.
The filter ash weight underestimates the true dust concentration because of the loss of soluble soil salts associated with the minerals during the extraction procedure (e.g., halides, CaCO3, CaSO4), and during the subsequent heating through the volatilization of bound water and various species and the breakdown of other components including organic matter. Based on an early assessment (Savoie & Prospero, 1977) we developed a canonical adjustment factor, 1.3, to compensate for those losses. In later work, the concentration of 19 elements was measured by neutron activation (Arimoto et al., 1995) in aliquots of 1349 dust-laden filter samples collected at Barbados. A scatter plot of Al vs the ash weights of aliquots of these same filters (obtained as described above) yields an average Al concentration of 10.4%. When we apply the canonical factor, 1.3, we obtain an Al concentration of 8.0%. This value is close to the average abundance of Al in the upper continental crust: 8.15%, (Rudnick & Gao, 2003). With this normalization method, a large suite of elements is found to be present in Barbados dust at ratios close to crustal abundances (e.g., Trapp et al., 2010; Bozlaker et al., 2017.
; PM10 Measurements.
Aerosol concentrations are measured with ThermoFisher Scientific TEOM 1400 series instruments. Measurements are made in two size classes: PM2.5 and PM10. The TEOM is a gravimetric instrument that continuously weighs the collection filter and measures particle mass concentrations. The TEOM is an EPA equivalent method for measuring PM2.5 and PM10 particulate matter and is used worldwide in air quality programs (Nosratabadi, Graff, Karlsson, Ljungman, & Leanderson, 2019). The instruments at Cayenne are operated and maintained by ATMO-Guyane, a non-profit organization (https://www.atmo-guyane.org/qui-sommes-nous/statuts/ ) responsible for monitoring air quality in FG following French national air quality requirements and protocols (https://www.atmo-guyane.org/). Sampling is carried out at various fixed sites in the Cayenne region and also at mobile sites as needed. Data is stored as 15-minute averages. In this work, we use the daily means of values between midnight to midnight. Data can be downloaded at https://www.atmo-guyane.org/donnees/open-data/; Column 1: Date of sample
Column 2. Mineral dust concentration as determined from filter samples. In the case of multi-day samples, the concentration is copied over the span of the days sampled. Colored blocks identify multi-day time span.
Column 3: The number of days that the filter sample was exposed for collection of aerosol.
Column 4: TEOM PM10 concentrations (ug/m3). This is the average of all stations in operation during that specific day.
Column 5: PM10 concentrations (ug/m3) where the daily sample values are averaged over the time range of the multi-day filter samples.