Water soluble organic aerosols in the Colorado Rocky Mountains, USA: composition, sources and optical properties

Atmospheric aerosols have been shown to be an important input of organic carbon and nutrients to alpine watersheds and influence biogeochemical processes in these remote settings. For many remote, high elevation watersheds, direct evidence of the sources of water soluble organic aerosols and their chemical and optical characteristics is lacking. Here, we show that the concentration of water soluble organic carbon (WSOC) in the total suspended particulate (TSP) load at a high elevation site in the Colorado Rocky Mountains was strongly correlated with UV absorbance at 254 nm (Abs254, r = 0.88 p < 0.01) and organic carbon (OC, r = 0.95 p < 0.01), accounting for >90% of OC on average. According to source apportionment analysis, biomass burning had the highest contribution (50.3%) to average WSOC concentration; SOA formation and motor vehicle emissions dominated the contribution to WSOC in the summer. The source apportionment and backward trajectory analysis results supported the notion that both wildfire and Colorado Front Range pollution sources contribute to the summertime OC peaks observed in wet deposition at high elevation sites in the Colorado Rocky Mountains. These findings have important implications for water quality in remote, high-elevation, mountain catchments considered to be our pristine reference sites.


PARAFAC model application
We applied PARAFAC modeling to further identify the individual fluorescent components in our corrected EEM dataset. PARAFAC modeling of EEMs was conducted in MATLAB using the "drEEM Toolbox" (ver 0.1.0) following the recommendations and procedures of Murphy et al. 1 . To prepare the data for PARAFAC modeling, regions of the spectrum influenced by Rayleigh scatter peaks were removed. After data preparation, outlier identification was performed for 2 to 7 models run with non-negativity constraints.
Two samples emerged as outliers and were removed. The model was fit using 3-, 4-, 5-  Supplementary Table S2.

OC-EC analysis.
A 1.5 cm 2 punch taken from each filter sample was loaded on a prebaked punch (1.5 cm 2 ) of quartz fiber filter and analyzed using the NIOSH method 5040 2,3 on a Sunset Thermal Optical Transmission (TOT) Laboratory ECOC analyzer. The total OC carbon includes OC1, OC2, OC3, OC4 and PC, representing the carbon measured at four distinct temperature steps (340, 500, 615, and 900 °C) with a pyrolized carbon (PC) adjustment in the first heating cycle of the method. The EC was made up of the carbon measured during the second heating cycle with a final temperature of 910 °C. The OC and EC amounts, total deposition area of the GFF, and sample volume were used to obtain the final concentration. Field blanks were collected and the blank values were more than one order of magnitude lower than the ambient sample with the lowest OC and EC loadings.

Analysis of WS-OMMs
Aliquots of each filter were extracted by 20 mL of methanol and methylene chloride mixture (1:1, v/v) ultrasonically two times (15 minutes each). The total extracts were filtered and evaporated to a final volume of ~0.5 mL. After that, the extracts were GC oven temperature was programmed from 80 °C (hold for 5 min) to 200 °C at 3 °C min −1 , and then increased to a final temperature of 300 °C (hold for 10 min) at 15 °C min −1 . Linear calibration curves were derived from five dilutions of quantification standards. Dicarboxylic acids and saccharides were quantified by authentic standards; 2methyltetrols (2-methylthreitol and 2-methylerythritol) were quantified using mesoerythritol; other SOA tracers (e.g., hydroxyl dicarboxylic acid) were quantified using cisketopinic acid (KPA). The species not quantified using authentic standards were identified by the comparison of mass spectra to previously reported data [4][5][6] . Field blanks were collected and no contamination was observed for identified species. The WS-OMMs quantified in this work were listed in Table 1. Recoveries of those WS-OMMs were obtained by spiking standards onto prebaked filters, followed by extraction and quantification in a same manner as that for ambient samples. Except glutaric acid and adipic acid, other species had average recoveries higher than 70% (70.3 ± 3.87 -97.9 ± 3.17%). The recoveries of glutaric acid and adipic acid were low (50.1 ± 6.05% and 45.6 ± 4.32%) but stable, and their concentrations were still given. The reported concentrations of WS-OMMs were not adjusted by their recoveries.

Absorbance correction
Absorbance spectra were corrected by subtracting the mean of the absorbance from 790 -800 nm according to Mitchell et al. 7 and Mladenov et al. 8 . Other studies recommend single point absorbance correction using the absorbance at 700 nm 9,10 . We compared the single point absorbance value at 700 nm with the mean for the 790-800 nm range for five random samples in our dataset and found that the difference between the two methods was negligible (< 0.0005 a.u.; Supplementary Table S3).