***** Dataset title *****

Number concentration and fluorescent class fraction of fluorescent and hyper-fluorescent aerosol particles measured during the Antarctic Circumnavigation Expedition

***** Dataset abstract *****

This dataset consists of a 5-minute time series of number concentrations of fluorescent and total aerosol particles that were measured by wideband integrated by aerosol sensor during the Antarctic Circumnavigation Expedition in the austral summer of 2016/2017. Furthermore, the dataset includes the fraction of fluorescent classes of aerosol particles, according to the ABC classification of Perring et al. (2015). The dataset provides information on fluorescent and hyper-fluorescent aerosols which were obtained by considering low and high fluorescence thresholds. For this dataset, aerosol particles that have optical diameter greater than 1 μm are considered. Since the ship’s exhaust could considerably affect the fluorescence properties of aerosol particles, in this dataset we removed the periods where it is likely that the samples were contaminated by the ship's exhaust.

***** Original data collection *****

The 5-minute time series data was obtained by processing the single aerosol measurements conducted by the wideband integrated by aerosol sensor (WIBS-4, University of Hertfordshire, Hatfield, UK) (see 10.5281/zenodo.5109386). The raw data was initially stored in CSV format on the data-collection computer.

***** Data processing *****

The first step in processing the single particle data was to calculate the fluorescent thresholds. The fluorescent thresholds were calculated by extracting the force triggering data from the single particle measurements, and resampling them into two-hour time intervals. In WIBS data analysis fluorescent thresholds have to be calculated for each fluorescent channel separately. In our analysis, the mean and standard deviation of the aerosol fluorescent signals in the three measured channels were calculated for the two-hour time intervals. The low fluorescent threshold was calculated by adding the mean value to three times the standard deviation. The high fluorescent threshold was calculated by adding the mean value to nine times the standard deviation, for each fluorescent channel.
Subsequently, the two-hour thresholds were merged with the time series of single particle data and the threshold at any given time was calculated by interpolating the two-hour threshold values. Aerosol particles whose fluorescent signals (in any given channel) exceeded the low thresholds were considered as fluorescent aerosols while if their fluorescent signal was higher than the high fluorescent threshold they were considered as hyper-fluorescent. For a given (hyper-)fluorescent aerosol, the ABC classification was obtained based on the fluorescent channels which the fluorescent signal exceeds the thresholds according to the scheme explained in Perring et al. (2015). A Python program was written to calculate the thresholds, while another Python program was used to merge the thresholds with single particle data and label them based on their ABC classes and whether the aerosol are fluorescent or hyper-fluorescent. The labelled data was then exported as .feather files.
 
After labelling single particle data, a separate program was written in Python to group the single particle data by 5-minute time intervals and conduct statistical analysis in the 5-minute groups. To obtain the number concentration of aerosol particles the number of particles were calculated for each 5-minute time interval and was divided by the 5-minute sample volume, which was equal to 1.15 l (i.e. instrument sample flow rate (0.23 lpm) multiplied by 5 minutes). Note that for (hyper-)fluorescent aerosol number concentration we only counted the (hyper-)fluorescent single particle data, while for total aerosol number concentration we counted all the single particles within the 5-minute time interval. The fraction of ABC classes for (hyper-)fluorescent aerosols were obtained by dividing the number concentration of each individual class by the total (hyper-)fluorescent number concentration. In our number concentration analysis, we only considered samples with a diameter larger than 1 μm. During the campaign, the atmospheric sample extracted by the instrument was suspected to be contaminated by the ship's exhaust. In this study, a data removal procedure was applied to remove periods that are thought to be contaminated by the the ship's exhaust. The data removal procedure is explained in more detail in section S1.3 of the supporting information in Moallemi et al. (2021). 

It should be noted that although the sampling frequency of WIBS is very high (~ 125 Hz), we expect that occurrence of bioaerosols and fluorescent aerosols have time scales in the range of hours. Based on this assumption we conducted most of our analysis on either 5-minute or 1-hour average time series. It should be noted that the dataset represents samples that were collected after 2017-01-06. We exclude older samples as we suspected that the instrument was not functioning properly prior to 2017-01-06.

***** Dataset contents *****

- N_fluorescent.csv, data file, comma-separated values
- N_hyper_fluorescent.csv, data file, comma-separated values
- data_file_header.txt, metadata, text
- README.txt, metadata, text

***** Dataset contact *****

Alireza Moallemi, Paul Scherrer Institute. ORCID: 0000-0002-9491-1993. Email: alireza.moallemi@psi.ch

***** Dataset license*****

This number concentration and fluorescent class fraction of fluorescent and hyper-fluorescent aerosol particles dataset from ACE is made available under the Creative Commons Attribution 4.0 International License (CC BY 4.0) whose full text can be found at https://creativecommons.org/licenses/by/4.0/

***** Dataset citation *****

Please cite this dataset as: 
Moallemi, A., J. Schmale, M. Schnaiter, M. Gysel-Beer, R. Modini and S. Landwehr. (2021). Number concentration and fluorescent class fraction of fluorescent and hyper-fluorescent aerosol particles measured during the Antarctic Circumnavigation Expedition. (1.0) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.5109382