• Research dataclose
• Other research productsclose
• Open Access close
• Swiss National Science Foundation close
• FI close
• RU close

This dataset contains CCN concentrations at five supersaturation levels, averaged to 1 min time resolution, measured during the year-long Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC) expedition from October 2019 to September 2020. The measurements were performed in the Swiss container on the D-deck of Research Vessel Polarstern, using the model CCN-100 from Droplet Measurement Technologies (DMT, Boulder, USA). Detailed description of the measurement principle can be found in e.g. Roberts & Nenes (2005). The instrument was located behind an automated valve, which switched hourly between a total and an interstitial air inlet, with upper cutoff sizes of 40 and 1 µm respectively (Heutte et al., Submitted; Beck et al., 2022; Dada et al., 2022). The measurements were performed in 1-h cycles, with a 0.5 L/min sample flow and a 2 L/min make up flow, where the supersaturations 0.15, 0.2, 0.3, 0.5 and 1.0 % were measured. The supersaturation of 0.15 % is measured for 20 min, as it takes longer to equilibrate, and the remaining supersaturations were measured for 10 min each. The instrument was calibrated in July 2019 before the campaign, and in March and April 2020 during the campaign. Based on the inter-variability of the calculated supersaturation levels during these calibrations, we can expect values ranging from 0.15-0.20, 0.20-0.25, 0.29-0.33, 0.43-0.5, 0.78-1.0 % for the nominal supersaturations of 0.15, 0.2, 0.3, 0.5 and 1.0 %, respectively. The counting error for the CCNC is associated with the error in the optical counting of particles and is about 10 %. Data were removed during the cooling cycle (i.e., the time when the measurement cycle starts again and the temperature is cooled to set the lowest supersaturation), which corresponds roughly to the first 10 min of each hour (so 50 % of the 0.15 % supersaturation period). Additionally, the first minute of the transition between supersaturations was removed before averaging the data to 1 min time resolution. During some time periods, a difference pattern of mean and standard deviation of the measurements between even and odd hours was observed, most probably caused by a persistent pressure drop in the inlet lines, resulting in a proportional reduction of the concentration measurements. For correction, the 1-h arithmetic mean of interstitial inlet measurements and the mean of the two adjacent hours of total inlet measurements were subtracted, and the resulting difference was added as a constant to the data points of the interstitial inlet measurements. The dataset contains a pollution mask for local pollution (predominantly exhaust from the Research Vessel Polarstern) with 0 indicating clean, and 1 indicating polluted periods (Beck et al., 2022; Beck et al., 2022).

These datasets contain the total particle number concentrations and normalized size distributions (dN/dlogDp) of excited, fluorescent, and hyper-fluorescent particles of sizes 0.5 to 20 μm (optical diameter). The normalized size distribution datasets are split into 20 size bins: 0.5 - 0.6 μm, 0.6 - 0.72 μm, 0.72 - 0.87 μm, 0.87 - 1.05 μm, 1.05 - 1.26 μm, 1.26 - 1.51 μm, 1.51 - 1.82 μm, 1.82 - 2.19 μm, 2.19 - 2.63 μm, 2.63 - 3.16 μm, 3.16 - 3.8 μm, 3.8 - 4.57 μm, 4.57 - 5.5 μm, 5.5 - 6.61 μm, 6.61 - 7.95 μm, 7.95 - 9.56 μm, 9.56 - 11.50 μm, 11.5 - 13.83 μm, 13.83- 16.63 μm and 16.63 - 20 μm. The data were measured by a WIBS-NEO (Wideband Integrated Bioaerosol Sensor, model New Electronics option) by droplet measurement techniques ltd. The data were processed using the IGOR WIBS toolkit V1.36 (DMT) and python version 3.9.7. These datasets have been averaged to 1 hour time resolution. The datasets were cleaned from local pollution sources by applying a pollution flag developed by Beck et al. (2022a,b), which is based on the rate of change in particle number concentration with 1 min time resolution. Data points with more than 10 polluted minutes within an hour were removed from the WIBS datasets. Time periods with zero filter measurements and time periods with unstable flow that affected number concentrations have been removed from the dataset. The WIBS measures the size, asymmetry and fluorescence of particles with an optical diameter of 0.5 – 20 µm. Detected particles are excited by two UV flashlamps at wavelengths of 280 and 370 nm and their emitted fluorescence is measured by two photomultipliers with bandwidths of 310 - 400 nm, and 420 - 650 nm. The WIBS counts excited particles at a maximum frequency of 125 Hz, which corresponds to a maximum concentration of 2.5*104 particles/L with a sample flow of 0.3 L/min. Excited particles were classified as fluorescent if their fluorescent intensity exceeded the background intensity by three standard deviations (3σ) and as hyper-fluorescent if the fluorescent intensity exceeded the background intensity by 9σ. Excited particles with a lower fluorescent intensity were considered to be non-fluorescent. The background fluorescence was determined by measuring the fluorescent signal in the measurement chamber in absence of particles. Background measurements were performed every 26 h. The combination of two excitation wavelengths and two detector wavebands allows the classification of fluorescent particles into seven types: A, B, C, AB, AC, BC, and ABC (Perring et al. (2015); Savage et al. (2017)). For further information about the instrumental setup, refer to Heutte et al. (Submitted).

These datasets contain the total particle number concentrations of excited particles of sizes 0.5 to 20 μm (optical diameter). The WIBS measures the size, asymmetry and fluorescence of particles with an optical diameter of 0.5 – 20 µm. Detected particles are excited by two UV flashlamps at wavelengths of 280 and 370 nm and their emitted fluorescence is measured by two photomultipliers with bandwidths of 310 - 400 nm, and 420 - 650 nm. The WIBS counts excited particles at a maximum frequency of 125 Hz, which corresponds to a maximum concentration of 2.5*104 particles/L with a sample flow of 0.3 L/min. Excited particles were classified as fluorescent if their fluorescent intensity exceeded the background intensity by three standard deviations (3σ) and as hyper-fluorescent if the fluorescent intensity exceeded the background intensity by 9σ. Excited particles with a lower fluorescent intensity were considered to be non-fluorescent. The background fluorescence was determined by measuring the fluorescent signal in the measurement chamber in absence of particles. Background measurements were performed every 26 h. The combination of two excitation wavelengths and two detector wavebands allows the classification of fluorescent particles into seven types: A, B, C, AB, AC, BC, and ABC (Perring et al. (2015); Savage et al. (2017)). For further information about the instrumental setup, refer to Heutte et al. (Submitted).

These datasets contain the total particle number concentrations of fluorescent particles of sizes 0.5 to 20 μm (optical diameter). The WIBS measures the size, asymmetry and fluorescence of particles with an optical diameter of 0.5 – 20 µm. Detected particles are excited by two UV flashlamps at wavelengths of 280 and 370 nm and their emitted fluorescence is measured by two photomultipliers with bandwidths of 310 - 400 nm, and 420 - 650 nm. The WIBS counts excited particles at a maximum frequency of 125 Hz, which corresponds to a maximum concentration of 2.5*104 particles/L with a sample flow of 0.3 L/min. Excited particles were classified as fluorescent if their fluorescent intensity exceeded the background intensity by three standard deviations (3σ) and as hyper-fluorescent if the fluorescent intensity exceeded the background intensity by 9σ. Excited particles with a lower fluorescent intensity were considered to be non-fluorescent. The background fluorescence was determined by measuring the fluorescent signal in the measurement chamber in absence of particles. Background measurements were performed every 26 h. The combination of two excitation wavelengths and two detector wavebands allows the classification of fluorescent particles into seven types: A, B, C, AB, AC, BC, and ABC (Perring et al. (2015); Savage et al. (2017)). For further information about the instrumental setup, refer to Heutte et al. (Submitted).

These datasets contain normalized size distributions (dN/dlogDp) of hyper-fluorescent particles of sizes 0.5 to 20 μm (optical diameter). The normalized size distribution datasets are split into 20 size bins: 0.5 - 0.6 μm, 0.6 - 0.72 μm, 0.72 - 0.87 μm, 0.87 - 1.05 μm, 1.05 - 1.26 μm, 1.26 - 1.51 μm, 1.51 - 1.82 μm, 1.82 - 2.19 μm, 2.19 - 2.63 μm, 2.63 - 3.16 μm, 3.16 - 3.8 μm, 3.8 - 4.57 μm, 4.57 - 5.5 μm, 5.5 - 6.61 μm, 6.61 - 7.95 μm, 7.95 - 9.56 μm, 9.56 - 11.50 μm, 11.5 - 13.83 μm, 13.83- 16.63 μm and 16.63 - 20 μm. The data were measured by a WIBS-NEO (Wideband Integrated Bioaerosol Sensor, model New Electronics option) by droplet measurement techniques ltd. The data were processed using the IGOR WIBS toolkit V1.36 (DMT) and python version 3.9.7. These datasets have been averaged to 1 hour time resolution. The datasets were cleaned from local pollution sources by applying a pollution flag developed by Beck et al. (2022a,b), which is based on the rate of change in particle number concentration with 1 min time resolution. Data points with more than 10 polluted minutes within an hour were removed from the WIBS datasets. Time periods with zero filter measurements and time periods with unstable flow that affected number concentrations have been removed from the dataset.

This dataset contains CCN concentrations at five supersaturation levels, averaged to 1 min time resolution, measured during the year-long Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC) expedition from October 2019 to September 2020. The measurements were performed in the Swiss container on the D-deck of Research Vessel Polarstern, using the model CCN-100 from Droplet Measurement Technologies (DMT, Boulder, USA). Detailed description of the measurement principle can be found in e.g. Roberts & Nenes (2005). The instrument was located behind an automated valve, which switched hourly between a total and an interstitial air inlet, with upper cutoff sizes of 40 and 1 µm respectively (Heutte et al., Submitted; Beck et al., 2022; Dada et al., 2022). The measurements were performed in 1-h cycles, with a 0.5 L/min sample flow and a 2 L/min make up flow, where the supersaturations 0.15, 0.2, 0.3, 0.5 and 1.0 % were measured. The supersaturation of 0.15 % is measured for 20 min, as it takes longer to equilibrate, and the remaining supersaturations were measured for 10 min each. The instrument was calibrated in July 2019 before the campaign, and in March and April 2020 during the campaign. Based on the inter-variability of the calculated supersaturation levels during these calibrations, we can expect values ranging from 0.15-0.20, 0.20-0.25, 0.29-0.33, 0.43-0.5, 0.78-1.0 % for the nominal supersaturations of 0.15, 0.2, 0.3, 0.5 and 1.0 %, respectively. The counting error for the CCNC is associated with the error in the optical counting of particles and is about 10 %. Data were removed during the cooling cycle (i.e., the time when the measurement cycle starts again and the temperature is cooled to set the lowest supersaturation), which corresponds roughly to the first 10 min of each hour (so 50 % of the 0.15 % supersaturation period). Additionally, the first minute of the transition between supersaturations was removed before averaging the data to 1 min time resolution. During some time periods, a difference pattern of mean and standard deviation of the measurements between even and odd hours was observed, most probably caused by a persistent pressure drop in the inlet lines, resulting in a proportional reduction of the concentration measurements. For correction, the 1-h arithmetic mean of interstitial inlet measurements and the mean of the two adjacent hours of total inlet measurements were subtracted, and the resulting difference was added as a constant to the data points of the interstitial inlet measurements. The dataset contains a pollution mask for local pollution (predominantly exhaust from the Research Vessel Polarstern) with 0 indicating clean, and 1 indicating polluted periods (Beck et al., 2022; Beck et al., 2022).

This dataset contains a pollution flag in 1 min time resolution and the corresponding particle number concentration data. The data columns include Event, Time, Latitude, Longitude, Particle number concentration and a pollution flag to indicate polluted periods (0=not polluted, 1=polluted). The pollution flag is derived from the Pollution Detection Algorithm (PDA), a python-based open access script to automatically detect contamination in remote atmospheric time series Beck et al. (2022). The following parameters were used in the PDA script to derive this pollution flag:• a= 0.35 cm-3s-1• m = 0.58 s-1• avg_time = 60 s• upper_threshold: 104 cm-3• lower_threshold: 60 cm-3• neighboring points filter: Yes/on• median deviation factor: 1.4• sparse window: 30• sparse threshold: 6Remark_1: The corrected particle number concentration may still contain some minor artefacts and a critical review of the data by an expert is required. The pollution flag is based on the aforementioned parameters. If needed, the PDA can be tuned to be stricter. The decision whether a single data point is affected by pollution is up to the user and requires an expert review.Remark_2: This pollution mask can be applied to other particle and trace measurements obtained during MOSAiC. Please see Beck et al. (2022), for a detailed discussion. This dataset contains particle number concentrations and a pollution flag in 1 min time resolution. It is derived by the pollution detection algorithm (PDA, doi:10.5281/zenodo.5761101) based on the corrected particle number concentration data of the CPC3776 measured during the year-long MOSAiC expedition from October 2019 to September 2020. With pollution, we refer to emission from the exhaust of the ship stack, snow groomers, diesel generators, ship vents, helicopters and other (primary pollution, not circulated, nor transported). Pollution hence reflects locally emitted particles and trace gases, which are not representative of the central Arctic ambient concentrations. The PDA identifies and flags periods of polluted data in the particle number concentration dataset five steps. The first and most important step identifies polluted periods based on the gradient (time-derivative) of a concentration over time. If this gradient exceeds a given threshold, data are flagged as polluted. Further pollution identification steps are a simple concentration threshold filter, a neighboring points filter (optional), a median and a sparse data filter (optional). The detailed methodology of the derivation of the pollution flag is described in Beck et al. (2022).

This dataset contains the bulk size-resolved chemical composition and mass concentration of non-refractory submicron aerosols (NR-PM1) measured during the MOSAiC expedition from October 2019 to July 2020. These include the mass concentrations of sulfate (SO42-), nitrate (NO3-), ammonium (NH4+), chloride (Chl), and organics (Org). The measurements were performed in the Swiss container on the D-deck of Research Vessel Polarstern, using a commercial High-Resolution Time-of-Flight Aerosol Mass Spectrometer (HR-ToF-AMS, Aerodyne Research, Inc.). One can refer to existing literature (e.g., DeCarlo et al. (2006) and Canagaratna et al. (2007)) for detailed description, functioning principles and field deployment procedures of the AMS. The instrument was located behind an automated valve, which switched hourly between a total and an interstitial air inlet, with upper cutoff sizes of 40 and 1 µm respectively (Heutte et al. (Submitted), Beck et al. (2022), and Dada et al. (2022)). Ambient air was hence sampled alternately every hour from the total and interstitial inlets into an aerodynamic lens with a 1 µm critical orifice and a flow of 0.07 L/min.All data were processed using SQUIRREL v1.65B and PIKA v1.25B within the IGOR Pro v9.00 software. This was done separately for the three distinct periods of available measurements, October to December 2019, March to May, and June to July 2020, as the instrument was each time in a different state (after long down times related to turbo pump failures). Regular on-site calibrations using monodisperse, number concentration-defined, ammonium nitrate (NH4NO3) and ammonium sulfate ((NH4)2SO4) particles were performed to determine the ionization efficiency of NO3- and relative ionization efficiencies of NH4+ and SO42- (Jimenez et al. (2003) and Allan et al. (2003)). An airbeam correction factor was applied to the dataset, along with a time and composition-dependent collection efficiency (CDCE, Middlebrook et al. (2012)). Several times per month, zero measurements were performed using High-Efficiency Particulate Absorbing (HEPA) filters. These were used to (1) adapt the AMS fragmentation table for air fragmentation patterns (in post processing), and (2) compute the detection limits (DL) for the main aerosol chemical species (SO42-, NO3-, NH4+, Chl, and Org) as 3 times the standard deviation from the mean species concentration during filter measurements. At the instrument's native time resolution of 90 s, the DL are equal to (for Oct-Dec, Mar-May, and Jun-Jul, respectively) 0.017, 0.102, and 0.084 µg/m3 for SO42-, 0.011, 0.069 and 0.152 µg/m3 for NO3-, 0.001, 0.027 and 0.261 µg/m3 for NH4+, 0.055, 0.055 and 0.071 µg/m3 for Chl, and 0.284, 0.718 and 1.029 µg/m3 for Org. A mass closure analysis was performed between the total NR-PM1 calculated from the AMS and aethalometer and the one approximated from the mobility diameter (dm) measured by the Scanning Mobility Particle Sizer (SMPS) located in a neighboring container. The mass closure analysis was performed independently for the three periods (Oct-Dec, Mar-May and Jun-Jul) and yielded the following scaling factors: 3.77, 0.65 and 0.35 for the three respective periods. These scaling factors are not applied by default on the AMS timeseries, and the user is left with the decision to apply them or compute new ones. The following periods were removed from the dataset: when the airbeam correction factor was larger than 2 or smaller than 0, outliers (defined as more than 3 times the standard deviation of half an hour moving average, that constitute < 1 % of the whole dataset), all calibrations, periods when a HEPA filter was placed in front of the instrument, and data non-representative of ambient conditions (e.g. container air).We applied two corrections. First, the switching valve caused data distortion, observed at every full hour (i.e., when the valve turns and the ambient sampling changes from one inlet to another, there is a brief moment with under-pressure in the inlet lines). Consequently, all data points within ± 2 min of the full hours were removed. Second, during some periods when the inlet switching valve was activated, we observed a difference pattern of mean and standard deviation of the measurements between even and odd hours, most probably caused by a persistent pressure drop in the inlet lines, resulting in a proportional reduction of the concentration measurements. The 1-h arithmetic mean of interstitial inlet measurements and the mean of the two adjacent hours of total inlet measurements were subtracted, and the resulting difference was added as a constant to the data points of the interstitial inlet measurements.This dataset contains a pollution flag ("Flag_pollution") to flag datapoints that were identified as directly influenced by fresh local pollution (e.g., Polarstern exhaust, on-ice diesel generators, skidoos), where a flag equal to 0 indicates clean data and 1 indicates polluted data. The identification method, based on the cosine similarity of the measured mass spectra with a known reference polluted spectrum, is described in Dada et al. (2022). Additionally, a sparse filter with a moving window spanning 60 datapoints (approx. 1h30) was applied to define as entirely polluted periods where more than 60% of the points were already classified as polluted by the cosine similarity method.Finally, the dataset also contains a quality flag for the ammonium timeseries ("Flag_NH4"), as turbo pump failures rendered ammonium measurements very noisy. A value of 1 indicates a quality assured measurement, while a value of 0 indicates a "bad" measure of NH4+ (basically all data after May 24th 2020). We encourage data users to refer to Heutte et al. (Submitted) for a more detailed description of the data acquisition, data processing and corrections applied.We thank the Laboratory for Atmospheric Chemistry at the Paul Scherrer Institute for providing the instrument and expertise.

This dataset contains the temperature and relative humidity in a bypass to the total inlet, measured during the MOSAiC expedition from October 2019 to September 2020. The measurements were performed in the bypass interstitial inlet of the Swiss container on the D-deck of Research Vessel Polarstern, using a commercial hygrometer model HC2 (Rotronic AG, Bassersdorf, Switzerland). The interstitial inlet had a flow of > 17 L/min and was equipped with a 1 µm cyclone, for sampling interstitial particles only, as opposed to the total inlet (flow of > 15 L/min) which sampled all particles and hydrometeors up to 40 µm in diameter. The total and interstitial inlets were located 3 meters apart and connected to a valve that switched hourly between the two inlets for a targeted set of instruments. The complete instrumental setup inside the Swiss container is presented in Heutte et al. (Submitted), Dada et al. (2022), and Beck et al. (2022). Inside both inlets, the temperature was kept constant around 20 °C and the relative humidity was maintained below 40 %, with a heating system following Global Atmosphere Watch (GAW) standards for aerosol sampling (WMO, 2016). This dataset contains measurements of temperature and relative humidity inside the bypass to the total inlet, averaged to 10 min time resolution (native resolution was 5 min), along with the position of the switching valve (1 = total inlet, 0 = interstitial inlet).