Peter Tunved, Roxana Cremer, Johan Ström
The dataset is a composition of longterm aerosol observations from Zeppelin Observatory, Svalbard, available on the database EBAS and a cluster analysis of those data. It was created to apply new analysis methods to existing longterm timeseries to provide additional information on atmospheric processes that control black carbon (BC) in the Arctic.
The cluster analysis is a tool to classify the avialable data under a different viewpoint by grouping them based on their microphysical properties. This classification can help to identify the atmospheric processes at play in the air mass which is measured at the station. The four major groups in this dataset are: washout, nucelation, intermediate, polluted. The washout group is likely to have experienced a high precipitation amount. The nucelation group shows signs of recent new particle formation or growth. Intermediate shows further growth and processing and the polluted group contains the most aged particles, which is related to long range transport.
The dataset contains hourly data of the number particle size distribution, the absorption coefficient and the categorisation of the cluster analysis from 2002-03-08 until 2010-11-12.
Download data
Citation
Peter Tunved, Roxana Cremer, Johan Ström (2022) Clustered aerosol particle size distributions at Zeppelin Observatory, Svalbard, 2002 – 2010. Dataset version 1. Bolin Centre Database. https://doi.org/10.17043/zeppelin-aerosol-cluster-1
References
Tunved P, Cremer RS, Zieger P, Ström J (2021) Using correlations between observed equivalent black carbon and aerosol size distribution to derive size resolved BC mass concentration: a method applied on long-term observations performed at Zeppelin station, Ny-Ålesund, Svalbard. Tellus B: Chemical and Physical Meteorology 73:1, 1 – 17. https://doi.org/10.1080/16000889.2021.1933775
Cremer RS, Tunved P, Ström J (2022) Airmass analysis of size-resolved black carbon particles observed in the Arctic based on cluster analysis. Atmosphere 13(5):648. https://doi.org/10.3390/atmos13050648
Data description
There is one datafile, in semicolon-separated (.csv) format, with data columns structured in the following way:
- Date and time in the format YYYY-MM-DD hh
- Number of particles per cubic centimeter for each size bins according to the DMPS system
- absorption coefficient measured by PSAP in m⁻¹
- Main clusters derived by kmeans cluster in numbers 1 – 12
- 4 cluster groups in numbers 1 – 4
- 2 subgroups for cluster group 4
The four major groups in this dataset are Washout (cluster 1, cluster group 1), Nucelation (clusters 2-7, cluster group 2), Intermediate (clusters 8-9, cluster group 3) and Polluted (clusters 10-12, cluster group 4). The Washout group is likely to have experienced a high precipitation amount, Nucelation group shows signs of recent new particle formation or growth, Intermediate shows further growth and processing and the Polluted group contains the most aged particles, which is related to long range transport.
The dataset contains hourly data of the number particle size distribution, the absorption coefficient and the categorisation of the cluster analysis from 2002-03-08 14:00 until 2010-11-12 16:00, collected at Zeppelin Observatory, Svalbard, Norway (78.9° N, 11.9° E, 474 m above ground level).
Comments
Please reference the according publication (Tunved et al. 2021) when using the data set.
The provided data are based on the aerosol particle size distribution and absorption coefficient data collected at Zeppelin Observatory published on the EBAS database.
The station ID in EBAS is NO0042G, Zeppelin Mountain, Svalbard. Datasets from EBAS used here include:
- RH: NO01L_NO42_aws,relative_humidity - met (i.e relative humidity)
- DMPS: NO01L_DMPS, DMPS particle_number_size_distribution NO01L_DMPS (particle number size distribution)
- PSAP: SE03L_ZEP-PSAP-1,filter_absorption_photometer-aerosol_absorption_coefficient (aerosol absorption coefficient)
The particle size distributions are measured by a differential mobility particle sizer (DMPS), the absorption coefficient by a particle soot/absorption photometer (PSAP), both instrument data sets were harmonised on the time scale and then screened for high relative humidity (RH) values. The RH treshold of 95% was chosen to exclude cloud influence from the data set. Next the dataset was clustered on base of the particle size distribution. The number of clusters (12) was chosen subjectively. After the 12 clusters, the most similar clusters were merged to 4 cluster groups based on their size distribution shape and magnitude. These 4 groups got the names 'Washout', 'Nucelation', 'Intermediate' and 'Polluted'. For the related study the cluster group 'Polluted' was further investigated and split into 2 subgroups.