Interpretation of Arctic aerosol properties using cluster analysis applied to observations in the Svalbard area

ABSTRACT Atmospheric aerosols play an important role in global climate change, directly through radiative forcing and indirectly through their effect on cloud properties. Numerous measurements have been performed in the last three decades in order to characterize polar aerosols. Information about aerosol characteristics is needed to calculate induced changes in the Earth's heat balance. However, this forcing is highly variable in space and time. Accurate quantification of forcing by aerosols will require combined efforts, assimilating information from different sources such as satellite, aircraft and surface‐based observations. Adding to the complexity of the problem is that the measurements themselves are often not directly comparable as they vary in spatial/temporal resolution and in the basic properties of the aerosol that they measure. Therefore it is desirable to close the gap between the differences in temporal and spatial resolution and coverage among the observational approaches. In order to keep the entire information content and to treat aerosol variability in a consistent and manageable way an approach has to be achieved which enables one to combine these data. This study presents one possibility for linking together a complex Arctic aerosol data set in terms of parameters, timescale and place of measurement as well as meteorological parameters. A cluster analysis was applied as a pattern recognition technique. The data set is classified in clusters and expressed in terms of mean statistical values, which represent the entire database and its variation. For this study, different time‐series of microphysical, optical and chemical aerosol parameters as well as meteorological parameters were analysed. The database was obtained during an extensive aerosol measurement campaign, the ASTAR 2000 (Arctic Study of Tropospheric Aerosol and Radiation) field campaign, with coordinated simultaneous ground‐based and airborne measurements in the vicinity of Spitsbergen (Svalbard). Furthermore, long‐term measurements at two ground‐based sites situated at different altitudes were incorporated into the analysis. The approach presented in this study allows the necessary linking of routine long‐term measurements with short‐term extensive observations. It also involves integration of intermittent vertical aerosol profile measurements. This is useful for many applications, especially in climate research where the required data coverage is large.