Characterizing cloud properties and synoptic variability over the Azores using self-organizing maps
Authors
David B. Mechem — University of Kansas
Matthew Allen Miller — North Carolina State University
Sandra Yuter — North Carolina State University
Simon Paul de Szoeke — Oregon State University
Category
Cloud Properties
Description
The ARM Mobile Facility (AMF) deployment on Graciosa Island in the Azores during 2009–2010 demonstrated that the northeast Atlantic exhibits a striking amount of variability in cloud system behavior and synoptic configuration. We apply the technique of self-organizing maps (SOMs) to characterize variability in synoptic conditions and cloud properties. SOMs are an unsupervised artificial neural network learning technique, which produce from input data a finite number of characteristic patterns: for example, the preferential synoptic configurations present in the input data. We apply the SOM technique to geopotential height and vertical motion fields from ERA-Interim reanalysis, specifically focusing on the joint variability between different meteorological variables. We use the resulting SOM products to investigate the relationship between synoptic configuration and cloud properties.
