Multispectral Classification of Solar Imagery

NOAA's Space Weather Prediction Center (SWPC) is the Nation's official source of space weather alerts, watches and warnings. In that role, the Center will be ingesting GOES-R Solar Ultraviolet Imager (SUVI) data beginning in the 2015-16 timeframe. Along with other NOAA and non-NOAA sources of solar imagery these observations are used by SWPC forecasters to inform their analysis (nowcast) and forecasts and also as sources of data to run empirical and numerical models. A supervised, multispectral, Bayesian pixel classifier has been developed and produces what are referred to as thematic maps to assist forecasters in their analysis. These maps represent classes of pixels including: space, coronal holes, quiet corona, filaments, active regions, and flares. The thematic maps product is currently undergoing operational test and evaluation at SWPC. The test uses synoptic data from the Atmospheric Imaging Array (AIA) on NASA's SDO mission in near real time as a proxy for SUVI data. This work presents the methodology and initial results of that operational test and evaluation.