A Community Dataset for Comparing Automated Coronal Hole Detection Schemes and its Imprint on Magnetic Models

It is now known that the fast solar wind streams originating from coronal holes can have a significant contribution to geomagnetic activity, particularly during periods of low solar activity. Moreover, coronal holes, many of which are long lived and overall time steady structures, have proven to be ideal hunting grounds for understanding the fast solar wind. In this regard, automated detection schemes are nowadays a standard approach for locating coronal holes in EUV images from the Solar Dynamics Observatory (SDO). However, several inevitable factors make this automatic identification challenging. While discrepancies between detection schemes have been noted in the literature, a comprehensive assessment of these discrepancies, which is still lacking, is of equal importance. Here we present the first community dataset for comparing automated coronal hole detection schemes. This dataset consists of 29 SDO images, all of which were selected by experienced observers to challenge automated schemes. We then use this dataset as input to 14 widely-applied automated schemes to study coronal holes and collect their detection results. From this, we select and study three SDO images that exemplify the most important lessons learned from this effort. We find that different detection schemes highlight significantly different physical properties of coronal holes. Motivated by these outcomes, we look into the effect of these results on the inferred magnetic connectivity in the corona by comparing the detected coronal hole boundaries to magnetic model solutions. These results, along with the database, will provide rich inputs to our understanding of coronal holes and their connection to the solar wind.