hide
Free keywords:
-
Abstract:
One of the main sources of solar wind disturbances are coronal holes which can be identified in extreme ultra-violet (EUV) images of the Sun. Previous research has shown the connection between coronal holes and an increase of the solar wind speed at Earth. In this study, we propose a new machine learning model predicting the solar wind speed originating from coronal holes. We detect coronal holes by applying a recently introduced segmentation algorithm to solar EUV images. Based on that, we derive time series of coronal hole characteristics, which are used as input to the model to predict the solar wind speed. Since coronal holes are structures that change over time, we also process their temporal evolution. We put a special focus on learning the geoeffective coronal hole areas, by splitting up the solar surface into multiple sectors of different latitudes and longitudes. This approach enables to predict the disturbances up to approximately 5 days in advance. We show that our model can accurately predict the solar wind speed with a temporal resolution of one hour during time periods when the solar wind is dominated by coronal hole activity. Moreover, we apply it to 10 years of data and compare our results to other state-of-the-art models.