Hierarchical approach to forecasting recurrent solar wind streams
- 51 Downloads
The hierarchical approach to predicting quasi-stationary, high-speed solar wind (SW) streams is described. This approach integrates various types of data into a single forecasting system by means of an ensemble of experts. The input data included the daily values of the coronal hole areas, which were calculated from the ultraviolet images of the Sun, and the speed of the SW streams during the previous solar rotations. The coronal hole areas were calculated from the images taken by the SWAP instrument aboard the PROBA2 satellite in the spectral interval centered at a wavelength of 17.4 nm and by the AIA instrument aboard the SDO spacecraft in the interval of wavelengths centered at 19.3 and 17.1 nm. The forecast was based on the data for 2010, corresponding to the rising phase of the 24th solar cycle. On the first hierarchical level, a few simple model estimates were obtained for the speed of the SW streams from the input data of each type. On the second level of hierarchy, the final 3 day ahead forecast of the SW velocity was formulated on the basis of the obtained estimates. The proposed hierarchical approach improves the accuracy of forecasting the SW velocity. In addition, in such a method of prediction, the data gaps in the records of one instrument do not crucially affect the final result of forecasting of the system as a whole.
KeywordsSolar Wind Root Mean Square Error Solar Cycle Coronal Hole Solar System Research
Unable to display preview. Download preview PDF.
- Brown, G., Ensemble Learning, in Encyclopedia of Machine Learning, Sammut, C. and Weeb, G.I., Eds., Springer Press, 2010, pp. 312–320.Google Scholar
- Chapman, S. and Bartels, J., Geomagnetism, Oxford: Clarendon, 1940.Google Scholar
- De Toma, G., Evolution of Coronal Holes and Implications for High-Speed Solar Wind during the Minimum between Cycles 23 and 24, Sol. Phys., 2011, in press.Google Scholar
- Hurlburt, N., Cheung, M., Schrijver, C., et al., Heliophysics Event Knowledgebase for the Solar Dynamics Observatory (SDO) and beyond, Sol. Phys., 2010. doi: 10.1007/s11207-010-9624-2Google Scholar
- Shugay, J.S., Guzhva, A.G., Dolenko, S.A., and Persiantsev, I.G., An Algorithm for Construction of a Hierarchical Neural Network Complex for Time Series Analysis and its Application for Studying Sun-Earth Relations, in Proc. 8th Int. Conf. Patt. Recognit. Image Analys.: New Information Technologies (PRIA-8-2007), Yoshkar-Ola, 2007, vol. 2, pp. 335–358.Google Scholar
- Shugay, Yu.S. and Veselovsky, I.S., Studying the Structure of Coronal Holes in Different Wavelength Ranges, in Tr. Konf. “Mnogovolnovye Issledovaniya Solntsa i Sovremennye Problemy Solnechnoi Aktivnosti” (Proc. Conf. on Multiwave Studies of the Sun and Modern Problems in Solar Activity), Nizhni Arkhyz, 2007, pp. 92–100.Google Scholar
- Terekhov, S.A., Brilliant Committees of Clever Machines, in Lektsii po neiroinformatike 10-i Vserossiiskoi nauchno-tekhnicheskoi konferentsii “Neiroinformatika-2008”. Sb. nauch. tr. (Lectures on Neuroinformatics, Proc. 10th All-Rus. Sci.-Tech. Conf. “Neuroinfiormatics-2008”), Moscow, 2008, part 2, pp. 11–42.Google Scholar