Content area

Abstract

The AE index has two components: one driven by the solar wind and one related to the magnetotail unloading process. We recall some past findings on this issue and present a new ANN algorithm for the AE forecasting at 5 and 60 min time scales, built by adding to a previous algorithm a further layer with a hyperbolic tangent transfer function and two more inputs, the output at time t and the difference between the input at time t and the output at time t - 1. We show that, at the 60 min time scale and for AE > 400 nT, the new algorithm performs better than the former one, while no improvement is obtained at the 5 min time scale. This result confirms that the AE component driven by the solar wind can be forecast, at least partially, while the unloading component may not be reproduced from solar wind inputs. (PUBLICATION ABSTRACT)

Details

Title
ANN forecast of hourly averaged AE index based on L1 IMF and plasma measurements
Author
Amata, Ermanno; Consolini, Giuseppe; Pallocchia, Giuseppe; Marcucci, Maria Federica
Pages
185-196
Publication year
2009
Publication date
Mar 2009
Publisher
Springer Nature B.V.
ISSN
18956572
e-ISSN
18957455
Source type
Scholarly Journal
Language of publication
English
ProQuest document ID
197570565
Copyright
© Versita Warsaw and Springer-Verlag Berlin Heidelberg 2009