Real-time neural network predictions of magnetospheric activity

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Monday, 18 January 2010
Ramkumar Bala, Rice University, Houston, TX; and P. H. Reiff

Handout (2.1 MB)

The Boyle potential, Φ (kV) = 10-4v2 + 11.7 |B|sin3(θ/2) where v is measured in km/s and B in nT, or the Boyle Index (BI), is an empirically-derived formula that can characterize the Earth's polar cap potential. It is derivable readily in real time using the solar wind data from ACE (Advanced Composition Explorer). The logarithms of both 3-hour and 1-hour averages of the Boyle Index correlate well with the subsequent Kp: Kp = 8.93 log10(BI) - 12.55. Artificial Neural Networks (ANNs) are useful for non-linear and temporal prediction problems in that they can be trained using the time history of data. Inputs to our ANN models have greatly benefitted from the BI and its proven record as a forecasting parameter since its initiation in October, 2003. We have developed a new set of algorithms suitable for short term space weather forecasts with an enhanced lead-time and better accuracy in predicting Kp over some leading models; the algorithms omit the time history of Kp to utilize only the solar wind data.

Kp is a 3-hour averaged global geomagnetic activity index representing midlatitude regions. Our algorithms can predict the upcoming full 3-hour Kp, updated each half-hour, purely from the solar wind data and achieve a linear correlation coefficient of 0.84, which means that it predicts the upcoming Kp value on average to within 1.1 step, which is approximately the resolution of the real-time Kp estimate. Also, when predicting an equivalent 'one hour Kp', the correlation coefficient is 0.86, meaning on average a prediction within 0.99 step. Furthermore, retrospective results and temporal correlations demonstrate that our new algorithms are as efficient as those that incorporates the time history of Kp in their inputs. Live results of the BI plot and Kp prediction can be obtained from http://space.rice.edu/ISTP/wind.html while members of the Rice Space Institute's free spacalrt system will receive email alerts whenever the value of the predicted Kp reaches 6 or higher, or when the 10-minute BI exceeds 200 kV.