Solar flare prediction

Prediction is difficult, especially the future
---variously attributed

March 2006: The daily predictions on these pages are no longer being updated.

Recently I have been working on a Bayesian method of solar flare prediction. The idea is to use the information in the numbers and sizes of solar flares already observed from the Sun together with the known phenomenological rules of flare statistics to make a prediction for the occurrence of a flare within a given period, or to improve an existing prediction. A paper on the theory of the method was published in the Astrophysical Journal (Wheatland 2004, ApJ 609, 1134-1139). A version of the paper is also available online from astro-ph.

The method has been applied to X-ray flare observations from the Geostationary Observational Environmental Satellites (GOES). GOES events (flares) are characterised by their peak soft X-ray flux in the 1-8 Angstrom band. Moderate-sized events are M class flares, which have a peak flux greater than 0.00001 Watts per square metre at the satellite. Large events are X class flares, which have a peak flux greater than 0.0001 Watts per square metre. There is interest in predicting M and X class events, in particular because of the effect of the enhanced flux of soft X-rays on the upper atmosphere. The GOES implementation is described in detail in a second paper on the method (Wheatland 2005, Space Weather Vol. 3, No. 7, S07003, 10.1029/2004SW000131). A version of this paper is also available from astro-ph.

An experimental automated prediction system for GOES flares is also being developed. Below are the current predictions for the occurrence of M and X class flares within 24 hours (the times are in Universal Time), based on a year's data prior to the stated time. This page should have new predictions each day. (For more information on how these predictions were made, see this page. For more information on the reliability of these predictions, see this page and this page.)

Current prediction:

Prediction for the probability of occurrence of a large flare on the Sun
within 24 hours, based on the previous year of GOES events.
 
Prob. of at least one flare of size M to X = 0.050+/-0.011
Prob. of at least one flare of X size or greater = 0.008+/-0.002
 
Prediction made at 2006/03/16 00:02:00 (UT)


Previous prediction:

For reference, here is the previous prediction. You can also examine what the Sun has actually been doing in the last few days in the GOES plot at the bottom. When activity increases, the method should give higher probabilities for flaring.

Prediction for the probability of occurrence of a large flare on the Sun
within 24 hours, based on the previous year of GOES events.
 
Prob. of at least one flare of size M to X = 0.050+/-0.011
Prob. of at least one flare of X size or greater = 0.008+/-0.002
 
Prediction made at 2006/03/15 00:02:00 (UT)


Additional information:

More detail on how today's predictions were made
Summary of all predictions made to date
Automatic check of the reliability of the predictions
The prediction for 4 November 2003
A test of the method on historical data
Links to other pages related to flare prediction
Mike Wheatland's home page


Acknowledgement: The predictions given here are based on information from the Space Environment Center, Boulder, CO, National Oceanic and Atmospheric Administration (NOAA), US Dept. of Commerce. The GOES plot is obtained from the same source.


Page maintained by m.wheatland@physics.usyd.edu.au Page last updated Tuesday, 25-Sep-2007 08:33:23 AEST