Forecasting Methods, Based On Integrated Information Model Of The Solar Activity

The space environment is one of the key factor that needs to think about in thespace technology and the wireless communication technology, and the driving sourcethat leads to the disturbance of the space environment is the solar activity. Studyingand mastering the transition law of the solar activity and further making accurateprediction is of great theoretical significance and wide application potential.As the moment the Human knowledge for the physical mechanism of solaractivity is still very limited. The predicting for solar activity mainly depends onmathematical and statistical methods. Available statistical method and functiondescription method were both designed to predict the overall solar activity cycleprofile. But their common shortcoming is the lack use of information, short predictinglead time and the dissatisfactory performance of predicting the peak of the cycle.Statistical precursor method can act as an effective complement to the above twomethods, which is designed to predict the cycle peak. The profile predicting methodsin conjunction with the predicting result of the cycle peak provided by the statisticalprecursor methods can not only improve the accuracy of prediction but also advancethe lead time.Statistical precursor method to predict the solar activity cycle peak is simple,with long lead time and widely used. However, it will appear some error in a certain cycleprediction by this method of the type based on independent information, which makespeople doubt the reliability of this method's predicting results.The research work of this thesis, sponsored by The National 10th Five-Year.Project Foundation, study the new solar activity cycle prediction method based onmulti-information.The major contributions of this thesis include:1. A synthesis prediction method based on multi-information is provided here. The newmethod is adding the hidden information and geometrical information of the solaractivity cycle profile to the traditional statistical precursor method based on thegeomagnetic activity indicators information, and making the cycle peak predictionusing multiple regression technology. It's shown by the experiment results that theprediction accuracy, adaptability and stability of the new method have improvedsignificantly. The different prediction models satisfying separately the medium-termand long-term prediction needs are provided here. According to the time when the information is available, we sum up the different prediction models satisfyingseparately the long-term, medium-term and short-term prediction needs. Using theprofile prediction methods in conjunction with the cycle peak prediction results fromthese above different prediction models, we can get a better prediction for thecomplete solar activity cycle profile.2. Tracking and predicting the solar activity cycle profile by using particle filtertechnique, which is the first attempt in this research field, has been done and has madesome progress in some local time range.