On The Prediction Of Hazardous Space Environment Events

Three kinds of disastrous space weather events were investigated in this work which consisted of solar proton event,severe geomagnetic storm and relativistic energetic electron storm.Several algorithms were employed in data analyzing and forecasting model construction.With the end of 23^rdsolar cycle,it was found the there were some problem in the early widely used long term solar proton event(SPE)forecast model on both the events distribution and the flux estimation.Latest proton data was include in our work,and contrasting to the early woke by Feymann(1990)and Nymmik(2002),the occurrence of the SPE in a whole cycle was divided into three stages.The first stage starts in the minimum of solar cycle and lasts 4 years called ascending phase.And second stage,peak phase,4 years.Third stage follows,declining phase,3 years.The frequency of SPE was significantly different in them.A Statistical model was built based on the assumption mentioned above.The test run showed that the flux prediction was well approved.In the short-term SPE forecast,our investigation focused on the characteristics and the variation trend of the sunspots.The area of the sunspot indicates the mean and maximum strength of local magnetic field.The larger the sunspot area,the higher the magnetic strength.Statistic showed that large sunspot was the favorite source of solar flares.The Mcintosh classification of the sunspot indicates the structure of the root area of the magnetic rope from the sunspot regions.The complexity of the structure means the mixture of the different polarities paten.The complexity of the magnetic field distribution decides the energy stored in the sunspot. The magnetic type is a more direct and further indicator of the magnetic field distribution than the McIntosh.The joint applying of the Mcintosh classification and the magnetic type enable us to describe the structure of the sunspot much clearly.We also chose two popular bands,10cm radio and 10-80nm soft X-ray.The sudden rise of the emission flux in the two bands indicates the coming or emerging of a strong AR.The artificial neural network(ANN)is employed.An important work was to do is to digitize the characteristic based on the analyzing and sorting of the data.The digitized morphology data is then arranged and"fed"the neural net.Through training the model based on ANN algorithm with the event data,a forecast model was constructed.With it,we can get the forecast of SPE from some referred area l-3days ahead.This model worked well and we were satisfied with it.Statistic showed that the accuracy of our forecast was about 80%.All severe geomagnetic storm were caused by CME.Solar flares are well known events on the solar disk while few of them being geo-effective.The key factors that make them geo-effective are weather they have CME accompanied and the features of CME as well.But among the hundreds of CME,only few of them could cause significant geomagnetic disturbances,which mainly depends on whether they move towards to the earth.In this work,the relationship between the geomagnetic disturbances and the energetic proton flux(ACE—EPAM)data,together with the parameter of the solar flares that related to the CME was carefully investigated.The preliminary result i8 that,more than 90%of the enhancement of the particle flux followed by shock could be measured by ACE.But the correlation between flux of the particles and magnitudes of the geomagnetic disturbances was not much clear.Other factors that related to the characteristics of the CME have also to be taken into consideration.The position of the flare,which may affect the direction of the CME, the flare scale,which may decide the velocity,and the duration,which could relate to the magnetic field strength.But through statistical work,it was found that the relationship between the magnitudes of the geomagnetic disturbance index and all those parameters mentioned above were non—liner,so neural network method was introduced to calculate the relation automatically.After the neural work being trained with the historical data range from 1986 to 2002,a model to predict the geomagnetic storms after the solar eruptive events was constructed.This result showed that the error of the model comparing with the measurement was less than 20%.Through studying the mechanism that the geomagnetosphere was affected by the condition of interplanetary media,the geomagnetic disturbance index Dst was found to have close,and complex relationship with both the solar wind parameters and IMF features.By employing the measured parameters from ACE spacecraft,these parameters were the solar wind Velocity,the density of solar wind plasma and the southward component of IMF.The most recent measured Dst was also figured to correlate to the Dst several hoursahead.We construct the relationship between interplanetary measured parameters and Dst index,fully connected neural network was introduced.This neural network could demonstrate the complex relationship through building up the internal connection between separate neurons in hidden layer.After a training process with historical data,the forecast model was built during which the neural network which arrange the internal connection between units automatically according to the input parameters.The storm time data of 1998 and 1999 was selected in the training process of model construction.The data set during the geomagnetic storm in July 24-29 was used to test the model and the error of the test data was 14.3%.The muon measurement data from Nagoya station,Japan,was employed and the characters of cosmic ray evolvement before geomagnetic storm were revealed by analyzing the references between the data just before the geomagnetic storms and the quiet days.It was found that fluctuations before geomagnetic storms increased due to the approaching of CME because the shock front and strong IMF induced by CME.An index to measure the fluctuation of data,D8(t),was used in the cosmic ray data processing.The result shows that D8(t)always increases monotonously several hours ahead the geomagnetic storm,which hopefully could become a useful factor for geomagnetic storm prediction.As it had been known that most of the large geomagnetic storms were caused by CME accompanying the solar Proton Event(SPE). the SPE were also chosen together with D8(t)in the prediction process.The mentioned algorithm was tested with the relative data of whole year 2001.The result turned out to be encouraging with the accuracy ratereach to 80%f8 out of 101 and false rate less than 18%(2 out of 11).The electron flux on geo-synchronous orbit varies in a large range even up to three orders accompanied the passage of interplanetary magnetic cloud and the following geomagnetic disturbances.Through time series analyzing,the fluctuation of electron flux was found to have a period of 27days.It was also reveal that the non periodic relativistic electron storm all occurred during the time range when the energetic electron flux was due to be high which was coincident with the theory of"seed electron".A model based on Max-entropy theory was built concerning the role of period and geomagnetic disturbance.The result indicated that the model could predict both periodic and non-periodic relativistic electron storm and the predicted flux was also well predicted.Upon the investigation of electron flux,interplanetary solar wind data,and geomagnetic data as well,we employed the artificial neural network to build the relationship between the electron flux and interplanetary parameters.The input parameter is:the solar wind velocity in km/s;the density of the plasma density;the total strength of interplanetary magnetic field.All of them were daily average value. To represent the influence of the variation of the interplanetary parameters influence, the ratio with those of the previous day of the mentioned parameters were also input. Through the training,the neural network can adjust the internal weight and the value of the nods to build the relationship between the electron flux and the interplanetary parameters and their variation.Preliminary result showed that the accuracy forecast of electron flux 1 day ahead can reach 80%.