Mapping Ionosphereric Medium And Short Term Variation Using GPS Obervation

With the increasing dependence on space technology, a new scientific discipline named space weather has developed rapidly in recent years. As a connecting link between near-earth atmosphere and outer space, the ionosphere is one of the main objects of space weather research. The history of ionospheric research has only gone through about a hundred years. The insufficiency of long-term continuous observations has led to certain restrictions of sample representativeness. The complexity of ionosphere temporal and spatial variation has given rise to considerable uncertainty of modeling. In this case, multi-perspective analyses of existing data are still the focuses of the current research of the ionosphere. Especially, statistical modeling is still the most practical approach in the routine application. According to the features of TEC diurnal short-term variation and monthly medium-term variation, the applicability of two statistical models were discussed.The main research results and innovation points can be summarized as followings.1. The implied periodic components of TEC diurnal variation were revealed by spectral analysis. Taking the TEC diurnal variation time series in Sichuan area provided by IGS as examples, the amplitude summed of the prior 4 ones in the identified 5 periodic components, which corresponding to 24,12,8,64 hours, has a magnitude of 98% of the total variation amplitude. This means the 4-order Chapman-Miller modeling in the analysis of TEC diurnal variation will only lead to 2% of the calculation error.2. In the short-term modeling analysis, the periodic stability of the ionospheric TEC diurnal variation was well utililized. So the short-term (1 to 7 days) prediction accuracy of the Chapman-Miller method would not significantly reduce along with the increase of time, while other statistical models would do.3. The influences on TEC diurnal variation of the sun and the moon were quantitatively measured. In the amplitude of TEC diurnal variation in Sichuan area in 2008, solar component took the dominate position of about 83.4 to 95.0%. Although the moon could not directly affect the TEC variation as energy radiation, the attractive force of the moon can affect the spatial and temporal distribution and movement of the Earth’s atmosphere to a certain extent.4. The fitting value of TEC before Wenchuan M8.0 earthquake was calculated. Compared with solar radiation flux F107, the TEC disturbances were found well correlated with the variation of F107. TEC presented an increase of 30% with the increase 3.0%of F107, a decrease of 28.3% with the decrease of 2.9% of F107. The disturbances have closer relation with the variation of F107 rather than earthquake-induced possibly ionospheric changes.5. The index, which called mean of average absolute percentage error in within-sample forecasts in the last years, was used to judge the X-12-ARIMA models. The results of 5183 time series models showed that scheme (212)(011) was significantly better than others in the 5 optional combinations of autoregression and moving average method. This was different than the default scheme (212)(011).6. The unified smoothing standard used by X-12-ARIMA model brought about the comparability of components decomposed from monthly averaged TEC among different months. Taking the solar activity maximum and minimum month as examples, the asymmetry of TEC between north and south hemisphere could be expressed by two components. One is symmetrical trend and cycle component, the other is asymmetrical seasonal component. This will benefit the visualized comparison of ionosphere status among different regions.7. Several pieces of characteristics could be concluded from spatial and temporal variation of monthly averaged TEC calculated by Chapman-Miller method. Semiannual anomaly was found obvious in low latitude area, while annual variation dominated in the middle and high latitude area. The influence of seasonal change with monthly averaged TEC in high latitude area, especially in Antarctic regions, was far bigger than middle and low latitude area. Trend and cycle component showed solar activity periodic variation rather than seasonal variation.