[MIS09-01] Prediction of Geomagnetic Daily Variations
キーワード：geomagnetic daily variation, Natural Orthogonal Method, disturbances
Since 2008, many geomagnetic observatories have been built in China. The variety and time span of observations are increasing, which has provided abundant and valuable geomagnetic data for scientific research. However, rapid urbanization and construction development has negatively affected the geomagnetic observation environment. The magnetic field observations are subject to disturbances of several nT to several hundreds nT caused by infrastructure construction, vehicles, subway and light rail, high-voltage DC power lines, factory operations and so on. It is important to identify and exclude disturbances from massive data.
In this study, we have analysed how the magnetic disturbances relate to various influencing factors mentioned above. In addition, Natural Orthogonal Method (NOC), Space Weighting Method and Long Short-Term Memory Recuttent Neural Network (LSTM RNN) are utilised to develop a geomagnetic daily variation model based on geomagnetic data obtained from 30 observatories of Geomagnetic Network of China from 2017 to 2018. We will compare the daily variations calculated by the model and those measured by the observatories. Finally, we hope the results could be used to correct the observatory data for a significant part of the disturbances.
In this study, we have analysed how the magnetic disturbances relate to various influencing factors mentioned above. In addition, Natural Orthogonal Method (NOC), Space Weighting Method and Long Short-Term Memory Recuttent Neural Network (LSTM RNN) are utilised to develop a geomagnetic daily variation model based on geomagnetic data obtained from 30 observatories of Geomagnetic Network of China from 2017 to 2018. We will compare the daily variations calculated by the model and those measured by the observatories. Finally, we hope the results could be used to correct the observatory data for a significant part of the disturbances.