The development of a regional geomagnetic daily variation model using neural networks

Abstract

Abstract. Global and regional geomagnetic ®eld models give the components of the geomagnetic ®eld as functions of position and epoch; most utilise a polynomial or Fourier series to map the input variables to the geomagnetic ®eld values. The only temporal variation generally catered for in these models is the long term secular variation. However, there is an increasing need amongst certain users for models able to provide shorter term temporal variations, such as the geomagnetic daily variation. In this study, for the ®rst time, arti®cial neural networks (ANNs) are utilised to develop a geomagnetic daily variation model. The model developed is for the southern African region; however, the method used could be applied to any other region or even globally. Besides local time and latitude, input variables considered in the daily variation model are season, sunspot number, and degree of geomagnetic activity. The ANN modelling of the geomagnetic daily variation is found to give results very similar to those obtained by the synthesis of harmonic coe cients which have been computed by the more traditional harmonic analysis of the daily variation.

Systems theory for geospace plasma dynamics

“… Jankovicova et al [2002] optimized the solar wind input to their Dst neural network by using a principal component basis. High‐latitude geomagnetic activity has also been studied with neural networks including the response of geomagnetic indices [ Gleisner and Lundstedt , 1997; Takalo and Timonen , 1999; Weigel et al , 1999], the regional magnetic field [ Sutcliffe , 2000], and substorm occurrence [ Sutcliffe , 1997].…”

The Embrace Magnetometer Network for South America: First Scientific Results

“…Basic description of the model Sutcliffe (1999) Empirical model developed for providing a geomagnetic daily variation over the Southern African region as a function of season, sunspot number, and degree of geomagnetic activity, based on artificial neural networks Alken and Maus (2007) Empirical model of the equatorial electrojet magnetic signature as a function of longitude, local time, season, solar flux, and lunar local time, derived from CHAMP, Ørsted, and SAC-C satellites Unnikrishnan (2012) Empirical model for providing a geomagnetic daily variation over the Alibag observatory, India, for solar quiet conditions, based on artificial neural networks Ouadfeul et al (2015) Empirical model for providing a geomagnetic daily variation over the German Wingst observatory, Germany, as function of degree of geomagnetic activity, based on artificial neural networks Jayapal et al (2016) Empirical model developed for the horizontal component of the Earth's magnetic field over Thiruvananthapuram, India, as a function of the solar cycle, seasonal, and degree of geomagnetic activity, based on Fourier analyzes Radio Science 10.1002/2018RS006540 (Rostoker, 1972). Therefore, we decided to investigate the use of equation (2) for computing on the range of the component over the 3 h period at each magnetic station (N = 1) instead of using the H variation (or either X and Y components) given by equation (1).…”

Nonlinear time series analysis of the fluctuations of the geomagnetic horizontal field

“…Although the Sq variations are the most regular of all the geomagnetic field variations, tending to repeat itself with a periodicity of 24 h, significant day-to-day differences do occur (Hibberd, 1981;Sutcliffe, 2000). At low and middle latitude stations H , is known to change drastically during a geomagnetic storm.…”

“…The analysis of the fluctuations in the geomagnetic field has many practical applications in magnetic navigation, orientation control, geophysical exploration, etc. (Newitt, 1993;Kerridge, 1993;Gonzalez et al, 1994;Sutcliffe, 2000). The analysis of storm morphology has been undertaken by several authors.…”