Equatorial predictions from a new neural network based global foF2 model L.A. McKinnell a,b, * , E.O. Oyeyemi c a Hermanus Magnetic Observatory, Hermanus, South Africa b Department of Physics & Electronics, Rhodes University, Grahamstown, South Africa c Department of Physics, University of Lagos, Lagos, Nigeria Received 14 September 2009; received in revised form 12 December 2009; accepted 12 December 2009 Abstract A new neural network (NN) based global empirical model for the foF2 parameter, which represents the peak ionospheric electron density, has been developed using extended temporal and spatial geophysical relevant inputs. It has been proposed that this new model be considered as a suitable replacement for the International Union of Radio Science (URSI) and International Radio Consultative Committee (CCIR) model options currently used within the International Reference Ionosphere (IRI) model for the purpose of F2 peak electron density predictions. The most recent version of the model has incorporated data from 135 global ionospheric stations including a number of equatorial stations. This paper concentrates on the ability of this new model to predict foF2 for the equatorial sector, an area that has been identified as problematic within the current IRI peak prediction setup. The improvement in the predictions of the foF2 parameter by the new model as compared to the URSI and CCIR model options of the IRI is demonstrated and the requirement for additional foF2 data from the equa- torial zone for the purpose of global modeling of foF2 is highlighted in this paper. Ó 2010 COSPAR. Published by Elsevier Ltd. All rights reserved. Keywords: Equatorial ionosphere; foF2; Neural networks; Modeling; IRI 1. Introduction Many researchers (Jones and Obitts, 1970; Rush et al., 1983, 1984; Fox and McNamara, 1988; Bilitza, 2001; Ful- ler-Rowell et al., 2000) have made significant efforts both at developing and improving the existing global models for ion- ospheric parameter predictions. In particular during recent years various researchers have demonstrated the applicabil- ity of the technique of neural networks (NNs) to ionospheric predictions (Williscroft and Poole, 1996; Wu and Lundstedt, 1996; Altinay et al., 1997; Weigel et al., 1999; Wintoft and Cander, 1999; Lamming and Cander, 1999; Poole and McKinnell, 2000; Tulunay et al., 2000; McKinnell, 2002). Their works have shown that highly non-linear and complex processes in the near Earth-space are advantageously dealt with using data-driven modeling techniques such as NNs. Previously (Oyeyemi et al., 2005a, 2005b, 2006; Oyeyemi and McKinnell, 2008) we have demonstrated that NNs can be successfully used in the development of a global model for the ionospheric parameter foF2 (F2 peak critical fre- quency representing the peak ionospheric electron density). In these works we used measured foF2 ionosonde data from a number of ionospheric stations across the globe for training NNs. In Oyeyemi and McKinnell (2008) and McKinnell and Oyeyemi (2009), we report on the status of the global foF2 model, which at that time was developed using data from 85 global ionospheric stations. The two papers appealed for more data from particular areas, such 0273-1177/$36.00 Ó 2010 COSPAR. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.asr.2010.06.003 * Corresponding author at: Department of Physics & Electronics, Rhodes University, Grahamstown, South Africa. E-mail address: lmckinnell@hmo.ac.za (L.A. McKinnell). www.elsevier.com/locate/asr Available online at www.sciencedirect.com Advances in Space Research 46 (2010) 1016–1023