A Spatial Reconstitution Model Based on Geographic Deep Echo State Networks for Ionospheric foF2 in East Asia

Developing a high-accuracy regional reconstitution model for the critical frequency of the ionospheric F2 layer (foF2), using data from ionospheric monitoring stations, is important for both high-frequency (HF) wireless communications and space remote sensing. To finish further accuracy reconstitution, we propose a novel spatial interpolation method, the Geo-deep echo state network (DESN) model, to generate the spatial distribution of ionospheric foF2. This model is based on the DESN's approach and is designed to construct a regional map of foF2, incorporating a geographic layer to integrate the spatial correlations observed in ionospheric monitoring station data. The reconstruction modeling and analysis are conducted using data from 13 monitoring stations in East Asia, covering the period from 2013 to 2017. The results indicate that the Geo-DESN model significantly outperforms the CCIR and International Union of Radio Science (URSI) models within the International Reference Ionosphere (IRI), as well as the Kriging method, during both high and low solar activity years, with an accuracy improvement ranging from 10% to 30%. During the severe magnetic storm of 2014, the Geo-DESN model, furthermore, demonstrates superior capability in tracking the trends of measured foF2 data. Additionally, we construct a spatial map of foF2 in East Asia with a resolution of 0.5(degrees)x0.5(degrees) . Comparative analysis confirms that the Geo-DESN model achieves high-precision map construction of foF2 in East Asia and exhibits greater robustness than other models.