.png){kind=icon}
.png){kind=icon}
Return
Multi-layer ionospheric model construction constrained with NeQuick-G model
.png){kind=icon}
PRE
AI
DOI：10.1007/s10291-026-02073-1.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The ionosphere significantly affects radio propagation, making accurate modeling essential for GNSS applications. For the multi-layer ionospheric model, the space partitioning and the form of the mathematical function should be predetermined. To this end, the number and heights of layers are determined by principal component analysis (PCA) and K-means cluster using the ionospheric electron density (IED) profile obtained by the NeQuick-G model, respectively. Subsequently, the order of polynomial of each layer and modeling period is predetermined by Akaike information criterion (AIC) using the simulated total electron contents (TEC) from the NeQuick-G model during one-day gap. Finally, the multi-layer ionospheric model is constructed by least square method using the real GNSS observations collected on days of year (DOYs) 090–094 and 109–113, 2024 in Hong Kong area. To validate the proposed model, the fitting errors at validation stations, differential slant TEC (dSTEC) observations and positioning accuracy of single-frequency precise point positioning (SF-PPP) with constraint of ionospheric model are calculated. Results show that the FMLM improves the mean bias and RMS of TEC fitting by about 6% and 35%, respectively, during DOYs 090–094, and yields better dSTEC consistency in both test periods. In SF-PPP, the proposed model achieves positioning accuracy of about 0.5 m in both horizontal and vertical components, with overall improvements of 20%–50% relative to the other two models. Gathering all results claims the validation of the proposed method in TEC and positioning domains.
Keywords:
Multi-layer ionospheric model
NeQuick-G model
Principal component analysis (PCA)
K-means cluster
Akaike information criterion