Verifications of a 3-D regional ionospheric physics-based model over the Korean peninsula

We have recently developed a Korean ionosphere physics-based prediction model (KIPM) that can predict the state of the ionosphere over and around the Korean peninsula as a function of latitude, longitude, altitude, and time. The KIPM model is an extended version of the SAMI2-CNU model (Kim et al., 2016), which extended the SAMI2 model into a three-dimensional space. The KIPM model has a spatial resolution of 1° for longitude (120-140°E) and latitude (20-40°N) near the Korean Peninsula, with a height resolution of 10 km between 85 km and 1000 km. We upgraded the solar irradiance model and the thermospheric wind field and density models used in the SAMI2-CNU model to the latest versions. To evaluate the performance of the KIPM model, we compared its predictions with various ionospheric measurements (Ionosonde, GPS TEC) observed in Korea, along with the SAMI2, SAMI3, and IRI-2016 models. Through this comparison, we found the KIPM model showed 37% (10.2%), 8% (-4.6%), and 14.7% (1.1%) better predictive performance in terms of NmF2 (hmf2) at the Jeju location than the SAMI2, SAMI3, and IRI-2016 models, respectively. For vertical TEC, our model showed improvements of 14.2%, 20.7%, and 28.7% over the SAMI3, IRI-2016, and SAMI2 models, respectively, on average overall seasons. The SAMI3 model, a three-dimensional global model, can calculate the effect of plasma transport in the longitudinal direction, but it takes too long (several hours) to predict 24 h. On the other hand, the KIPM model only takes about 5-7 min, though it does not include the effect of longitudinal plasma transport. The fast calculation and performance of the KIPM should be significantly advantageous when estimating the 3-D regional ionosphere with data assimilation in near real time.