Performance evaluation of the IRI 2016 and IRI-Plas 2017 models over central Asian mid-latitude regions in the descending phase of solar cycle 24

This paper scrutinizes the performance of the latest versions of the IRI model (IRI 2016 with NeQuick, IRI01-corr and IRI2001 options for the topside electron density) and the IRI extended to the plasmasphere (IRI-Plas 2017) models in the estimation of the Vertical TEC (VTEC) variation over the central Asian mid-latitude regions in the descending phase of solar cycle 24 (2014-2016). The GPS dual frequency receivers installed at Kurchatov, KRTV (geog 50.71 degrees N, 78.62 degrees E, Geom. 41.84 degrees N), Khantau, SUMK (44.21 degrees N, 74.00 degrees E, Geom. 35.73 degrees N), Talas, TALA (42.45 degrees N, 72.21 degrees E, Geom.34.13 degrees N) and Kazarman, KAZA (41.38 degrees N, 73.94 degrees E, Geom.32.92 degrees N) have been used to derive the estimate of the vertical TEC (GPS-VTEC) for the comparison of the monthly and seasonal performance of the models. The modelled VTEC values generally tend to be larger than the GPS-VTEC values during periods of high solar irradiance (daytime hours), with the highest overestimation being observed by IRI-Plas 2017 model followed by IRI 2016 model with the IRI2001 topside option. However, the differences between the models and between the models and the GPS-VTEC values become diminished as the solar irradiance decreases, with the highest underestimation being observed by the IRI 2016 model with NeQuick topside option. It has also been shown that the smallest root-mean-square deviations between the GPS-VTEC and modelled VTEC are observed generally in the June solstice months, showing that the models perform best during local summer. On the contrary, the largest root-mean-square deviations between the modelled VTEC and GPS-VTEC are observed during high solar irradiance on the surface of the Earth (especially in the time interval between 05:00 and 10:00 UT which corresponds to the daytime hours 10:00 and 15:00 LT), showing that the models perform poorly during high solar irradiance. In addition, both the IRI 2016 and IRI-Plas 2017 models show a progressive decline in VTEC during a negative storm, but do not adequately estimate the storm time VTEC variation. (C) 2019 COSPAR. Published by Elsevier Ltd. All rights reserved.