Magnetospheric substorms have been in the center of keen interest since the early 1960s (Kepko et al., 2015;Meng & Liou, 2004). An immense number of studies have been published, focused on many aspects of substorm effects in the magnetosphere and at ionospheric altitudes; see Chu et al. (2021) for an ample list of references. With respect to the modeling, the substorms still pose a great challenge, due to the large variety of disturbance scenarios and the stubborn problem of predicting their onsets. First-principle simulations reproduce the substorms in fair resemblance to reality; however, results of the existing codes are widely different from each other (Gordeev et al., 2017).In the empirical data-based modeling, the most recent significant advance in reconstructing the substorm effects was made by Stephens et al. (2019) andSitnov et al. (2019) (referred below as SST19) on the basis of the nearest-neighbor (NN) data-mining (DM) in the space of magnetospheric state parameters. Equatorial sources were represented by a sum of quasi-orthogonal vector potentials (Tsyganenko & Sitnov, 2007), corresponding to a set of embedded planar current sheets (CSs) of different fixed thickness, while the field-aligned currents (FACs) were included as a superposition of deformed conical harmonics (Tsyganenko, 1991(Tsyganenko, , 2002a(Tsyganenko, , 2002b. The NN selection was based on values of AL and Sym-H indices and their time derivatives, slide-averaged over time intervals of 1 and 6 h, respectively. Also, the interplanetary s E vB driver was taken into account using the L1 upstream data. Based on the above premises, a successful reconstruction of the entire empirical picture of the substorm cycle was made in SST19, including the magnetotail flux