Models
MAGE  CHIMP | GAMERA | IPWM | RCM | REMIX | SAMI3 | TS07D | WACCM-X

TS07D model

Tsyganenko & Sitnov '07 Data Mining

DM-based reconstruction of the July 2013 storm: (a-b) Evolution of the indices Sym-H and AL, solar wind/IMF parameter vBzIMF and the dynamic pressure Pdyn. (c) The color-coded azimuthal component of the equatorial current density at the moment marked in (a-b) by the vertical magenta line with blue arrows showing the projections of the current density vectors on the equatorial plane and geosynchronous orbit shown as a white circle.

TS07d is a data-mining algorithm including the most flexible and expandable magnetic field model TS07 (Tsyganenko&Sitnov, 2007). It allows one to produce from historical spaceborne magnetometer data 3-D snapshots of the empirical geomagnetic field, underlying electric currents and derived plasma pressure. The latter can be directly ingested into a global MHD model to improve its ad hoc equation of state and it helps it to properly describe the storm-time ring current. The main idea of the data mining (DM) is to form out of historical data a swarm of virtual probes whose observations were made when the state of the magnetosphere was similar to that of the query moment.

The number of probes in the swarm must be:

  1. large enough to fit a sufficiently detailed description of the magnetic field without overfitting, and at the same time,
  2. small enough to provide sensitivity to storm and substorm variations, including their different phases

At present, the number of records in the model database covering the period 1995-2017 with 5- to 15-minute cadence is ~4 • 106, while the size of DM bins varies from 8 • 103 to 3 • 104. This allows one to use for the magnetosphere reconstruction a magnetic field model with > 103 degrees of freedom to capture both storm (Sitnov et al., 2008, 2010; Stephens et al., 2016) and substorm (Sitnov et al., 2019; Stephens et al., 2019) features. DM is provided by the K Nearest Neighbor method or KNN (Cover&Hurt, 1967), whose advanced version with radially weighted NNs allows (Mitchell, 1997) to fit the model using effectively as few as 200 virtual probes. It is proven to be able to restore the plasma pressure distributions for superstorms, such as the Bastille Day event (Sitnov et al., 2018). TS07d has been validated using available in-situ measurements for many cases studies, the statistical analysis of NN bins, comparison with AMPERE observations of low-altitude field-aligned current distributions, energetic neutral atom images and storm-time plasma pressure evaluation using Van Allen Probes data (Sitnov et al., 2018).

References

Cover, T., & Hart, P., Nearest neighbor pattern classification. IEEE Transactions on Information Theory, 13(1), 21–27. https://doi.org/10.1109/TIT.1967.1053964, 1967.

Mitchell, T., Machine learning. McGraw Hill, 1997.

Sitnov, M. I., N. A. Tsyganenko, A. Y. Ukhorskiy, and P. C. Brandt, Dynamical data‐based modeling of the storm‐time geomagnetic field with enhanced spatial resolution, J. Geophys. Res., 113, A07218, https://doi.org/10.1029/2007JA013003, 2008.

Sitnov, M. I., Tsyganenko, N. A., Ukhorskiy, A. Y., Anderson, B. J., Korth, H., Lui, A. T. Y., & Brandt, P. C., Empirical modeling of a CIR‐driven magnetic storm. Journal of Geophysical Research, 115, A07231. https://doi.org/10.1029/2009JA015169, 2010.

Sitnov, M. I., Stephens, G. K., Gkioulidou, M., Merkin, V., Ukhorskiy, A. Y., Korth, H., Brandt, P. C., & Tsyganenko, N. A., Chapter 11 ‐ Empirical modeling of extreme events: Storm‐time geomagnetic field, electric current, and pressure distributions. In N. Buzulukova (Ed.), Extreme Events in Geospace (pp. 259–279): Elsevier. https://doi.org/10.1016/B978-0-12-812700-1.00011-X, 2018.

Sitnov, M. I., Stephens, G. K., Tsyganenko, N. A., Miyashita, Y., Merkin, V. G., Motoba, T., et al., Signatures of nonideal plasma evolution during substorms obtained by mining multimission magnetometer data. Journal of Geophysical Research: Space Physics, 124. https://doi.org/10.1029/2019JA027037, 2019.

Stephens, G. K., Sitnov, M. I., Ukhorskiy, A. Y., Roelof, E. C., Tsyganenko, N. A., & Le, G., Empirical modeling of the storm time innermost magnetosphere using Van Allen Probes and THEMIS data: Eastward and banana currents. Journal of Geophysical Research: Space Physics, 121, 157–170. https://doi.org/10.1002/2015JA021700, 2016.

Stephens, G. K., Sitnov, M. I., Korth, H., Tsyganenko, N. A., Ohtani, S., Gkioulidou, M., & Ukhorskiy, A. Y. Global empirical picture of magnetospheric substorms inferred from multimission magnetometer data. Journal of Geophysical Research: Space Physics, 124, 1085– 1110. https://doi.org/10.1029/2018JA025843, 2019.

Tsyganenko N. A., M. I. Sitnov, Magnetospheric configurations from a high-resolution data-based magnetic field model, J. Geophys. Res., 112, A06225, https://doi.org/10.1029/2007JA012260, 2007.