11:45 AM - 12:00 PM
[SEM13-10] Benchmark study toward DGRF2020 for IGRF14 - Reproducing DGRF2015 using magnetic data from ground observatories and Swarm satellites
Keywords:IGRF, magnetic field, satellite data, magetnic observatory
The 14th generation of the International Geomagnetic Reference Field (IGRF14) is scheduled for release on the 1st of January, 2025. Each new generation of IGRF consists of three models: DGRF, IGRF, and SV. DGRF serves as the definitive model for the epoch five years prior to the release, while IGRF is the tentative model for the epoch. SV, on the other hand, is the first time derivatives valid for next five years. These models are presented by Gauss coefficients of the spherical harmonic (SH) expansion of the scalar potential with truncation degrees of 13 for DGRF and IGRF, and 8 for SV.
Japanese groups have never contributed to either DGRF or IGRF, although an SV model has been submitted for the first time to the IGRF community (Minami et al., 2020). To contribute to DGRF2020 in IGRF14 from Japan, we are currently conducting a benchmark study wherein we aim to generate a magnetic field model at 2015.0 sufficiently close to the DGRF2015 model (Alken et al., 2021). We utilize magnetic data from both on-land magnetic observatories and Swarm satellites.
We adopt the geomagnetic virtual observatory (GVO) method proposed by Hammer et al. (2021) to evaluate satellite data. The Earth’s surface is divided into 300 circles with a radius of 700 km, and a cylinder with semi-infinite height from the surface is imaged for each circle. Magnetic scalar potential within each cylinder is estimated using Swarm satellite data measured within the cylinder during a specific time window, with local Cartesian coordinates centered at the GVO point (the center of the circle at an altitude of 490 km). The magnetic scalar potential is expressed using third-order polynomials, and the coefficients are determined to fit the three component satellite magnetic data within the cylinder via a robust least square method with Huber weight (Constable, 1988). We divide twelve months into three time-windows and successfully calculate the three components of the magnetic field at each GVO point for each time window. Ultimately, we compute the annual mean of the magnetic field at each GVO point, by which its seasonal variations are eliminated.
The SH expansion of the Earth's main magnetic field at 2015.0 is calculated along with the annual mean of ground observatory data. Three components of magnetic data are utilized following data selection criteria of Kp < 2, the absolute value of Dst index < 30 nT, and | geomagnetic latitude | < 55 degrees. As for local time, 23 to 5 LT is adopted. Currently, the root mean square power of the differences in Gauss coefficients between our model and the DGRF2015 is about 300 nT, which implies that there is still a room for improvement in our method. In the presentation, we explain our methodology to build the DGRF model and report details of the comparison between our model at 2015.0 with DGRF2015.
Japanese groups have never contributed to either DGRF or IGRF, although an SV model has been submitted for the first time to the IGRF community (Minami et al., 2020). To contribute to DGRF2020 in IGRF14 from Japan, we are currently conducting a benchmark study wherein we aim to generate a magnetic field model at 2015.0 sufficiently close to the DGRF2015 model (Alken et al., 2021). We utilize magnetic data from both on-land magnetic observatories and Swarm satellites.
We adopt the geomagnetic virtual observatory (GVO) method proposed by Hammer et al. (2021) to evaluate satellite data. The Earth’s surface is divided into 300 circles with a radius of 700 km, and a cylinder with semi-infinite height from the surface is imaged for each circle. Magnetic scalar potential within each cylinder is estimated using Swarm satellite data measured within the cylinder during a specific time window, with local Cartesian coordinates centered at the GVO point (the center of the circle at an altitude of 490 km). The magnetic scalar potential is expressed using third-order polynomials, and the coefficients are determined to fit the three component satellite magnetic data within the cylinder via a robust least square method with Huber weight (Constable, 1988). We divide twelve months into three time-windows and successfully calculate the three components of the magnetic field at each GVO point for each time window. Ultimately, we compute the annual mean of the magnetic field at each GVO point, by which its seasonal variations are eliminated.
The SH expansion of the Earth's main magnetic field at 2015.0 is calculated along with the annual mean of ground observatory data. Three components of magnetic data are utilized following data selection criteria of Kp < 2, the absolute value of Dst index < 30 nT, and | geomagnetic latitude | < 55 degrees. As for local time, 23 to 5 LT is adopted. Currently, the root mean square power of the differences in Gauss coefficients between our model and the DGRF2015 is about 300 nT, which implies that there is still a room for improvement in our method. In the presentation, we explain our methodology to build the DGRF model and report details of the comparison between our model at 2015.0 with DGRF2015.