4:15 PM - 4:30 PM
[PEM10-20] Reconstruction of precipitating electrons and three-dimensional structure of a pulsating auroral patch from monochromatic auroral images obtained from multiple observation points
Keywords:Pulsating aurora, Computed tomography, Ionoshpere, EISCAT radar
One of promising methods to examine the energy and spatial distributions of precipitating electrons is Aurora Computed Tomography (ACT). ACT reconstructs the three-dimensional (3-D) volume emission rate (VER) and two-dimensional (2-D) precipitating electron flux from monochromatic auroral images obtained at multipoint ground-based stations by solving an inverse problem. The ACT method has been applied to discrete auroras, but there have been no reports of 3-D reconstruction of pulsating auroras (PsAs) so far. This is because PsAs generally have a low contrast and faint structure. Recently, on the other hand, remote operation of many high-sensitivity cameras via internet and an archive system capable of storing a huge amount of aurora data make it possible to observe PsA emission with high signal-to-noise ratio and with high temporal resolution at multiple observation points simultaneously. Therefore, this study aims to reconstruct the 3-D VER and 2-D precipitating electrons in PsA patches using the highly sensitive cameras for the first time and to evaluate the reconstruction result.
We used three all-sky cameras at Skibotn (69.35 degrees north, 20.36 degrees east), Kilpisjarvi (69.05 degrees north, 20.78 degrees east), and Abisko (68.36 degrees north, 18.82 degrees east) to obtain monochromatic auroral images. The observation wavelength was 427.8 nm and temporal resolution was 2 s. During the substorm recovery phase in 0–2 UT on February 18, 2018, PsA patches were observed by the three cameras. We conducted ACT for PsA patches observed for 12 s at the observation point of the European Incoherent Scatter (EISCAT) radar within the field-of-view of the three cameras. To apply the ACT method to diffuse and dimmer PsA auroral patches, we improved the previous ACT used for discrete auroras in the following three points: the subtraction of background diffuse aurora from the auroral images prior to ACT, the estimation of the relative sensitivity between all-sky cameras, and the determination of the hyper-parameters of the smoothness term.
As a result, we succeeded to reconstruct 3-D VER of the PsA patches and the horizontal distribution of precipitating electrons for the first time. The characteristic energy of the reconstructed precipitating electron flux ranges from 6 keV to 23 keV and the peak altitude of the reconstructed VER ranges from 90 to 104 km. These results are consistent with previous studies. We found that the horizontal distribution of precipitating electron’s characteristic energy was neither uniform nor stable in the PsA patch during the pulsation. These spatiotemporal variations indicate changes in the cyclotron resonance energy of whistler mode chorus waves, which precipitate electrons from the magnetosphere to the ionosphere. Therefore, the observed spatiotemporal variations of PsAs are important to understand the background magnetic and plasma conditions in the magnetospheric source region. Our results of reconstruction of precipitating electrons and VER are a great advantage of multiple ground-based data because such 2- and 3-D distributions cannot be obtained by rockets and satellites.
We quantitatively evaluated the reconstructed results using a PsA patch model with adding artificial noises. In addition, the reconstructed VER was converted to the ionospheric electron density by solving the continuity equation with the Runge–Kutta method. We confirmed that the electron density was reconstructed with sufficient accuracy by comparing the reconstructed electron density with that obtained from the EISCAT radar.
We used three all-sky cameras at Skibotn (69.35 degrees north, 20.36 degrees east), Kilpisjarvi (69.05 degrees north, 20.78 degrees east), and Abisko (68.36 degrees north, 18.82 degrees east) to obtain monochromatic auroral images. The observation wavelength was 427.8 nm and temporal resolution was 2 s. During the substorm recovery phase in 0–2 UT on February 18, 2018, PsA patches were observed by the three cameras. We conducted ACT for PsA patches observed for 12 s at the observation point of the European Incoherent Scatter (EISCAT) radar within the field-of-view of the three cameras. To apply the ACT method to diffuse and dimmer PsA auroral patches, we improved the previous ACT used for discrete auroras in the following three points: the subtraction of background diffuse aurora from the auroral images prior to ACT, the estimation of the relative sensitivity between all-sky cameras, and the determination of the hyper-parameters of the smoothness term.
As a result, we succeeded to reconstruct 3-D VER of the PsA patches and the horizontal distribution of precipitating electrons for the first time. The characteristic energy of the reconstructed precipitating electron flux ranges from 6 keV to 23 keV and the peak altitude of the reconstructed VER ranges from 90 to 104 km. These results are consistent with previous studies. We found that the horizontal distribution of precipitating electron’s characteristic energy was neither uniform nor stable in the PsA patch during the pulsation. These spatiotemporal variations indicate changes in the cyclotron resonance energy of whistler mode chorus waves, which precipitate electrons from the magnetosphere to the ionosphere. Therefore, the observed spatiotemporal variations of PsAs are important to understand the background magnetic and plasma conditions in the magnetospheric source region. Our results of reconstruction of precipitating electrons and VER are a great advantage of multiple ground-based data because such 2- and 3-D distributions cannot be obtained by rockets and satellites.
We quantitatively evaluated the reconstructed results using a PsA patch model with adding artificial noises. In addition, the reconstructed VER was converted to the ionospheric electron density by solving the continuity equation with the Runge–Kutta method. We confirmed that the electron density was reconstructed with sufficient accuracy by comparing the reconstructed electron density with that obtained from the EISCAT radar.