2:10 PM - 2:25 PM
[PEM09-15] Reconstruction of the three-dimensional structure and precipitating electrons of pulsating auroral patches by Aurora Computed Tomography
Keywords:Pulsating aurora, Computed Tomography
Aurora Computed Tomography (ACT) is an inversion problem analysis method to reconstruct the auroral three-dimensional (3-D) structure from monochromatic auroral images obtained with multiple ground-based optical instruments which have common field of view by applying the Computed Tomography method (e.g., Aso et al., 1990). In addition, Generalized-Aurora Computed Tomography (G-ACT) is the analysis method developed from ACT to reconstruct the energy and spatial distribution of precipitating electrons by combining ionospheric electron density data observed by an Incoherent Scatter radar (Aso et al., 2008; Tanaka et al., 2011). These methods have been applied to discrete auroras whose structures have discrete boundaries such as auroral arcs. On the other hand, it is relatively difficult to reconstruct the 3-D structure of pulsating aurora (PsA) by ACT since the structure boundaries of PsA is generally blurred. Therefore, ACT and G-ACT have not been applied to PsA yet.
In this study, we evaluate the possibility applying ACT and G-ACT to the PsA patch and will finally investigate the temporal changes of the 3-D structure and precipitating electrons of PsA patches. We used monochromatic auroral images obtained with three all-sky images in Scandinavia (Skibotn (69.35°N, 18.82°E), Kilpisjärvi (69.05°N, 20.36°E), Abisko (68.36°N, 18.82°E)) during the substorm recovery phase in 0 – 2 UT on 18 February 2018. The observation wavelength is 427.8 nm and temporal resolution is 2 s. We selected the isolated PsA patch observed by the three all-sky imagers to make the problem simple though PsA patches often occur in close proximity to each other. Then, we carried out to reconstruct its 3-D structure from only auroral images by ACT. The background diffuse auroral emission was subtracted from the auroral images as uniform emission to reduce noises before conducting ACT.
As a result, we obtained the auroral 3-D structure whose widths in the east-west, north-south, and altitude directions were 60 km, 42 km, and 14 km respectively. The accuracy of the reconstruction was evaluated by using a model aurora. We compared the auroral 3-D structure reconstructed from the pseudo images obtained by integrating model 3-D aurora in the line-of-sight directions to the model auroral 3-D structure and found that the peak altitude was correctly reconstructed but the peak value and altitude width of the reconstructed aurora were respectively ~18 % smaller and ~33 % larger than those of the model aurora. These errors would be caused by the horizontally wide structure of PsA patch which makes ambiguity in determining the lower- and upper-limit altitude from auroral images.
Next, we examined the events that a PsA patch was observed at the observation point of the EISCAT UHF radar in Tromsø (69.58°N, 19.23°E) and reconstructed the auroral 3-D structures from the three auroral images. The reconstructed 3-D volume emission rates were converted to electron densities to compare with the altitude profiles of electron densities observed by the EISCAT UHF radar. As a result, the maximum values of reconstructed electron densities were ~32 – 40 % smaller than those of observed ones. These errors might be caused by the subtraction of the background diffuse auroral emission from auroral images before conducting ACT, the error of reconstructing the 3-D volume emission rate by ACT, and the errors of the effective recombination coefficient and neutral atmospheric model which used to convert the volume emission rate to electron density.
We give the talk on the reconstruction of 3-D structure of PsA patches by G-ACT, and evaluate their accuracy. We also show the results on the temporal changes of 3-D structure and energy distribution of precipitating electrons that cause PsA patches.
In this study, we evaluate the possibility applying ACT and G-ACT to the PsA patch and will finally investigate the temporal changes of the 3-D structure and precipitating electrons of PsA patches. We used monochromatic auroral images obtained with three all-sky images in Scandinavia (Skibotn (69.35°N, 18.82°E), Kilpisjärvi (69.05°N, 20.36°E), Abisko (68.36°N, 18.82°E)) during the substorm recovery phase in 0 – 2 UT on 18 February 2018. The observation wavelength is 427.8 nm and temporal resolution is 2 s. We selected the isolated PsA patch observed by the three all-sky imagers to make the problem simple though PsA patches often occur in close proximity to each other. Then, we carried out to reconstruct its 3-D structure from only auroral images by ACT. The background diffuse auroral emission was subtracted from the auroral images as uniform emission to reduce noises before conducting ACT.
As a result, we obtained the auroral 3-D structure whose widths in the east-west, north-south, and altitude directions were 60 km, 42 km, and 14 km respectively. The accuracy of the reconstruction was evaluated by using a model aurora. We compared the auroral 3-D structure reconstructed from the pseudo images obtained by integrating model 3-D aurora in the line-of-sight directions to the model auroral 3-D structure and found that the peak altitude was correctly reconstructed but the peak value and altitude width of the reconstructed aurora were respectively ~18 % smaller and ~33 % larger than those of the model aurora. These errors would be caused by the horizontally wide structure of PsA patch which makes ambiguity in determining the lower- and upper-limit altitude from auroral images.
Next, we examined the events that a PsA patch was observed at the observation point of the EISCAT UHF radar in Tromsø (69.58°N, 19.23°E) and reconstructed the auroral 3-D structures from the three auroral images. The reconstructed 3-D volume emission rates were converted to electron densities to compare with the altitude profiles of electron densities observed by the EISCAT UHF radar. As a result, the maximum values of reconstructed electron densities were ~32 – 40 % smaller than those of observed ones. These errors might be caused by the subtraction of the background diffuse auroral emission from auroral images before conducting ACT, the error of reconstructing the 3-D volume emission rate by ACT, and the errors of the effective recombination coefficient and neutral atmospheric model which used to convert the volume emission rate to electron density.
We give the talk on the reconstruction of 3-D structure of PsA patches by G-ACT, and evaluate their accuracy. We also show the results on the temporal changes of 3-D structure and energy distribution of precipitating electrons that cause PsA patches.