Auroral Images Research Articles

A statistical study of Ring Current Aurora

The occurrence of Ring Current Aurora (RCA) due to precipitating ring current ions by pitch angle diffusion at sub-auroral latitudes was estimated by analyzing the TIMED/GUVI auroral images. RCA occurred at all magnetic local time (MLT) with a peak occurrence around 15:00 MLT and ∼65° magnetic latitudes. This is consistent with the EMIC wave occurrence pattern during geomagnetic active time (Meredith et al., 2014). The occurrences were similar in both hemispheres. However, the northern hemisphere occurrence was about 3 times of that in the southern hemisphere. The hemispheric difference may be due to the differences in the loss cone sizes on conjugated magnetic flux tubes and remains to investigated.

Far ultraviolet mirrors for aurora imaging: design and fabrication.

The emission lines of 140-180nm are auroral bands of N 2 Lyman-Birge-Hopfield, and they have been imaging targets of many satellites that need reflective mirrors. To obtain good imaging quality, the mirrors also should have excellent out-of-band reflection suppression as well as high reflectance at working wavelengths. We designed and fabricated non-periodic multilayer L a F 3/M g F 2 mirrors with working wave bands of 140-160nm and 160-180nm, respectively. We used a match design method and deep search method to design the multilayer. Our work has been utilized in the new wide-field auroral imager of China, and the application of these notch mirrors with excellent out-of-band suppression reduces the utilization of corresponding transmissive filters in the optical system of space payload. Furthermore, our work provides new routes for the design of other reflective mirrors in the far ultraviolet region.

Leroy L. Cogger: A Pioneer in Optical Observations From Space

AbstractThis perspective is in memory of Leroy Leonard Cogger, a pioneer in optical imaging of the aurora and the airglow, who passed away in October 2022 in Canada at the age of 85. In his 40‐plus‐year career, which spans both space and ground‐based optical instrumentation and observations, Dr. Cogger has made significant contributions to the development of instruments for several space missions and ground observatories, including the world's first truly digital (charge‐coupled device) camera to be used in the field, the UV auroral imagers for the Swedish Viking and Freja satellites, the NASA IMAGE satellite, and the Russian INTERBALL‐2 satellite, respectively, and the fast near‐infrared auroral imager on the Canadian CASSIOPE/Enhanced Polar Outflow Probe satellite. He has also collaborated and mentored the next generation of space scientists in Canada and Europe in many areas of auroral and airglow research; he was particularly adept at combining diverse observations to provide insight into a wide array of atmospheric phenomena. His legacies go well beyond his contributions in the peer‐reviewed literature and innovative instrumentation: his scientific vision and leadership played an indispensable role in the Canadian space science research agenda, and his commitment to mentoring has left an indelible mark on many of his students and associates. He exemplified AGU's motto of “Unselfish cooperation in research.” His gentle and quiet spirit, his passion for research and in making a difference in life will be very much missed by his many friends and colleagues in Canada and abroad alike, as will his great sense of humor.

Automated Classification of Auroral Images with Deep Neural Networks

Terrestrial auroras are highly structured that visualize the perturbations of energetic particles and electromagnetic fields in Earth’s space environments. However, the identification of auroral morphologies is often subjective, which results in confusion in the community. Automated tools are highly valuable in the classification of auroral structures. Both CNNs (convolutional neural networks) and transformer models based on the self-attention mechanism in deep learning are capable of extracting features from images. In this study, we applied multiple algorithms in the classification of auroral structures and performed a comparison on their performances. Trans-former and ConvNeXt models were firstly used in the analysis of auroras in this study. The results show that the ConvNeXt model can have the highest accuracy of 98.5% among all of the applied algorithms. This study provides a direct comparison of deep learning tools on the application of classifying auroral structures and shows promising capability, clearly demonstrating that auto-mated tools can help to minimize the bias in future auroral studies.

Multipoint Observations of Dispersive Scale Alfvénic Field‐Line Resonances Associated With Substorm Auroral Beads

AbstractWe present a case study of the field‐aligned current (FAC) systems that transpire within the high‐altitude auroral acceleration region of an “auroral bead” initiated double oval substorm observed on 23 February 2001 by the Cluster fleet. Conjunctive Cluster measurements and auroral images from IMAGE reveal that auroral bead current system formation and evolution is a multi‐scale, injection‐mediated process. The FACs at large scales vary on substorm evolution time scales (∼minutes) in response to the injection and evolution of hotter denser magnetospheric plasma. Embedded within the large‐scale FACs are intense short‐scale (≲ few 10s of km) currents comprising dispersive scale Alfvén wave (DAW) fluctuations. The DAWs are a complex mixture of ingoing and reflected components that regularly interfere to form a broad spectrum of kinetic (dispersive) scale Alfvénic field‐line resonances (KFLRs). The Alfvénic currents appear as a nested series of upward and downward FAC densities with amplitudes reaching a few 100 nA/m2. Energized field‐aligned or counterstreaming electrons near keV energies and below are observed with parallel skews that vary in concert with variations in the DAW current sense. Positive correlations between DAW electric field energy densities and the energies of energized H+, He+, and O+ outflow are observed, indicative of ion energization within the DAW fields. Due to their L‐shell location (L ∼5.8–7.0) and associations with injections, the KFLRs are interpreted as the high‐altitude auroral zone analog of KFLRs observed in the equatorial inner magnetosphere.

Clustering of Ultraviolet Auroral Oval Images Based on Deep Representation Learning

极光受太阳风驱动的地磁亚暴等大尺度动力学影响，其形态及演化因不同的太阳风-磁层-电离层耦合作用可能表现不同。目前，极光卵及其形态的归类大多依据极光演化理论作主观定性分析，没有明确的分类标准，故难以借助统计分析方法和有监督分类模型开展客观定量研究。建立了基于深度表征学习的紫外极光卵图像聚类模型（MoCo-GMM），并利用空间环境参数设计了评估模型物理合理性的方法，在大规模POLAR卫星紫外极光卵图像数据上进行了实验，聚类结果不仅具有良好的簇内凝聚性和簇间分散性，且具备一定的物理可解释性，有效实现了基于图像的极光卵及其形态的客观归类。

Clustering of Ultraviolet Auroral Oval Images Based on Deep Representation Learning

极光受太阳风驱动的地磁亚暴等大尺度动力学影响，其形态及演化因不同的太阳风-磁层-电离层耦合作用可能表现不同。目前，极光卵及其形态的归类大多依据极光演化理论作主观定性分析，没有明确的分类标准，故难以借助统计分析方法和有监督分类模型开展客观定量研究。建立了基于深度表征学习的紫外极光卵图像聚类模型（MoCo-GMM），并利用空间环境参数设计了评估模型物理合理性的方法，在大规模POLAR卫星紫外极光卵图像数据上进行了实验，聚类结果不仅具有良好的簇内凝聚性和簇间分散性，且具备一定的物理可解释性，有效实现了基于图像的极光卵及其形态的客观归类。

Reconstruction of electron precipitation spectra at the top of the upper atmosphere using 427.8 nm auroral images

We present an innovative method to reconstruct the characteristics of precipitated electrons in auroral regions from optical measurements. This method is based on an optimization implemented between numerical simulations of the Transsolo code and tomographic maps made from the Auroral Large Imaging System (ALIS) network. We focus on the Volume Emission Rate (VER) of the blue line $ {\mathrm{N}}_2^{+}$ 1NG 427.8 nm, which is the most representative line of the energy deposition by electrons. The optimization is tested with the ALIS measurements carried out on March 05, 2008, at 18:41:30 UT and 18:42:40 UT. The reconstruction is performed by extracting the energy flux and the mean energy of the precipitating particles. Both Maxwellian and quasi-monoenergetic energy distributions are considered. Calculations performed with a Maxwellian energy distribution yielded a mean energy ranging from 1.8 to 5.2 keV with energy flux from 0.1 to 44.3 erg·cm−2·s−1 for 18:41:30 UT, and a mean energy from 2.2 to 9.5 keV with energy flux from 2.1 to 136.7 erg·cm−2·s−1 for 18:42:40 UT. Assuming a quasi-monoenergetic energy distribution, we find a mean energy ranging from 4.2 to 11.8 keV with energy flux ranging from 0.1 to 45 erg·cm−2·s−1 for 18:41:30 UT, and 8 to 17.1 keV with energy flux ranging from 2.2 to 110.1 erg·cm−2·s−1 for 18:42:40 UT. Moreover, we show this method allows us to reconstruct the energy characteristic of the precipitating electrons on a large region covering approximately 150 km × 150 km. This study also shows that some VER profiles of the maps are better fitted by quasi mono-energetic distributions while some others correspond to broadband distributions. It appears clearly that the energy flux is linked to the column integrated intensity, the mean energy is linked with the peak altitude of the emission, and the width of the energy distribution with the altitude thickness of the emissions.

Identifying the Structure and Propagation of Dawnside Pc5 ULF Waves Using Space‐Ground Conjunctions

AbstractPc5 ultralow frequency waves are important for transferring energy between the magnetosphere and ionosphere. While many observations have been performed on Pc5 waves properties, it has been difficult to determine the source region, signal propagation path, and the two‐dimensional structure of Pc5 waves beyond coverage by a small number of satellites. Pc5 waves often show a dawn‐dusk asymmetry, but the cause of the asymmetry is under debate. To address these issues, we used conjunction events between the THEMIS satellites and all‐sky imagers and analyzed two Pc5 wave events that were stronger on the dawnside. For both events, the Pc5 waves propagated from dawnside magnetopause toward the nightside magnetosphere. The Pc5 waves were also associated with dawnside magnetopause surface waves, which were probably induced by the Kelvin‐Helmholtz instability. The ionospheric equivalent currents identified multiple vortices on the dawnside associated with quasi‐periodic auroral arcs and much weaker perturbations on the duskside. Global auroral imaging also presented a similar dawn‐dusk asymmetry with multiple arcs on the dawnside, while only one or two major arcs existed on the duskside. Pc5 waves in the magnetosphere had an anti‐phase relation between the total magnetic field and thermal pressure, with a slower propagation velocity compared with magnetohydrodynamic waves. The Poynting flux was anti‐sunward with an oscillating field‐aligned component. These properties suggest that Pc5 waves were slow or drift mirror mode waves coupled with standing Alfven waves. The ground‐based and multi‐satellite observations provide crucial information for determining the Pc5 waves properties, possible source region, and signal propagation path.

Thermospheric Conditions Associated With the Loss of 40 Starlink Satellites

AbstractWe analyzed far ultraviolet data from Defense Meteorological Satellite Program (DMSP)/Special Sensor Ultraviolet Spectrographic Imager (SSUSI) and Thermosphere Ionosphere Mesosphere Energetics and Dynamics (TIMED)/Global Ultraviolet Imager (GUVI) and found significant changes in the thermospheric density and composition during the 3–5 February 2022 storm when 40 Starlink satellites started to re‐enter the atmosphere associated with increased neutral drag at an altitude around 210 km. The standard NRLMSISE‐00 model predicts only ∼5% increase in neutral density at 210 km. TIMED/GUVI observations showed a clear increase in the thermospheric N2/O column density ratio and an increase in the nitric oxide (NO) column density, indicating high thermospheric density, and temperature. DMSP SSUSI 130.4 nm limb data showed a clear increase in neutral density in the low latitude region (−30°–30°). On the duskside (17:32 LT), the density increase (compared to the pre‐storm density) is around 11%–18% at 210 km and around 18%–26% at 520 km. On the dawn‐side (05:19 LT), the density increase is more significant, 40%–59% at 210 km and ∼300% at 520 km. There are a few reasons for the dawn‐dusk asymmetry in the relative density increase: (a) the pre‐storm background density or heat capacity is lower on the dawn‐side than that on the duskside, (b) stronger auroral heating was observed on the dawnside based on analysis of auroral images, (c) the thermosphere density enhancement partially recovered when the disturbance co‐rotated with the Earth from dawn to dusk side due to enhanced infrared cooling by NO. Note that the neutral density increase is expected to be much higher at high latitudes where the neutral disturbances are initiated.

On the Relation Between Auroral Morphologies and Compression Conditions of Jupiter's Magnetopause: Observations From Juno and the Hubble Space Telescope

AbstractJupiter displays the most powerful auroral emissions in our solar system, which result from strong energy dissipation in Jupiter's surrounding space environment. Although mass and energy in Jupiter's magnetosphere mostly come from the innermost Galilean moon Io's volcanic activity and Jupiter's rotation, solar wind perturbations can play crucial roles in releasing magnetospheric energy. The systematic response of the aurora to a solar wind compression remains poorly understood because of timing uncertainties. Here we report the analysis of a set of auroral images from the Hubble Space Telescope with contemporaneous in situ magnetopause detections from Juno, allowing for a more direct comparison. By analyzing the dawn side main auroral emission, we distinguish two non‐mutually exclusive types of auroral enhancements: auroral dawn storm (ADS) featured with latitudinal extension in limited longitudes, and a long‐lasting main auroral brightening with limited extension in latitudes while extending over a large longitude range. Only the latter systematically appears under a compressed magnetopause, while the dawn storms could occur whatever the state of the magnetopause. The results could provide important constraints to improve theoretical models and numerical simulations. During expanded magnetopause conditions, Jupiter's aurora displayed either quiet or dawn storm morphology. The result is consistent with recent discovery of the initiation of ADSs in midnight and post‐midnight, possibly driven by magnetic reconnection plasma instabilities in night magnetotail. Our results show that some typical auroral morphologies could be used as a diagnostic of solar wind conditions at Jupiter.

AuroraX, PyAuroraX, and aurora-asi-lib: A user-friendly auroral all-sky imager analysis framework

Within the context of the Heliophysics System Observatory, optical images of the aurora are emerging as an important resource for exploring multi-scale geospace processes. This capability has never been more critical as we are on the cusp of a new era of geospace research, by which we mean studying the overall system as a system of systems. Historically, the patchwork of ground-based instrumentation has required customized solutions for accessing data, assessing data relevance, and then ultimately using each individual network alongside other assets. Here we introduce a new and comprehensive approach for data discovery and utilization for one type of data, namely auroral images. The AuroraX project (https://aurorax.space/) is a cyberinfrastructure platform for the discovery of scientific opportunities with access to optical auroral data. The program has broad objectives, so we focus on one key thread. In particular, we focus on describing the AuroraX platform and its API and web-based tools for all-sky imager (ASI) data. As a practical example, we demonstrate how to identify conjunctions using the AuroraX conjunction finder or PyAuroraX, a Python library that interfaces with the AuroraX platform. We then demonstrate how aurora-asi-lib, a Python library for interacting with and analyzing high-resolution ASI data, can be used for detailed conjunction analysis on a personal computer. Together, these tools enable a rapid and streamlined end-to-end exploration of auroral data.

In Situ Observation of a Magnetopause Indentation that Is Correspondent to Throat Aurora and Is Caused by Magnetopause Reconnection

AbstractThroat auroras have been suggested to correspond to magnetopause indentations, but how such indentations can be generated is an open question. Using coordinated magnetopause in situ and two‐dimensional auroral imaging observations, we showed that a transient magnetopause crossing observed by Time History of Events and Macroscale Interactions during Substorms near the dayside magnetopause corresponds to a throat aurora observed by Defense Meteorological Satellite Program/Special Sensor Ultraviolet Spectrographic Imager. The magnetopause crossing was confirmed to present a localized magnetopause indentation rather than arise from motion of the whole magnetopause. Using the minimum variance analysis and Walén test methods, we captured clear reconnection signatures at either boundary of the indentation. Also, an upstream solar wind discontinuity and a magnetosheath variation were identified to correlate well with the magnetopause indentation occurrence. These observations suggest that the magnetopause indentation was generated by reconnection and lend solid support to an assumption that throat aurora is produced by a particular reconnection that can develop inward the magnetosphere at a much‐localized region.

Reconstruction of precipitating electrons and three-dimensional structure of a pulsating auroral patch from monochromatic auroral images obtained from multiple observation points

Abstract. In recent years, aurora observation networks using high-sensitivity cameras have been developed in the polar regions. These networks allow dimmer auroras, such as pulsating auroras (PsAs), to be observed with a high signal-to-noise ratio. We reconstructed the horizontal distribution of precipitating electrons using computed tomography with monochromatic PsA images obtained from three observation points. The three-dimensional distribution of the volume emission rate (VER) of the PsA was also reconstructed. The characteristic energy of the reconstructed precipitating electron flux ranged from 6 to 23 keV, and the peak altitude of the reconstructed VER ranged from 90 to 104 km. We evaluated the results using a model aurora and compared the model's electron density with the observed one. The electron density was reconstructed correctly to some extent, even after a decrease in PsA intensity. These results suggest that the horizontal distribution of precipitating electrons associated with PsAs can be effectively reconstructed from ground-based optical observations.

Ionospheric Boundaries Derived From Auroral Images

AbstractThis paper presents updated methods for locating the Poleward and Equatorward Auroral Luminosity Boundaries (PALB and EALB) directly from IMAGE Far UltraViolet (FUV) images of the Northern Hemisphere auroral oval. Separate boundaries are determined from images measured at different FUV wavelengths. In addition, new methods for indirectly estimating the Open‐Closed magnetic field line Boundary (OCB) and the Equatorward Precipitation Boundary (EPB) locations are presented; these new boundaries are derived from a combination of the auroral luminosity boundary estimates with statistical latitudinal offsets derived from comparisons with low‐altitude spacecraft Particle Precipitation Boundaries (PPBs). Subsequently, we derive new circle model fits for all these boundary data sets, as well as new quality control criteria for these model fits. The suitability of circle fits for each of the data sets is discussed, and the OCB and PALB circle fits are validated against the Convection Reversal Boundary (CRB), as measured by low‐altitude in situ spacecraft. All the new boundary data sets, covering the epoch May 2000 to October 2002, are freely available online.

Northeastward‐Moving Auroral Forms From Possible High‐Latitude Reconnection

AbstractThis study presents a number of previously unreported auroral features that form shortly after the IMF –component turns sharply northward. The auroral images were captured by the Boreal Aurora Camera Constellation (BACC) All‐sky camera at the Kjell Henriksen Observatory, located in Longyearbyen, Svalbard. These observations are different from those previously reported. There are several differences: (a) The Eastward‐Moving Auroral Forms (EMAFs) do not originate from type 1 or type 2 aurora. The EMAFs originate from a diffuse vertical auroral form in the far west part of the ASC image. Vertical refers to the auroral form oriented in the north‐south direction in the frame of the ASC image. (b) The EMAFs form crescent shapes after they begin to drift eastward. (c) The EMAFs are discrete auroral forms; previously reported ionospheric signatures of high‐latitude reconnection are described as patches of brightness that form within a pre‐existing auroral band. (d) The EMAFs are observed to first move eastward and then slightly poleward during their journey. The movements of previously reported Ionospheric signatures of high‐latitude reconnection are described as expanding patches of brightness within a pre‐existing auroral band. (e) Some of the EMAFs brighten again after fading, during their northeastward drift across the field of view. (f) The EMAFs begin their eastward motion aligned in the north‐south direction. During their northeast drift, they begin to tilt toward the east. Some of the EMAFs keep tilting until they are aligned in the east‐west direction.

Lobe Reconnection and Cusp‐Aligned Auroral Arcs

AbstractFollowing the St. Patrick's Day (17 March) geomagnetic storm of 2013, the interplanetary magnetic field had near‐zero clock angle for almost two days. Throughout this period multiple cusp‐aligned auroral arcs formed in the polar regions; we present observations of, and provide a new explanation for, this poorly understood phenomenon. The arcs were observed by auroral imagers onboard satellites of the Defense Meteorological Satellite Program. Ionospheric flow measurements and observations of energetic particles from the same satellites show that the arcs were produced by inverted‐V precipitation associated with upward field‐aligned currents (FACs) at shears in the convection pattern. The large‐scale convection pattern revealed by the Super Dual Auroral Radar Network and the corresponding FAC pattern observed by the Active Magnetosphere and Planetary Electrodynamics Response Experiment suggest that dual‐lobe reconnection was ongoing to produce significant closure of the magnetosphere. However, we propose that once the magnetosphere became nearly closed complicated lobe reconnection geometries arose that produced interleaving of regions of open and closed magnetic flux and spatial and temporal structure in the convection pattern that evolved on timescales shorter than the orbital period of the DMSP spacecraft. This new model naturally explains many features of cusp‐aligned arcs, including why they focus in from the nightside toward the cusp region.

An automated auroral detection system using deep learning: real-time operation in Tromsø, Norway

The activity of citizen scientists who capture images of aurora borealis using digital cameras has recently been contributing to research regarding space physics by professional scientists. Auroral images captured using digital cameras not only fascinate us, but may also provide information about the energy of precipitating auroral electrons from space; this ability makes the use of digital cameras more meaningful. To support the application of digital cameras, we have developed artificial intelligence that monitors the auroral appearance in Tromsø, Norway, instead of relying on the human eye, and implemented a web application, “Tromsø AI”, which notifies the scientists of the appearance of auroras in real-time. This “AI” has a double meaning: artificial intelligence and eyes (instead of human eyes). Utilizing the Tromsø AI, we also classified large-scale optical data to derive annual, monthly, and UT variations of the auroral occurrence rate for the first time. The derived occurrence characteristics are fairly consistent with the results obtained using the naked eye, and the evaluation using the validation data also showed a high F1 score of over 93%, indicating that the classifier has a performance comparable to that of the human eye classifying observed images.

Automatic classification of mesoscale auroral forms using convolutional neural networks

Convolutional neural networks (CNNs) in deep learning enable the extraction of features in image data. Through the multi-layer superposition of a convolutional neural network, we can better capture the essential characteristics of different auroral subclasses and further classify auroral images in detail. Because the auroral morphological features often present abstract characteristics, our study compares different CNN architectures and different layering in order to test the best neural network model for mesoscale aurora classification. Although the classification models and subclasses used by us are both more complex, the highest F1 score of aurora classification of the test set reaches 99.6% (ResNet-50), which performs best comparing with previous works. Our classification models work also quite well when applied to an independent auroral image sequence, declaring our approach can automatically select images of various mesoscale auroral forms using CNNs, and allow the time sequence of auroral evolution to be seen automatically through the mesoscale auroral feature recognitions.

Bionic Design of the Vertical Bracket of Wide Angle Auroral Imager by Additive Manufacturing

In the aerospace field, lightweight design is a never-ending pursuit. By integrating structural bionics and structural optimization, the vertical bracket of a wide angle auroral imager is designed and manufactured by additive manufacturing technology in this work. Initially, the classical topology optimization is utilized for the vertical bracket to find the optimal material layout and primary load carrying paths. Drawing on the width-to-diameter ratio and the bone mineral density distribution of human femur, the vertical support is designed as a bionic structure with a solid middle section and thin wall in other parts. Afterwards, size optimization is maintained for the bionic design model to obtain the optimal model. The simulation results show that the three-way eigenfrequencies of bionic optimized structure are 320 Hz, 303 Hz, and 765 Hz, respectively, which are closely approximate to the original structure. However, the mass of bionic optimized structure is reduced by 23%. Benefiting from Selective laser melting, the complex optimized design can be rapidly manufactured. The three-way eigenfrequencies of the optimized structure measured by the 0.2 g sweep tests are 307 Hz, 292 Hz, and 736 Hz, respectively. The vibration test of bionic optimized structure verifies the accuracy of the simulation results. This study indicates that the combination of structural bionics and structural optimization provides a powerful tool kit to the design of similar support structure for space applications.