The ionospheric disturbances induced by the Falcon 9 rocket launch on 17 January 2016 were reconstructed by three-dimensional computerized ionospheric tomography using observations from the North American global navigation satellite system networks. The results showed that concentric traveling ionospheric disturbances (CTIDs) occurred ∼18 min after the rocket launch and were remarkable at 200–300 km altitudes. The vertical phase velocities of the CTIDs were consistent with the inclinations of the U-shaped structures. At specific azimuth directions of 350°, 30°, and 105°, the estimated vertical phase velocities between 100 and 200 km altitudes were ∼222.2, ∼208.3, and ∼242.4 m/s, respectively. When the CTIDs propagated upward to 400–500 km altitudes, their vertical velocities increased to ∼566.7, ∼966.7, and ∼944.4 m/s. CTIDs traveling northward (azimuths 350°, 30°) had periods of ∼11 min. At 200 and 300 km altitudes, their horizontal phase velocities were ∼309.9–323.3 and ∼309.4–330.9 m/s, respectively, with horizontal wavelengths of ∼204.5–213.4 and ∼204.2–218.4 km. In contrast, CTIDs propagating eastward (azimuth 105°) displayed a period of ∼15 min. At 200 and 300 km altitudes, their horizontal phase velocities were ∼223.2 and ∼241.1 m/s, respectively, with horizontal wavelengths of ∼200.9 and ∼217.0 km. These CTIDs propagation characteristics agreed well with the theory of gravity waves.

The spatiotemporal variations and mechanisms of the ionospheric storms over the American sector during the August 2018 geomagnetic storm are investigated. One positive and one negative ionospheric storm occurred in North America and two positive storms were observed in South America. The ionosphere showed prominent hemispheric asymmetries during the four storms. The maximum VTEC (vertical total electron content) variation was more than 15 TECU during the positive storms and about −10 during the negative storm. The GUVI (Global Ultraviolet Imager) oxygen (O) to nitrogen (N₂) column density ratio (∑O/N₂) and SuperDARN (Super Dual Auroral Radar Network) polar cap potential results showed that the electric field variations played a decisive role in generating the North American negative storm while the thermspheric composition changes were responsible for the North American positive storm and the two South America positive storms.

The eruption of Tonga submarine volcano on 15 January 2022 caused significant ionospheric oscillations, and the ionospheric disturbance exhibited rare interhemispheric coupling in the magnetic conjugate regions. The concentric traveling ionospheric disturbances (CTIDs) with different frequencies were detected using GNSS observations from Japan and Australia. The three-dimensional computerized ionospheric tomography (3DCIT) was used to investigate the four-dimensional evolution of the CTIDs caused by Tonga volcanic eruption. The VTEC spectrograms indicate that the disturbances over Japan occurred around 08:30 UT, consistent with those obtained over Australia, but with a smaller magnitude. The 3DCIT results show that the upside-down cone-shaped CTIDs propagated to altitudes exceeding 500 km. The horizontal phase velocities (~357.1–374.7 m/s) of the conjugate CTIDs over Japan were comparable to those over Australia, with a period of ~42 minutes. They propagated faster at higher altitudes and showed consistency between opposite hemispheres. The vertical phase velocity of the CTIDs can vary from ~100 m/s to beyond 600 m/s (downward) with altitude increasing.

Detecting ionospheric anomalies constitutes a key research focus in the field of early seismic warnings and responses. This study investigates pre- and post-seismic ionospheric anomalies in relation to strong earthquakes, utilizing Global Ionospheric Map (GIM) data and applying three detection methods: the interquartile range (IQR), first-order and second-order numerical difference (FSND), and butterworth bandpass filter (BBF). The study focuses on two cases of strong earthquakes (inland and offshore earthquakes): the Mw 7.1 earthquake in Aksu, Xinjiang, China (January 23, 2024), and the Mw 7.4 earthquake off the west coast of Honshu, Japan (January 1, 2024). The ionospheric total electron content (TEC) disturbances in these events were detected and analyzed, and the detection performances of the three methods were compared. The results show that all three methods effectively capture ionospheric TEC disturbances and provide valuable insights into earthquake precursors and epicentral anomalies. The IQR method detected pre-seismic disturbances for both inland and offshore earthquakes. The FSND method consistently identifies “N”-type disturbances in the Differential TEC (DTEC) for both events. The FSDN method can process long-term TEC time series, whereas the BBF method is restricted to smaller datasets. In practice, the BBF experiment examined TEC data on a daily basis to identify ionospheric dis-turbances.The BBF method also highlights “N”-type disturbances by employing 30-minute TEC data, unlike the other two methods that used 1-hour sampled data. These findings offer theoretical insights into ionospheric responses to earthquakes and have practical implications for earthquake prediction.

Hyperspectral image classification (HSI) faces challenges from high-dimensional redundancy and spatial heterogeneity. Existing methods struggle with computational inefficiency in classifier selection due to diverse architectures and hyperparameters. We propose DEDES, a dynamic ensemble framework that addresses this via spectral-spatial embedding and density-based clustering to partition features into homogeneous regions, enabling localized classifier activation. Experiments on Botswana and Purdue datasets demonstrate state-of-the-art accuracy (overall accuracy of 86.25% and 94.40%) compared to conventional methods. The framework eliminates exhaustive aggregation, demonstrating optimal accuracy through adaptive minimal classifier selection, offering a scalable solution for real-world HSI interpretation.