Multipoint observation of the response of the magnetosphere and ionosphere related to the sudden impulse event on 19 November 2007

Segarra, Antoni; Nosé, M.; Curto, J. J.; Araki, Tohru

doi:10.1051/swsc/2015016

Cited by 3 publications

(3 citation statements)

References 37 publications

(40 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An additional constraint of the identification of the sudden commencement is that it happens quasi-simultaneously (<1 min) at all latitudes (e.g. Engebretson et al, 1999;Segarra et al, 2015). In our study we have used magnetometer stations at Abisko (ABK, 68 • N), Lerwick (LER, 60 • N), Fürstenfeldbruck (FUR, 48 • N), Bangui (BNG, 4 • N), Ascension Island (ASC, 8 • S), and Mawson (MAW, 67 • S) which are part of the INTERMAGNET consortium (Love and Chulliat, 2013).…”

Section: Delay Measurement and Databasementioning

confidence: 99%

Timing of the solar wind propagation delay between L1 and Earth based on machine learning

Baumann

McCloskey

2021

J. Space Weather Space Clim.

View full text Add to dashboard Cite

Erroneous GNSS positioning, failures in spacecraft operations and power outages due to geomagnetically induced currents are severe threats originating from space weather. Having knowledge of potential impacts on modern society in advance is key for many end-user applications. This covers not only the timing of severe geomagnetic storms but also predictions of substorm onsets at polar latitudes. In this study we aim at contributing to the timing problem of space weather impacts and propose a new method to predict the solar wind propagation delay between Lagrangian point L1 and the Earth based on machine learning, specifically decision tree models. The propagation delay is measured from the identification of interplanetary discontinuities detected by the Advanced Composition Explorer (ACE) and their subsequent sudden commencements in the magnetosphere recorded by ground-based magnetometers. A database of the propagation delay has been constructed on this principle including 380 interplanetary shocks with data ranging from 1998 to 2018. The feature set of the machine learning approach consists of six features, namely the three components of each the solar wind speed and position of ACE around L1. The performance assessment of the machine learning model is examined on the basis of a 10-fold cross-validation. The machine learning results are compared to physics-based models, i.e., the flat propagation delay and the more sophisticated method based on the normal vector of solar wind discontinuities (vector delay). After hyperparameter optimization, the trained gradient boosting (GB) model is the best machine learning model among the tested ones. The GB model achieves an RMSE of 4.5 min with respect to the measured solar wind propagation delay and also outperforms the physical flat and vector delay models by 50 % and 15 % respectively. To increase the confidence in the predictions of the trained GB model, we perform a operational validation, provide drop-column feature importance and analyse the feature impact on the model output with Shapley values. The major advantage of the machine learning approach is its simplicity when it comes to its application. After training, values for the solar wind speed and spacecraft position from only one datapoint have to be fed into the algorithm for a good prediction.

show abstract

Section: Delay Measurement and Databasementioning

confidence: 99%

Timing of the solar wind propagation delay between L1 and Earth based on machine learning

Baumann

McCloskey

2021

J. Space Weather Space Clim.

View full text Add to dashboard Cite

show abstract

“…From the vast number of solar windmagnetosphere studies the commonly used parameters for predictions are solar wind magnetic field component B z , or both B y and B z , particle density n, and speed V, or expressed as the dynamic pressure p ∝ nV 2 and the convective electric field component E y ≈ ÀVB z . The dynamic pressure plays an important role in the creation of the global geomagnetic disturbances known as sudden impulses (SI) (Segarra et al, 2015), which becomes classified as a SSC if there is a subsequent storm, while the solar wind electric field E y is generally considered to control the reconnection rate (Pulkkinen et al, 2007), although it has been suggested that it is not the solar wind electric field that controls the reconnection rate but it works due to a coincidence (Borovsky and Birn, 2014). As the neural network can approximate continuous functions arbitrarily well (Cybenko, 1989) the individual solar wind parameters can be provided as inputs from which the relations are found.…”

Section: Algorithmsmentioning

confidence: 99%

Forecasting Kp from solar wind data: input parameter study using 3-hour averages and 3-hour range values

Wintoft

Wik

Matzka

et al. 2017

J. Space Weather Space Clim.

View full text Add to dashboard Cite

-We have developed neural network models that predict Kp from upstream solar wind data. We study the importance of various input parameters, starting with the magnetic component B z , particle density n, and velocity Vand then adding total field B and the B y component. As we also notice a seasonal and UT variation in average Kp we include functions of day-of-year and UT. Finally, as Kp is a global representation of the maximum range of geomagnetic variation over 3-hour UT intervals we conclude that sudden changes in the solar wind can have a big effect on Kp, even though it is a 3-hour value. Therefore, 3-hour solar wind averages will not always appropriately represent the solar wind condition, and we introduce 3-hour maxima and minima values to some degree address this problem. We find that introducing total field B and 3-hour maxima and minima, derived from 1-minute solar wind data, have a great influence on the performance. Due to the low number of samples for high Kp values there can be considerable variation in predicted Kp for different networks with similar validation errors. We address this issue by using an ensemble of networks from which we use the median predicted Kp. The models (ensemble of networks) provide prediction lead times in the range 20-90 min given by the time it takes a solar wind structure to travel from L1 to Earth. Two models are implemented that can be run with real time data: (1) IRF-Kp-2017-h3 uses the 3-hour averages of the solar wind data and (2) IRF-Kp-2017 uses in addition to the averages, also the minima and maxima values. The IRF-Kp-2017 model has RMS error of 0.55 and linear correlation of 0.92 based on an independent test set with final Kp covering 2 years using ACE Level 2 data. The IRF-Kp-2017-h3 model has RMSE = 0.63 and correlation = 0.89. We also explore the errors when tested on another two-year period with real-time ACE data which gives RMSE = 0.59 for IRF-Kp-2017 and RMSE = 0.73 for IRF-Kp-2017-h3. The errors as function of Kp and for different years are also studied.

show abstract

“…The interaction of a shock with the magnetosphere could have global effects, leading to magnetic/electric field and plasma perturbations in the magnetosphere and ionosphere. Sudden magnetic field changes have been investigated using multisatellite observations and numerical MHD simulations [e.g., Samsonov et al, 2014;Segarra et al, 2015]; transient electric field has also been observed in the dayside magnetosphere which was directed westward [Shinbori et al, 2004]. In accordance, plasma motion has been reported to be earthward around the geosynchronous orbit, based on in situ observations of 1.2 keV electron [Baumjohann et al, 1983] and ions with energies from a few eV/e to tens of keV/e [Keika et al, 2008;Zhang et al, 2012].…”

Section: Introductionmentioning

confidence: 99%

Prompt GPS TEC response to magnetospheric compression

et al. 2017

View full text Add to dashboard Cite

A new type of total electron content (TEC) variation was observed when an interplanetary shock impacted on the Earth's magnetosphere on 17 March 2015. With hundreds of ground‐based Global Positioning System (GPS) receivers, instantaneous TEC impulses were detected in the signals from four GPS satellites cruising in the dayside equatorial magnetosphere. Despite the small amplitude (0.3 TECU), the impulses were synchronously registered by receivers spreading from low to middle latitudes on the ground; they lasted for several minutes, during which period the interplanetary magnetic field (IMF) kept northward. The TEC impulses therefore discriminate themselves from usual traveling ionospheric disturbances and exclude possible effects from the southward IMF‐driven magnetic storm. We suggest that the TEC variation is caused by shock‐induced magnetospheric compression, which moves plasma earthward in the dayside plasmasphere. As a result, some plasma outside of GPS orbit (4.2 RE) is moved to the inside and contributes to the plasma content traversed by GPS raypath. The GPS TEC technique thus exhibits an unprecedented capability to capture small tremor of the magnetosphere, and with the dense receiver network on the ground it can be a feasible tool for remote sensing of the plasma dynamics around 4.2 RE.

show abstract

Multipoint observation of the response of the magnetosphere and ionosphere related to the sudden impulse event on 19 November 2007

Cited by 3 publications

References 37 publications

Timing of the solar wind propagation delay between L1 and Earth based on machine learning

Timing of the solar wind propagation delay between L1 and Earth based on machine learning

Forecasting Kp from solar wind data: input parameter study using 3-hour averages and 3-hour range values

Prompt GPS TEC response to magnetospheric compression

Contact Info

Product

Resources

About