Space Weather in the Machine Learning Era

Camporeale, Enrico; Wing, Simon; Johnson, Jay

doi:10.1029/2018eo101897

Cited by 18 publications

(21 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The SHAP values can disentangle the effects of different interplanetary parameters. Typically, a machine learning model is often treated as a “black box” (Camporeale et al., 2018) because we do not know how a specific input leads to the current output. The interpretable machine learning procedure can help us to unravel the complex dependence, transparentize the relations in the machine learning model, and make it intelligible to the human brain.…”

Section: Discussionmentioning

confidence: 99%

“…We construct a deep neural network (DNN) magnetopause model that uses the significant parameters recommended by the interpretable procedure as inputs. The machine learning model is often treated as a “black box” (Camporeale et al., 2018). However, in this study, the machine learning model is treated as a “white box” since the interpretable procedure introduces interpretability.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

An Interpretable Machine Learning Procedure Which Unravels Hidden Interplanetary Drivers of the Low Latitude Dayside Magnetopause

Sun

Chen

2023

Space Weather

View full text Add to dashboard Cite

In this study, we propose an interpretable machine learning procedure to unravel the importance of multiple interplanetary parameters to the Earth's magnetopause standoff distance (MSD). We construct the interpretable procedure based on SHapley Additive exPlanations. A magnetopause crossings database from the Time History of Events and Macroscale Interactions during Substorms satellites and the multiple interplanetary parameters from OMNI during the period of 2007–2016 are utilized. The solar wind dynamic pressure and the interplanetary magnetic field (IMF) BZ are widely suggested as the important two interplanetary parameters that drive the MSD. However, the examination of the interpretable procedure suggests that the magnitude of the IMF is the second significant parameter after the dynamic pressure. Although the magnetic pressure, which is the function of the IMF magnitude was considered in previous studies, the importance of the IMF magnitude was underestimated. The interpretable procedure also reveals that the IMF magnitude and the BZ have different effects on the MSD. Their joint effect is the formation of the MSD sag near BZ = 5 nT. This is for the first time the interpretable concept is being applied to construct a machine‐learning magnetopause model.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

An Interpretable Machine Learning Procedure Which Unravels Hidden Interplanetary Drivers of the Low Latitude Dayside Magnetopause

Sun

Chen

2023

Space Weather

View full text Add to dashboard Cite

show abstract

“…Current models of the ionosphere can be categorized as physical, empirical, or mathematical (Farzaneh & Forootan, 2018). State-of-the-art methods utilize artificial intelligence, specifically machine learning techniques, to identify nonlinear relationships among the variables to improve forecasting, especially factors related to space-weather processes (Camporeale et al, 2018;Natras & Schmidt, 2021). Physical ionosphere models are based on physical and chemical processes in the ionosphere, as shown by the Global Assimilation of Ionospheric Measurements (GAIM) model (Schunk et al, 2004) and the Global Ionosphere-Thermosphere Model (GITM) (Ridley et al, 2006).…”

Section: Introductionmentioning

confidence: 99%

Regional Ionosphere Delay Models Based on CORS Data and Machine Learning

Natraš

Goss

Halilovic

et al. 2023

navi

View full text Add to dashboard Cite

The ionospheric refraction of GNSS signals can have an impact on positioning accuracy, especially in cases of single-frequency observations. Ionosphere models that are broadcasted by the satellite systems (e.g., Klobuchar, NeQuick-G) do not include enough details to permit them to correct single-frequency observations with sufficient accuracy. To address this issue, regional ionosphere models (RIMs) have been developed in several countries in the western Balkans based on dense Continuous Operating Reference Stations (CORS) observations. Subsequently, a RIM for the western Balkans was built using an artificial neural network that combined regional ionosphere parameters estimated from the CORS data with spatiotemporal (latitude, longitude, hour of day), solar (F10.7) and geomagnetic (Kp, Dst) parameters. The RIMs were tested at the solar maximum (March 2014), a geomagnetic storm (March 2015), and the solar minimum (March 2018). The new RIMs mimic the integrated electron density much more effectively than the Klobuchar model. Furthermore, RIMs significantly reduce the ionospheric effects on single-frequency positioning, indicating their necessity for use in positioning applications.

show abstract

“…That indicates the SIR could convey a strong magnetic field that disturbs the Earth's magnetic field and can potentially cause malfunctions in the communication systems, electric power transmission systems, and GPS‐based navigation systems (Lucas et al., 2018; Thomson et al., 2005). Moreover, SIRs increase the high energetic particles in the space environment around the Earth which in turn raises the risk of ionizing the insulating layers of satellite's components and causes single event effects (Alielden, 2020; Camporeale et al., 2018; Koons & Fennell, 2006). In many cases, software and duplicate circuits are used to correct these effects, but such high‐speed solar wind (HSSW) causes a challenging environment with many malfunctions in a short period of time (Horne et al., 2013).…”

Section: Introductionmentioning

confidence: 99%

Prediction Interval of Interface Regions: Machine Learning Nowcasting Approach

et al. 2023

View full text Add to dashboard Cite

Stream interaction regions (SIRs) form in interplanetary space where the fast solar wind stream overtakes the preceding slow solar wind stream (Allen et al., 2021;Pizzo, 1980). The fast wind piles up the upstream density and forms a compressed plasma along the leading edge of the stream interface region (IR), which is typically identified as the peak in total pressure between the two solar wind populations, and is followed by a rarefaction region in the fast solar wind (Richardson, 2018). The hydrodynamic models of solar wind describe that as the fast streams flow away from the Sun, the solar wind velocity transits steeply from slow to fast and the pile-up density of the IR upstream grows (Alielden & Taroyan, 2022;Hundhausen, 1973). The fast solar wind is from coronal holes and meets the slow wind at a far distance from the Sun. If the fast wind travels significantly faster than the slow wind, a shock wave forms in the interface region which in turn leads to a discontinuity of the total pressure and an increase of the plasma speed upstream of the interface region. Sometimes two pairs of shock waves

show abstract

Space Weather in the Machine Learning Era

Abstract: Space Weather: A Multi-disciplinary Approach; Leiden, Netherlands, 25–29 September 2017

Cited by 18 publications

References 0 publications

An Interpretable Machine Learning Procedure Which Unravels Hidden Interplanetary Drivers of the Low Latitude Dayside Magnetopause

An Interpretable Machine Learning Procedure Which Unravels Hidden Interplanetary Drivers of the Low Latitude Dayside Magnetopause

Regional Ionosphere Delay Models Based on CORS Data and Machine Learning

Prediction Interval of Interface Regions: Machine Learning Nowcasting Approach

Contact Info

Product

Resources

About