An official website of the United States government
Abstract The prediction of large fluctuations in the ground magnetic field (dB/dt) is essential for preventing damage from Geomagnetically Induced Currents. Directly forecasting these fluctuations has proven difficult, but accurately determining the risk of extreme events can allow for the worst of the damage to be prevented. Here we trained Convolutional Neural Network models for eight mid‐latitude magnetometers to predict the probability thatdB/dtwill exceed the 99th percentile threshold 30–60 min in the future. Two model frameworks were compared, a model trained using solar wind data from the Advanced Composition Explorer (ACE) satellite, and another model trained on both ACE and SuperMAG ground magnetometer data. The models were compared to examine if the addition of current ground magnetometer data significantly improved the forecasts ofdB/dtin the future prediction window. A bootstrapping method was employed using a random split of the training and validation data to provide a measure of uncertainty in model predictions. The models were evaluated on the ground truth data during eight geomagnetic storms and a suite of evaluation metrics are presented. The models were also compared to a persistence model to ensure that the model using both datasets did not over‐rely ondB/dtvalues in making its predictions. Overall, we find that the models using both the solar wind and ground magnetometer data had better metric scores than the solar wind only and persistence models, and was able to capture more spatially localized variations in thedB/dtthreshold crossings.
more »
« less
This content will become publicly available on August 1, 2026
Using Machine Learning Explainability Techniques to Examine Drivers of Ground Magnetic Field Localization
Abstract Solar wind particles interact with the Earth's magnetic field and can cause rapid changes in the magnetic field on the ground. This can result in Geomagnetically Induced Currents capable of causing significant damage to infrastructure, making it vital to predict when and where the fluctuations will occur so the impact can be limited. The fluctuations can occur on both a large and highly localized scale, further complicating precise predictions. Machine learning (ML) techniques have emerged as an effective method of predicting space weather phenomena, with their largest complication being their lack of explainability. Here we seek to use such ML methods, combined with a model explainability technique called SHapley Additive exPlanation to both predict and times of extreme localization. Using L1 solar wind data and magnetometer data from SuperMAG, we train two different types of models, one predicting extreme and one predicting large Region‐to‐Specific Difference (RSD). We are seeking to forecast the maximum of RSD and within a rolling 60‐min window, beginning 30 min in the future. The models perform well across a variety of latitudes and Magnetic Local times. While traditional drivers of space weather ( and ) are important drivers of the ML models, other not often examined parameters (particularly ) exhibit non‐uniform spatial and latitudinal dependencies which cannot be attributed to correlation with more influential parameters. Additionally, the inertia of the internal geomagnetic field on a regional scale exhibits a more nuanced behavior compared to previous studies on individual magnetometer stations.
more »
« less
- PAR ID:
- 10657845
- Publisher / Repository:
- Wiley
- Date Published:
- Journal Name:
- Space Weather
- Volume:
- 23
- Issue:
- 8
- ISSN:
- 1542-7390
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract Extreme (>20 nT/s) geomagnetic disturbances (GMDs, also denoted as MPEs—magnetic perturbation events)—impulsive nighttime disturbances with time scale ∼5–10 min, have sufficient amplitude to cause bursts of geomagnetically induced currents (GICs) that can damage technical infrastructure. In this study, we present occurrence statistics for extreme GMD events from five stations in the MACCS and AUTUMNX magnetometer arrays in Arctic Canada at magnetic latitudes ranging from 65° to 75°. We report all large (≥6 nT/s) and extreme GMDs from these stations from 2011 through 2022 to analyze variations of GMD activity over a full solar cycle and compare them to those found in three earlier studies. GMD activity between 2011 and 2022 did not closely follow the sunspot cycle, but instead was lowest during its rising phase and maximum (2011–2014) and highest during the early declining phase (2015–2017). Most of these GMDs, especially the most extreme, were associated with high‐speed solar wind streams (Vsw >600 km/s) and steady solar wind pressure. All extreme GMDs occurred within 80 min after substorm onsets, but few within 5 min. Multistation data often revealed a poleward progression of GMDs, consistent with a tailward retreat of the magnetotail reconnection region. These observations indicate that extreme GIC hazard conditions can occur for a variety of solar wind drivers and geomagnetic conditions, not only for fast‐coronal mass ejection driven storms.more » « less
-
Geomagnetically induced currents (GICs) pose a significant space weather hazard, driven by geomagnetic field variation due to the coupling of the solar wind to the magnetosphere-ionosphere system. Extensive research has been dedicated to understanding ground-level geomagnetic field perturbations as a GIC proxy. Still, the non-uniform aspect of geomagnetic fluctuations make it difficult to fully characterize the ground-level magnetic field across large regions of the globe. Here, we focus on localized geomagnetic disturbances (LGMDs) in the North American region and specify the degree to which these disturbances are localized. Employing the electrodynamics-informed Spherical Elementary Current Systems (SECS) method, we spatially interpolate magnetic field perturbations between ground-based magnetometer stations. In this way, we represent the ground magnetic field as a series of heatmaps at high temporal and spatial resolution. We leverage heatmaps from storm time during solar cycle 24 to automatically identify LGMDs. We build a statistical picture of the frequency with which LGMDs occur, their scale sizes, and their latitude-longitude aspect ratios. Additionally, we use an information theory approach to quantify the dependence of these three attributes on the phase of the solar cycle. We find no clear influence of the solar cycle on any of the three attributes. We offer some avenues toward explaining why LGMDs might behave broadly the same whether they arise during solar maximum or solar minimum.more » « less
-
Abstract In the ion foreshock, hot flow anomalies (HFAs) and foreshock bubbles (FBs) are two types of foreshock transients that have the strongest fluctuations, which can disturb the magnetosphere‐ionosphere system and increase shock acceleration efficiency. They form due to interaction between the foreshock ions and solar wind discontinuities: the direction of the foreshock ion‐driven current and whether it decreases or increases the magnetic field strength behind the discontinuity determine whether the transient's formation can be promoted or suppressed. Thus, to predict the HFA and FB formation and forecast their space weather effects, it is necessary to predict the foreshock ion‐driven current direction. In this study, we derive analytical equations of foreshock ion velocities within discontinuities to estimate foreshock ion‐driven current direction, which provides a quantitative criterion of HFA and FB formation. To validate the criterion, we use Acceleration Reconnection Turbulence & Electrodynamics of Moon's Interaction with the Sun to observe pristine solar wind discontinuities and calculate discontinuity parameters. We use Magnetospheric Multiscale to observe the foreshock ion motion around the discontinuities and show that the data support our model. This study is another step toward a predictive model of HFA and FB formation so that we can forecast their space weather effects at Earth using solar wind observations at lunar orbit or L1.more » « less
-
Abstract We determine the primary modes of field‐aligned current (FAC) variability and their hemispheric asymmetry by nonlinear regression analysis of a multiyear global data set of Iridium constellation engineering‐grade magnetometer data from the Active Magnetosphere and Planetary Electrodynamics Response Experiment program. The spatial and temporal FAC variability associated with three major categories of solar wind drivers, (1) slow flow, (2) high‐speed streams (HSS), (3) transient flow related to coronal mass ejections (CMEs), and (4) a combination of these, is characterized as empirical orthogonal functions (EOFs) and their time‐varying amplitude. For the combined solar wind category, the order of the modes of variability are strengthening/weakening of (1) EOF1—all FACs; (2) EOF2—Region 2 (R2) FACs; and (3) EOF3—dayside/nightside FACs. The first two EOFs are associated with solar wind coupling; EOF3 is associated with the ecliptic components of the interplanetary magnetic field (IMF). We also find hemispheric asymmetry in FACs. Northern Hemisphere EOFs show clearer spatial features and higher correlation coefficients with solar wind drivers. The Northern Hemisphere also shows higher correlation coefficients in all seasons except winter. We find transient flow EOFs to be better correlated with solar wind drivers such as IMFBzand coupling functions, while HSS EOFs are better correlated with solar wind plasma parameters. CME‐related transient flow EOFs also show R2 FAC variabilities that are not found in other separate wind drivers. Application of the EOF analysis to the Iridium magnetometer data shows significant promise for greater understanding of geoeffectiveness of solar wind interactions with geospace.more » « less
