ESSI1.5 | Machine Learning in Planetary Sciences and Heliophysics
EDI
Machine Learning in Planetary Sciences and Heliophysics
Co-organized by PS4/ST1
Convener: Hannah Theresa Rüdisser | Co-conveners: Justin Le Louëdec, Ute Amerstorfer, Simon Bouriat
Orals
| Mon, 15 Apr, 16:15–18:00 (CEST)
 
Room -2.16
Posters on site
| Attendance Mon, 15 Apr, 10:45–12:30 (CEST) | Display Mon, 15 Apr, 08:30–12:30
 
Hall X4
Orals |
Mon, 16:15
Mon, 10:45
The recent growing number of probes in the heliosphere and future missions in preparation led to the current decade being labelled as "the golden age of heliophysics research". With more viewpoints and data downstreamed to Earth, machine learning (ML) has become a precious tool for planetary and heliospheric research to process the increasing amount of data and help the discovery and modelisation of physical systems. Recent years have also seen the development of novel approaches leveraging complex data representations with highly parameterised machine learning models and combining them with well-defined and understood physical models. These advancements in ML with physical insights or physically informed neural networks inspire new questions about how each field can respectively help develop the other. To better understand this intersection between data-driven learning approaches and physical models in planetary sciences and heliophysics, we seek to bring ML researchers and physical scientists together as part of this session and stimulate the interaction of both fields by presenting state-of-the-art approaches and cross-disciplinary visions of the field.

The "ML for Planetary Sciences and Heliophysics" session aims to provide an inclusive and cutting-edge space for discussions and exchanges at the intersection of machine learning, planetary and heliophysics topics. This space covers (1) the application of machine learning/deep learning to space research, (2) novel datasets and statistical data analysis methods over large data corpora, and (3) new approaches combining learning-based with physics-based to bring an understanding of the new AI-powered science and the resulting advancements in heliophysics research.
Topics of interest include all aspects of ML and heliophysics, including, but not limited to: space weather forecasting, computer vision systems applied to space data, time-series analysis of dynamical systems, new machine learning models and data-assimilation techniques, and physically informed models.

Orals: Mon, 15 Apr | Room -2.16

Chairpersons: Hannah Theresa Rüdisser, Justin Le Louëdec, Ute Amerstorfer
16:15–16:20
16:20–16:40
|
EGU24-6558
|
ECS
|
solicited
|
On-site presentation
Andong Hu and Enrico Camporeale

We present an innovative method, ProBoost (Probabilistic Boosting), for forecasting extreme space weather events using ensemble machine learning (ML). Ensembles enhance prediction accuracy, but applying them to ML faces challenges as ML models often lack wellcalibrated uncertainty estimates. Moreover, space weather problems are typically affected by very imbalanced datasets (i.e., extreme and rare events) To overcome these difficulties, we developed a method that incorporates uncertainty quantification (UQ) in neural networks, enabling simultaneous forecasting of prediction uncertainty.
Our study applies ProBoost to the following space weather applications:
• One-to-Six-Hour Lead-Time Model: Predicting Disturbance Storm Time (Dst) values using solar wind data.
• Two-Day Lead-Time Model: Forecasting Dst probability using solar images.
• Geoelectric Field Model: Multi-hour lead time, incorporating solar wind and SuperMag data.
• Ambient Solar Wind Velocity Forecast: Up to 5 days ahead.
ProBoost is model-agnostic, making it adaptable to various forecasting applications beyond space weather.

How to cite: Hu, A. and Camporeale, E.: A New Machine Learning Approach for Predicting Extreme Space Weather, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6558, https://doi.org/10.5194/egusphere-egu24-6558, 2024.

16:40–16:50
|
EGU24-1604
|
ECS
|
On-site presentation
|
Juan Esteban Agudelo Ortiz, Germain Nicolás Morales Suarez, Santiago Vargas Domínguez, and Sergiy Shelyag

The arrival of new and more powerful spectropolarimetric instruments such as DKIST, the development of better magnetohydrodinamic (MHD) simulation codes and the creation of newly inversion methods, are coming with the demands of increasing amounts of computational time and power. This, with increasing generation of data, will come with even years of processing that will stop the advance of scientific investigations on mid-late stages. The arrival of Machine Learning models able to replicate patterns in data come with the possibilites of them to adapt to different types of datasets, such as those for classification or for creation of sequences like the seq2seq models, that once trained, they are able to give results according to previous methods that differ on order of magnitude in time processing, being a lot faster. Some work has been done within this field for creating machine learning inversion methods using data obtained from actual inversion codes applied on observational data, and using data from radiative transfer codes for synthesis, reducing both computational demands and time processing. This work attempts to follow onto this steps, using in this case datasets obtained from simulation codes like MURaM and their correspondent Stokes parameters obtained from non-lte radiative transfer codes like NICOLE, training forward (synthesis) and backward (inversion) some neural network models to test whether or not they can learn their physical behaviours and at what accuracy, for being used in the future to process actual data obtained from newly simulation codes and for real solar observations, being another step into the future for creating a new paradigm on how to invert and sunthesize quantities in Physics in general.

How to cite: Agudelo Ortiz, J. E., Morales Suarez, G. N., Vargas Domínguez, S., and Shelyag, S.: Machine Learning Synthesis and inversion method for Stokes Parameters in the solar context, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1604, https://doi.org/10.5194/egusphere-egu24-1604, 2024.

16:50–17:00
|
EGU24-15813
|
ECS
|
On-site presentation
Christoph Schirninger, Astrid Veronig, Robert Jarolim, J. Emmanuel Johnson, Anna Jungbluth, Richard Galvez, Lilli Freischem, and Anne Spalding

Various instruments are used to study the Sun, including ground-based observatories and space telescopes. These data products are constantly changing due to technological improvements, different instrumentation, or atmospheric effects. However, for certain applications such as ground-based solar image reconstruction or solar cycle studies, enhanced and combined data products are necessary.

We present a general AI tool called Instrument-to-Instrument (ITI; Jarolim et al. 2023) translation, which is capable of translating datasets between two different image domains. This approach enables instrument intercalibration, image enhancement, mitigation of quality degradations, and super-resolution across multiple wavelength bands. The tool is built on unpaired image-to-image translation, which enables a wide range of applications, where no spatial or temporal overlap is required between the considered datasets.

In this presentation, we highlight ITI as a general tool for Heliospheric applications and demonstrate its capabilities by applying it to data from Solar Orbiter/EUI, PROBA2/SWAP, and the Solar Dynamics Observatory/AIA in order to achieve a homogenous, machine-learning ready dataset that combines three different EUV imagers. 

The direct comparison of aligned observations shows the close relation of ITI-enhanced and real high-quality observations. The evaluation of light-curves demonstrates an improved inter-calibration.

ITI is provided open-source to the community  and can be easily applied to novel datasets and various research applications. 

This research is funded through a NASA 22-MDRAIT22-0018 award (No 80NSSC23K1045) and managed by Trillium Technologies, Inc (trillium.tech)

How to cite: Schirninger, C., Veronig, A., Jarolim, R., Johnson, J. E., Jungbluth, A., Galvez, R., Freischem, L., and Spalding, A.: Instrument-to-Instrument translation: An AI tool to intercalibrate, enhance and super-resolve solar observations, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15813, https://doi.org/10.5194/egusphere-egu24-15813, 2024.

17:00–17:10
|
EGU24-420
|
ECS
|
On-site presentation
Filipa Barros, João José Graça Lima, Rui F. Pinto, and André Restivo

In previous work, an Artificial Neural Network (ANN) was developed to automate the estimation of solar wind profiles used as initial conditions in MULTI-VP simulations. This approach, coupled with profile clustering, reduced the time previously required for estimation by MULTI-VP, enhancing the efficiency of the simulation process. It was observed that generating initial estimates closer to the final simulation led to reduced computation time, with a mean speedup of 1.13. Additionally, this adjustment yielded a twofold advantage: it minimized the amplitude of spurious transients, reinforcing the numerical stability of calculations and enabling the code to maintain a more moderate integration time step.

However, upon further analysis, it became evident that the physical model inherently required a relaxation time for the final solution to stabilize. Therefore, while refining initial conditions offered improvements, there was a limit to how much it could accelerate the process. Consequently, attention turned towards the development of a surrogate model focused on the upper corona (from 3 solar radii to 30 solar radii). This range was chosen because the model can avoid learning the initial phases of wind acceleration, which are hard to accurately predict. Moreover, in order to connect the model to heliospheric models and for space weather applications, more than 3 radii is more than sufficient and guarantees that the physics remain consistent within the reproducible domain.

This surrogate model aims at delivering faster forecasts, with MULTI-VP running in parallel (eventually refining the solutions). The surrogate model for MULTI-VP was tested using a heliospheric model and data from spacecraft at L1, validating its efficacy beyond Mean Squared Error (MSE) evaluations and ensuring physical conservation principles were upheld.

This work aims at simplifying and accelerating the process of establishing boundary conditions for heliospheric models without dismissing the physical models for both extreme events and for more physically accurate results. 

How to cite: Barros, F., Lima, J. J. G., F. Pinto, R., and Restivo, A.: Neural Networks for Surrogate Models of the Corona and Solar Wind, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-420, https://doi.org/10.5194/egusphere-egu24-420, 2024.

17:10–17:20
|
EGU24-4471
|
ECS
|
On-site presentation
Federico Sabbatini and Catia Grimani

The unprecedented predictive capabilities of machine learning models make them inestimable tools to perform data forecasting and other complex tasks. Benefits of these predictors are even more precious when there is the necessity of surrogating unavailable data due to the lack of dedicated instrumentation on board space missions. For instance, the future ESA space interferometer LISA for low-frequency gravitational wave detection will host, as part of its diagnostics subsystem, particle detectors to measure the galactic cosmic-ray flux and magnetometers to monitor the magnetic field intensity in the region of the interferometer mirrors. No instrumentation dedicated to the interplanetary medium parameter monitoring will be placed on the three spacecraft constituting the LISA constellation. However, important lessons about the correlation between galactic cosmic-ray flux short-term variations and the solar wind speed profile have been learned with the ESA LISA precursor mission, LISA Pathfinder, orbiting around the L1 Lagrange point. In a previous work, we have demonstrated that for LISA Pathfinder it was possible to reconstruct with an uncertainty of 2 nT the interplanetary magnetic field intensity for interplanetary structure transit monitoring. Machine learning models are proposed here to infer the solar wind speed that is not measured on the three LISA spacecraft from galactic cosmic-ray measurements. This work is precious and necessary since LISA, scheduled to launch in 2035, will trail Earth on the ecliptic at 50 million km distance, too far from the orbits of other space missions dedicated to the interplanetary medium monitoring to benefit of their observations.

We built an interpretable machine learning predictor based on galactic cosmic-ray and interplanetary magnetic field observations to obtain a solar wind speed reconstruction within ±65 km s-1 of uncertainty. Interpretability is achieved by applying the CReEPy symbolic knowledge extractor to the outcomes of a k-NN regressor. The extracted knowledge consists of linear equations aimed at describing the solar wind speed in terms of four statistical indices calculated for the input variables.

Details about the model workflow, performance and validation will be presented at the conference, together with the advantages, drawbacks and possible future enhancements, to demonstrate that our model may provide the LISA mission with an effective and human-interpretable tool to carry out reliable solar wind speed estimates and recognise the transit of interplanetary structures nearby the LISA spacecraft, as a support to the data analysis activity for the monitoring of the external forces acting on the spectrometer mirrors.

How to cite: Sabbatini, F. and Grimani, C.: Solar Wind Speed Estimation via Symbolic Knowledge Extraction from Opaque Models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4471, https://doi.org/10.5194/egusphere-egu24-4471, 2024.

17:20–17:30
|
EGU24-6899
|
On-site presentation
Enrico Camporeale

We use the framework of Physics-Informed Neural Network (PINN) to solve the inverse problem associated with the Fokker-Planck equation for radiation belts' electron transport, using 4 years of Van Allen Probes data. Traditionally, reduced models have employed a diffusion equation based on the quasilinear approximation. We show that the dynamics of “killer electrons” is described more accurately by a drift-diffusion equation, and that drift is as important as diffusion for nearly-equatorially trapped ∼1 MeV electrons in the inner part of the belt. Moreover, we present a recipe for gleaning physical insight from solving the ill-posed inverse problem of inferring model coefficients from data using PINNs. Furthermore, we derive a parameterization for the diffusion and drift coefficients as a function of L only, which is both simpler and more accurate than earlier models. Finally, we use the PINN technique to develop an automatic event identification method that allows identifying times at which the radial transport assumption is inadequate to describe all the physics of interest.

How to cite: Camporeale, E.: Data-Driven Discovery of Fokker-Planck Equation for the Earth's Radiation Belts Electrons Using Physics-Informed Neural Networks, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6899, https://doi.org/10.5194/egusphere-egu24-6899, 2024.

17:30–17:40
|
EGU24-19558
|
On-site presentation
Tom Andert, Benedikt Aigner, Fabian Dallinger, Benjamin Haser, Martin Pätzold, and Matthias Hahn

In Precise Orbit Determination (POD), employing proper methods for pre-processing tracking data is crucial not only to mitigate data noise but also to identify potential unmodeled effects that may elude the prediction model of the POD algorithm. Unaccounted effects can skew parameter estimation, causing certain parameters to assimilate the unmodeled effects and deviate from their true values. Therefore, enhancing the pre-processing of tracking data ultimately contributes to refining the prediction model.

The Rosetta spacecraft, during its two-year mission alongside comet 67P/Churyumov-Gerasimenko, collected a substantial dataset of tracking data. In addition to this data, also tracking data from the Mars Express spacecraft, orbiting Mars since 2004, will serve as a use case to assess and compare diverse data pre-processing methods. Both traditional and AI-based techniques are explored to examine the impact of various strategies on the accuracy of orbit determination. This aims to enhance POD, thereby yielding a more robust scientific outcome.

How to cite: Andert, T., Aigner, B., Dallinger, F., Haser, B., Pätzold, M., and Hahn, M.: Comparative Analysis of Data Preprocessing Methods for Precise Orbit Determination, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19558, https://doi.org/10.5194/egusphere-egu24-19558, 2024.

17:40–17:50
|
EGU24-9174
|
ECS
|
On-site presentation
Salome Gruchola, Peter Keresztes Schmidt, Marek Tulej, Andreas Riedo, Klaus Mezger, and Peter Wurz

The efficient use of the provided downlink capacity for scientific data is a fundamental aspect of space exploration. The use thereof can be optimised through sophisticated data reduction techniques and automation of processes on board that otherwise require interaction with the operations centres on Earth. Machine learning-based autonomous methods serve both purposes; yet space-based ML applications remain relatively rare compared to the application of ML on Earth to data acquired in space.

In this contribution, we present a potential application of unsupervised machine learning to cluster mass spectrometric data on-board a spacecraft. Data were acquired from a phoscorite rock [1] using a prototype of a laser ablation ionisation mass spectrometer (LIMS) for space research [2]. Two unsupervised dimensionality reduction algorithms, UMAP and densMAP [3,4], were employed to construct low-dimensional representations of the data. Clusters corresponding to different mineral phases within these embeddings were found using HDBSCAN [5]. The impact of data pre-processing and model parameter selection on the classification outcome was investigated through varying levels of pre-processing and extensive grid searches.

Both UMAP and densMAP effectively isolated major mineral phases present within the rock sample, but densMAP additionally found minor inclusions present only in a small number of mass spectra. However, densMAP exhibited higher sensitivity to data pre-processing, yielding lower scores for minimally treated data compared to UMAP. For highly processed data, both UMAP and densMAP exhibited high stability across a broad model parameter space.

Given that the data were recorded using a miniature mass spectrometric instrument designed for space flight, these methods demonstrate effective strategies for substantial reduction of data similarly to what is anticipated on future space missions. Autonomous clustering of data into groups of different chemical composition, followed by the downlink of a representative mass spectrum of each cluster, aids in identifying relevant data. Mission return can therefore be enhanced through the selective downlink of data of interest. As both UMAP and densMAP, coupled with HDBSCAN, are relatively complex algorithms compared to more traditional techniques, such as k-means, it is important to evaluate the benefits and drawbacks of using simpler methods on-board spacecraft.

 

[1] Tulej, M. et al., 2022, https://doi.org/10.3390/universe8080410.

[2] Riedo, A. et al., 2012, https://doi.org/10.1002/jms.3104.

[3] McInnes, L. et al., 2018, https://doi.org/10.48550/arXiv.1802.03426.

[4] Narayan, A., et al., 2021, https://doi.org/10.1038/s41587-020-00801-7.

[5] McInnes, L., et al., 2017, https://doi.org/10.21105/JOSS.00205.

How to cite: Gruchola, S., Keresztes Schmidt, P., Tulej, M., Riedo, A., Mezger, K., and Wurz, P.: Enhancing Space Mission Return through On-Board Data Reduction using Unsupervised Machine Learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9174, https://doi.org/10.5194/egusphere-egu24-9174, 2024.

17:50–18:00
|
EGU24-4231
|
ECS
|
On-site presentation
A Machine-learning-based Model of the Three-dimensional Ion Flux in the Earth’s Northern Cusp.
(withdrawn after no-show)
Gonzalo Cucho-Padin, David Sibeck, and Xueyi Wang

Posters on site: Mon, 15 Apr, 10:45–12:30 | Hall X4

Display time: Mon, 15 Apr, 08:30–Mon, 15 Apr, 12:30
Chairpersons: Simon Bouriat, Ute Amerstorfer, Justin Le Louëdec
X4.100
|
EGU24-21463
|
ECS
Lucas Mandl, Apostolous Christou, and Andreas Windisch

In this work we conduct a thorough examination of utilizing machine learning and computer
vision techniques for classifying meteors based on their characteristics. The focus of the re-
search is the analysis of light curves emitted by meteors as they pass through the Earth’s atmo-
sphere, including aspects such as luminosity, duration, and shape. Through extracting features
from these light curves and comparing them to established meteors orbits, valuable informa-
tion about the meteor’s origin and chemical composition is sought to be obtained. A significant
contribution of the thesis is the development of methods for classifying meteors by extracting
features from the light curve shape through the usage of unsupervised classification algorithms.
This approach allows for the automatic classification of meteors into various groups based on
their properties. Data for the research is collected by a three-camera setup at the Armagh observatory,
comprising one medium-angle camera and
two wide-angle cameras. This setup enables the capturing of detailed images of meteor light
curves, as well as various other observations such as coordinate and angular data. The research
also involves the use of machine learning algorithms for data reduction and classification tasks.
By applying these techniques to the data collected from the camera setup, the identification of
parent objects based on chemical composition and meteor path is facilitated, along with the
acquisition of other valuable information about the meteors.

How to cite: Mandl, L., Christou, A., and Windisch, A.: A machine learning approach to meteor light curve analysis, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-21463, https://doi.org/10.5194/egusphere-egu24-21463, 2024.

X4.101
|
EGU24-19248
Oleg Stepanyuk and Kamen Kozarev

Solar eruptive events are complex phenomena, which most often include coronal mass ejections (CME), flares, compressive/shock waves, and filament eruptions. CMEs are large eruptions of magnetized plasma from the Sun’s outer atmosphere or corona, that propagate outward into the interplanetary space. Solar Energetic Particles (SEP) are produced through particle acceleration in flares or CME-driven shocks. Exact mechanisms behind SEP production are yet to be understood, but it is thought that most of their acceleration occurs in shocks starting in the low corona. Over the last several decades a large amount of remote solar eruption observations have become available from ground-based and space-borne instruments. This has required the development of software approaches for automated characterization of eruptive features. Most solar feature detection and tracking algorithms currently in use have restricted applicability and complicated processing chains, while the complexities in engineering machine learning (ML) training sets limit the use of data-driven approaches for tracking or solar eruptive related phenomena. Recently, we introduced a hybrid algorithmic—data driven approach for characterization and tracking of solar eruptive features with the improved wavelet-based, multi-instrument Wavetrack package (Stepanyuk et.al, J. Space Weather Space Clim. (2024)), which was used to produce training datasets for data driven image segmentation with convolutional neural networks (CNN). Its perfomance was shown on a limited set of SDO AIA 193A instrument data perfoming segmentation of EUV and shock waves. Here we extend this approach and present an ensemble of more general CNN models for data-driven segmentation of various eruptive phenomena for the set of ground-based and remote instruments data. We discuss our approach to engineering training sets and data augmentation, CNN topology and training techniques. 

How to cite: Stepanyuk, O. and Kozarev, K.: Segmentation and Tracking of Solar Eruptive Phenomena with Convolutional Neural Networks (CNN), EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19248, https://doi.org/10.5194/egusphere-egu24-19248, 2024.

X4.102
|
EGU24-2046
|
ECS
Filip Arnaut, Aleksandra Kolarski, and Vladimir Srećković

In our previous publication (Arnaut et al. 2023), we demonstrated the application of the Random Forest (RF) algorithm for classifying disturbances associated with solar flares (SF), erroneous signals, and measurement errors in VLF amplitude data i.e., anomaly detection in VLF amplitude data. The RF algorithm is widely regarded as a preferred option for conducting research in novel domains. Its advantages, such as its ability to avoid overfitting data and its simplicity, make it particularly valuable in these situations. Nevertheless, it is imperative to conduct thorough testing and evaluation of alternative algorithms and methods to ascertain their potential advantages and enhance the overall efficiency of the method. This brief communication demonstrates the application of the XGBoost (XGB) method on the exact dataset previously used for the RF algorithm, along with a comparative analysis between the two algorithms. Given that the problem is framed as a machine learning (ML) problem with a focus on the minority class, the comparative analysis is exclusively conducted using the minority (anomalous) data class. The data pre-processing methodology can be found in Arnaut et al. (2023). The XGB tuning process involved using a grid search method to optimize the hyperparameters of the model. The number of estimators (trees) was varied from 25 to 500 in increments of 25, and the learning rate was varied from 0.02 to 0.4 in increments of 0.02. The F1-Score for the anomalous data class is similar for both models, with a value of 0.508 for the RF model and 0.51 for the XGB model. These scores were calculated using the entire test dataset, which consists of 19 transmitter-receiver pairs. Upon closer examination, it becomes evident that the RF model exhibits a higher precision metric (0.488) than the XGB model (0.37), while the XGB model demonstrates a higher recall metric (0.84) compared to the RF model (0.53). Upon examining each individual transmitter-receiver pair, it was found that XGB outperformed RF in terms of F1-Scores in 10 out of 19 cases. The most significant disparities are observed in cases where the XGB model outperformed by a margin of 0.15 in terms of F1-Score, but conversely performed worse by approximately -0.16 in another instance for the anomalous data class. The XGB models outperformed the RF model by approximately 6.72% in terms of the F1-score for the anomalous data class when averaging all the 19 transmitter-receiver pairs. When utilizing a point-based evaluation metric that assigns rewards or penalties for each entry in the confusion matrix, the RF model demonstrates an overall improvement of approximately 5% compared to the XGB model. Overall, the comparison between the RF and XGB models is ambiguous. Both models have instances where one is superior to the other. Further research is necessary to fully optimize the method, which has benefits in automatically classifying VLF amplitude anomalous signals caused by SF effects, erroneous measurements, and other factors.

References:

Arnaut, F., Kolarski, A. and Srećković, V.A., 2023. Random Forest Classification and Ionospheric Response to Solar Flares: Analysis and Validation. Universe9(10), p.436.

How to cite: Arnaut, F., Kolarski, A., and Srećković, V.: Comparative Analysis of Random Forest and XGBoost in Classifying Ionospheric Signal Disturbances During Solar Flares, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2046, https://doi.org/10.5194/egusphere-egu24-2046, 2024.

X4.103
|
EGU24-4181
Everton Frigo and Italo Gonçalves

The solar activity has various direct and indirect impacts on human activities. During periods of high solar activity, the harmful effects triggered by solar variability are maximized. On a decadal to multidecadal time scale, solar variability exhibits a main cycle of around 11 years known as the Schwabe solar cycle, leading to a solar maximum approximately every 11 years. The most commonly used variable for measuring solar activity is the sunspot number. Over the last few decades, numerous techniques have been employed to predict the time evolution of the solar cycle for subsequent years. Recently, there has been a growing number of studies utilizing machine learning methods to predict solar cycles. One such method is the Gaussian process, which is well-suited for working with small amounts of data and can also provide an uncertainty measure for predictions. In this study, the Gaussian process technique is employed to predict the sunspot number between 2024 and 2050. The dataset used to train and validate the model comprises monthly averages of sunspots relative to the period 1700-2023. According to the results, the current solar cycle, currently at its maximum, is anticipated to last until 2030. The subsequent solar maximum is projected to occur around the end of 2033, with an estimated maximum sunspot number of approximately 150. If this prediction holds true, the next solar cycle's maximum will resemble that observed in the current one.

How to cite: Frigo, E. and Gonçalves, I.: Prediction of sunspot number using Gaussian processes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4181, https://doi.org/10.5194/egusphere-egu24-4181, 2024.

X4.104
|
EGU24-12885
|
ECS
Zoe Faes, Laura Hayes, Daniel Müller, and Andrew Walsh

This study aims to identify sets of in-situ measurements of the solar wind which sample the same volume of plasma at different times and locations as it travels through the heliosphere using ensemble machine learning methods. Multiple observations of a single volume of plasma by different spacecraft - referred to here as conjunctions - are becoming more frequent in the current “golden age of heliophysics research” and are key to characterizing the expansion of the solar wind. Specifically, identifying these related observations will enable us to test the current understanding of solar wind acceleration from the corona to the inner heliosphere with a more comprehensive set of measurements than has been used in previous analyses.

Using in-situ measurements of the background solar wind from Solar Orbiter, Parker Solar Probe, STEREO-A, Wind and BepiColombo, we identify a set of criteria based on features of magnetic field, velocity, density and temperature timeseries of known conjunctions and search for other instances for which the criteria are satisfied, to find previously unknown conjunctions. We use an ensemble of models, including random forests and recurrent neural networks with long short-term memory trained on synthetic observations obtained from magnetohydrodynamic simulations, to identify candidate conjunctions solely from kinetic properties of the solar wind. Initial results show a previously unidentified set of conjunctions between the spacecraft considered in this study. While this analysis has thus far only been performed on observations obtained since 2021 (start of Solar Orbiter science operations), the methods used here can be applied to other datasets to increase the potential for scientific return of existing and future heliophysics missions.

The modular scientific software built over the course of this research includes methods for the retrieval, processing, visualisation, and analysis of observational and synthetic timeseries of solar wind properties. It also includes methods for feature engineering and integration with widely used machine learning libraries. The software is available as an open-source Python package to ensure results can be easily reproduced and to facilitate further investigation of coordinated in-situ data in heliophysics.

How to cite: Faes, Z., Hayes, L., Müller, D., and Walsh, A.: Finding Hidden Conjunctions in the Solar Wind, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12885, https://doi.org/10.5194/egusphere-egu24-12885, 2024.

X4.105
|
EGU24-14186
Sumiaya Rahman, Hyun-Jin Jeong, Ashraf Siddique, and Yong-Jae Moon

Magnetohydrodynamic (MHD) models provide a quantitative 3D distribution of the solar corona parameters (density, radial velocity, and temperature). However, this process is expensive and time-consuming. For this, we apply deep learning models to reproduce the 3D distribution of solar coronal parameters from 2D synoptic photospheric magnetic fields. We consider synoptic photospheric magnetic fields as an input to obtain 3D solar coronal parameters simulated by the MHD Algorithm outside a Sphere (MAS) from June 2010 to January 2023. Each parameter is individually trained using 150 deep learning models, corresponding to 150 solar radial distances ranging from 1 to 30 solar radii. Our study yields significant findings. Firstly, our model accurately reproduces 3D coronal parameter structures across the 1 to 30 solar radii range, demonstrating an average correlation coefficient value of approximately 0.96. Secondly, the 150 deep-learning models exhibit a remarkably shorter runtime (about 16 seconds for each parameter), with an NVIDIA Titan XP GPU, in comparison to the conventional MAS simulation time. As the MAS simulation is a regularization model, we may significantly reduce the simulation time by using our results as an initial magnetic configuration to obtain an equilibrium condition. In the future, we hope that the generated solar coronal parameters can be used for near real-time forecasting of heliospheric propagation of solar eruptions.

How to cite: Rahman, S., Jeong, H.-J., Siddique, A., and Moon, Y.-J.: Near real-time construction of Solar Coronal Parameters based on MAS simulation by Deep Learning , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14186, https://doi.org/10.5194/egusphere-egu24-14186, 2024.

X4.106
|
EGU24-18534
|
ECS
|
Mariia Drozdova, Vitaliy Kinakh, Francesco Ramunno, Erica Lastufka, and Slava Voloshynovskiy

Operating continuously for over three years, Solar Orbiter's STIX has observed more than 43 thousand X-ray flares. This study presents a compelling visualization of this publicly available database, using self-supervised learning to organize reconstructed flare images by their visual properties. Networks designed for self-supervised learning, such as Masked Siamese Networks or Autoencoders, are able to learn latent space embeddings which encode core characteristics of the data. We investigate the effectiveness of various pre-trained vision models, fine-tuning strategies, and image preparation. This visual representation offers a valuable starting point for identifying interesting events and grouping flares based on shared morphological characteristics, useful for conducting statistical studies or finding unique flares in this rich set of observations.

How to cite: Drozdova, M., Kinakh, V., Ramunno, F., Lastufka, E., and Voloshynovskiy, S.: Visualizing three years of STIX X-ray flare observations using self-supervised learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18534, https://doi.org/10.5194/egusphere-egu24-18534, 2024.

X4.107
|
EGU24-15981
|
ECS
Sophia Köhne, Brecht Laperre, Jorge Amaya, Sara Jamal, Simon Lautenbach, Rainer Grauer, Giovanni Lapenta, and Maria Elena Innocenti

When deriving fluid equations from the Vlasov equation for collisionless plasmas, one runs into the so-called closure problem: each equation for the temporal evolution of one particle moment (density, current, pressure, heat flux, …) includes terms depending on the next order moment. Therefore, when choosing to truncate the description at the nth order, one must approximate the terms related to the (n+1)th order moment included in the evolution equation for the nth order moment. The order at which the hierarchy is closed and the assumption behind the approximations used determine how accurately a fluid description can reproduce kinetic processes.

In this work, we aim at reconstructing specific particle moments from kinetic simulations, using as input the electric and magnetic field and the lower moments. We use fully kinetic Particle In Cell simulations, where all physical information is available, as the ground truth. The approach we present here uses supervised machine learning to enable a neural network to learn how to reconstruct higher moments from lower moments and fields.

Starting from the work of Laperre et al., 2022 we built a framework which makes it possible to train feedforward multilayer perceptrons on kinetic simulations to learn to predict the higher moments of the Vlasov equation from the lower moments, which would also be available in fluid simulations. We train on simulations of magnetic reconnection in a double Harris current sheet with varying background guide field obtained with the semi-implicit Particle-in-Cell code iPiC3D (Markidis et al, 2010). We test the influence of data preprocessing techniques, of (hyper-)parameter variations and of different architectures of the neural networks on the quality of the predictions that are produced. Furthermore, we investigate which metrics are most useful to evaluate the quality of the outcome.

How to cite: Köhne, S., Laperre, B., Amaya, J., Jamal, S., Lautenbach, S., Grauer, R., Lapenta, G., and Innocenti, M. E.: Addressing the closure problem using supervised Machine Learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15981, https://doi.org/10.5194/egusphere-egu24-15981, 2024.

X4.108
|
EGU24-1494
Yi Bi

we show the evolutions of the separated strands within the apparent single coronal loops observed in Atmospheric Imaging Assembly (AIA) images. The loop strands are detected on  the upsampled AIA 193 equation.pdf images, which are   generated using a super-resolution convolutional neural  network, respectively. The architecture of the network is designed to map the AIA images to unprecedentedly high spatial resolution coronal images taken by  High-resolution Coronal Imager (Hi-C) during its brief flight. At some times, pairs of individual strands appeared to braid with each other and subsequently evolved to become pairs of almost parallel ones with their segments having exchanged totally.  These evolutions provide  morphological evidence supporting occurrences of magnetic reconnections between the braiding strands, which are further confirmed by  the occurrences of the transient hot emissions (>5 MK)  located at the footpoints of  the braiding structures. 

How to cite: Bi, Y.: The coronal braiding structures detected in the machine-learning upscaled SDO/AIA images, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1494, https://doi.org/10.5194/egusphere-egu24-1494, 2024.

X4.109
|
EGU24-10715
|
ECS
|
Margherita Lampani, Sabrina Guastavino, Michele Piana, Federico Benvenuto, and Anna Maria Massone

Typical supervised feature-based machine learning approaches to flare forecasting rely on descriptors extracted from magnetograms, as from Helioseismic and Magnetic Imager (HMI) images, and standardized before being used in the training phase of the machine learning pipeline. However, this artificial intelligence (AI) model does not take into account the physical nature of the features and their role in the plasma physics equations. This talk proposes to generate novel features according to simple physics-driven combinations of the original descriptors, and to show whether this original physically explainable AI model leads to a more predictive solar flare forecasting.

How to cite: Lampani, M., Guastavino, S., Piana, M., Benvenuto, F., and Massone, A. M.: Physics-driven feature combination for an explainable AI approach to flare forecasting, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10715, https://doi.org/10.5194/egusphere-egu24-10715, 2024.

X4.110
|
EGU24-5545
|
ECS
Limits of solar flare forecasting models and new deep learning approach
(withdrawn)
Grégoire Francisco, Michele Berretti, Simone Chierichini, Ronish Mugatwala, João Fernandes, Teresa Barata, and Dario Del Moro
X4.111
|
EGU24-8018
Victor Bacu

The detection of asteroids involves the processing of sequences of astronomical images. The main challenges arise from the huge volume of data that should be processed in a reasonable amount of time. To address this, we developed the NEARBY platform [1], [2] for efficiently automatic detection of asteroids in sequence of astronomical images. This platform encompasses multidimensional data processing capabilities, human-verified visual analysis, and cloud-based adaptability. This paper outlines the enhancements we have made to this automated asteroid detection system by integrating a machine learning-based classifier known as the CERES module. The integration of the CERES module [3] into the NEARBY platform substantially enhances its performance by automatically reducing the number of false positive detections. Consequently, this leads to a more reliable and efficient system for asteroid identification, while also reducing the time and effort required by human experts to validate detected candidates (asteroids). The experiments highlight these improvements and their significance in advancing the field of asteroid tracking. Additionally, we explore the applicability of the asteroid classification model, initially trained using images from a specific telescope, across different telescopes.

Acknowledgment:

  • This work was supported by a grant of the Romanian Ministry of Education and Research, CCCDI - UEFISCDI, project number PN-III-P2-2.1-PED-2019-0796, within PNCDI III. (the development of the dataset and CNN models)
  • This research was partially supported by the project 38 PFE in the frame of the programme PDI-PFE-CDI 2021.

References:

  • Bacu, V., Sabou, A., Stefanut, T., Gorgan, D., Vaduvescu, O., NEARBY platform for detecting asteroids in astronomical images using cloud-based containerized applications, 2018 IEEE 14th International Conference on Intelligent Computer Communication and Processing (ICCP), pp. 371-376
  • Stefanut, T., Bacu, V., Nandra, C., Balasz, D., Gorgan, D., Vaduvescu, O., NEARBY Platform: Algorithm for automated asteroids detection in astronomical images, 2018 IEEE 14th International Conference on Intelligent Computer Communication and Processing (ICCP), pp. 365-369
  • Bacu, V.; Nandra, C.; Sabou, A.; Stefanut, T.; Gorgan, D. Assessment of Asteroid Classification Using Deep Convolutional Neural Networks. Aerospace 2023, 10, 752. https://doi.org/10.3390/aerospace10090752

 

How to cite: Bacu, V.: Enhancement of the NEARBY automated asteroid detection platform with a machine learning-based classifier, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8018, https://doi.org/10.5194/egusphere-egu24-8018, 2024.

X4.112
|
EGU24-12961
|
ECS
Robert Jarolim, Benoit Tremblay, Matthias Rempel, Julia Thalmann, Astrid Veronig, Momchil Molnar, and Tatiana Podladchikova

Physics-informed neural networks (PINNs) provide a novel approach for data-driven numerical simulations, tackling challenges of discretization and enabling seamless integration of noisy data and physical models (e.g., partial differential equations). In this presentation, we discuss the results of our recent studies where we apply PINNs for coronal magnetic field extrapolations of the solar atmosphere, which are essential to understand the genesis and initiation of solar eruptions and to predict the occurrence of high-energy events from our Sun.
We utilize our PINN to estimate the 3D coronal magnetic fields based on photospheric vector magnetograms and the force-free physical model. This approach provides state-of-the-art coronal magnetic field extrapolations in quasi real-time. We simulate the evolution of Active Region NOAA 11158 over 5 continuous days, where the derived time profile of the free magnetic energy unambiguously relates to the observed flare activity.
We extend this approach by utilizing multi-height magnetic field measurements and combine them in a single magnetic field model. Our evaluation shows that the additional chromospheric field information leads to a more realistic approximation of the solar coronal magnetic field. In addition, our method intrinsically provides an estimate of the height corrugation of the observed magnetograms.
We provide an outlook on our ongoing work where we use PINNs for global force-free magnetic field extrapolations. This approach enables a novel understanding of the global magnetic topology with a realistic treatment of current carrying fields.
In summary, PINNs have the potential to greatly advance the field of numerical simulations, accelerate scientific research, and enable advanced space weather monitoring.

How to cite: Jarolim, R., Tremblay, B., Rempel, M., Thalmann, J., Veronig, A., Molnar, M., and Podladchikova, T.: Physics-informed neural networks for advanced solar magnetic field extrapolations, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12961, https://doi.org/10.5194/egusphere-egu24-12961, 2024.