Abstract The critical frequency of ionospheric F2 layer (foF2) is an important ionospheric characteristic parameter. In this paper, a deep learning model based on Bidirectional long short‐term memory (BiLSTM) and attention mechanism is implemented for predicting the foF2 parameter. The inputs of models are the foF2 of globally available ionospheric ionosonde stations, geographic longitude and latitude, world time (UT), geomagnetic activity index, and solar activity index from 2015 to 2017. The superiority of the model is analyzed from different latitudes, seasons, and geomagnetic conditions. The results show that the prediction performance of the Bidirectional long short‐term memory model based on attention mechanism (BiLSTM‐Attention) is better than other models. The performance of the prediction model is optimal at high latitudes. The root mean square error (RMSE) and correlation coefficient (R) of the BiLSTM‐Attention model are 0.539 MHZ and 0.908 MHz at high latitudes, respectively. In terms of RMSE, it is 25.243%, 18.209%, and 11.203% lower than those of the international reference ionosphere (IRI), LSTM, and BiLSTM models, respectively. The prediction results of the four seasons show that the models are more applicable in winter. Compared with the IRI model, the RMSE of the BiLSTM‐Attention model in spring, summer, autumn, and winter is reduced by 24.344%, 21.181%, 25.058%, and 30.948%, respectively. The prediction effect of the BiLSTM‐Attention model is improved in the magnetic quiet period, the magnetic moderate period and the magnetic storm period. Also, the improvement effect is more obvious in the magnetostatic day, and the RMSE is reduced by 27.462% compared with the IRI model.
{"title":"Forecasting Ionospheric foF2 Using Bidirectional LSTM and Attention Mechanism","authors":"Jun Tang, Dengpan Yang, Mingfei Ding","doi":"10.1029/2023sw003508","DOIUrl":"https://doi.org/10.1029/2023sw003508","url":null,"abstract":"Abstract The critical frequency of ionospheric F2 layer (foF2) is an important ionospheric characteristic parameter. In this paper, a deep learning model based on Bidirectional long short‐term memory (BiLSTM) and attention mechanism is implemented for predicting the foF2 parameter. The inputs of models are the foF2 of globally available ionospheric ionosonde stations, geographic longitude and latitude, world time (UT), geomagnetic activity index, and solar activity index from 2015 to 2017. The superiority of the model is analyzed from different latitudes, seasons, and geomagnetic conditions. The results show that the prediction performance of the Bidirectional long short‐term memory model based on attention mechanism (BiLSTM‐Attention) is better than other models. The performance of the prediction model is optimal at high latitudes. The root mean square error (RMSE) and correlation coefficient (R) of the BiLSTM‐Attention model are 0.539 MHZ and 0.908 MHz at high latitudes, respectively. In terms of RMSE, it is 25.243%, 18.209%, and 11.203% lower than those of the international reference ionosphere (IRI), LSTM, and BiLSTM models, respectively. The prediction results of the four seasons show that the models are more applicable in winter. Compared with the IRI model, the RMSE of the BiLSTM‐Attention model in spring, summer, autumn, and winter is reduced by 24.344%, 21.181%, 25.058%, and 30.948%, respectively. The prediction effect of the BiLSTM‐Attention model is improved in the magnetic quiet period, the magnetic moderate period and the magnetic storm period. Also, the improvement effect is more obvious in the magnetostatic day, and the RMSE is reduced by 27.462% compared with the IRI model.","PeriodicalId":49487,"journal":{"name":"Space Weather-The International Journal of Research and Applications","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135714654","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Zheng Wang, Meiyi Zhan, Pengdong Gao, Guojun Wang, Chu Qiu, Quan Qi, Jiankui Shi, Xiao Wang
Abstract An intelligent Spread‐F image detection and classification method is presented in this paper based on an ionogram image set using deep learning models. The ionogram images from the Hainan station, spanning from 2002 to 2015, have been manually labeled into five categories, resulting in a unique ionogram image set for supervised learning models. To balance the number of different types, simulated noises were added to these images. Based on 80,000 samples with Spread‐F and 20,000 samples without, numerous experiments have been conducted to train VGG, ResNet, EfficientNet, ViT, MobileNet, and other networks. The results on the test set indicate that these models except VGG have a good ability of exacting features of different types, leading to a high level of accuracy in detecting Spread‐F and a relatively accurate classification of it. The ionogram images in 2016 are then employed as another test set to further examine the performance of the trained models. Both quantitative and qualitative analyses have demonstrated the results obtained by deep learning models are highly consistent with manual identification.
{"title":"Automatic Detection and Classification of Spread‐F From Ionosonde at Hainan With Image‐Based Deep Learning Method","authors":"Zheng Wang, Meiyi Zhan, Pengdong Gao, Guojun Wang, Chu Qiu, Quan Qi, Jiankui Shi, Xiao Wang","doi":"10.1029/2023sw003498","DOIUrl":"https://doi.org/10.1029/2023sw003498","url":null,"abstract":"Abstract An intelligent Spread‐F image detection and classification method is presented in this paper based on an ionogram image set using deep learning models. The ionogram images from the Hainan station, spanning from 2002 to 2015, have been manually labeled into five categories, resulting in a unique ionogram image set for supervised learning models. To balance the number of different types, simulated noises were added to these images. Based on 80,000 samples with Spread‐F and 20,000 samples without, numerous experiments have been conducted to train VGG, ResNet, EfficientNet, ViT, MobileNet, and other networks. The results on the test set indicate that these models except VGG have a good ability of exacting features of different types, leading to a high level of accuracy in detecting Spread‐F and a relatively accurate classification of it. The ionogram images in 2016 are then employed as another test set to further examine the performance of the trained models. Both quantitative and qualitative analyses have demonstrated the results obtained by deep learning models are highly consistent with manual identification.","PeriodicalId":49487,"journal":{"name":"Space Weather-The International Journal of Research and Applications","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135410145","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
D. Conde, F. L. Castillo, C. Escobar, C. García, J. E. García, V. Sanz, B. Zaldívar, J. J. Curto, S. Marsal, J. M. Torta
Abstract Severe space weather produced by disturbed conditions on the Sun results in harmful effects both for humans in space and in high‐latitude flights, and for technological systems such as spacecraft or communications. Also, geomagnetically induced currents (GICs) flowing on long ground‐based conductors, such as power networks, potentially threaten critical infrastructures on Earth. The first step in developing an alarm system against GICs is to forecast them. This is a challenging task given the highly non‐linear dependencies of the response of the magnetosphere to these perturbations. In the last few years, modern machine‐learning models have shown to be very good at predicting magnetic activity indices. However, such complex models are on the one hand difficult to tune, and on the other hand they are known to bring along potentially large prediction uncertainties which are generally difficult to estimate. In this work we aim at predicting the SYM‐H index characterizing geomagnetic storms multiple‐hour ahead, using public interplanetary magnetic field (IMF) data from the Sun‐Earth L1 Lagrange point and SYM‐H data. We implement a type of machine‐learning model called long short‐term memory (LSTM) network. Our scope is to estimate the prediction uncertainties coming from a deep‐learning model in the context of forecasting the SYM‐H index. These uncertainties will be essential to set reliable alarm thresholds. The resulting uncertainties turn out to be sizable at the critical stages of the geomagnetic storms. Our methodology includes as well an efficient optimization of important hyper‐parameters of the LSTM network and robustness tests.
{"title":"Forecasting Geomagnetic Storm Disturbances and Their Uncertainties Using Deep Learning","authors":"D. Conde, F. L. Castillo, C. Escobar, C. García, J. E. García, V. Sanz, B. Zaldívar, J. J. Curto, S. Marsal, J. M. Torta","doi":"10.1029/2023sw003474","DOIUrl":"https://doi.org/10.1029/2023sw003474","url":null,"abstract":"Abstract Severe space weather produced by disturbed conditions on the Sun results in harmful effects both for humans in space and in high‐latitude flights, and for technological systems such as spacecraft or communications. Also, geomagnetically induced currents (GICs) flowing on long ground‐based conductors, such as power networks, potentially threaten critical infrastructures on Earth. The first step in developing an alarm system against GICs is to forecast them. This is a challenging task given the highly non‐linear dependencies of the response of the magnetosphere to these perturbations. In the last few years, modern machine‐learning models have shown to be very good at predicting magnetic activity indices. However, such complex models are on the one hand difficult to tune, and on the other hand they are known to bring along potentially large prediction uncertainties which are generally difficult to estimate. In this work we aim at predicting the SYM‐H index characterizing geomagnetic storms multiple‐hour ahead, using public interplanetary magnetic field (IMF) data from the Sun‐Earth L1 Lagrange point and SYM‐H data. We implement a type of machine‐learning model called long short‐term memory (LSTM) network. Our scope is to estimate the prediction uncertainties coming from a deep‐learning model in the context of forecasting the SYM‐H index. These uncertainties will be essential to set reliable alarm thresholds. The resulting uncertainties turn out to be sizable at the critical stages of the geomagnetic storms. Our methodology includes as well an efficient optimization of important hyper‐parameters of the LSTM network and robustness tests.","PeriodicalId":49487,"journal":{"name":"Space Weather-The International Journal of Research and Applications","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135714650","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
C. Kay, T. Nieves‐Chinchilla, S. J. Hofmeister, E. Palmerio, V. E. Ledvina
Abstract Coronal mass ejections (CMEs) and high speed streams (HSSs) are large‐scale transient structures that routinely propagate away from the Sun. Individually, they can cause space weather effects at the Earth, or elsewhere in space, but many of the largest events occur when these structures interact during their interplanetary propagation. We present the initial coupling of Open Solar Physics Rapid Ensemble Information (OSPREI), a model for CME evolution, with Mostly Empirical Operational Wind with a High Speed Stream, a time‐dependent HSS model that can serve as a background for the OSPREI CME. We present several improvements made to OSPREI in order to take advantage of the new time‐dependent, higher‐dimension background. This includes an update in the drag calculation and the ability to determine the rotation of a yaw‐like angle. We present several theoretical case studies, describing the difference in the CME behavior between a HSS background and a quiescent one. This behavior includes interplanetary CME propagation, expansion, deformation, and rotation, as well as the formation of a CME‐driven sheath. We also determine how the CME behavior changes with the HSS size and initial front distance. Generally, for a fast CME, we see that the drag is greatly reduced within the HSS, leading to faster CMEs and shorter travel times. The drag reappears stronger if the CME reaches the stream interaction region or upstream solar wind, leading to a stronger shock with more compression until the CME sufficiently decelerates. We model a CME–HSS interaction event observed by Parker Solar Probe in January 2022. The model improvements create a better match to the observed in situ profiles.
{"title":"A Series of Advances in Analytic Interplanetary CME Modeling","authors":"C. Kay, T. Nieves‐Chinchilla, S. J. Hofmeister, E. Palmerio, V. E. Ledvina","doi":"10.1029/2023sw003647","DOIUrl":"https://doi.org/10.1029/2023sw003647","url":null,"abstract":"Abstract Coronal mass ejections (CMEs) and high speed streams (HSSs) are large‐scale transient structures that routinely propagate away from the Sun. Individually, they can cause space weather effects at the Earth, or elsewhere in space, but many of the largest events occur when these structures interact during their interplanetary propagation. We present the initial coupling of Open Solar Physics Rapid Ensemble Information (OSPREI), a model for CME evolution, with Mostly Empirical Operational Wind with a High Speed Stream, a time‐dependent HSS model that can serve as a background for the OSPREI CME. We present several improvements made to OSPREI in order to take advantage of the new time‐dependent, higher‐dimension background. This includes an update in the drag calculation and the ability to determine the rotation of a yaw‐like angle. We present several theoretical case studies, describing the difference in the CME behavior between a HSS background and a quiescent one. This behavior includes interplanetary CME propagation, expansion, deformation, and rotation, as well as the formation of a CME‐driven sheath. We also determine how the CME behavior changes with the HSS size and initial front distance. Generally, for a fast CME, we see that the drag is greatly reduced within the HSS, leading to faster CMEs and shorter travel times. The drag reappears stronger if the CME reaches the stream interaction region or upstream solar wind, leading to a stronger shock with more compression until the CME sufficiently decelerates. We model a CME–HSS interaction event observed by Parker Solar Probe in January 2022. The model improvements create a better match to the observed in situ profiles.","PeriodicalId":49487,"journal":{"name":"Space Weather-The International Journal of Research and Applications","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135810216","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Abstract The solar wind which arrives at any location in the solar system is, in principle, relatable to the outflow of solar plasma from a single source location. This source location, itself usually being part of a larger coronal hole, is traceable to 1 R S along the Sun's magnetic field, in which the entire path from 1 R S to a location in the heliosphere is referred to as the solar wind connectivity. While not directly measurable, the connectivity between the near‐Earth solar wind is of particular importance to space weather. The solar wind solar source region can be obtained by leveraging near‐sun magnetic field models and a model of the interplanetary solar wind. In this article, we present a method for making an ensemble forecast of the connectivity presented as a probability distribution obtained from a weighted collection of individual forecasts from the combined Air Force Data Assimilative Photospheric Flux Transport‐Wang Sheeley Arge (ADAPT‐WSA) model. The ADAPT model derives the photospheric magnetic field from synchronic magnetogram data, using flux transport physics and ongoing data assimilation processes. The WSA model uses a coupled set of potential field type models to derive the coronal magnetic field, and an empirical relationship to derive the terminal solar wind speed observed at Earth. Our method produces an arbitrary 2D probability distribution capable of reflecting complex source configurations with minimal assumptions about the distribution structure, prepared in a computationally efficient manner.
{"title":"Ensemble Forecasts of Solar Wind Connectivity to 1 <i>R</i><sub><i>s</i></sub> Using ADAPT‐WSA","authors":"D. E. da Silva, S. Wallace, C. N. Arge, S. Jones","doi":"10.1029/2023sw003554","DOIUrl":"https://doi.org/10.1029/2023sw003554","url":null,"abstract":"Abstract The solar wind which arrives at any location in the solar system is, in principle, relatable to the outflow of solar plasma from a single source location. This source location, itself usually being part of a larger coronal hole, is traceable to 1 R S along the Sun's magnetic field, in which the entire path from 1 R S to a location in the heliosphere is referred to as the solar wind connectivity. While not directly measurable, the connectivity between the near‐Earth solar wind is of particular importance to space weather. The solar wind solar source region can be obtained by leveraging near‐sun magnetic field models and a model of the interplanetary solar wind. In this article, we present a method for making an ensemble forecast of the connectivity presented as a probability distribution obtained from a weighted collection of individual forecasts from the combined Air Force Data Assimilative Photospheric Flux Transport‐Wang Sheeley Arge (ADAPT‐WSA) model. The ADAPT model derives the photospheric magnetic field from synchronic magnetogram data, using flux transport physics and ongoing data assimilation processes. The WSA model uses a coupled set of potential field type models to derive the coronal magnetic field, and an empirical relationship to derive the terminal solar wind speed observed at Earth. Our method produces an arbitrary 2D probability distribution capable of reflecting complex source configurations with minimal assumptions about the distribution structure, prepared in a computationally efficient manner.","PeriodicalId":49487,"journal":{"name":"Space Weather-The International Journal of Research and Applications","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136009491","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Abstract Reliable short‐time prediction of thermospheric mass density along the satellite orbit is always essential but challenging for the operation of Low‐Earth orbit satellites. In this paper, three machine‐learning prediction algorithms are investigated, including the Bidirectional Long Short‐Term Memory, the Transformer, and the Light Gradient Boosting Machine (LightGBM) ensemble model of the above models. We use satellite data from CHAMP, GOCE, and SWARM‐C to evaluate the robustness and accuracy of different density variations. The comparison demonstrates that all models achieve compelling predictions and are much better than NRLMSISE‐00. The LightGBM ensemble model (LE‐model) consistently outperforms others in accuracy and stability. Furthermore, when the obtained density data from the newly launched satellites are limited, the trained LE‐model can provide a valid prediction for the new satellite orbit by transfer learning. This study offers a promising insight into the short‐time prediction of thermospheric mass density using ensemble‐transfer learning and may be advantageous to future research on space whether.
{"title":"The Short‐Time Prediction of Thermospheric Mass Density Based on Ensemble‐Transfer Learning","authors":"Peian Wang, Zhou Chen, Xiaohua Deng, Jing‐Song Wang, Rongxing Tang, Haimeng Li, Sheng Hong, Zhiping Wu","doi":"10.1029/2023sw003576","DOIUrl":"https://doi.org/10.1029/2023sw003576","url":null,"abstract":"Abstract Reliable short‐time prediction of thermospheric mass density along the satellite orbit is always essential but challenging for the operation of Low‐Earth orbit satellites. In this paper, three machine‐learning prediction algorithms are investigated, including the Bidirectional Long Short‐Term Memory, the Transformer, and the Light Gradient Boosting Machine (LightGBM) ensemble model of the above models. We use satellite data from CHAMP, GOCE, and SWARM‐C to evaluate the robustness and accuracy of different density variations. The comparison demonstrates that all models achieve compelling predictions and are much better than NRLMSISE‐00. The LightGBM ensemble model (LE‐model) consistently outperforms others in accuracy and stability. Furthermore, when the obtained density data from the newly launched satellites are limited, the trained LE‐model can provide a valid prediction for the new satellite orbit by transfer learning. This study offers a promising insight into the short‐time prediction of thermospheric mass density using ensemble‐transfer learning and may be advantageous to future research on space whether.","PeriodicalId":49487,"journal":{"name":"Space Weather-The International Journal of Research and Applications","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135568489","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A. Waszewski, J. S. Morgan, R. Chhetri, R. Ekers, M. C. M. Cheung, N. D. R. Bhat, M. Johnston‐Hollitt
Abstract We have conducted a blind search in 49 consecutive days of interplanetary scintillation observations made by the Murchison Widefield Array from mid‐2019, with overlapping daily observations approximately East and South‐East of the Sun at an elongation of ∼30° and a field of view of 30°. These observations detect an unprecedented density of sources. In spite of these observations being taken at sunspot minimum, this search has revealed several interesting transitory features characterized by elevated scintillation levels. One solar wind enhancement is captured in two observations several hours apart, allowing its radial movement away from the Sun to be measured. We present here a methodology for measuring the plane‐of‐sky velocity for the moving heliospheric structure. The plane‐of‐sky velocity was inferred as 0.66 ± 0.147 hr −1 , or 480 ± 106 kms −1 assuming a distance of 1AU. After cross‐referencing our observed structure with multiple catalogs of heliospheric events, we propose that the likely source of our observed structure is a stream‐interaction region originating from a low‐latitude coronal hole. This work demonstrates the power of widefield interplanetary scintillation observations to capture detailed features in the heliosphere which are otherwise unresolvable and go undetected.
{"title":"Resolving Moving Heliospheric Structures Using Interplanetary Scintillation Observations With the Murchison Widefield Array","authors":"A. Waszewski, J. S. Morgan, R. Chhetri, R. Ekers, M. C. M. Cheung, N. D. R. Bhat, M. Johnston‐Hollitt","doi":"10.1029/2023sw003570","DOIUrl":"https://doi.org/10.1029/2023sw003570","url":null,"abstract":"Abstract We have conducted a blind search in 49 consecutive days of interplanetary scintillation observations made by the Murchison Widefield Array from mid‐2019, with overlapping daily observations approximately East and South‐East of the Sun at an elongation of ∼30° and a field of view of 30°. These observations detect an unprecedented density of sources. In spite of these observations being taken at sunspot minimum, this search has revealed several interesting transitory features characterized by elevated scintillation levels. One solar wind enhancement is captured in two observations several hours apart, allowing its radial movement away from the Sun to be measured. We present here a methodology for measuring the plane‐of‐sky velocity for the moving heliospheric structure. The plane‐of‐sky velocity was inferred as 0.66 ± 0.147 hr −1 , or 480 ± 106 kms −1 assuming a distance of 1AU. After cross‐referencing our observed structure with multiple catalogs of heliospheric events, we propose that the likely source of our observed structure is a stream‐interaction region originating from a low‐latitude coronal hole. This work demonstrates the power of widefield interplanetary scintillation observations to capture detailed features in the heliosphere which are otherwise unresolvable and go undetected.","PeriodicalId":49487,"journal":{"name":"Space Weather-The International Journal of Research and Applications","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134934305","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Abstract Machine learning (ML) has been increasingly applied to space weather and ionosphere problems in recent years, with the goal of improving modeling and forecasting capabilities through a data‐driven modeling approach of nonlinear relationships. However, little work has been done to quantify the uncertainty of the results, lacking an indication of how confident and reliable the results of an ML system are. In this paper, we implement and analyze several uncertainty quantification approaches for an ML‐based model to forecast Vertical Total Electron Content (VTEC) 1‐day ahead and corresponding uncertainties with 95% confidence intervals (CI): (a) Super‐Ensemble of ML‐based VTEC models (SE), (b) Gradient Tree Boosting with quantile loss function (Quantile Gradient Boosting, QGB), (c) Bayesian neural network (BNN), and (d) BNN including data uncertainty (BNN + D). Techniques that consider only model parameter uncertainties (a and c) predict narrow CI and over‐optimistic results, whereas accounting for both model parameter and data uncertainties with the BNN + D approach leads to a wider CI and the most realistic uncertainties quantification of VTEC forecast. However, the BNN + D approach suffers from a high computational burden, while the QGB approach is the most computationally efficient solution with slightly less realistic uncertainties. The QGB CI are determined to a large extent from space weather indices, as revealed by the feature analysis. They exhibit variations related to daytime/nightime, solar irradiance, geomagnetic activity, and post‐sunset low‐latitude ionosphere enhancement.
{"title":"Uncertainty Quantification for Machine Learning‐Based Ionosphere and Space Weather Forecasting: Ensemble, Bayesian Neural Network, and Quantile Gradient Boosting","authors":"Randa Natras, Benedikt Soja, Michael Schmidt","doi":"10.1029/2023sw003483","DOIUrl":"https://doi.org/10.1029/2023sw003483","url":null,"abstract":"Abstract Machine learning (ML) has been increasingly applied to space weather and ionosphere problems in recent years, with the goal of improving modeling and forecasting capabilities through a data‐driven modeling approach of nonlinear relationships. However, little work has been done to quantify the uncertainty of the results, lacking an indication of how confident and reliable the results of an ML system are. In this paper, we implement and analyze several uncertainty quantification approaches for an ML‐based model to forecast Vertical Total Electron Content (VTEC) 1‐day ahead and corresponding uncertainties with 95% confidence intervals (CI): (a) Super‐Ensemble of ML‐based VTEC models (SE), (b) Gradient Tree Boosting with quantile loss function (Quantile Gradient Boosting, QGB), (c) Bayesian neural network (BNN), and (d) BNN including data uncertainty (BNN + D). Techniques that consider only model parameter uncertainties (a and c) predict narrow CI and over‐optimistic results, whereas accounting for both model parameter and data uncertainties with the BNN + D approach leads to a wider CI and the most realistic uncertainties quantification of VTEC forecast. However, the BNN + D approach suffers from a high computational burden, while the QGB approach is the most computationally efficient solution with slightly less realistic uncertainties. The QGB CI are determined to a large extent from space weather indices, as revealed by the feature analysis. They exhibit variations related to daytime/nightime, solar irradiance, geomagnetic activity, and post‐sunset low‐latitude ionosphere enhancement.","PeriodicalId":49487,"journal":{"name":"Space Weather-The International Journal of Research and Applications","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134935936","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Diptiranjan Rout, S. Patra, S. Kumar, D. Chakrabarty, G. D. Reeves, C. Stolle, K. Pandey, S. Chakraborty, E. A. Spencer
Abstract The total energy transfer from the solar wind to the magnetosphere is governed by the reconnection rate at the magnetosphere edges as the Z‐component of interplanetary magnetic field (IMF B z ) turns southward. The geomagnetic storm on 21–22 January 2005 is considered to be anomalous as the SYM‐H index that signifies the strength of ring current, decreases and had a sustained trough value of −101 nT lasting more than 6 hr under northward IMF B z conditions. In this work, the standard WINDMI model is utilized to estimate the growth and decay of magnetospheric currents by using several solar wind‐magnetosphere coupling functions. However, it is found that the WINDMI model driven by any of these coupling functions is not fully able to explain the decrease of SYM‐H under northward IMF B z . A dense plasma sheet along with signatures of a highly stretched magnetosphere was observed during this storm. The SYM‐H variations during the entire duration of the storm were only reproduced after modifying the WINDMI model to account for the effects of the dense plasma sheet. The limitations of directly driven models relying purely on the solar wind parameters and not accounting for the state of the magnetosphere are highlighted by this work.
当行星际磁场的Z分量(IMF B Z)向南转变时，太阳风向磁层的总能量转移受磁层边缘重联率的控制。2005年1月21日至22日的地磁风暴被认为是异常的，因为在北向的IMF B - z条件下，表征环电流强度的SYM - H指数下降，并且持续槽值为- 101 nT，持续时间超过6小时。在这项工作中，利用标准的WINDMI模型，通过几个太阳风-磁层耦合函数来估计磁层电流的增长和衰减。然而，我们发现由这些耦合函数驱动的WINDMI模型都不能完全解释北移的IMF B z下SYM‐H的减少。在这次风暴中观测到密集的等离子体层以及高度拉伸的磁层的特征。在整个风暴期间的SYM - H变化只有在修改了WINDMI模型以考虑到致密等离子体层的影响后才能重现。直接驱动模式的局限性仅仅依赖于太阳风参数，而不考虑磁层的状态。
{"title":"The Growth of Ring Current/SYM‐H Under Northward IMF <i>B</i><sub><i>z</i></sub> Conditions Present During the 21–22 January 2005 Geomagnetic Storm","authors":"Diptiranjan Rout, S. Patra, S. Kumar, D. Chakrabarty, G. D. Reeves, C. Stolle, K. Pandey, S. Chakraborty, E. A. Spencer","doi":"10.1029/2023sw003489","DOIUrl":"https://doi.org/10.1029/2023sw003489","url":null,"abstract":"Abstract The total energy transfer from the solar wind to the magnetosphere is governed by the reconnection rate at the magnetosphere edges as the Z‐component of interplanetary magnetic field (IMF B z ) turns southward. The geomagnetic storm on 21–22 January 2005 is considered to be anomalous as the SYM‐H index that signifies the strength of ring current, decreases and had a sustained trough value of −101 nT lasting more than 6 hr under northward IMF B z conditions. In this work, the standard WINDMI model is utilized to estimate the growth and decay of magnetospheric currents by using several solar wind‐magnetosphere coupling functions. However, it is found that the WINDMI model driven by any of these coupling functions is not fully able to explain the decrease of SYM‐H under northward IMF B z . A dense plasma sheet along with signatures of a highly stretched magnetosphere was observed during this storm. The SYM‐H variations during the entire duration of the storm were only reproduced after modifying the WINDMI model to account for the effects of the dense plasma sheet. The limitations of directly driven models relying purely on the solar wind parameters and not accounting for the state of the magnetosphere are highlighted by this work.","PeriodicalId":49487,"journal":{"name":"Space Weather-The International Journal of Research and Applications","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135810772","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Saba Javed, Nazish Rubab, Sadia Zaheer, Stefaan Poedts, Ghulam Jaffer
Abstract Surface charging at geosynchronous altitude is one of the major concerns for satellites and spacecrafts. Spacecraft anomalies are often associated with extreme surface charging events, especially during substorms in which the GEO plasma is better modeled as two temperatures non‐Maxwellian plasma. In such case, we employ two temperature q‐non‐extensive distribution function to determine the onset of spacecraft surface charging which becomes complex since many parameters control the surface charging. We developed a current balance equation which better explains the charging threshold in comparison to a Maxwellian distribution function. The effect of non‐extensive parameters, temperature and density ratio on the current balance equation has been explained. The modified current balance equation predicts the critical and anti‐critical temperatures for various space‐grade materials both analytically and numerically. A significant change is observed in the quantities characterizing the charging current, average yield and density ratio in the presence of non‐extensive two temperature electrons. The mechanism underlying different charging behaviors at or near the threshold is also indicated at various plasma parametric domains. Furthermore, the general conditions of potential jump are also obtained theoretically which predicts the sudden or smooth potential transition.
{"title":"Numerical Calculations of Charging Threshold at GEO Altitudes With Two Temperature Non‐Extensive Electrons","authors":"Saba Javed, Nazish Rubab, Sadia Zaheer, Stefaan Poedts, Ghulam Jaffer","doi":"10.1029/2022sw003412","DOIUrl":"https://doi.org/10.1029/2022sw003412","url":null,"abstract":"Abstract Surface charging at geosynchronous altitude is one of the major concerns for satellites and spacecrafts. Spacecraft anomalies are often associated with extreme surface charging events, especially during substorms in which the GEO plasma is better modeled as two temperatures non‐Maxwellian plasma. In such case, we employ two temperature q‐non‐extensive distribution function to determine the onset of spacecraft surface charging which becomes complex since many parameters control the surface charging. We developed a current balance equation which better explains the charging threshold in comparison to a Maxwellian distribution function. The effect of non‐extensive parameters, temperature and density ratio on the current balance equation has been explained. The modified current balance equation predicts the critical and anti‐critical temperatures for various space‐grade materials both analytically and numerically. A significant change is observed in the quantities characterizing the charging current, average yield and density ratio in the presence of non‐extensive two temperature electrons. The mechanism underlying different charging behaviors at or near the threshold is also indicated at various plasma parametric domains. Furthermore, the general conditions of potential jump are also obtained theoretically which predicts the sudden or smooth potential transition.","PeriodicalId":49487,"journal":{"name":"Space Weather-The International Journal of Research and Applications","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136159990","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：