We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Predicting Equatorial Ionospheric Convective Instability Using Machine Learning.
- Authors
Garcia, D.; Rojas, E. L.; Hysell, D. L.
- Abstract
The numerical forecast methods used to predict ionospheric convective plasma instabilities associated with Equatorial Spread‐F (ESF) have limited accuracy and are often computationally expensive. We test whether it is possible to bypass first‐principle numeric simulations and forecast irregularities using machine learning models. The data are obtained from the incoherent scatter radar at the Jicamarca Radio Observatory located in Lima, Peru. Our models map vertical plasma drifts, time, and solar activity to the occurrence and location of clusters of echoes telltale of ionospheric irregularities. Our results show that these models are capable of identifying the predictive power of the tested inputs, obtaining accuracies around 75%. Plain Language Summary: Equatorial Spread‐F (ESF) is a phenomenon that happens in the ionosphere and it can create disturbances that affect communication and navigation systems, such as GPS or satellite phones. Scientists have been trying to predict when and where ESF will happen using numerical forecast methods, but these methods are not very accurate and can be expensive. In this study, we tested a new way to predict ESF using machine learning models. We used data from a radar in Lima, Peru, and looked at information such as how the plasma in the ionosphere moves vertically and how active the sun was. The models were able to predict where and when ESF would occur with about 75% accuracy, which is a promising development. Key Points: Clusters of radar echo bins associated with Equatorial Spread‐F can be identified using the DBSCAN algorithm from coherent scatter dataA random forest model showed that the day of year and vertical plasma drifts previous to sunset are good predictors of high SNR clustersA convolutional neural network was able to predict ionospheric irregularities with an accuracy of more than 70% despite scarce data
- Subjects
LIMA (Peru); CONVOLUTIONAL neural networks; ECHO; MACHINE learning; PLASMA instabilities; IONOSPHERIC plasma; SOLAR activity; INCOHERENT scattering; THERMAL instability
- Publication
Space Weather: The International Journal of Research & Applications, 2023, Vol 21, Issue 12, p1
- ISSN
1539-4956
- Publication type
Article
- DOI
10.1029/2023SW003505