Search for: All records

Abstract We model the electron density in the topside of the ionosphere with an improved machine learning (ML) model and compare it to existing empirical models, specifically the International Reference Ionosphere (IRI) and the Empirical‐Canadian High Arctic Ionospheric Model (E‐CHAIM). In prior work, an artificial neural network (NN) was developed and trained on two solar cycles worth of Defense Meteorological Satellite Program data (113 satellite‐years), along with global drivers and indices to predict topside electron density. In this paper, we highlight improvements made to this NN, and present a detailed comparison of the new model to E‐CHAIM and IRI as a function of location, geomagnetic condition, time of year, and solar local time. We discuss precision and accuracy metrics to better understand model strengths and weaknesses. The updated neural network shows improved mid‐latitude performance with absolute errors lower than the IRI by 2.5 × 10⁹to 2.5 × 10¹⁰e⁻/m³, modestly improved performance in disturbed geomagnetic conditions with absolute errors reduced by about 2.5 × 10⁹ e⁻/m³at high Kp compared to the IRI, and high Kp percentage errors reduced by >50% when compared to E‐CHAIM. 

Abstract We demonstrate a methodology for utilizing measurements from very low frequency (VLF, 3−30 kHz) transmitters and lightning emissions to produce 3D lower electron density maps, and apply it to multiple geophysical disturbances. The D‐region lower ionosphere (60−90 km) forms the upper boundary of the Earth‐ionosphere waveguide which allows VLF radio waves to propagate to global distances. Measurements of these signals have, in many prior studies, been used to infer path‐average electron density profiles within the D region. Historically, researchers have focused on either measurements of VLF transmitters or radio atmospherics (sferics) from lightning. In this work, we build on recently published methods for each and present a method to unify the two approaches via tomography. The output of the tomographic inversion produces maps of electron density over a large portion of the United States and Gulf of Mexico. To illustrate the benefits of this unified approach, daytime and nighttime maps are compared between a sferic‐only model and the new approach suggested here. We apply the model to characterize two geophysical disturbances: solar flares and lower ionospheric changes associated with thunderstorms. 

Abstract In this work, convolutional neural networks (CNN) are developed to detect and characterize sporadic E (E_s), demonstrating an improvement over current methods. This includes a binary classification model to determine ifE_sis present, followed by a regression model to estimate theE_sordinary mode critical frequency (foEs), a proxy for the intensity, along with the height at which theE_slayer occurs (hEs). Signal‐to‐noise ratio (SNR) and excess phase profiles from six Global Navigation Satellite System (GNSS) radio occultation (RO) missions during the years 2008–2022 are used as the inputs of the model. Intensity (foEs) and the height (hEs) values are obtained from the global network of ground‐based Digisonde ionosondes and are used as the “ground truth,” or target variables, during training. After corresponding the two data sets, a total of 36,521 samples are available for training and testing the models. The foEs CNN binary classification model achieved an accuracy of 74% and F1‐score of 0.70. Mean absolute errors (MAE) of 0.63 MHz and 5.81 km along with root‐mean squared errors (RMSE) of 0.95 MHz and 7.89 km were attained for estimating foEs and hEs, respectively, when it was known thatE_swas present. When combining the classification and regression models together for use in practical applications where it is unknown ifE_sis present, an foEs MAE and RMSE of 0.97 and 1.65 MHz, respectively, were realized. We implemented three other techniques for sporadic E characterization, and found that the CNN model appears to perform better. 

Abstract We present a tomographic imaging technique for the D‐region electron density using a set of spatially distributed very low frequency (VLF) remote sensing measurements. The D‐region ionosphere plays a critical role in many long‐range and over‐the‐horizon communication systems; however, it is unreachable by most direct measurement techniques such as balloons and satellites. Fortunately, the D region, combined with Earth's surface, forms what is known as the Earth‐Ionosphere waveguide allowing VLF and low frequency (LF) radio waves to propagate to global distances. By measuring these signals, we can estimate a path measurement of the electron density, which we assume to be a path‐averaged electron density profile of the D region. In this work, we use path‐averaged inferences from lightning‐generated radio atmospherics (sferics) with a tomographic inversion to produce 3D models of electron density over the Southeastern United States and the Gulf of Mexico. The model begins with two‐dimensional great circle path observations, each of which is parameterized so it includes vertical profile information. The tomography is then solved in two dimensions (latitude and longitude) at arbitrarily many altitude slices to construct the 3D electron density. We examine the model's performance in the synthetic case and determine that we have an expected percent error better than 10% within our area of interest. We apply our model to the 2017 “Great American Solar Eclipse” and find a clear relationship between sunlight percentage and electron density at different altitudes.