Machine learning (ML) has been applied to space weather problems with increasing frequency in recent years, driven by an influx of in-situ measurements and a desire to improve modeling and forecasting capabilities throughout the field. Space weather originates from solar perturbations and is comprised of the resulting complex variations they cause within the numerous systems between the Sun and Earth. These systems are often tightly coupled and not well understood. This creates a need for skillful models with knowledge about the confidence of their predictions. One example of such a dynamical system highly impacted by space weather is the thermosphere, the neutral region of Earth’s upper atmosphere. Our inability to forecast it has severe repercussions in the context of satellite drag and computation of probability of collision between two space objects in low Earth orbit (LEO) for decision making in space operations. Even with (assumed) perfect forecast of model drivers, our incomplete knowledge of the system results in often inaccurate thermospheric neutral mass density predictions. Continuing efforts are being made to improve model accuracy, but density models rarely provide estimates of confidence in predictions. In this work, we propose two techniques to develop nonlinear ML regression models to predict thermospheric density while providing robust and reliable uncertainty estimates: Monte Carlo (MC) dropout and direct prediction of the probability distribution, both using the negative logarithm of predictive density (NLPD) loss function. We show the performance capabilities for models trained on both local and global datasets. We show that the NLPD loss provides similar results for both techniques but the direct probability distribution prediction method has a much lower computational cost. For the global model regressed on the Space Environment Technologies High Accuracy Satellite Drag Model (HASDM) density database, we achieve errors of approximately 11% on independent test data with well-calibrated uncertainty estimates. Using an in-situ CHAllenging Minisatellite Payload (CHAMP) density dataset, models developed using both techniques provide test error on the order of 13%. The CHAMP models—on validation and test data—are within 2% of perfect calibration for the twenty prediction intervals tested. We show that this model can also be used to obtain global density predictions with uncertainties at a given epoch.
At altitudes below about 600 km, satellite drag is one of the most important and variable forces acting on a satellite. Neutral mass density predictions in the upper atmosphere are therefore critical for (a) designing satellites; (b) performing adjustments to stay in an intended orbit; and (c) collision avoidance maneuver planning. Density predictions have a great deal of uncertainty, including model biases and model misrepresentation of the atmospheric response to energy input. These may stem from inaccurate approximations of terms in the Navier‐Stokes equations, unmodeled physics, incorrect boundary conditions, or incorrect parameterizations. Two commonly parameterized source terms are the thermal conduction and eddy diffusion. Both are critical components in the transfer of the heat in the thermosphere. Determining how well the major constituents (N2, O2, and O) are as heat conductors will have effects on the temperature and mass density changes from a heat source. This work shows the effectiveness of using the retrospective cost model refinement (RCMR) technique at removing model bias caused by different sources within the Global Ionosphere Thermosphere Model. Numerical experiments, Challenging Minisatellite Payload and Gravity Recovery and Climate Experiment data during real events are used to show that RCMR can compensate for model bias caused by both inaccurate parameterizations and drivers. RCMR is used to show that eliminating model bias before a storm allows for more accurate predictions throughout the storm.
more » « less- NSF-PAR ID:
- 10419840
- Publisher / Repository:
- DOI PREFIX: 10.1029
- Date Published:
- Journal Name:
- Space Weather
- Volume:
- 21
- Issue:
- 3
- ISSN:
- 1542-7390
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
more » « less
Abstract -
more » « less
Abstract The geospace environment is volatile and highly driven. Space weather has effects on Earth's magnetosphere that cause a dynamic and enigmatic response in the thermosphere, particularly on the evolution of neutral mass density. Many models exist that use space weather drivers to produce a density response, but these models are typically computationally expensive or inaccurate for certain space weather conditions. In response, this work aims to employ a probabilistic machine learning (ML) method to create an efficient surrogate for the Thermosphere Ionosphere Electrodynamics General Circulation Model (TIE‐GCM), a physics‐based thermosphere model. Our method leverages principal component analysis to reduce the dimensionality of TIE‐GCM and recurrent neural networks to model the dynamic behavior of the thermosphere much quicker than the numerical model. The newly developed reduced order probabilistic emulator (ROPE) uses Long‐Short Term Memory neural networks to perform time‐series forecasting in the reduced state and provide distributions for future density. We show that across the available data, TIE‐GCM ROPE has similar error to previous linear approaches while improving storm‐time modeling. We also conduct a satellite propagation study for the significant November 2003 storm which shows that TIE‐GCM ROPE can capture the position resulting from TIE‐GCM density with <5 km bias. Simultaneously, linear approaches provide point estimates that can result in biases of 7–18 km.
-
more » « less
Abstract The ratio of O to N2number densities (O/N2) at different altitudes is an important parameter in describing thermospheric neutral composition changes and their effects on the ionosphere during geomagnetic storms. However, storm‐induced vertical variations in O/N2and its dependence on the O and N2perturbations are still not fully understood. Here, the Thermosphere/Ionosphere Electrodynamics General Circulation Model simulations were used to investigate the responses of thermospheric composition at different pressure levels to the super geomagnetic storm occurred on November 20 and 21 in 2003. Our analysis shows that the behaviors of O/N2perturbations on different pressure levels are similar above ∼180 km altitude. In the middle and low thermosphere of below ∼300 km, the storm‐time O/N2decrease is mainly caused by a large reduction of O number density. However, N2enhancement plays a vital role in O/N2decreases in the upper thermosphere. The O/N2enhancement is mainly attributed to the N2decreases at all pressure levels. The changes of O and N2number densities at a constant pressure level can be explained by the perturbations of their mass mixing ratio (mmr) and total mass density (
ρ ). The regions of the O/N2decrease are characterized by the O mmr decrease and N2mmr enhancement, whereas the regions of the O/N2increase are characterized by the O mmr increase and N2mmr decrease. Theρ value that shows the decrease globally at most pressure levels during the storm either enhance or reduce the O and N2perturbations. -
The Response of Middle Thermosphere (∼160 km) Composition to the November 20 and 21, 2003 Superstormmore » « less
Abstract TIMED/GUVI limb measurements and first‐principles simulations from the Thermosphere Ionosphere Electrodynamics Global Circulation Model (TIEGCM) are used to investigate thermospheric atomic oxygen (O) and molecular nitrogen (N2) responses in the middle thermosphere on a constant pressure surface (∼160 km) to the November 20 and 21, 2003 superstorm. The consistency between GUVI observations and TIEGCM simulated composition changes allows us to utilize TIEGCM outputs to investigate the storm‐time behaviors of O and N2systematically. Diagnostic analysis shows that horizontal and vertical advection are the two main processes that determine the storm‐induced perturbations in the middle thermosphere. Molecular diffusion has a relatively smaller magnitude than the two advection processes, acting to compensate for the changes caused by the transport partly. Contributions from chemistry and eddy diffusion are negligible. During the storm initial and main phases, composition variations at high latitudes are determined by both horizontal and vertical advection. At middle‐low latitudes, horizontal advection is the main driver for the composition changes where O mass mixing ratio
decreases (N2mass mixing ratio increases); whereas horizontal and vertical advection combined to dominate the changes in the regions where increases ( decreases). Over the entire storm period, horizontal advection plays a significant role in transporting high‐latitude composition perturbations globally. Our results also demonstrate that storm‐time temperature changes are not the direct cause of the composition perturbations on constant pressure surfaces. -
more » « less
Abstract Thermospheric mass density perturbations are commonly observed during geomagnetic storms and fundamental to upper atmosphere dynamics, but the sources of these perturbations are not well understood. Large neutral density perturbations during storms greatly affect the drag experienced by low Earth orbit. We investigated the thermospheric density perturbations at all latitudes observed along the CHAMP and GRACE satellite trajectories during the August 24–25, 2005 geomagnetic storm. Observations show that large neutral density enhancements occurred not only at high latitudes, but also globally. Large density perturbations were seen in the equatorial regions away from the high‐latitude, magnetospheric energy sources. We used the high‐resolution Multiscale Atmosphere Geospace Environment (MAGE) model to simulate consecutive neutral density changes observed by satellites during the storm. The MAGE simulation, which resolved mesoscale high‐latitude convection electric fields and field‐aligned currents, and included physics‐based specification of auroral precipitation, was contrasted with a standalone ionosphere‐thermosphere simulation driven by a high‐latitude electrodynamics empirical model. The comparison demonstrates that first‐principles representations of highly dynamic and localized Joule heating events in a fully coupled whole geospace model is critical to accurately capture both generation and propagation of traveling atmospheric disturbances (TADs) that produce neutral density perturbations globally. The MAGE simulation shows that larger density peaks in the equatorial region observed by CHAMP and GRACE are the result of TADs generated at high‐latitudes in both hemispheres, and intersect at low‐latitudes. This study reveals the importance of investigating thermospheric density variations at all latitudes in a fully coupled geospace model with sufficiently high resolving power.
