Simulations of the effects of vertical transport on the thermosphere and ionosphere using two coupled models

This paper presents our effort to assimilate FORMOSAT‐3/COSMIC (F3/C) GPS Occultation Experiment (GOX) observations into the National Center for Atmospheric Research (NCAR) Thermosphere Ionosphere Electrodynamics General Circulation Model (TIE‐GCM) by means of ensemble Kalman filtering (EnKF). The F3/C electron density profiles (EDPs) uniformly distributed around the globe which provide an excellent opportunity to monitor the ionospheric electron density structure. The NCAR TIE‐GCM simulates the Earth's thermosphere and ionosphere by using self‐consistent solutions for the coupled nonlinear equations of hydrodynamics, neutral and ion chemistry, and electrodynamics. The F3/C EDP are combined with the TIE‐GCM simulations by EnKF algorithms implemented in the NCAR Data Assimilation Research Testbed (DART) open‐source community facility to compute the expected value of electron density, which is ‘the best’ estimate of the current ionospheric state. Assimilation analyses obtained with real F3/C electron density profiles are compared with independent ground‐based observations as well as the F3/C profiles themselves. The comparison shows the improvement of the primary ionospheric parameters, such as NmF2 and hmF2. Nevertheless, some unrealistic signatures appearing in the results and high rejection rates of observations due to the applied outlier threshold and quality control are found in the assimilation experiments. This paper further discusses the limitations of the model and the impact of ensemble member creation approaches on the assimilation results, and proposes possible methods to avoid these problems for future work.

We examine the accuracy of density prediction by the first principles model Thermosphere Ionosphere Electrodynamics General Circulation Model (TIEGCM) developed by the National Center for Atmospheric Research and compare it to the accuracy of three empirical models: Jacchia 71, the Naval Research Laboratory Mass Spectrometer Incoherent Scatter Extended 2000 (NRLMSIS), Jacchia 1971, and Jacchia‐Bowman 2008. Comparisons are made for three large storms: the October 2003 storm, the March 2013 storm, and the March 2015 storm. To evaluate the accuracy of these models we use tracking data for nine space objects in low Earth orbit. Additionally, we evaluate the accuracy of the TIEGCM and NRLMSIS with data from high precision accelerometers on the Challenging Minisatellite Payload (CHAMP) and Gravity field and Circulation Explorer (GOCE) satellites. The goal is to assess the use of a first principles model as a potential tool for forecasting satellite drag during large magnetic storms. For the storms considered, we found the TIEGCM, JB2008, and NRLMSIS models to be substantially more accurate than the Jacchia 71 model. The accuracies of the TIEGCM and JB2008 models were similar, but overall, the TIEGCM was more accurate. We found smaller differences for TIEGCM versus CHAMP than for NRLMIS for the Halloween Storm, and smaller differences than results published for JB2008 and the assimilative model HASDM. The empirical models are at present more practical for operational purposes, but the TIEGCM, developed as a research model, with a greater focus on operational use offers the potential for improved utility during stressing conditions.

The first midlatitude conjugate thermospheric wind observations in the American sector showed various degrees of conjugacy between Palmer (64°S, 64°W, magnetic latitude (MLAT) 50°S) and Millstone Hill (42.82°N, 71.5°W, MLAT 53°N) under three different geomagnetic conditions (recovery after a substorm, moderately active, and quiet). The agreement with the National Center for Atmospheric Research's Thermosphere Ionosphere Electrodynamics General Circulation Model (TIEGCM) simulations also varies with the geomagnetic activity level. During substorm recovery, the observations at Palmer (PA) and Millstone Hill (MH) both showed strong westward zonal winds, which the standard TIEGCM greatly underestimated. Inadequate ion convection pattern size and lack of effect from Subauroral Polarization Streams (SAPS) may be the cause of the large discrepancy. The TIEGCM with a SAPS model produced stronger westward zonal winds near PA but did not change the zonal wind near MH. The empirical SAPS model needs further refinements. In general, there is better conjugacy with moderate geomagnetic activity levels. The TIEGCM also agrees better with the observations. Under geomagnetically quiet conditions, the meridional winds appear to be less conjugate. The agreement between the observations and model is reasonable. Optical conjugate observations are severely limited by the seasons and weather conditions in the two hemispheres. Yet they are necessary to understanding the thermospheric dynamics in the subauroral region and its relationship with geomagnetic activity levels. The comparisons with TIEGCM are necessary for future model improvements.

Using Fabry‐Perot interferometers at five midlatitude stations (Boulder, Palmer, Millstone Hill, Mount John, and Kelan) in both hemispheres, we examine the interhemispheric and seasonal variations of midlatitude thermospheric dynamics. We also use the National Center for Atmospheric Research Thermosphere Ionosphere Electrodynamics General Circulation Model (TIEGCM) to simulate the seasonal changes of winds and the effects from Sub‐Auroral Polarization Streams. The observations and TIEGCM simulations show a clear seasonal variation with more westward and equatorward summer winds. The TIEGCM runs overestimate the westward zonal winds and underestimate the electron densities in the northern summer. We believe that the underestimated TIEGCM electron density leads to a weak ion drag effect in the model, and strong westward zonal winds. TIEGCM overestimates the Sub‐Auroral Polarization Stream effects on neutral winds in most cases, probably because the empirical Sub‐Auroral Polarization Stream model used by the TIEGCM applies an unrealistic persistent electric field for a long period of time (over 3 hr) due to the low temporal resolution of the Kp index.

Results from four first‐principle models are compared with Millstone Hill incoherent scatter radar and Fabry‐Perot interferometer measurements taken during January 24–26, 1993, a period which included a minor geomagnetic storm. The models used in this study are the thermosphere ionosphere electrodynamics general circulation model (TIEGCM) with and without forcings from the assimilative mapping of ionospheric electrodynamics (AMIE) technique, the coupled thermosphere ionosphere model (CTIM), and the field line interhemispheric plasma (FLIP) model. The present study is the first time the AMIE inputs have been used in the TIEGCM model. TIEGCM and CTIM both underestimate the neutral temperature because of an underestimation of the Joule heating rate. An increase in the high latitude Joule heating would modify the thermospheric circulation. This could result in increases in N2 and O2 density above Millstone Hill, which would decrease the AMIE TIEGCM peak electron density (NmF2) to agree better with the observations, but would result in poorer agreement between CTIM and the data. The FLIP model NmF2 is a little low compared to the data, perhaps because of an inadequacy of the mass spectrometer incoherent scatter (MSIS) 86 model composition or the H+ flux in the model. Good agreement is obtained between atomic oxygen density [O] given by MSIS and [O] obtained from the radar data using a heat balance equation, provided an O+–O collision frequency factor of 1.3 is used. While the TIEGCM underestimates the electron and ion temperatures, the FLIP model reproduces major features of the data, apart from a large nighttime enhancement in Te. During the minor storm interval the observed neutral winds show alternating equatorward surges and abatements apparently due to passage of traveling atmospheric disturbances (TADs) seen in the model results. These are associated with a late evening increase observed in NmF2 accompanied by a large increase in F2 peak height (hmF2). These perturbations in NmF2 and hmF2 are not reproduced by the TIEGCM or CTIM. The NmF2 increase may be due to a decrease in O+ recombination rate caused by the higher hmF2, combined with compressional effects of a TAD and an enhanced downward flux of O+ ions.

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.

<p><span>The neutral density in the thermosphere is directly related to the atmospheric drag acceleration acting on satellites. In fact, the atmospheric drag acceleration, is the largest non-gravitational perturbation for satellites below 1000 km that has to be considered for precise orbit determination. There are several global empirical and physical models providing the neutral density in the thermosphere. However, there are significant differences between the modeled neutral densities and densities observed via accelerometers. More precise thermospheric density models are required for improving drag modeling as well as orbit determination. We study the coupling between ionosphere and thermosphere based on observations and model outputs of the Thermosphere Ionosphere Electrodynamics General Circulation Model (TIE-GCM). At first, we analyse the model’s representation of the coupling using electron and neutral densities. In comparison, we study the coupling based on observations, i.e., accelerometer-derived neutral densities and electron densities from a 4D electron density model based on GNSS and satellite altimetry data as well as radio occultation measurements. We expect that increased electron densities can be related to increased neutral densities. This is indicated for example by a correlation of approximately 55% between the neutral densities and the electron densities computed by the TIE-GCM. Finally, we investigate whether neutral density simulations fit better to in-situ densities from accelerometry when electron densities are assimilated.</span></p>

We use the National Center for Atmospheric Research TIEGCM (Thermosphere Ionosphere Electrodynamics General Circulation Model) model to investigate the eddy diffusion and tidal effects on the ionosphere SAO (semiannual oscillation). We also use the COSMIC (Constellation Observing System for Meteorology, Ionosphere, and Climate) satellite GPS radio occultation observations to validate the simulation results. The TIEGCM is driven at the 97 km lower boundary by tidal and gravity wave (eddy diffusion coefficient) inputs. The eddy diffusion input can be constant or with a SAO modulation, and the tidal input has on and off options. The TIEGCM simulation with a SAO modulated eddy diffusion (with tidal input) agrees better with the COSMIC observation than that without the SAO. Turning off the tides at the lower boundary makes the TIEGCM‐simulated ionospheric density much higher than the COSMIC observation. The simulations showed two results: (1) the need to add the SAO modulation to the eddy diffusion and (2) how tides reduce the ionospheric density and SAO. As to how much of the SAO should be added to the eddy diffusion is dependent on the amplitudes of the tides since both can have effects on the ionospheric density. The TIEGCM results also demonstrate that the ionospheric density diurnal signal is mostly in situ excited, while the semidiurnal signal comes from lower atmosphere.

Far ultraviolet observations of Earth's dayglow from the National Aeronautics and Space Administration (NASA) Global‐scale Observations of the Limb and Disk (GOLD) mission presents an unparalleled opportunity for upper atmosphere radiance data assimilation. Assimilation of the Lyman‐Birge‐Hopfield (LBH) band emissions can be formulated in a similar fashion to lower atmosphere radiance data assimilation approaches. To provide a proof‐of‐concept for such an approach, this paper presents assimilation experiments of simulated LBH emission data using an ensemble filter measurement update step implemented with National Oceanic and Atmospheric Administration (NOAA)'s Whole Atmosphere Model (WAM) and National Center for Atmospheric Research (NCAR)'s Global Airglow (GLOW) model. Primary findings from observing system simulation experiments (OSSEs), wherein “truth” atmospheric conditions simulated by NCAR's Thermosphere Ionosphere Electrodynamic General Circulation Model (TIEGCM) are used to generate synthetic GOLD data, are as follows: (1) Assimilation of GOLD LBH disk emission data can reduce the bias in model temperature specification (ensemble mean) by 60% under both geomagnetically quiet conditions and disturbed conditions. (2) The reduction in model uncertainty (ensemble spread) as a result of assimilation is about 20% in the lower thermosphere and 30% in the upper thermosphere for both conditions. These OSSEs demonstrate the potential for far ultraviolet radiance data assimilation to dramatically reduce the model biases in thermospheric temperature specification and to extend the utility of GOLD observations by helping to resolve the altitude‐dependent global‐scale response of the thermosphere to geomagnetic storms.

Using the total electron content (TEC) observations from GPS, and simulations from the Thermosphere Ionosphere Electrodynamic General Circulation Model (TIEGCM), this work investigates the large-scale traveling ionospheric disturbances (LSTIDs) and the possible involved drivers during the geomagnetic storm on January 14-15, 2022. Based on the term analysis of O+ continuity equation in TIEGCM, it is found that the traveling atmospheric disturbances in equatorward winds are responsible for the LSTIDs, with minor contributions from plasma drifts owing to the prompt penetration electric field. A strong interhemispheric asymmetry of the LSTIDs is observed, which might be attributed to both the equatorward wind disturbances and background plasma. The stronger wind (plasma) disturbances occurs in the winter hemisphere than that in the summer hemisphere. The maximum magnitude of LSTIDs in electron density disturbances occurs at ∼250 and ∼270 km in the northern and southern hemispheres, respectively, owing to both the thermospheric equatorward winds and background plasma. An interesting phenomenon that tail-like LSTIDs occur at the dip equator and low latitudes might be related to the eruption of the Tonga volcano, but it is not well reproduced in TIEGCM that deserves further exploration in a future study.

Simulation of nonmigrating tide influences on the thermosphere and ionosphere with a TIMED data driven TIEGCM

Presented is an analysis of the occurrence of postsunset Equatorial Plasma Bubbles (EPBs) detected using a Global Positioning System (GPS) receiver at Vanimo. The three year data set shows that the EPB occurrence maximizes (minimizes) during the equinoxes (solstices), in good agreement with previous findings. The Vanimo ionosonde station is used with the GPS receiver in an analysis of the day‐to‐day EPB occurrence variability during the 2000 equinox period. A superposed epoch analysis (SEA) reveals that the altitude, and the change in altitude, of the F layer height is ∼1 standard deviation (1σ) larger on the days for which EPBs were detected, compared to non‐EPB days. These results are then compared to results from the Thermosphere Ionosphere Electrodynamics General Circulation Model (TIEGCM), which show strong similarities with the observations. The TIEGCM is used to calculate the flux‐tube integrated Rayleigh‐Taylor (R‐T) instability linear growth rate. A SEA reveals that the modeled R‐T growth rate is 1σ higher on average for EPB days compared to non‐EPB days, and that the upward plasma drift is the most dominant contributor. It is further demonstrated that the TIEGCM's success in describing the observed daily EPB variability during the scintillation season resides in the variations caused by geomagnetic activity (as parameterized by Kp) rather than solar EUV flux (as parameterized by F10.7). Geomagnetic activity varies the modeled high‐latitude plasma convection and the associated Joule heating that affects the low‐latitude F region dynamo, and consequently the equatorial upward plasma drift.

In this study, we employed the National Center for Atmospheric Research Thermosphere Ionosphere Electrodynamics General Circulation Model (TIEGCM) and the extended Canadian Middle Atmosphere Model (eCMAM) to investigate the role of the migrating terdiurnal tide on the formation and variation of the thermosphere midnight temperature maximum (MTM) and midnight mass density maximum (MDM). The migrating terdiurnal tide from the eCMAM was applied at the TIEGCM's lower boundary, along with the migrating diurnal and semidiurnal tides from the Global‐Scale Wave Model. Several numerical experiments with different combinations of tidal forcing at the TIEGCM's lower boundary were carried out to determine the contribution of each tide to MTM/MDM. We found that the interplay between diurnal, semidiurnal, and terdiurnal tides determines the formation of MTM/MDM and their structure in the upper thermosphere. The decrease of thermospheric mass density after MDM reaches its maximum at ~02:00 local time is mainly controlled by the terdiurnal tide. Furthermore, we examined the generation mechanisms of the migrating terdiurnal tide in the upper thermosphere and found that they come from three sources: upward propagation from the lower thermosphere, in situ generation via nonlinear interaction, and thermal excitation.

Improvement of TIE-GCM thermospheric density predictions via incorporation of helium data from NRLMSISE-00

The purpose of this study is to appreciate the estimation of TIEGCM (Thermosphere Ionosphere Electrodynamics General Circulation Model) and that of the 2012 version of IRI (International Reference Ionosphere) in African Equatorial Ionization Anomaly (EIA) region through the diurnal variation of F2 layer critical frequency (foF2).The comparison is made between data and theoretical values carried out from TIEGCM and IRI-2012 during solar cycle minimum and maximum phases and under quiet time condition over seasons. Data concern solar cycle 22 foF2 data of Ouagadougou station (Lat: 12.4° N; Long: 358.5°E, dip: 1.43°N for 2013) provided by Telecom Bretagne. Quiet time condition is determined by Aa inferior or equal to 20 nT and solar cycle maximum and minimum phases correspond to sunspot number Rz superior to 100 and Rz inferior to 20, respectively. Seasons are estimated by considering December as winter month, March as spring month, June as summer month and September as autumn month. The seasonal Hourly quiet time foF2 is given by the arithmetic mean values of the five quietest day hourly values. Data profiles show noon bite out profile with more and less pronounced morning or afternoon peak in equinox and that during solar maximum and that also in Original Research Article Physical Science International Journal, 4(6): 892-902, 2014 893 solar minimum except during solstice where the profile fairly is dome or plateau. During solar minimum, both models present more or less pronounced afternoon peak with more or less deep trough between 1000 LT and 1400 LT. During solar maximum, in general, TIEGCM shows afternoon peak and IRI-2012 present plateau profile. This result exhibits the non-well estimation of the dynamic process of this region. Model accuracy is highlighted by the Mean Relative Error (MRE) values. These values show better prediction for IRI-2012 except in September for both solar cycle phases involved. The non-good prediction of TIEGCM is observed in December during solar minimum and in June during solar maximum. Models predictions are better during solar maximum than during solar minimum and strongly dependent on pre-sunrise and post sunset periods.