Search for: All records

Machine learning (ML) models are universal function approximators and—if used correctly—can summarize the information content of observational data sets in a functional form for scientific and engineering applications. A benefit to ML over parametric models is that there are no a priori assumptions about particular basis functions which can potentially limit the phenomena that can be modeled. In this work, we develop ML models on three data sets: the Space Environment Technologies High Accuracy Satellite Drag Model (HASDM) density database, a spatiotemporally matched data set of outputs from the Jacchia‐Bowman 2008 Empirical Thermospheric Density Model (JB2008), and an accelerometer‐derived density data set from CHAllenging Minisatellite Payload (CHAMP). These ML models are compared to the Naval Research Laboratory Mass Spectrometer and Incoherent Scatter radar (NRLMSIS 2.0) model to study the presence of post‐storm cooling in the middle‐thermosphere. We find that both NRLMSIS 2.0 and JB2008‐ML do not account for post‐storm cooling and consequently perform poorly in periods following strong geomagnetic storms (e.g., the 2003 Halloween storms). Conversely, HASDM‐ML and CHAMP‐ML do show evidence of post‐storm cooling indicating that this phenomenon is present in the original data sets. Results show that density reductions up to 40% can occur 1–3 days post‐storm depending on the location and strength of the storm.

The EXospheric TEMperatures on a PoLyhedrAl gRid (EXTEMPLAR) method predicts the neutral densities in the thermosphere. The performance of this model has been evaluated through a comparison with the Air Force High Accuracy Satellite Drag Model (HASDM). The Space Environment Technologies (SET) HASDM database that was used for this test spans the 20 years 2000 through 2019, containing densities at 3 hr time intervals at 25 km altitude steps, and a spatial resolution of 10° latitude by 15° longitude. The upgraded EXTEMPLAR that was tested uses the newer Naval Research Laboratory MSIS 2.0 model to convert global exospheric temperature values to neutral density as a function of altitude. The revision also incorporated time delays that varied as a function of location, between the total Poynting flux in the polar regions and the exospheric temperature response. The density values from both models were integrated on spherical shells at altitudes ranging from 200 to 800 km. These sums were compared as a function of time. The results show an excellent agreement at temporal scales ranging from hours to years. The EXTEMPLAR model performs best at altitudes of 400 km and above, where geomagnetic storms produce the largest relative changes in neutral density. In addition to providing an effective method to compare models that have very different spatial resolutions, the use of density totals at various altitudes presents a useful illustration of how the thermosphere behaves at different altitudes, on time scales ranging from hours to complete solar cycles.

This study explores intra‐annual oscillations (IAOs) in upper thermospheric winds using GOCE cross‐track wind measurements between 70°S and 70°N. Due to the Sun‐synchronous dawn‐dusk orbit of GOCE, the cross‐contamination between seasonality and local time variations in winds is minimal, which makes it a unique space‐based data set to extract IAOs and study their latitudinal variation. Our analysis reveals that the annual (AO), semiannual (SAO), and terannual (TAO) oscillations are robust features in thermospheric winds. The AO is strongest at middle latitudes; SAO and TAO amplitudes increase with increasing latitude. The latitudinally averaged amplitudes of the AO, SAO, and TAO for dusk/dawn are 30.0/35.0, 8.5/11.3, and 6.0/6.6 m/s, respectively. The phase of AO reverses around the equator. SAO and TAO phases vary with latitude but do not reverse like the AO. For both the SAO and TAO, the average phase at dusk and dawn differs by∼30 days.

We present an empirical model of thermospheric winds (High‐latitude Thermospheric Wind Model [HL‐TWiM]) that specifiesFregion high‐latitude horizontal neutral winds as a function of day of year, latitude, longitude, local time, and geomagnetic activity. HL‐TWiM represents the large‐scale neutral wind circulation, in geomagnetic coordinates, for the given input conditions. The model synthesizes the most extensive collection to date of historical high‐latitude wind measurements; it is based on statistical analyses of several decades ofFregion thermospheric wind measurements from 21 ground‐based stations (Fabry‐Perot Interferometers and Scanning Doppler Imaging Fabry‐Perot Interferometers) located at various northern and southern high latitudes and two space‐based instruments (UARS WINDII and GOCE). The geomagnetic latitude and local time dependences in HL‐TWiM are represented using vector spherical harmonics, day of year and longitude variations are represented using simple harmonic functions, and the geomagnetic activity dependence is represented using quadratic B splines. In this paper, we describe the HL‐TWiM formulation and fitting procedures, and we verify the model against the neutral wind databases used in its formulation. HL‐TWiM provides a necessary benchmark for validating new wind observations and tuning our physical understanding of complex wind behaviors. Results show stronger Universal Time variation in winds at southern than northern high latitudes. Model‐data intra‐annual comparisons in this study show semiannual oscillation‐like behavior of GOCE winds, rarely observed before in wind data.