Search

Observations of thermospheric neutral winds and temperatures obtained during a geomagnetic storm on 2 October 2013 from a network of six Fabry‐Perot interferometers (FPIs) deployed in the Midwest United States are presented. Coincident with the commencement of the storm, the apparent horizontal wind is observed to surge westward and southward (toward the equator). Simultaneous to this surge in the apparent horizontal winds, an apparent downward wind of approximately 100 m/s lasting for 6 h is observed. The apparent neutral temperature is observed to increase by approximately 400 K over all of the sites. Observations from an all‐sky imaging system operated at the Millstone Hill observatory indicate the presence of a stable auroral red (SAR) arc and diffuse red aurora during this time. We suggest that the large sustained apparent downward winds arise from contamination of the spectral profile of the nominal thermospheric 630.0 nm emission by 630.0 nm emission from a different (nonthermospheric) source. Modeling demonstrates that the effect of an additional population of 630.0 nm photons, with a distinct velocity and temperature distribution, introduces an apparent Doppler shift when the combined emissions from the two sources are analyzed as a single population. Thus, the apparent Doppler shifts should not be interpreted as the bulk motion of the thermosphere, calling into question results from previous FPI studies of midlatitude storm time thermospheric winds. One possible source of contamination could be fast O related to the infusion of low‐energy O+ions from the magnetosphere. The presence of low‐energy O+is supported by observations made by the Helium, Oxygen, Proton, and Electron spectrometer instruments on the twin Van Allen Probes spacecraft, which show an influx of low‐energy ions during this period. These results emphasize the importance of distributed networks of instruments in understanding the complex dynamics that occur in the upper atmosphere during disturbed conditions. Coordinated observations made by five Fabry‐Perot interferometers are shownAn unexpected strong and sustained vertical thermospheric wind is observedThe vertical wind is interpreted to be due to spectral contamination

A study of 37 large solar storms occurring in the equinox periods from 2000 to 2018 is presented to quantify their effects on total electron content (TEC) variations in the ionosphere over the North American sector. The dependence of storm‐related TEC enhancements and depletions is studied as a function of various storm parameters using machine‐learning and statistical algorithms to quantify their relative importance. The variables under consideration include the onset time and duration of each storm, as well as characteristic scalar values derived from ring current variations detected by a network of ground‐based magnetometers. The TEC data used in the study are taken from the Continuously Operating Reference Station network of ground‐based GPS measurements, archived in the Madrigal database. The agreement and disagreement of the results across the different methods studied, along with the relative uncertainties in model prediction capability of the machine‐learning study, show that the geo‐effectiveness of large storms in the equinox periods is more closely related to the storm onset time and to ionospheric preconditioning than to any other parameters used in characterizing the storm. The effects in the higher latitude regions are shown to be very different from the lower latitude regions over the North American sector. This behavior can possibly be explained through storm‐induced changes in the atmospheric chemistry and the related temperature‐dependent production and loss processes at the higher latitudes. Large geomagnetic storms reveal which characteristics of the measured Symmetric H‐component response are closely related to total electron content (TEC) perturbations over the USMachine learning and other statistical methods are independently applied to regions with distinct geomagnetic geometriesThe onset time of the storms, especially with respect to local sunrise, is the parameter most closely linked to the TEC response Large geomagnetic storms reveal which characteristics of the measured Symmetric H‐component response are closely related to total electron content (TEC) perturbations over the US Machine learning and other statistical methods are independently applied to regions with distinct geomagnetic geometries The onset time of the storms, especially with respect to local sunrise, is the parameter most closely linked to the TEC response