Hours Of Local Time Research Articles

GPS normalization and preliminary modeling results of total electron content during a midlatitude space weather event

On November 22–23, 1997, a geomagnetic storm occurred during a period of excellent viewing conditions over the Arecibo Observatory in Puerto Rico. Here we explore the total electron content (TEC) registered by Global Positioning System (GPS) receivers located close to the Cornell All‐Sky Imager (CASI) at the Arecibo Observatory. The storm began with the equatorward surge of a very high (100% increase) TEC enhancement stretching for many hours of local time on the dayside. At dusk the TEC over the Caribbean remained elevated with levels equal to the noontime monthly averages. During the event the TEC was highly structured and clearly correlated with high and low airglow emission levels. In one fortuitous instance a common ionospheric penetration point (15 km apart), shared by two GPS satellites viewed from two receiving stations, registered an 8 TEC unit difference during the active period. We show that a GPS station can be calibrated using the pseudorange method and a reliable data‐driven technique during quiet conditions and still have absolute TEC capability within 2 TEC units (RMS) 5 days later. We compare the observations to a climatological model which, although reasonable for quiet times, is very poor during the storm period. We also present an independent evaluation of the GPS TEC. This study is an initial step toward quality control of this database, needed before it is used in an assimilation model.

Two-dimensional electric field measurements in the ionospheric footprint of a flux transfer event

Line-of-sight Doppler velocities from the SuperDARN CUTLASS HF radar pair have been combined to produce the first two-dimensional vector measurements of the convection pattern throughout the ionospheric footprint of a flux transfer event (a pulsed ionospheric flow, or PIF). Very stable and moderate interplanetary magnetic field conditions, along with a preceding prolonged period of northward interplanetary magnetic field, allow a detailed study of the spatial and the temporal evolution of the ionospheric response to magnetic reconnection. The flux tube footprint is tracked for half an hour across six hours of local time in the auroral zone, from magnetic local noon to dusk. The motion of the footprint of the newly reconnected flux tube is compared with the ionospheric convection velocity. Two primary intervals in the PIF's evolution have been determined. For the first half of its lifetime in the radar field of view the phase speed of the PIF is highly variable and the mean speed is nearly twice the ionospheric convection speed. For the final half of its lifetime the phase velocity becomes much less variable and slows down to the ionospheric convection velocity. The evolution of the flux tube in the magnetosphere has been studied using magnetic field, magnetopause and magnetosheath models. The data are consistent with an interval of azimuthally propagating magnetopause reconnection, in a manner consonant with a peeling of magnetic flux from the magnetopause, followed by an interval of anti-sunward convection of reconnected flux tubes.Key words: Magnetospheric physics (magnetosphere · ionosphere interactions; plasma convection; solar wind · magnetosphere interactions)

Two-dimensional electric field measurements in the ionospheric footprint of a flux transfer event

Abstract. Line-of-sight Doppler velocities from the SuperDARN CUTLASS HF radar pair have been combined to produce the first two-dimensional vector measurements of the convection pattern throughout the ionospheric footprint of a flux transfer event (a pulsed ionospheric flow, or PIF). Very stable and moderate interplanetary magnetic field conditions, along with a preceding prolonged period of northward interplanetary magnetic field, allow a detailed study of the spatial and the temporal evolution of the ionospheric response to magnetic reconnection. The flux tube footprint is tracked for half an hour across six hours of local time in the auroral zone, from magnetic local noon to dusk. The motion of the footprint of the newly reconnected flux tube is compared with the ionospheric convection velocity. Two primary intervals in the PIF's evolution have been determined. For the first half of its lifetime in the radar field of view the phase speed of the PIF is highly variable and the mean speed is nearly twice the ionospheric convection speed. For the final half of its lifetime the phase velocity becomes much less variable and slows down to the ionospheric convection velocity. The evolution of the flux tube in the magnetosphere has been studied using magnetic field, magnetopause and magnetosheath models. The data are consistent with an interval of azimuthally propagating magnetopause reconnection, in a manner consonant with a peeling of magnetic flux from the magnetopause, followed by an interval of anti-sunward convection of reconnected flux tubes.Key words: Magnetospheric physics (magnetosphere · ionosphere interactions; plasma convection; solar wind · magnetosphere interactions)

Ionospheric disturbances observed by DMSP at middle to low latitudes during the magnetic storm of June 4–6, 1991

This paper extends a recent study of electric field penetration into the inner magnetosphere observed by the Combined Release and Radiation Effects (CRRES) satellite and the Defense Meteorological Satellite Program (DMSP) satellite F8 during the magnetic storm of June 4–6, 1991, to consider its ionospheric consequences. Effects include the development of > 1 km/s subauroral ion drift (SAID) structures, the formation of midlatitude density troughs, and the vertical transport of equatorial plasma, bubbles. Nearly simultaneous auroral electron and plasma drift measurements were acquired by three DMSP satellites with F8 and F9 in one hemisphere and FlO in the other. Moderate to strong SAID structures were consistently detected for ∼10 hours during the early main phase of the storm. Weak SAIDs were encountered during ∼8 hours of the early recovery phase. DMSP data show that SAIDs with similar characteristics developed at magnetically conjugate locations and extended for at least 3 hours in local time. Simultaneous measurements show that the SAIDs spanned temporally grooving but latitudinally narrow plasma density troughs. These observations suggest that the magnetospheric sources of SAIDs act more like voltage than current generators. Energetic electron fluxes, electric fields, and plasma waves measured by CRRES indicate that during this storm the ring current shielding charges and SAID sources were located in regions of high plasma density characteristic of the plasmasphere. The sequence in which DMSP detected equatorial plasma density irregularities is consistent with model predictions that stormtime electrodynamics at low latitudes operate on distinctive fast and slow timescales [Fejer and Schcrliess, 1997].

Ionospheric current signatures of transient plasma sheet flows

The plasma flow in the central plasma sheet of the magnetospheric tail often includes short impulsive bursts. Here we investigate the ionospheric signatures of such bursts. The Wind satellite recorded several transient fast flows in the plasma sheet (at geocentric distances of ∼ 12 RE) on December 21–22, 1995. The data are compared with magnetic field observations made in the Scandinavian sector, at the ionospheric conjugate point of the satellite. Superposed epoch analysis of the satellite data suggests that most of the events are Earthward flow bursts accompanied by magnetic dipolarizations, increases in the convection electric field, and drops in the plasma pressure and density. Occasionally, also isolated tailward flow bursts within closed flux tubes can be observed. We demonstrate that in both cases the transient plasma sheet flows are systematically associated with distinct ground magnetic field variations which (after 90° rotation) have specific vortex‐like spatial distributions. The vortex patterns have similar duration to that of the flows at Wind and their longitudinal extent (≤1 hour in local time) is consistent with the azimuthal scale sizes (∼3 RE) of the transient flows reported earlier. In many cases the sense of flow rotation observed at Wind and at Wind's ionospheric footpoint agree with our expectation. Despite the care taken in accounting for the instantaneous and local currents that affect the mapping, uncertainties in the footpoint location may still be responsible of the absence of a higher degree for compliance with theory.

Sino-Magnetic Array at Low Latitudes (SMALL) including initial results from the sister sites in the United States

The Sino-Magnetic Array at Low Latitudes (SMALL) is an ongoing project for the development of a two-dimensional (2-D) ground-based magnetometer array in China. The ultimate array will consist of 24 stations spanning over 3 hours in local time in the low-latitude (L < 2) region. We present some initial results from the Beijing station as well as several U.S. sister sites using the same instrumentation. It is shown that the coverage, high sensitivity, and accurate timing of the SMALL array will advance our understanding of low-latitude geomagnetic phenomena.

Mapping daily and monthly radiation components using METEOSAT data

The aim of the study is to present actual daily and climatically representative monthly maps of the surface radiation balance components derived from METEOSAT data. The encrypted digital images are now available in every three hours in Hungary. Hence, the first problem to solve is the estimation of diurnal mean radiation components from the archived 1–3 observations of the day between 8 and 15 UTC (+1–2 hours in local time). This computation is performed by Lagrange-interpolation. Since the skill of these day-to-day estimations is not at all obvious, the estimation of monthly averages is performed in two steps. First, the hourly observations from the 1992–1996 period are averaged for each hour separately, and these, more regularly varying values are smoothed by the Lagrange-interpolation, in the second step. For the monthly radiation maps, our purpose was also to improve their climatic representativity by using stratified sampling, based on macrosynoptic classification, first described by Mika et al. (1994). This classification considers the surface pressure patterns in Central Europe and the reference climatology for them is originated from the 1961–1990 period. The differences between sample means and stratified means are in some cases considerable, as presented in the parallel figures of some months and components presented in the paper.

Simultaneous observations of solar wind plasma entry from FAST and POLAR

On November 15, 1996, NASA FAST and POLAR satellites obtained data in the cusp/cleft region for extended intervals. POLAR sampled a narrow range of local times. FAST observed cusp/cleft plasma for more than 5 hours in local time. At 07:51 UT the satellites were in near magnetic conjunction, and the interplanetary magnetic field was steady for an extended interval before this time. We confirm prior observations that show solar wind plasma often enters the magnetosphere over extended regions of local time. For the one hour interval of steady solar wind conditions when POLAR encountered cusp/cleft plasma, irregularly spaced ion injections interspersed with intervals of nearly constant cusp/cleft ion fluxes were observed. The data are consistent with temporal variations in the reconnection rate. At the time of the near conjugate observations the data suggest that solar wind plasma entered the magnetosphere over an extended region of the magnetopause, i.e. not in small discontinuous patches.

Equatorial density irregularity structures at intermediate scales and their temporal evolution

We examine high resolution measurements of ion density in the equatorial ionosphere from the AE‐E satellite during the years 1977–1981. Structure over spatial scales from 18 km to 200 m is characterized by the spectrum of irregularities at larger and smaller scales and at altitudes above 350 km and below 300 km. In the low‐altitude region, only small amplitude large‐scale (λ&gt;5 km) density modulations are often observed, and thus the power spectrum of these density structures exhibits a steep spectral slope at kilometer scales. In the high‐altitude region, sinusoidal density fluctuations, characterized by enhanced power near 1‐km scale, are frequently observed during 2000–0200 LT. However, such fluctuations are confined to regions at the edges of larger bubble structures where the average background density is high. Small amplitude irregularity structures, observed at early local time hours, grow rapidly to high‐intensity structures in about 90 min. Fully developed structures, which are observed at late local time hours, decay very slowly producing only small differences in spectral characteristics even 4 hours later. The local time evolution of irregularity structure is investigated by using average statistics for low‐(1%&lt;σ&lt;5%) and high‐intensity (σ&gt;10%) structures. At lower altitudes, little change in the spectral slope is seen as a function of local time, while at higher altitudes the growth and maintenance of structures near 1 km scales dramatically affects the spectral slope.

October 1995 magnetic cloud and accompanying storm activity: Ring current evolution

The passage at Earth of the October 1995 magnetic cloud and the high‐speed corotating stream overtaking it, monitored by the Global Geospace Science (GGS) spacecraft Wind, caused two consecutive geomagnetic storms: a major one during the strong Bz &lt; 0 nT phase of cloud passage and a moderate one during the intermittent Bz &lt; 0 activity in the fast corotating stream. Large dynamic pressure changes were observed in the sheath region ahead of the cloud and in the cloud‐stream interface region at its rear, resulting in substantial corrections to the measured Dst index. A burst of superdense plasma sheet extending over ∼2 hours in local time was observed at geostationary orbit during the second storm. We simulate the ring current development during this storm period using our kinetic model and calculate the magnetic field perturbation caused by the ring current. The plasma inflow on the nightside is modeled throughout the investigated period using data measured at geosynchronous orbit. The modeled Dst index is compared with the observed Dst values corrected for magnetopause and telluric currents. The temporal evolution of the ring current H+ and O+ distribution functions is computed, considering losses due to charge exchange, Coulomb collisions, and ion precipitation. We find that (1) the storm time enhancement of the plasma sheet ion population contributed significantly to the ring current buildup; (2) an additional ∼12 nT decrease in Dst is achieved when the symmetry line of the plasma convection paths is rotated eastward from the dawn‐dusk direction with 3 hours during the first storm; (3) the major loss process is charge exchange, followed by Coulomb collisions and ion precipitation; (4) however, the energy losses due to ion precipitation increase monotonically during the more active periods, reaching the level of Coulomb losses at peak storm intensity. We argue that the losses due to ion precipitation considered in this study are closely related to the enhanced convection electric field, which in our model is parameterized with the planetary Kp index. Correspondingly, we find that (5) there is a very good correlation between the variations in time of this index and the magnitude of the ion precipitation losses.

Substorm precipitation in the polar cap and associated Pc 5 modulation

This paper reports substorm‐related cosmic radio noise absorption events observed deep into the polar cap, well poleward of the typical nightside auroral oval. We find that such events usually occur in the late expansion/recovery phase of a substorm during moderate‐to‐strong planetary disturbances (Kp=3–8' and at times of high solar wind speed conditions. We present riometer data from U.S. manned and automatic geophysical observatories in Antarctica to track the morphology and progression of auroral absorption features over 5 hours of local time at ≥80° magnetic latitude. The azimuthal motions observed, westward pre‐midnight, eastward post‐midnight, are similar to the dynamics expected for lower latitude substorm‐related auroral absorption. ULF modulation of the absorption (and hence of the causative fluxes of precipitating energetic electrons' are often observed. These modulations are usually dominated by spectral peaks between 1–4 mHz and, in the one case examined in detail, were confined to a narrow, ∼1° latitude band embedded within more uniform precipitation. Pc 5 (150 to 600 s periods' magnetic variations with similar spectral peaks, and auroral arcs are often observed concurrently. Several mechanisms for producing the particle modulations are examined within the context of these observations.

Multipoint analysis of a bursty bulk flow event on April 11, 1985

In an attempt to place short‐lived, high‐speed magnetotail flows termed bursty bulk flow events (BBFs) in the context of substorm phenomenology we analyze one such event that took place on April 11, 1985, using data from several spacecraft and many ground stations. The substorm onset, which took place at 0127 UT, had a meridian 2 hours of local time east of AMPTE/IRM. The satellite did not detect high‐speed flows at that time. A high‐latitude (∼70° corrected geomagnetic) substorm intensification took place at 0202 UT centered ∼0.5 hour of local time west of the AMPTE/IRM meridian. The ISEE 2 satellite at the magnetotail lobe and the LANL 019 satellite at geosynchronous altitude were both at the same meridian as AMPTE/IRM at the time. The 0202 UT substorm intensification was associated with (1) a dipolarization at the ISEE 2 satellite at 0200:30 UT, (2) a BBF oiiset at AMPTE/IRM at 0202 UT accompanied by an intense dipolarization consistent with current wedge formation, (3) an energetic particle injection at geosynchronous altitude that took place at 0204 UT. The plasma acceleration region associated with this substorm intensification was estimated to be ∼ 8 RE tailward of AMPTE/IRM. Thus, during this activity the BBF event was due to an observed tail collapse Earthward of X ∼ −26 RE. The Earthward energy transport measured at AMPTE/IRM can account for the expected magnetospheric power consumption if the BBF has a cross‐sectional area of only 1–2 RE2 in the Y–Z direction. Similarly, the Earthward magnetic flux transport rate measured at AMPTE/IRM during the BBF event can result in a potential drop comparable to the expected transpolar cap potential if the BBF event has a size of 1–2 RE in the Y direction. The large amounts of flux transport measured past the satellite necessitate the existence of lobe flux reconnection tailward of AMPTE/IRM. The above results assume the validity of the frozen‐in condition over the ∼10‐min duration of the BBF event. Although activity continued in the ionosphere and the ring current for well over 1.5 hours after the 0202 UT substorm intensification, most of the earthward energy and magnetic flux transport past IRM had ceased ∼10 min after the BBF onset. We propose that the fast flows transport and pile up magnetic flux through a very narrow (a few RE in Y extent) flow channel in the midtail to the edge of an expanding dipolarization front in the near‐Earth region. After the plasma sheet dipolarizes at a given location enhanced flux transport ceases, resulting in an apparent short (10‐min timescale) duration of the fast flows. Unlike the near‐Earth plasma sheet, which dipolarizes across many hours of local time, the midtail plasma sheet may exhibit longitudinally localized dipolarization. This may explain the often observed lack of one‐to‐one correlation between midtail activity and substorms.

Observations of a solar‐wind‐driven modulation of the dayside ionospheric DPY current system

We report here a collection of complementary measurements near local magnetic noon of a unique ionospheric convection variation observed by the Sondrestrom radar which is related to a nearly periodic variation (≈ 25 ‐ 30 min period) in the interplanetary magnetic field By component. We observe also a poleward phase propagation of magnetic pulsations over a region limited in longitude and latitude near the dayside polar cusp. A series of poleward propagating radio absorption enhancements are observed in the Greenland and South Pole imaging riometer data. The pulsations and absorption enhancements are associated with a latitudinally narrow and longitudinally limited intensification of the westward convection and associated eastward Hall current, which propagates poleward over the magnetometers and radar field of view. For the cases presented the interplanetary Bz is strongly negative, while the ionospheric variations are associated with the low‐frequency component of variations in the interplanetary By component. In contrast to the previously discovered traveling convection vortices, these features exhibit a poleward phase motion rather than one along lines of invariant latitude. The propagation velocity is slower (≈ 0.5 ‐ 1.0 km/s) and the structures cover 2 to 3 hours of local time. We interpret the observations as a poleward propagation of the DPY current system intensification associated with enhancements in the IMF By component. Our observations indicate that the DPY field‐aligned current system is propagating poleward and may be moving independent of the convection motion of the plasma and associated field lines.

Locating the polar cap boundary from observations of 6300 Å auroral emission

We examine ground‐based observations of the meridional profile of 6300 Å atmospheric emission from 67.3° to 80.7° invariant latitude for the signature of the polar cap boundary, the ionospheric boundary between open and closed magnetic field lines. The open‐closed field line boundary is assumed to lie at the boundary between polar rain and plasma sheet precipitation. We assume that nonprecipitation‐dependent sources of 6300 Å emission cause a spatially uniform luminosity in the polar cap and that auroral zone luminosity is also spatially uniform. Therefore we determine the location of the polar cap boundary from the auroral emission data at each time by finding the best fit of the observations to a step function in latitude. Thus we produce a time series of the location of the polar cap boundary. We have developed criteria on the step function fit that identify when a reliable boundary identification has been obtained. Generally, where these criteria are not satisfied, the boundary is outside the latitudinal range of the optical observations. We compare the boundary identified from the emissions to the boundary in precipitating particle observations made by DMSP as it passes along a meridian within 1 1/2 hours of local time of the photometer. The latitudes of the two boundaries are highly correlated. During the expansion phase of substorms, however, there are large discrepancies apparently arising from longitudinal structure of the polar cap boundary associated with auroral surges. We conclude that 6300 Å emissions provide a good means for monitoring the polar cap boundary continuously with an estimated precision of ±0.9° invariant latitude.

High‐latitude F region electron concentration measurements near noon: A case study

A case study is presented of a ridge of enhanced F region plasma in the afternoon sector near the equatorward edge of the auroral oval. The study for September 26, 1982, principally exploits data from three ionosondes near 62° invariant latitude in the southern hemisphere separated by 5½ hours of local time but only 3½ hours of magnetic time. These observations demonstrate that the ridge structure is better ordered in magnetic time than in local time. Each station observes the ridge for about 4 hours. The poleward part of the ridge is rich in small‐scale (1–10 km) ionospheric irregularities. It is suggested that the entire ridge is formed by the mechanism described by Foster (1993) in which plasma initially corotates past noon at lower latitudes, before being entrained in the convection pattern in the afternoon sector, and brought back toward noon. The plasma thus has an extended trajectory in sunlight leading to high concentrations. The irregularities probably result from the combined effects of high electric fields and energetic particle precipitation. The contribution which this plasma is likely to make to plasma structures in the polar cap is discussed briefly.

Electric potential patterns in the northern and southern polar regions parameterized by the interplanetary magnetic field

Electric potential patterns have been obtained from the IZMIRAN electrodynamic model (IZMEM) for the northern and southern polar regions during summer, winter, and equinox. The model is derived from a large quantity of high‐latitude ground‐based geomagnetic data (above ± 57° corrected geomagnetic latitude) at all magnetic local time hours. A linear regression analysis technique has been used to obtain the quantitative response of each magnetic observatory to changes of interplanetary magnetic field (IMF) components. Since no ionospheric conductivity model exists specifically for the southern polar region, the statistical model of Wallis and Budzinski (1981) has been applied in both hemispheres. A cross‐polar “background” potential of ∼35 kV, derived by Reiff et al. (1981), is used to calibrate IZMEM's potential patterns. The model's responses to changes in the IMF By and Bz components are analyzed to obtain a set of “elementary” convection patterns in both polar regions for each season of the year. Asymmetry in the potential pattern geometry in both hemispheres can be attributed either to the influence of the “northern” ionospheric conductivity model which was applied to the southern polar region, or to some natural phenomena. The modeled background cross‐polar potential for the condition when Bz = By = 0 is found to be ∼37 kV. Average values of the modeled potential drop caused by each nanotesla of the IMF are the following: ∼14 kV for southward Bz; ∼ −4 kV for northward Bz; and ∼ ±4.5 kV for By components. The latter is not applicable to the “dawn‐dusk” potential drop; it may be applied across the cusp region only. Nevertheless, a combination of the background and elementary potential patterns in the case studies gives a certain estimation of the cross‐polar potential drop, which may be strongly distorted during time of large By. It is concluded that IZMEM provides realistic convection patterns parameterized by the IMF component directions and magnitudes and may be used to provide routine estimates of convection patterns and electric potentials if IMF data are available.

Mid‐latitude spread F occurrence related to geomagnetic activity at preferred local times

Superposed‐epoch methods have been used to investigate the level of spread F occurrence relative to high geomagnetic activity. The controls used were determined from high values of the AE index during two 6‐hour periods representing early morning intervals at two high midlatitude stations separated by 10 hours of local time. Delayed spread F occurrences at these stations (Lannion in France and Canberra in Australia) were considered using these controls. The results show that spread F occurrence is enhanced (delayed, however, sometimes by as much as 3 days) by using controls for the early morning periods of the station being investigated, but not when controls for the early morning periods of the other station are used. It is suggested that in examining relationships between other geophysical phenomena and geomagnetic activity, it may be profitable to consider geomagnetic activity at specific local times rather than employing the use of a daily index such as ΣKp.

Solar flare effects at Ebre: Regular and reversed solar flare effects, statistical analysis (1953 to 1985), A global case study and a model of elliptical ionospheric currents

This paper presents the results of the analysis of geomagnetic effects of solar flares (sfe) recorded at Ebre observatory (40.8° latitude N, 0.5° longitude E) during 33 years (1953‐1985). At Ebre, located near the focus latitude, two types of sfe can be observed: regular and reversed sfe. Regular sfe are those which have phase differences less than 90° with the regular diurnal magnetic variation of the day, SR. Reversed sfe are those which have phase differences greater than 90° with SR. From these 33 years, 140 sfe events were selected and a statistical study was performed. We found a local time dependence of the phase differences between the sfe and SR vectors. Morning hours have slightly positive values and afternoon hours have slightly negative ones. Reversed sfe, with a phase difference exceeding 90°, concentrate between 10 and 12 hours. Reversed sfe show a dominant equinoctial character. Also, a weaker correlation was found between solar activity with reversed sfe (r=0.47) than with regular sfe (r=0.68). Using data from 67 observatories, we performed a global study of a sfe case, seen at Ebre as reversed sfe. In this case, in the northern hemisphere, the sfe system was about 1 hour of local time eastward of the SR system and formed 4° higher in latitude. Finally, we present a model of two elliptical ionospheric equivalent current systems with focus offset about 1 hour in local time to explain the phase difference between the sfe and Sq magnetic vectors observed at Ebre. The parameters of this model have been fitted from the results of a previous statistical analysis from Ebre data. Spatial and temporal distribution of the sfe and Sq vector phases are calculated with this model, and conditions for reversed sfe occurrence are predicted.

Relationship between Birkeland current regions, particle precipitation, and electric fields

Data from eight DMSP F7 satellite passes coincident with Sondrestrom radar observations have been examined to determine how the large‐scale dayside Birkeland currents are related to the particle precipitation regions and to the convection pattern. The classification schemes recently developed from DMSP particle data were adopted. The observations were limited to the prenoon local time hours and led to the following conclusions: (1) The local time of the mantle currents (which were traditionally called cusp currents) is not limited to the longitude of the cusp proper, but covers a larger local time extent. (2) The mantle currents flow entirely on open field lines (where “open field lines” is defined as a region where the ion precipitation and electron precipitation have the characteristics of plasma mantle, cusp, or polar rain.) This confirms and extends to all local times similar results obtained from other observations. (3) About half of region 1 currents flow on open field lines. This is consistent with the assumption that the region 1 currents are generated by the solar wind dynamo and flow within the surface that separates open and closed field lines. (4) More than 80% of the Birkeland current boundaries do not correspond to particle precipitation boundaries. Region 2 currents extend beyond the plasma sheet poleward boundary; region 1 currents flow in part on open field lines; mantle currents and mantle particles are not coincident. (5) On most passes when a triple current sheet is observed (region 2, region 1, and mantle currents), the convection reversal is located on closed field lines. When only two current sheets are observed (either region 2/region 1, or region 1/mantle currents), the convection reversal is on open field lines. (6) The data appear to be more consistent with a topology such that mantle currents are an extension of region 1 currents, rather than a separate system located poleward of the region 1 current system.

Observations of nightside auroral cavities

Radar observations of field‐aligned auroral F region density depletions (cavities) have been identified in a portion of the Sondre Stromfjord incoherent scatter radar (ISR) data base covering the period February 1986 to January 1988. These “auroral cavities” are nightside phenomena with localized, field‐aligned F region density depletions of 20 to 70 percent below surrounding values. They occur during moderate to quiet geomagnetic conditions when the poleward edge of the auroral oval is within view of Sondre Stromfjord. Seasonally, they are a wintertime phenomena occurring just poleward of the statistical auroral oval. Unlike the previously reported “polar hole,” the average width of the cavities is less than 100 km. Case studies show that the cavities closely track the poleward edge of the most poleward auroral arc. Sequential radar scans show that cavities appear on time scales as short as several minutes, suggestive of local electrodynamic formation or rapid transport. Data from January 24, 1987, collected during coordinated optical, radar, and satellite observations spanning an hour of local time, were examined for possible cavity formation mechanisms. The cavity formation processes examined herein include locally enhanced chemical loss, vertical diffusion, drifting horizontal gradients in the background plasma, and evacuation as a result of field‐aligned currents. The formation time scales, calculated evacuation fluxes, close proximity to E region aurora, and field‐aligned current signatures seen in magnetometer and radar observations suggest a strong association of the cavities with upward flowing electrons carrying region 1 downward field‐aligned currents.