1. Introduction
The equatorial and low latitude ionospheric region hosts a variety of processes, including the occurrence of ionospheric irregularities, commonly known as equatorial spread F (ESF), as well as the enhancements in the neutral temperature near the local midnight—the midnight temperature maximum (MTM). The crests of the equatorial ionization anomaly (EIA), typically located between 10° and 20° magnetic latitude, depending among other factors on solar activity, can be used to define the boundaries of the low latitude region. Poleward from the crests, other processes are observed, like medium-scale travelling ionospheric disturbances (MSTIDs). One then can expect the region between ~15° and 25° magnetic latitude to be a transition zone from the dominance of equatorial to mid-latitude physics.
All-sky imagers (ASIs) are optical instruments that provide information about these processes by measuring airglow emissions at different wavelengths. When looking at the 630.0 nm emission, the result of the dissociative recombination of O2+, followed by the relaxation of the excited atomic oxygen state O (1D), an all-sky imaging system can sample an area of ~106 km2 at 250 km, and thus one can measure these processes when they do not occur directly overhead. For example, from an instrument located at ~20° geographic latitude, like the one at the Arecibo Observatory, structures seen in 630.0 nm to the south at a ~75° zenith angle are occurring at ~10° geographic latitude, very close to the region where the northern crest of the EIA can be found.
Boston University has established a network of ASIs with sites at opposite ends of the same geomagnetic field lines in each hemisphere, i.e., geomagnetic conjugate sites [
1]. The ASIs are autonomous instruments that operate in mini-observatories situated at four conjugate pairs in North and South America, plus one pair linking Europe and South Africa. In this paper, we describe results from two pairs of ASIs in the American longitude sector, Arecibo (18.3° N, 66.7° W) and Mercedes (34.7° S, 59.4° W), and El Leoncito (31.8° S, 69.3° W) and Villa de Leyva (5.6° N, 73.5° W). Raw images need to be processed in order to assign a latitude and longitude to every pixel. This implies that an emission height needs to be assumed, and during low solar activity, 250 km is typically used. For a more quantitative way to determine the emission altitude, the Boston University Airglow Model [
2] is used. In this model the airglow emission altitude is calculated using ionospheric profiles from IRI-2016 [
3] and thermospheric profiles from NRLMSISE-00 [
4].
Figure 1 shows a map with the fields of view of the imagers, assuming an emission height of 250 km and zenith angles less than 80°. The dashed lines represent apex height contours of 750 and 2400 km, corresponding to magnetic latitudes of 15° and 30°, respectively. During low solar activity, the common feature observed at these sites is the wave-like pattern of medium-scale travelling ionospheric disturbances (MSTIDs). During mid and high solar activity, ESF structures tend to be more common, with some reaching apex heights (height of a magnetic field line at the geomagnetic equator) as high as ~2000 km. Thus, in addition to seeing ESF signatures frequently at El Leoncito and Villa de Leyva, the “low latitude sites”, close to ~15° magnetic latitude, they can also be detected at Mercedes and Arecibo, the “mid-latitude sites”, close to ~25° magnetic latitude. The zenith locations of the Mercedes and Arecibo ASIs are poleward from the crests of the equatorial ionization anomaly and the processes typically observed at the equatorial and low-latitude sites (like ESF) do not frequently reach zenith.
The work by [
5] showed examples of airglow depletions reaching the field of view of the Arecibo all-sky imager that were associated with equatorial spread F. A case shown in that study used the El Leoncito ASI, in the southern hemisphere, with a field of view overlapping with a small region of the mapped Arecibo ASI, and detected a depletion that was also observed in the Arecibo ASI. This was evidence of the geomagnetically conjugate behavior of the structures associated with ESF; but, as stated above, ESF-related airglow depletions are not very common at the Arecibo latitudes. Reference [
6] analyzed the occurrence rate of MSTIDs, a more common phenomenon at these latitudes, during the 2002–2007 period and showed that peaks of ~60% were observed during the winter and summer solstices. The unusually high winter peak, not observed at other longitude sectors [
7], prompted speculation on the influence of the opposite hemisphere on the formation of MSTIDs. Driven by these previous studies, an all-sky imager was installed at Mercedes, near the conjugate point of Arecibo. As a result, reference [
8] showed simultaneous and conjugate observations of MSTIDs at the two sites and proposed that the particular conditions in this longitude sector in the southern hemisphere, like the presence of the South Atlantic Magnetic Anomaly and a high occurrence of sporadic E layer structures, were responsible for the patterns observed.
This study will investigate the occurrence of ESF-related airglow depletions at Arecibo and will compare magnetically conjugate observations using the imagers at Arecibo and Mercedes and, to the west, at El Leoncito and Villa de Leyva.
3. Discussion
Different characteristics of ESF-related airglow depletions have been investigated: the seasonal occurrence of conjugate ESF depletions and the latitudinal extent and zonal velocity of depletions observed simultaneously in both hemispheres. ESF effects are observed at magnetic latitudes as high as ~25°. Previous case studies of conjugate ESF signatures using all-sky imagers showed an exact one-to-one correlation in the latitudinal extent at both hemispheres [
19,
20]. We have observed depletions in the northern hemisphere that extend to higher latitudes than the corresponding conjugate depletions in the southern hemisphere.
Extreme cases with depleted structures observed in GPS-TEC and airglow data were reported recently [
21,
22,
23]. These structures were associated with ESF and reached magnetic latitudes as high as 40°. Very active geomagnetic conditions were present during these observations. In this work, we show that airglow depletions can reach latitudes that usually are considered outside the region of influence of equatorial processes. Even though the percentage rates are low, as shown in
Figure 2 and
Figure 4, the detection at mid latitudes of a process—typically constrained to occur at low latitudes—reflects the dynamical coupling between the equatorial and mid-latitude regions. Modeling efforts have not been successful in reproducing ESF-structures reaching apex heights higher than 1500–2000 km. While semi-empirical modeling offered a buoyancy mechanism to account for the >1000 km apex height plumes [
24], subsequent full 3D numerical modeling did not achieve similar results [
25,
26].
In order to determine in a more quantitative way the latitudinal extent of the ESF-related airglow depletions, images from the Villa de Leyva–El Leoncito pair are calibrated to get airglow brightness in Rayleigh units (1 R = 10
6 photons/cm
2secstr) and horizontal cuts are taken through a Villa de Leyva image on 10 March 2015 at two fixed geographic latitudes, 7° and 9°, indicated in
Figure 11 (left) as the red horizontal lines. The conjugate locations for every point along the lines have been computed using the IGRF-12 magnetic field model [
27] to create the red curves at El Leoncito (right).
At this time the only common depletions observed are labelled B and B*. Depletion B bifurcates into two structures right above zenith in the Villa de Leyva image. We compared the brightness of the two branches of this depletion. Peak airglow at Villa de Leyva is around 1 kR near zenith due to the presence of a very strong northern crest of the EIA. El Leoncito airglow emission is much lower, peaking only at about 70 R. The two conjugate cuts plotted over El Leoncito show that although the depletions reach over 1100 km at Villa de Leyva they do not extend that far in the southern hemisphere.
Figure 12 shows the brightness values along the fixed latitudes on the Villa de Leyva ASI (blue line). The values for El Leoncito are shown in orange and are mapped along magnetic field lines back to the northern hemisphere in order to display them in the same coordinates as Villa de Leyva.
The data has been normalized to the maximum value along the lines. The left plot corresponds to the 9° N cut, showing no depleted airglow in the El Leoncito data, while the right plot, corresponding to the 7° N cut, shows very clearly the depletions at Villa de Leyva and El Leoncito (an average curve, in green, was overlaid to help identify the two depletions observed at El Leoncito).
We next explore the potential reasons for the different latitude reach: height chosen to unwarp the images, observational limitations due to weak airglow, and different background conditions due to meridional winds. The altitude of the airglow emission could contribute to the different latitudinal extent of the depletions. Recently, reference [
2] showed how variations in the assumed airglow emission altitude can lead to apparent interhemispheric asymmetries in the of ESF depletions, in particular when looking at the tilt of the depletions. For the cases presented in
Figure 3 and
Figure 7, in order to have the depletions reaching the same apex heights in both hemispheres, the images from the southern hemisphere sites would have to be unwarped ~200 km higher than the heights used in the analysis, an unrealistic situation because this would imply a peak emission height close to 500 km. Changes in the 630.0 nm airglow emission can be a result of changes in electron density or changes in the height of the F layer. Another wavelength used in all-sky imagers is 777.4 nm, sensitive only to changes in total electron content (TEC). In this study, similar results are observed in the 777.4 nm depletions, proving that the lower extent of the El Leoncito depletions is not due to limitations in detectability at 630.0 nm. While the low levels of airglow emission at El Leoncito are not affecting the detectability of the latitudinal extent of the depletions, the difference in brightness between the two sites may offer some insight into why the depletions extent is different. A difference in the background emission at two conjugate sites is a regular occurrence. The gray color scale next to each image in
Figure 11 shows that on 29 November 2014 the background 630.0 nm emission is greater at Villa de Leyva than at El Leoncito by a factor of about 15. Using the Boston University airglow model [
2] with NRLMSISE-00 and IRI-2016 as inputs, the brightness computed over El Leoncito was 70 R and over Villa de Leyva it was 170 R, only a factor of ~2.5 higher. In order to further investigate the large asymmetry observed in the brightness we looked at the total electron content (TEC) from GPS receivers. At El Leoncito, 20 TECu (1 TECu = 10
16 el/m
2) was measured at 04:00 UT and over Bogota, a station close to Villa de Leyva, a value of 80 TECu was measured. These measured TEC values can be compared with the TEC outputs from IRI-2016: 19 TECu over El Leoncito, close to the measured value, and 12 TECu over Bogota, ~6.5 times smaller than the measured value. The ratio of measured and modeled TEC over Bogota is similar to the ratio of measured and modeled emission (~7) and thus it accounts for most of the discrepancy between the modeled emission (170 R) and the measured emission (1.2 kR). Plasma redistribution is occurring and this is related to meridional neutral winds that have been previously shown to impact plasma distribution at equatorial and low latitudes [
28]. A trans-equatorial wind blowing from the summer hemisphere to the winter hemisphere drags the ions due to ion-neutral collisions and the presence of the magnetic field forces them to move up first and, after they cross the magnetic equator, down along the magnetic field lines, producing an enhancement in TEC and, consequently, in the 630-nm airglow intensity. Meridional winds that are in the same direction in both hemispheres have been shown to modify the equatorial plasma distribution [
25]. Average results from Horizontal Wind Model 14 (HWM-14) [
29] show that during the December solstice a strong northward wind is present during the night. Specific runs confirm these results, with meridional winds close to 100 m/s at El Leoncito and around 30 m/s at Villa de Leyva. Therefore, northward winds in both hemispheres are likely the major cause of the differences in airglow brightness, TEC, and the latitudinal extent of the airglow depletions.
The second topic studied here included the zonal plasma velocities at conjugate locations in both hemispheres. Several cases using the Arecibo–Mercedes pair have consistently shown the southern hemisphere values larger than the northern hemisphere ones. In addition, case studies using the Villa de Leyva–El Leoncito pair, located closer to the magnetic equator, also show higher velocities in the Southern hemisphere. During the COPEX (Conjugate Point Equatorial Experiment) campaign, conjugate imagers in the Brazilian region at ~ ±10° magnetic latitude were used to compare the characteristics of the conjugate airglow depletions and the conjugate zonal velocities [
15,
20]. The campaign included GPS receivers that were used to obtain the drifts by measuring how scintillation patterns moved at each location. The main results indicated that there was a perfect symmetry between the depletions observed at both hemispheres. When comparing the zonal drifts inferred from the motion of airglow depletions, they found that they were very similar. These results are different to the ones obtained here.
Model calculations shown in [
30] can be used to explain the observations presented in this paper. Using Quasi Dipolar coordinates (a non-orthogonal system that defines magnetic latitudes and longitudes such that they are constant along field lines) and the IGRF-12 magnetic field model, the zonal plasma velocity
v is calculated from
v =
E ×
B/
B2, where
E is the electric field and
B is the Earth’s magnetic field. The electric field is mainly due to the F-region dynamo given by the following equation:
E = −
U ×
B, where
U is the Pedersen conductivity-weighted neutral zonal wind and
B is the Earth’s magnetic field. These equations can be used to determine the ratio of the velocities at the two sites. The magnetic field is easily calculated at both sites. The electric field requires a more complex analysis, as discussed in [
30] where the authors set up a system of basis vectors that are used to map the electric field along magnetic field lines. Mapping electric fields in a QD coordinate system allows for velocities to be mapped along field lines as well using the equation defining
v. In their paper they show that the drift velocity can be separated into two components, magnetic zonal and magnetic meridional, both constant along magnetic field lines, by using their basis vectors. Using this technique, they create a map showing the difference in magnetic east and magnetic north velocities between conjugate locations, meaning that it is possible to compute, for example, how much the eastward component varies when it is mapped into the conjugate hemisphere. This is shown in
Figure 13 where the map for the zonal component is reproduced. For a purely dipolar geomagnetic field on a spherical Earth all the values in
Figure 13 would be 1. Values greater than these idealized dipolar values are plotted with red contours, while values less than the dipolar values are plotted with blue contours. If we consider a horizontal 100 m/s ExB velocity toward magnetic east at magnetic coordinates 30° N, 0° E, 250 km altitude, its horizontal component at the conjugate point is about 145 m/s (reading a scale factor of about 1.45 at the southern conjugate point). If we start with a 100 m/s horizontal magnetic eastward ExB velocity at magnetic coordinates 30° S, 0° E, we see that the velocity in the conjugate hemisphere is about 70 m/s magnetic eastward. These are the deviations from a purely dipolar mapping. This is the result of assuming that the electric potential along a field line is constant, so the potential difference between the two adjacent field lines in the zonal direction is also constant along a magnetic field line. The electric field (or negative of the gradient of the electric potential) and the magnetic field are geometry dependent, so we can conclude that the ExB velocity in one hemisphere has a different magnitude and direction than the velocity at the conjugate location. The contour lines show that zonal velocities near El Leoncito and Mercedes are on average about 1.3–1.4 times greater than the velocities at their conjugate locations. If the Earth’s magnetic field were a pure dipole then the model would show no difference in velocity. This ratio of velocities is consistent with our results, i.e., higher velocities in the Southern hemisphere.