HYSELL ET AL.: ROCKET INVESTIGATIONS OF EQUATORIAL SPREAD F 1

Rocket and radar investigation of background electrodynamicsand bottom-type scattering layers at the onset of equatorialspread F

D. L. Hysell

Earth and Atmospheric Science, Cornell University, Ithaca, New York

M. F. Larsen

Physics and Astronomy, Clemson University, Clemson, South Carolina

C. M. Swenson and A. Barjatya

Electrical and Computer Engineering, Utah State University, Logan, Utah

T. F. Wheeler

Electrical Engineering, Pennsylvania State University, University Park, Pennsylvania

T. W. Bullet

Space Vehicles Directorate, Air Force Research Laboratory, Hanscom AFB, Massachusetts

M. F. Sarango, R. F. Woodman, and J. L. Chau

Jicamarca Radio Observatory, Instituto Geofısico del Peru, Lima

D. Sponseller

Kwajalein Range Services LLC., ALTAIR Radar, United States Army Kwajalein Atoll - Reagan Test Site,Republic of the Marshall Islands

Abstract.Sounding rocket experiments were conducted during the NASA EQUIS II cam-paign on Kwajalein Atoll designed to elucidate the electrodynamics and layerstructure of the postsunset equatorial F region ionosphere prior to the onset ofequatorial spread F (ESF). Experiments took place on August 7 and 15, 2004,each comprised of the launch of an instrumented and two chemical release sound-ing rockets. The instrumented rockets measured plasma number density, vectorelectric fields, and other parameters to an apogee of about 450 km. The chemicalrelease rockets deployed trails of trimethyl aluminum (TMA) which yielded windprofile measurements through ground-based photography and triangulation. TheAltair radar was used to monitor coherent and incoherent scatter in UHF and VHFbands. Electron density profiles were also measured with rocket beacons and anionosonde. Strong plasma shear flow was evident in both experiments. Bottom-type scattering layers were observed below the shear nodes in westward-driftingplasma strata. The layers were patchy and distributed periodically in space. Theirhorizontal structuring was reproduced in the large-scale plasma depletions thatformed later during ESF conditions. The morphology of the bottom-type layerirregularities, which appear to have been produced by wind-driven gradient drift in-stabilities, was therefore predictive of the ESF that followed. We discuss a scenariowhere the large-scale plasma waves, present at the onset of ESF and highlightedby the bottom-type layer structure, arose from collisional shear instabilities.

Introduction

This manuscript describes a NASA sounding rocket in-vestigation into the electrodynamics of the postsunset equa-torial F region ionosphere and the bearing this has on the thin“bottom-type” scattering layers that form there. Coherentscatter from such layers has been observed at the JicamarcaRadio Observatory in Peru for many years and generally oc-curs as a precursor to more fully developed equatorial spreadF (ESF) [Woodman and La Hoz, 1976; Hysell and Burcham,1998; Kudeki and Bhattacharyya, 1999]. The layers have notbeen investigated with sounding rockets before and are diffi-cult to study with satellites, being typically no more than 50km thick and residing at altitudes below 400 km and as lowas 200 km during solar minimum. We study them and theenvironment in which they reside to establish the causal linkwith ESF and the communications and navigation systemsoutages that result.

The investigation was carried out in August, 2004, aspart of the NASA EQUIS II campaign from Kwajalein Atollin the Marshall Islands using a combination of soundingrockets and with the support of the Altair VHF/UHF radar,which functioned both in coherent and incoherent scattermode [Tsunoda et al., 1979; Close et al., 2002]. Two sets

Figure 1. Experiment geometry. The instrumented andchemical release rocket trajectories are shown, respectively,by the blue and cyan lines radiating from Roi Namur at thenorthern tip of Kwajalein Atoll. Solid red contours representthe altitude in kilometers for field aligned backscatter fromthe Altair radar on Roi Namur. Dashed contours indicate thegeomagnetic dip angle in degrees.

of launches took place on separate nights from the Roi Na-mur range at the northern tip of the atoll. Each set consistedof an instrumented payload launched north-westward andtwo chemical release payloads, one launched along nearlythe same azimuth as the instrumented payload, and the otherlaunched to the northeast. The instrumented payloads mea-sured plasma density, electron temperature, electron colli-sion frequency, and electric field profiles to an altitude ofabout 450 km while the chemical release payloads permit-ted the measurement of neutral wind profiles at three dis-tinct locations up to about 220 km altitude though pho-tographing and triangulation of chemiluminiscent trimethylaluminum (TMA) trails [Larsen and Odom, 1997]. In addi-tion, the chemical release payloads were outfitted with dualfrequency beacons. In a horizontally stratified ionosphere,electron density profiles can be inferred from beacon signalsreceived on the ground (on Roi Namur) using an inversionalgorithm. Figure 1 shows a plan view of the experiments.

The Air Force Research Laboratory (AFRL) Space WeatherCenter of Excellence also installed a Digisonde PortableSounder (DPS-4) on Roi-Namur to support all EQUIS IIlaunches. The ionosonde was operating by June 31, 2004,with modes adjusted and system performance optimizedthrough the first two weeks of July. The system made iono-grams every 5 minutes throughout the campaign, providinga general overview of ionosphere conditions, critical fre-quency measurements, and derived electron density profilemeasurements. AFRL intends to operate the ionosonde onRoi Namur indefinitely.

The objectives of the investigation were to 1) understandand quantify vertical shear in the horizontal plasma drift inthe bottomside F region around sunset, 2) identify the mech-anism producing the bottom-type scattering layers that formin westward-drifting strata and serve as precursors for ESF,and 3) assess the influence of shear flow on the overall stabil-ity of the postsunset equatorial F region. The remainder ofthe manuscript summarizes the preliminary findings of theinvestigation based on data from the instrumented rockets,the Altair radar, and the ionosonde.

Observations

The first experiment was conducted on August 7, 2004following the appearance of F region coherent scatter in theAltair radar scans at about 1930 SLT. Figure 2 shows UHF(422 MHz) radar data for a west-to-east scan lasting pre-cisely 8 min. and ending at 0840 UT. Note that SLT ≈ UT+ 11 hr on Kwajalein. Local sunset occurred at about 0710UT in early August, and E region sunset about 40 min. later.For this scan, the radar pointing was programmed so that themain beam of the antenna was directed perpendicular to the

2

HYSELL ET AL.: ROCKET INVESTIGATIONS OF EQUATORIAL SPREAD F 3

Figure 2. Altair radar perpendicular scan for August 7,2004, 0840 UT. Top panel: backscatter power, range cor-rected and scaled to electron density on a logarithmic scale.Top right panel: electron density for beam position closest tozenith. Bottom panel: Coherent scatter Doppler velocity ona color scale spanning ± 890 m/s. In this panel, blue (red)tones denote blue shifts (red shifts).

geomagnetic field at F region altitudes at all times. A 300 µs,3-bit coded pulse was transmitted. The received signal wasdecoded, noise corrected, range corrected, and converted toelectron density estimates using a calibrated system constant[Bowles et al., 1962]. No further corrections are required atnight when the electrons and ions are in thermal equilibrium.Peak electron densities thus calculated from Altair radar datacompare well with estimates derived from the ionosonde aswell as with in situ data from the sounding rockets in a com-mon volume.

Figure 2 exemplifies a number of features common tothe ten nights of observations falling between August 3–15that constituted our experiments. There is a tilt in the post-sunset F region arising from the local time variation acrossthe radar field of view that persists until about 2000 SLT.Note that the tilt is oriented so as to be unstable to eastwardwind-driven gradient drift instabilities [Kudeki and Bhat-tacharyya, 1999]. In addition, large-scale waves (with a pre-dominant wavelength of about 200 km in this case) werealso frequently observed. Such waves were visible amongthe first of the radar scans, starting at about 1840 SLT each

Figure 3. Kwajalein ionogram from August 7, 2004, 0900UT. The vertical axis is range delay in km, and the horizon-tal axis is plasma frequency in MHz. Ordinary and extraor-dinary mode echoes are shown in red and green. The virtualheight trace and derived electron density profile are shown asblack lines. Intense sporadic E layers mask the F layer be-low 5 MHz. The F layer peak is visible and shows signs ofboth range and frequency type spread developing. Multiplecolors in the spread trace represent a wide range of observedDoppler shifts.

evening, on every night when spread F eventually occurred.These waves could be viewed either as “seed irregularities”for Rayleigh-Taylor type instabilities or as the early stagesof instability. Either way, as they reliably preceded the on-set of the broadband irregularities associated with ESF, thewaves by themselves suggest a forecast strategy.

Figure 2 also shows strong coherent backscatter from asporadic E layer (Es) at about 105 km altitude. Sporadic Elayers occurred regularly at sunset but tended to break upand vanish as the evenings progressed. The Es developedhere around 0400 UT and persisted until local midnight. Itbecame particularly intense between 0630 UT and 0930 UT,with blanketing Es frequencies (fbEs) sometimes exceeding6 MHz and critical frequencies (foEs) exceeding 15 MHz.The presence of sporadic E severely limited our ability todetermine minimum F layer heights and bottomside electrondensity profiles from ionograms, but accurate foF2 estimateswere still obtainable (see Figure 3), and it was clear that theF layer was rising prior to the rocket flight.

Note that strong coherent echoes are plotted in Figure 2using a different color scale when they exceed a certain SNRthreshold; this and the presence of range sidelobes due tocoding tend to obscure the sporadic E layer in Figure 2, al-though the layer is clear and distinct in scans performed ear-lier with finer range resolution. Those observations are plot-

HYSELL ET AL.: ROCKET INVESTIGATIONS OF EQUATORIAL SPREAD F 4

Figure 4. Altair radar scans of the sporadic E layer in theplane perpendicular to B. The upper and lower panels showUHF data taken with an 88 baud binary phase coded pulseand VHF data taken with a 13 bit Barker code, respectively.The baud length was 1 µs for both frequencies.

ted in Figure 4, which shows the echoes recorded both atUHF and VHF (158 MHz) frequencies. Whereas the UHFsignals are a combination of coherent and incoherent scat-ter, the VHF signals are purely coherent scatter. The latterare plotted here on a signal-to-noise ratio scale. The formerare scaled and plotted to electron densities, which is appro-priate for incoherent scatter but not for the coherent scat-ter, which is presumably co-located with the VHF echoes.The layer appears to be continuous but highly structured andis reminiscent of those observed and documented by Millerand Smith [1975, 1978] at Arecibo. Layer densities increasefrom west to east, and double layers are evident at some az-imuths. Density profiles recovered from the beacons on boththe chemical release rockets also indicated strong sporadic Elayers during the experiment, with a double layer observedon the beacon launched to the northeast.

Finally and most relevant to this investigation, Figure 2shows a bottom-type scattering layer between 200–250 kmaltitude and mainly to the west of the radar. Coherent scatterfrom the layer was relatively weak — weaker than the inco-herent scatter from 300 km altitude, for example. We knowthat it was coherent scatter because the correlation time ofthe echoes was very long. The lower panel of Figure 2 rep-

resents the Doppler shift derived from the 200 µs lag prod-uct computed from the radar echoes prior to decoding. Val-ues are plotted only when the coherence exceeds a certainthreshold. The bottom-type layers exhibited small negativeDoppler shifts, implying a combination of upward and west-ward drifts perpendicular to B, depending on the zenith an-gle. As was generally the case, the layer was patchy ratherthan continuous.

Rocket launches into this event began with the launchof the instrumented rocket (Terrier Malemute 29.036) at0852:56 UT, followed by the launch of the two chemical re-lease rockets (Terrier Orions 41.043 and 41.045). Figure 5depicts the ionosphere during the time of the instrumentedrocket flight. In the time elapsed from the previous scan,the large-scale wave amplitude increased, and the bottom-type layer began to dissipate. The distance between the twomost distinct crests of the large-scale wave can be seen tobe about 200 km, a distance replicated by the distance be-tween the bottom-type layer patches. The arc in Figure 5represents one way of comparing radar and rocket data. Ittraces the trajectory of the rocket, mapped along geomag-netic field lines to the plane perpendicular to B scanned bythe radar. This mapping is most appropriate for comparingirregularities in the two datasets. The arc suggests that therocket passed through a weak patch of irregularities on itsupleg between 200–250 km altitude and may have done soagain on the downleg between 225–275 km. The figure alsoshows a topside depletion drifting into the radar field of viewfrom the west.

Figure 6 shows upleg data from the August 7 rocket ex-periment. The electron densities were derived by the UtahState University (USU) Swept Langmuir Probe, and the vec-tor electric fields from the Penn State University (PSU) E-field experiment. The density profiles, normalized to theUSU Plasma Impedance Probe for absolute calibration, sug-gest a postsunset F region ionosphere with a steep bottom-side density gradient and a kink in the density at 350 kmassociated with the upwelling seen in Figure 5. A strongsporadic E layer is evident in the density profile, as are ad-ditional layers in the valley that may be termed intermediatelayers. The zonal electric field was relatively small through-out the upleg except around 275 km where the vehicle en-tered the upwelling. The tendency for vertical plasma driftsin the postsunset F region to show structure, with ascentrates peaking in the bottomside, is a feature routinely ob-served at Jicamarca and associated with the demands of in-compressible flow in a layer with zonal gradients in the zonalplasma drifts [Pingree and Fejer, 1987]. The 1.5 mV/m fieldimplies ascent at the rate of 50 m/s. Finally, the vertical elec-tric field profile is indicative of strong shear flow, with theionospheric plasma moving eastward at up to 190 m/s above

HYSELL ET AL.: ROCKET INVESTIGATIONS OF EQUATORIAL SPREAD F 5

Figure 5. Altair radar perpendicular scan for August 7, 2004, 0858 UT. The white arc represents the location of the instrumentpayload, mapped along geomagnetic field lines to the plane perpendicular to B scanned by the radar. Echoes between 150–200 km apparent altitude are due to tropospheric clutter.

Figure 6. Upleg data from the August 7 rocket flight. Leftpanel: electron density. Center panel: zonal electric field.Right panel: vertical electric field.

250 km and westward at up to 50 m/s below 250 km. Arapid eastward plasma jet at 275 km is co-located with theupwelling and presumably associated with the electric fieldof the growing large-scale wave.

All of the profiles demonstrate that the payload inter-cepted intermediate- and small-scale bottom-type plasma ir-regularities in the interval between 200–250 km altitude.The irregularities existed in the rarefied valley region, wherethe density profile was nearly flat, rather than the steep bot-tomside, where Rayleigh-Taylor type instabilities are ex-pected to occur. While the relative RMS density fluctuationswere significant, of the order of 50%, the absolute densityfluctuations were consequently small. Moreover, the irregu-

Figure 7. Downleg data from the August 7 rocket flight.Left panel: electron density. Center panel: zonal electricfield. Right panel: vertical electric field.

larities were confined to the westward-drifting strata belowthe shear node. Finally, the irregularities were anisotropic,with the vertical electric field components being signifi-cantly stronger than the zonal field components.

The downleg data shown in Figure 7 are substantiallysimilar to the upleg data. The steep bottomside F regiondensity gradient along with layers in the valley and E re-gion remain present. The zonal electric field profile showsthat the vehicle flew through a region of downwelling above300 km. Strong shear flow is again present, with the shearnode falling at about 280 km. Plasma irregularities wereencountered between 225–275 km altitude, although thesewere weaker than the upleg irregularities in both relative and

HYSELL ET AL.: ROCKET INVESTIGATIONS OF EQUATORIAL SPREAD F 6

Figure 8. Power spectra of the bottom-type layer electricfields measured on August 7, 2004. A frequency of 1 Hzcorresponds roughly to a wavelength of 1.8 km.

Figure 9. Altair radar oblique scan for August 7, 2004, 0952UT.

absolute senses.

Figure 8 shows the power spectra of the zonal and verticalelectric fields measured in the bottom-type layers interceptedon the upleg and downleg. The spectra seem roughly to obeypower laws at high frequencies and are clearly anisotropic,with the vertical field fluctuations at the largest (kilometric)scales (i.e., lower frequencies) being significantly strongerthan the zonal field components.

Figure 9 shows an Altair radar UHF scan conducted ap-proximately one hour after the rocket launches. This oneused a 300 µs uncoded long pulse and was made with theradar beam directed 5◦ off perpendicular so that only inco-herent scatter was received. The scan shows the state of the

Figure 10. Altair radar perpendicular scan for August 15,2004, 0813 UT.

topside depletions from Figure 5 by the time they driftednearly overhead. The depletions are tilted steeply westwardand show evidence of bifurcation. There is considerable finestructure on the depletions superimposed on a dominant 200km wavelength undulation. The precursor waves from theearliest Altair scans seem therefore to have anticipated themorphology of the ESF that followed.

The second group of rockets was launched on August15. The conditions that initiated the experiment are depictedin Figure 10. The ionosphere was tilted and already pop-ulated by large-scale irregularities. Strong sporadic E layerechoes were also present. A topside depletion was beginningto form several hundred kilometers to the east of the radar.Most importantly, a strong bottom-type scattering layer wasforming directly overhead. The layer was patchy rather thancontinuous and displayed regular, 30 km spacing betweenpatches. Doppler information shows that the topside deple-tion was ascending rapidly and that the layer was driftingslowly upward and/or westward.

The Es conditions leading up to the second launch werefar less severe than for the first, affording the ionosondemuch better bottomside F layer electron density profiles (seeFigure 11). Blanketing Es frequencies were typically below3 MHz, although foEs reached 10 MHz during the secondinstrumented rocket flight. The F layer heights rose contin-uously from 0630 to 0815 UT. The ionogram F trace began

HYSELL ET AL.: ROCKET INVESTIGATIONS OF EQUATORIAL SPREAD F 7

Figure 11. Kwajalein ionogram from August 15, 2004, 0815UT. The F layer trace is extensively bifurcated and clearlya separate or secondary trace. The brighter red color of thesecond trace is indication of positive Doppler shift, distinctlydifferent from the consistently negative Doppler shift of theprimary trace.

Figure 12. Altair radar perpendicular scan for August 15,2004, 0824 UT.

to show disturbances at 0745 UT, significant bifurcation at0800 UT, and distinct multiple traces at the F layer peak by0815 UT. By 0830 UT, there would be significant frequency-type spread of the entire F layer which would developed intorange-type spread by 0845 UT. Range-type spread wouldcontinue until 1330 UT.

Within 10 minutes of the acquisition of Figure 10, thetopside depletion accent had accelerated, and the bottom-type layer had expanded horizontally and become somewhatstronger while maintaining its 30 km periodicity (see Fig-ure 12). An instrumented rocket (29.037) was launched at0821:53 UT, followed shortly thereafter by chemical releaserockets (41.044 and 41.046). The trajectory of the instru-mented rocket, mapped along magnetic field lines to themost closely related Altair scan, is highlighted in Figure 13.This figure suggests that the rocket intercepted bottom-typelayer patches on the upleg at about 250–300 km altitude andon the downleg at somewhat lower altitude. Large-scalewaves with a wavelength close to 200 km were starting toform in the bottomside F region during the flight, and therocket measurements should reflect the influence of thesewaves on the background ionospheric structure.

Upleg rocket measurements are shown in Figure 14, wherethe presence of irregularities is evident between about 240–300 km altitude. These irregularities were qualitatively dif-ferent from the ones encountered in the first experiment, ex-isting both in the valley and the bottomside and demonstrat-ing a greater degree of isotropy in their electric fields. Theirrelative RMS amplitude was smaller, but because the back-ground density was an order of magnitude greater than be-fore, the absolute density fluctuations were stronger. Thisis consistent with the detection of stronger coherent scatterfrom these layers.

These irregularities moreover resided in a slowly ascend-ing ionospheric layer that spanned the vertical shear node,occupying both eastward- and westward-drifting strata. Thevertical electric field profile appears to be highly distorted,having jets moving eastward and westward with respect tothe plasma at the F peak at altitudes just below and above300 km respectively. We attribute this as well as the kinks inplasma density to the large-scale wave forming in the region.Note also that the valley region was much flatter than in theearlier experiment and relatively free of layer structure. Asingle sporadic E layer was also present.

Downleg data for the second instrumented rocket flightappear in Figure 15. This time, the irregularities residedmainly in the valley region where the vertical density gra-dient was relatively small and the plasma drifts were upwardand westward. The relative RMS amplitude of the irregular-ities was again small. A strong westward jet in the plasmaflow existed at about 300 km, a feature we associate with the

HYSELL ET AL.: ROCKET INVESTIGATIONS OF EQUATORIAL SPREAD F 8

Figure 13. Altair radar perpendicular scan for August 15, 2004, 0824 UT. The white arc represents the location of theinstrument payload, mapped along geomagnetic field lines to the plane perpendicular to B scanned by the radar.

Figure 14. Upleg data from the August 15 rocket flight. Leftpanel: electron density. Center panel: zonal electric field.Right panel: vertical electric field.

growing large-scale waves in the bottomside. The sporadicE layer had a density comparable to the F region critical den-sity.

Figure 16 shows the power spectra of the zonal and ver-tical electric fields measured in the bottom-type layers inter-cepted on the upleg and downleg. The electric fields are notas clearly anisotropic as they were in the August 7 experi-ment.

Finally, Figure 17 shows the state of the ionosphere ap-proximately an hour after the rocket experiments began.The figure represents an off-perpendicular scan of the Al-tair UHF radar with its 300 µs long pulse mode. Patternsin the electron density are indicative of a large number ofsmall depleted channels (radar plumes) along with one or

Figure 15. Downleg data from the August 15 rocket flight.Left panel: electron density. Center panel: zonal electricfield. Right panel: vertical electric field.

two larger depletions that penetrated to the topside. The re-markable feature of these depletions is their regular, 30-kmspacing, mimicking the spacing of the patchy bottom-typelayer. Radar scans made looking perpendicular to B indi-cate the presence of strong coherent scatter from the bound-aries of the depletions. Had this event been observed bythe Jicamarca fixed-beam radar in conventional range-time-intensity (RTI) mode, the coherent scatter from the closely-spaced depletions would have tended to blend together toform a dense scattering layer confined mainly to the bot-tomside. Such “bottomside” layers are likely the manifes-tation of Rayleigh Taylor-type instabilities with decametricprimary wave wavelengths [Hysell and Burcham, 1998].

HYSELL ET AL.: ROCKET INVESTIGATIONS OF EQUATORIAL SPREAD F 9

Figure 16. Power spectra of the bottom-type layer electricfields measured on August 15, 2004. A frequency of 1 Hzcorresponds roughly to a wavelength of 1.8 km.

Figure 17. Altair radar oblique scan for August 15, 2004,0959 UT.

Analysis and Discussion

These experiments were conducted as a theory of bottom-type layers, shear flow, and ionospheric preconditioningleading to ESF was emerging from studies conducted formany years at Jicamarca, Altair, and elsewhere. The the-ory rests on several earlier findings, including that strongplasma shear flow exists in the postsunset equatorial F re-gion ionosphere [Kudeki et al., 1981; Tsunoda et al., 1981]and that bottom-type scattering layers exist exclusively inlow-altitude, westward-drifting strata [Hysell and Burcham,1998; Kudeki and Bhattacharyya, 1999]. The first of thesefindings suggests an additional, overlooked source of freeenergy for ionospheric instability. The second was puz-zling; Rayleigh-Taylor instabilities are expected to form inthe bottomside where there is a steep vertical density gra-dient and not in the valley, and irregularities forming in thevalley would tend to be “shorted out” by the high fraction ofthe field aligned Pedersen conductivity existing in the E re-gion. Layers produced by Rayleigh Taylor-type instabilitiesshould also exhibit vertical development, whereas bottom-type layers do not.

Carrying out highly accurate measurements of vectorplasma drifts associated with the evening vortex, Kudeki andBhattacharyya [1999] surmised that the bottom-type layerswere excited by the large differential neutral-plasma driftsthat exist at low altitudes, below the shear node, where theplasma streams rapidly westward at altitudes where the neu-tral flow is rapidly eastward. The wind-driven gradient driftinstability they described requires zonal plasma density gra-dients to function, but these are easily generated by the con-vection associated with the evening reversal of the zonalelectric field and the prereversal enhancement. This sameconvection pattern, combined with shear flow, are what pro-duce the vortex. The growth rate for the instability could besubstantial in view of the large differential drifts in question.

Direct evidence for the Kudeki and Bhattacharyya theorycame from Hysell et al. [2004], who observed bottom-typelayers using high-resolution aperture synthesis radar imag-ing techniques at Jicamarca. Their images revealed that theprimary plasma waves in the layer had kilometric wave-lengths and phase fronts that were elongated and alignedzonally. Zonal phase fronts are expected for zonal wind-driven gradient drift instabilities, whereas vertical frontsshould be produced by Rayleigh Taylor-type processes. Hy-sell et al. [2004] also showed that nonlocal effects could iso-late the primary waves from resistive loading by the E regionif the primary wave wavelengths are a few kilometers or less.

A third finding of Hysell et al. [2004] was that, on nightswhen fully-developed ESF occurred, the irregularities in theprecursor bottom-type layers were often clustered together

HYSELL ET AL.: ROCKET INVESTIGATIONS OF EQUATORIAL SPREAD F 10

into patches spaced by approximately 30 km in the zonal di-rection. The irregularities were meanwhile continuous andbeam-filling on nights when ESF did not occur. The authorssurmised that the 30 km periodicity was telltale of the pres-ence of a large-scale wave with ∼30 km wavelength. Differ-ent phases of the wave would present zonal plasma densitygradients alternately stable or unstable to Kudeki and Bhat-tacharyya’s wind-driven instability. The large-scale waveswould then be available to serve as seed waves for RayleighTaylor-type instabilities. If this picture is correct, then itshould be possible to discern the presence of large-scalewaves and forecast ESF relatively early in the evening byidentifying periodic structure in coherent backscatter fromthe bottom-type layers.

The necessity of seed or precursor waves for initiat-ing ESF has long been appreciated. Using the flux-tube-integrated formalism introduced by Haerendel [1973], it canreadily be shown that, in the absence of unusually strongstorm-time electric fields at the magnetic equator, the e-folding growth time for the generalized Rayleigh Taylor in-stability is seldom less than about 15 min. in the postsunsetbottomside F region and seldom for more than one or twoe-folding times (see for example Sultan [1996].) The timeof most rapid wave growth is generally about 1900 SLT, de-pending on season, solar cycle, and longitude. How is it pos-sible then for large-scale depletions and radar plumes to startappearing by 2000 SLT and earlier? The answer appears tobe the presence of preexisting seed waves.

The importance of large-scale waves in the behavior ofthe postsunset equatorial ionosphere has been highlighted byTsunoda and White [1981] and Kil and Heelis [1998]. Theirsource is often attributed to gravity waves [Rottger, 1973;Kelley et al., 1981; Singh et al., 1997; McClure et al., 1998].This hypothesis is difficult to test experimentally, however,since only the effects of gravity waves and not the gravitywaves themselves can be detected in the thermosphere usingISRs. Recently, Vadas and Fritts [2004] examined the is-sue theoretically, showing that a spectrum of gravity waveslaunched by mesoscale convection cells could survive windshears and viscous and conductive damping and penetrateinto the lower thermosphere. However, the preferred rangeof horizontal wavelengths for the surviving gravity waveswas about 50–150 km. This barely includes but does not fa-vor either the 30 km or the 200+ km large scale waves seenin Jicamarca and EQUIS II experiments.

However, Hysell and Kudeki [2004] considered whethershear flow itself could be the source of the large-scale waves.They followed the formalism developed by Keskinen et al.[1988] who assessed the viability of electrostatic KelvinHelmholtz instabilities in the auroral F region. That studysuggested that ion-neutral collisions damp the instability

in the ionosphere. Hysell and Kudeki [2004] adapted thework for equatorial application, incorporating the effectsof altitude-dependent collisions. They found a collisionalbranch of the instability that could operate in the collisionalregime in regions of strong, retrograde plasma motion. Thegrowth rate of the instability could be comparable to thatof the Rayleigh Taylor instability but with potentially ear-lier onset. Nonlocal analysis predicted that the growth rateshould be a maximum for kL ∼ 1/2, where k is the hori-zontal wavenumber and L is the vertical scale length of theshear. Taking L ∼ 15 km on the basis of Figure 7 there-fore implies a preferred wavelength of about 200 km. Aninitial value analysis, meanwhile, suggested that the insta-bility would exhibit a much shorter dominant wavelength inits early stages, of the order of 30 km in simulation. Theysuggested that the transient response of the instability couldtherefore account for the decakilometric large-scale waves inthe postsunset ionosphere and that the steady-state responsecould account for the large scale (L >

∼ 200 km) waves.

The new campaign data from Kwajalein place a numberof components of the shear instability theory of firm experi-mental ground. The vertical electric fields measured in situprove the existence of strong shear flow and retrograde driftsin the postsunset bottomside F region. Low plasma densityin the bottomside makes incoherent scatter difficult to re-ceive and process, and shear flow has been inferred in largepart (although not entirely) from coherent scatter from thebottom-type scattering layers that inhabit the region. Sincethere is no formal, rigorous relationship between coherentscatter signals and the state parameters of the plasma, how-ever, there was some cause for uncertainty prior to these ex-periments.

The rocket experiments further show that bottom-typelayers reside mainly in the valley region, below the altitudeswhere the vertical plasma density gradient is steepest and in-stead in the altitude range where the retrograde plasma driftis fastest. This, the fact that the vertical RMS perturbed elec-tric fields in the layers are stronger than the zonal fields, andthe widespread presence of zonal plasma density gradientsin the Altair data, support the theory of Kudeki and Bhat-tacharyya [1999] attributing the layers to wind-driven gra-dient drift instabilities. The outer-scale of the electric fieldsin the layers is consistent with the kilometric primary wavewavelengths observed in aperture-synthesis imaging experi-ments at Jicamarca and also with the nonlocal theory whichshows how the waves can exist at low magnetic apex alti-tudes [Hysell et al., 2004]. That the small-scale irregular-ities detected by the instrumented rocket on the upleg ofthe August 15 experiments were mainly in the bottomsidecan be interpreted as meaning that the transition to RayleighTaylor-type instability was underway. This interpretation is

HYSELL ET AL.: ROCKET INVESTIGATIONS OF EQUATORIAL SPREAD F 11

supported by the rapid growth of the large-scale waves in thevicinity in Figure 13 as well as by the jets in the upleg anddownleg electric field data in the vicinity of the irregularities.

Moreover, coherent scatter data from the Altair radar con-firm the observation that the bottom-type layers often consistof patches regularly distributed in space. This phenomenonwas discovered using aperture-synthesis imaging and val-idated by conventional radar data like those in Figure 13.That the regular spacing of the patches was later replicatedin the morphology of the depletions during full-blown ESF(Figures 8 and 15) argues that the layer patches are telltaleof preconditioning or seed waves in the bottomside existingwell in advance of ESF. Indeed, patchy bottom-type layerspreceded the onset of ESF on every night that it occurredduring our experiments on Kwajalein. Patchy bottom-typelayers seem to be reliable precursors of ESF.

Future work

Two important questions remain to be addressed with theEQUIS II dataset. The first concerns the cause of the shearflow. The factors that can contribute have been analyzed in-dividually in theoretical and experimental contexts by Anan-darao et al. [1978]; Fejer [1981]; Stenning [1981]; Takedaand Maeda [1983]; Farley et al. [1986] and by Haerendelet al. [1992], Haerendel and Eccles [1992], and again by Ec-cles [1998] in more theoretical treatments. They include 1)zonal electric fields on flux tubes with significant Hall con-ductivity, as are responsible for driving the equatorial elec-trojet, 2) zonal winds on flux tubes with significant Peder-sen conductivity, as drive the E and F region dynamos, 3)vertical winds, a largely unknown quantity, and 4) verticalboundary currents forced from above or below the flux tubein question. In the bottomside and valley regions aroundtwilight, this last factor could result from the closure of theequatorial electrojet, which must turn partly vertical at theboundary of the evening terminator. It could also result fromthe F region dynamo operating in flux tubes near the F peak.The finite efficiency of the dynamo implies the existence ofan upward vertical current there. The demands imposed bythis current can only be supported at lower altitudes, wherethe conductivity is smaller, by a potentially large verticalelectric field [Haerendel et al., 1992; Haerendel and Eccles,1992].

The second question concerns the viability of the colli-sional shear instability in the postsunset bottomside equato-rial F region and its ability to produce the large-scale wavespresent at the start of the spread F events described here.An approximate expression for the linear growth rate of the

instability can be written as [Hysell et al., 2005]:

γ ≈κφκp〈

hφhq

hpνin(u − v◦)n◦

′〉

κ2

φ〈hphq

hφνinn◦〉 + κ2

p〈hφhq

hpνinn◦〉

(1)

where u is the zonal neutral wind speed, v is the zonalplasma drift speed, n◦ is the background plasma numberdensity, and νin is the ion-neutral collision frequency. Theequation is expressed in magnetic dipole coordinates (p, q, φ)as defined in the reference above, and the h factors are scalefactors for this system. The κ factors are dimensionlesswavenumber components. The prime notation is shorthandfor h−1

p d/dp. Finally, the overbars and angle brackets de-note averages parallel to B and normal to B and the hori-zontal, respectively. Equation (1) is solved by carrying outthe necessary averages and then finding and substituting theratio of wavenumber components that maximize the growthrate. Growth rates found to exceed the growth rate of theflux-tube-integrated generalized Rayleigh Taylor instabilitywould point to the importance of shear instability in initiat-ing ESF.

Assessing both the causes of shear flow and the viabilityof shear instability require the specification of ionosphericconductivities, electric fields, and horizontal winds over awide sector of the low-latitude ionosphere. We are presentlyassembling the required empirical and physics-based mod-els, operating them with parameterizations chosen to pro-duce results maximally consistent with our measurementswherever they exist. It will be especially important to cor-rectly model both the steep bottomside F region density gra-dients and the valley region plasma number density in viewof the form of (1). The formula suggests that shear insta-bility is most likely if strata can be found with both steepvertical density gradients and rapid retrograde drifts. If theretrograde drifts are caused by significant electrical loadingin the E region, however, then growth will be suppressed inaccordance with the denominator in (1). Only if the shear isproduced mainly by one of the other aforementioned mech-anisms would we expect shear instability to occur.

Acknowledgments. This work was supported by the NationalAeronautics and Space Administration through grant no. NAG5-5380 to Cornell University. The authors are indebted to the staffsof the Reagan Test Site on Kwajalein and of the Wallops FlightFacility who made this campaign possible and also to Prof. ErhanKudeki and Dr. Robert Pfaff for many valuable discussions.

References

Anandarao, B. G., R. Raghavararao, J. N. Desai, and G. Haeren-del, Vertical winds and turbulence over Thumba, J. Atmos. Terr.Phys., 40, 157, 1978.

HYSELL ET AL.: ROCKET INVESTIGATIONS OF EQUATORIAL SPREAD F 12

Bowles, K. L., G. R. Ochs, and J. L. Green, On the absolute inten-sity of incoherent scatter echoes from the ionosphere, J. of Res.NBS - D. Rad. Prop., 66D, 395, 1962.

Close, S., M. M. Oppenheim, S. Hunt, and L. P. Dyrud, Scatteringcharacteristics of high-resolution meteor head echoes detectedat multiple frequencies, J. Geophys. Res., 107, 1295, 2002.

Eccles, J. V., A simple model of low-latitude electric fields, J. Geo-phys. Res., 103, 26,699, 1998.

Farley, D. T., E. Bonelli, B. G. Fejer, and M. F. Larsen, The prere-versal enhancement of the zonal electric field in the equatorialionosphere, J. Geophys. Res., 91, 13,723, 1986.

Fejer, B. G., The equatorial ionospheric electric fields. A review, J.Atmos. Terr. Phys., 43, 377, 1981.

Haerendel, G., Theory of equatorial spread F, Max-Planck Institutefur Physik und Astrophysik, Garching, West Germany, 1973,preprint.

Haerendel, G., and J. V. Eccles, The role of the equatorial electrojetin the evening ionosphere, J. Geophys. Res., 97, 1181, 1992.

Haerendel, G., J. V. Eccles, and S. Cakir, Theory for modeling theequatorial evening ionosphere and the origin of the shear in thehorizontal plasma flow, J. Geophys. Res., 97, 1209, 1992.

Hysell, D. L., and J. Burcham, JULIA radar studies of equatorialspread F, J. Geophys. Res., 103, 29,155, 1998.

Hysell, D. L., and E. Kudeki, Collisional shear instability inthe eqautorial F region ionosphere, J. Geophys. Res., 109,(A11,301), 2004.

Hysell, D. L., J. Chun, and J. L. Chau, Bottom-type scattering lay-ers and equatorial spread F, Ann. Geophys., 22, 4061, 2004.

Hysell, D. L., E. Kudeki, and J. L. Chau, Possible ionospheric pre-conditioning by shear flow leading to equatorial spread F, Ann.Geophys., 2005, submitted.

Kelley, M. C., M. F. Larsen, and C. La Hoz, Gravity wave interac-tion of equatorial spread F: A case study, J. Geophys. Res., 86,9087, 1981.

Keskinen, M. J., H. G. Mitchell, J. A. Fedder, P. Satyanarayana,S. T. Zalesak, and J. D. Huba, Nonlinear evolution of the Kelvin-Helmholtz instability in the high-latitude ionosphere, J. Geo-phys. Res., 93, 137, 1988.

Kil, H., and R. A. Heelis, Global distribution of density irregular-ities in the equatorial ionosphere, J. Geophys. Res., 103, 407,1998.

Kudeki, E., and S. Bhattacharyya, Post-sunset vortex in equatorialF-region plasma drifts and implications for bottomside spread-F, J. Geophys. Res., 104, 28,163, 1999.

Kudeki, E., B. G. Fejer, D. T. Farley, and H. M. Ierkic, Interfer-ometer studies of equatorial F region irregularities and drifts,Geophys. Res. Lett., 8, 377, 1981.

Larsen, M. F., and C. D. Odom, Observations of altitudinal and lat-itudinal E-region neutral wind gradients near sunset at the mag-netic equator, Geophys. Res. Lett., 24, 1711, 1997.

McClure, J. P., S. S, D. K. Bambogye, F. S. Johnson, and H. Kil,Occurrence of equatorial F region irregularities: Evidence fortropospheric seeding, J. Geophys. Res., 103, 29,119, 1998.

Miller, K. L., and L. G. Smith, Horizontal structure of midlatitudesporadic E layers observed by incoherent scatter radar, RadioSci., 10, 271, 1975.

Miller, K. L., and L. G. Smith, Incoherent scatter radar observa-tions of irregular structure in mid-latitude sporadic E layers, J.Geophys. Res., 83, 3761, 1978.

Pingree, J. E., and B. G. Fejer, On the height variation of the equa-torial F region vertical plasma drifts, J. Geophys. Res., 92, 4763,1987.

Rottger, J., Wavelike structures of large scale equatorial spread Firregularities, J. Atmos. Terr. Phys., 35, 1195, 1973.

Singh, S., F. S. Johnson, and R. A. Power, Gravity wave seeding ofequatorial plasma bubbles, J. Geophys. Res., 102, 7399, 1997.

Stenning, R. J., A two-layer ionospheric dynamo circulation, J.Geophys. Res., 86, 3564, 1981.

Sultan, P. J., Linear theory and modeling of the Rayleigh-Taylorinstability leading to the occurrence of equatorial spread F, J.Geophys. Res., 101, 26,875, 1996.

Takeda, M., and H. Maeda, F region dynamo in the evening —Interpretation of equatorial ∆D anomaly found by MAGSAT, J.Atmos. Sol. Terr. Phys., 45, 401, 1983.

Tsunoda, R. T., and B. R. White, On the generation and growth ofequatorial backscatter plumes. 1. Wave structure in the bottom-side F layer, J. Geophys. Res., 86, 3610, 1981.

Tsunoda, R. T., M. J. Barons, J. Owen, and D. M. Towle, Altair: Anincoherent scatter radar for equatorial spread F studies, RadioSci., 14, 1111, 1979.

Tsunoda, R. T., R. C. Livingston, and C. L. Rino, Evidence of avelocity shear in bulk plasma motion associated with the post-sunset rise of the equatorial F layer, Geophys. Res. Lett., 8, 807,1981.

Vadas, S. L., and D. C. Fritts, Thermospheric response to gravitywaves arising from mesoscale convective complexes, J. Atmos.Sol. Terr. Phys., 66, 781, 2004.

Woodman, R. F., and C. La Hoz, Radar observations of F regionequatorial irregularities, J. Geophys. Res., 81, 5447, 1976.

D. L. Hysell, Department of Earth and AtmosphericSciences, Cornell University, Ithaca, NY 14853, USA(daveh@geology.cornell.edu)

M. F. Larsen, Department of Physics and Astron-omy, Clemson University, Clemson, SC 29634, USA(mlarsen@hubcap.clemson.edu)

C. M. Swenson and A. Barjatya, Department of Elec-trical and Computer Engineering, Utah State Univer-sity, Logan, UT 84322, USA (charles.swenson@usu.edu,arohb@cc.usu.edu)

T. F. Wheeler, Department of Electrical Engineering,Pennsylvania State University, University Park, PA 16802,USA (tfw1@psu.edu)

T. W. Bullett, Space Vehicles Directorate, Air ForceResearch Laboratory, Hanscom AFB, MA 01731, USA(Terry.Bullett@noaa.gov)

M. F. Sarango, R. F. Woodman, and J. L. Chau, Jica-marca Radio Observatory, Instituto Geof ısico del Per u, Lima13, Per u (msarango@jro.igp.gob.pe, ronw@geo.igp.gob.pe,jchau@jro.igp.gob.pe)

D. Sponseller, Kwajalein Range Services LLC., AL-TAIR Radar, United States Army Kwajalein Atoll -Reagan Test Site, Republic of the Marshall Islands(dale.sponseller@us.army.mil)