Document Sample
qp Powered By Docstoc
1 Summary of Personnel and Commitments                                                                                                                                    1

2 Objectives and Expected Signi£cance of Proposed Work                                                                                                                   2
  2.1 Science overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .                                                                   3
  2.2 Experiment overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .                                                                    6

3 Technical Approach — science details                                                                     8
  3.1 Neutral stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
  3.2 Sporadic E layer patches and waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
  3.3 Other Plasma instabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

4 Methodology – experiment details                                                                                                                                       13
  4.1 Rocket measurements . . . . . . .      .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   13
      4.1.1 Instrumented rockets . . .       .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   14
      4.1.2 Chemical Release Rockets         .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   16
      4.1.3 Vehicle requirements . . .       .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   16
  4.2 Ground based radar imager . . . .      .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   16
  4.3 Wallops Island ionosonde . . . . .     .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   17

5 Scenarios Leading to Closure                                                                                                                                           17

6 Impact of Proposed Work                                                                                                                                                18

7 Relevance to Past, Present, and Future NASA Programs                                                                                                                   18

8 Management Approach                                                                                                                                                    18

9 Budget justi£cation                                                                                                                                                    19

The discovery of quasiperiodic (QP) echoes in Japan in the late 1980’s created new interest in the midlatitude
E-region ionosphere and in the sporadic E layers with which they are associated. The echoes have now been
observed over a broad range of longitudes at midlatitudes in the Northern Hemisphere and have been the
focus of extensive coherent scatter radar observations, as well as more limited incoherent scatter radar and
rocket investigations. We now know considerably more about the morphology of the echoes, including the
vertical extent, range of quasiperiods, range rates, Doppler spectra characteristics, and typical orientation
in the horizontal plane and direction of propagation. Those observations that have included neutral wind
pro£le measurements have also shown that large wind shears and dynamical instabilities in the vicinity of
the layers are consistent features of the QP environment.
    A number of the theories that have been proposed to explain the echoes rely on the existence of the shears
in that altitude range. Some rely on the Kelvin-Helmholtz billows created by the large shears to initiate the
plasma instabilities responsible for the layers. Others rely on gradients in the conductivities across the shear
layer. While it seems clear that the neutral forcing is the primary driver, the details of the interaction between
the neutrals and the plasma in the early stages of QP development are still largely unknown, and the main
plasma instabilities at work responsible for creating the small-scale £eld-aligned irregularities detected by
radar have not been identi£ed.
    We propose to carry out two sets of rocket launches from the Wallops Island Flight Facility in the
summer of 2007 to observe the details of the atmospheric ionospheric interaction and the plasma dynamics
that result with supporting measurements from state-of-the-art ground-based instrumentation, including a
high-resolution VHF coherent scatter imaging radar for diagnosing the plasma structure. The goal of the
study is to obtain a time history of the echoes and the neutral dynamics prior to the launches and during the
in situ rocket measurements, including measurements of the stability characteristics of the atmosphere at
and below the altitude of the sporadic E layer. The rocket measurements will include electric £elds, plasma
densities, neutral temperatures, and horizontal structure in the airglow on a low apogee rocket. A chemical
tracer trail will be released from the low-apogee rocket as well to characterize the horizontal structure in the
neutral ¤ow and turbulence in the vicinity of the layer. A second, higher apogee rocket will release upleg
and downleg chemical tracer trails to provide measurements of the vertical wind pro£les across the altitude
of the sporadic layer and to provide information about the horizontal gradient in the winds.
    The QP echoes represent some of the strongest midlatitude irregularities and are clearly an important
part of the space weather of that region of the ionosphere. In addition to affecting communications, the
echoes are also symptomatic of the dynamics and electrodynamics responsible for the irregularities, i.e., of
the basic interaction and coupling between the neutrals and ionization in the lower E region at midlatitudes.
Our objective is to understand that coupling in more detail, and in particular the processes that operate as
the instabilities develop.
1 Summary of Personnel and Commitments
Summary of Personnel: 1st year

 1. Senior personnel
          David Hysell (PI)              WY 0.2    $ 15,000
          Miguel Larsen (Co-I)           WY 0.04   $ 4,500
          Timothy Wheeler (Co-I)         WY 0.08   $ 4,488
          Charles Swenson (Co-I)         WY 0.02   $ 1,962
 2. Post-doctoral associate (1)          WY 0.12   $ 5,100
 3. Students (1.75)                      WY 1.75   $ 33,317
 4. Technical support staff              WY 0.1    $ 9,272
 5. Other                                WY        $
 6. TOTAL                                WY 2.31   $ 73,639
   Summary of Personnel: 2nd year

 1. Senior personnel
          David Hysell (PI)              WY 0.2    $ 15,000
          Miguel Larsen (Co-I)           WY 0.08   $ 9,000
          Timothy Wheeler (Co-I)         WY 0.33   $ 18,579
          Charles Swenson (Co-I)         WY 0.02   $ 2,045
 2. Post-doctoral associate (1)          WY 0.12   $ 5,100
 3. Students (1.75)                      WY 1.75   $ 33,622
 4. Technical support staff              WY 0.5    $ 30,921
 5. Other                                WY        $
 6. TOTAL                                WY 3      $ 114,267
   Summary of Personnel: 3rd year

 1. Senior personnel
          David Hysell (PI)              WY 0.2    $ 15,000
          Miguel Larsen (Co-I)           WY 0.08   $ 9,000
          Timothy Wheeler (Co-I)         WY 0.08   $ 4,807
          Charles Swenson (Co-I)         WY 0.06   $ 8,573
 2. Post-doctoral associate (1)          WY 0.12   $ 5,100
 3. Students (1.75)                      WY 1.75   $ 33,940
 4. Technical support staff              WY        $
 5. Other                                WY        $
 6. TOTAL                                WY 2.29   $ 76,420

2 Objectives and Expected Signi£-                            the structure is tied to the sporadic E layers rather
                                                             than to a speci£c geographic location. The contin-
  cance of Proposed Work                                     ued level of interest in quasi-periodic echo struc-
This is a proposal to investigate quasiperiodic struc-       tures is indicated in part by the SEEK-2 sounding
tures in midlatitude sporadic E layers and the               rocket campaign carried out in August, 2003, in
plasma irregularities generated within them respon-          Japan and by the special session of the MST 10 con-
sible for so-called quasiperiodic (QP) radar echoes.         ference held in May, 2003, in Peru devoted to the
We will also investigate the stability of the neu-           subject.
tral mesosphere and lower thermosphere (MLT) and                  Much more is now known about the occur-
the emergence of Kelvin Helmholtz instabilities              rence frequency and morphology of QP echoes
thought to play a role in producing the quasiperi-           than was known even just a few years ago, and
odic E layer structures. The investigation is to be          it is currently believed that QP radar echoes arise
carried out using instrumented and chemical release          from patchy irregularities in sporadic E layers like
sounding rockets with ground-based radar support             those £rst observed at Arecibo by Miller and Smith
and will involve a substantial data analysis effort.         (1978) and Smith and Miller (1980). Larsen (2000)
Our experiments will be conducted from the Wal-              has argued that the patches could be produced by
lops Flight Facility in the summer of 2007 by re-            Kelvin Helmholtz instabilities in the neutral at-
searchers from Cornell University, Clemson Univer-           mosphere driven by the same steep wind shears
sity, Utah State University, and the Aerospace Cor-          thought to create the layers, and Bernhardt (2002)
poration.                                                    has presented numerical simulations of the pro-
     The existence of midlatitude sporadic E layers          cess which seem to demonstrate its viability. At
has been known almost as long as HF and VHF ra-              the same time, however, Cosgrove and Tsunoda
dio communications have been used (see, e.g., the            (2003) have argued that large-scale plasma instabil-
comprehensive review by Whitehead (1972)). In                ities associated with neutral shear alone can cause
the 1970’s and 1980’s, a considerable effort was             sporadic E layers to break up. In either case,
made to improve our understanding of the mecha-              the plasma density gradients, electric £elds, and
nisms responsible for generating the enhanced ion-           wind shears within the patches would provide the
ization layers and the instabilities responsible for         free energy for intermediate-scale plasma instabil-
small-scale irregularity structures within those lay-        ities to occur which could in turn drive the small-
ers. Considerable progress was made although                 scale £eld-aligned irregularities detected by radar.
many questions remain, as discussed in the reviews           Intermediate-scale plasma waves are observed in
by Whitehead (1989) and Mathews (1998), for ex-              the layers, but the particular instabilities responsi-
ample. Interest in the dynamics of the layers in-            ble have not been identi£ed.
creased considerably in the 1990’s due to the dis-                It is clear in any case that sporadic E and QP
covery by Yamamoto et al. (1991, 1992) that quasi-           echoes are the result of complex interactions be-
periodic (QP) radar echo structure is often associ-          tween the neutral atmosphere and ionization in the
ated with sporadic E layers observed by the MU               mesosphere and lower thermosphere region, but in
radar in Japan, but it was still not known until             spite of the growth in the observational database and
recently if the phenomenon was peculiar to the               of numerous modeling efforts, fundamental ques-
Japanese sector or if it was a more universal phe-           tions remain unanswered, including:
nomenon.                                                        1. Is the MLT region dynamically unstable dur-
     In the intervening years, more extensive ob-                  ing QP echo events?
servations were made, and the QP structures have
now been detected in the American sector (Hysell                2. Are neutral shear instabilities or large-scale
and Burcham, 1999; Tsunoda et al., 1999; Chau                      plasma instabilities mainly responsible for the
and Woodman, 1999; Swartz et al., 2002) and in                     formation of patches and quasiperiodic struc-
other locations in the Asian sector (Pan et al., 1998;             tures in sporadic E layers?
Choudhary and Mahajan, 1999; Pan and Larsen,                    3. What additional, intermediate-scale plasma
2000; Pan and Rao, 2002), thus demonstrating that                  instabilities are involved in the production of

      QP radar echoes?                                             Figure 2 shows coherent scatter data obtained
                                                              from a radar located on St. Croix, USVI, and
    These questions can only be addressed using               situated so as to observe £eld-aligned irregulari-
sounding rockets, since ground-based radars and               ties in the ionosphere immediately over Arecibo.
satellite-based imagers lack the sensitivity and reso-        The quasiperiodic echoes shown in Figure 2 started
lution required to identify neutral wind irregularities       just as the sporadic E layer over Arecibo began to
and collocate them with wind shears and large- and            break into patchy structures. The £ve clusters of
small-scale plasma irregularities.                            QP echoes which appear starting at 2330, 0015,
    Sporadic E layers, layer patches, and E region            0100, 0200, and 0230 UT in the coherent scat-
plasma irregularities can both enhance and degrade            ter data are coincident with the passage of £ve
HF and VHF radio communications, and study-                   groups of patches passing over Arecibo. Further-
ing the geophysical factors that cause and in¤uence           more, using an interferometric radar imaging tech-
them contributes to the Living with a Star initiative.        nique (discussed later in the proposal), Hysell et al.
Since atmospheric ionospheric coupling appears to             (2004) showed unambiguously that the coherent
be an important controlling factor, we will also be           scatter arose precisely from patchy, drifting regions
addressing that key aspect of the Geospace Sciences           of plasma density enhancements seen by Arecibo.
discipline of NASA’s Sun-Earth Connection science             The patches are evidently inherently unstable and
theme. Finally, our emphasis on the structure and             produce small-scale £eld-aligned plasma irregulari-
effects of winds in the lower thermosphere supports           ties that can be detected using coherent scatter. No
the mission of the TIMED satellite.                           coherent echoes were received after 23 LT when the
                                                              layer became undifferentiated.
2.1   Science overview                                             The morphology of the QP sporadic E structures
                                                              and the associated coherent echoes is therefore un-
Figure 1 shows a sporadic E layer observed by the
                                                              derstood: QP echoes arise from sporadic E layer
Arecibo incoherent scatter radar in the summer of
                                                              patches, and the striations in coherent scatter RTI
2002. The plasma densities are plotted in units of
                                                              plots are mainly indicative of the proper motion of
cm−3 on a logarithmic scale. The zenith angle of
                                                              the patches toward or away from the radar. How-
the antenna feed was 15◦ , and the azimuth angle,
                                                              ever, this picture is incomplete and will remain so
which scanned in time, is plotted below the density
                                                              until three fundamental questions can be answered.
data. Local sunset was at 1904 LT, and most of the
E region can be seen to have recombined by about
1920 LT.                                                      Is the MLT region dynamically unstable during
     Before about 20 LT, the scene was dominated              QP echo events?
by a strong, narrow layer at about 105 km with only           The dynamical instability of the strati£ed neutral
a very weak layer beneath it at the base of the E             atmosphere in the presence of shear ¤ow can be
region. The stronger layer appears to be neither              roughly diagnosed by calculating the Richardson
tilted nor deeply modulated. If it were, regular vari-        number parameter, which is proportional to the ra-
ations in its altitude would accompany and be syn-            tio of the change in kinetic to potential energy of a
chronized with the azimuth scan. Moreover, there              vertically displaced ¤uid parcel:
is no signi£cant E region ionization above the layer.
Several preliminary theories of QP echo generation                                       N2
                                                                            Ri =
relied on multiple dense layers, tilted layers, deeply                                (du/dz)2
modulated layers, and density perturbations high in                                   g    dT   g
the E regions (e.g. Woodman et al. (1991), Tsunoda                         N2 =               +
                                                                                      T    dz   cp
et al. (1994), Maruyama et al. (2000), Cosgrove and
Tsunoda (2001), and Ogawa et al. (2002)). How-                where N is the Brunt-V¨ al¨ frequency which is
                                                                                        ais¨ a
ever, none of those features are present in Figure 1          real when the atmosphere is convectively stable. We
despite the fact that intense QP echoes were ob-              expect shear instability for low Richardson number
served from the same ionospheric volume by a co-              and stability for high Richardson number. The tran-
herent scatter radar beginning at 1930 LT.                    sition is often taken to occur at Ri = 1/4, at which

Figure 1: Electron densities measured with the Arecibo linefeed antenna. The azimuth of the antenna feed,
which was scanning, is plotted below.

Figure 2: Range-time-intensity (RTI) plot of coherent echoes received by a 30 MHz radar coherent scat-
ter radar sharing Arecibo’s scattering volume. Grayscales depict signal-to-noise ratio in dB. Note that
UT=LT+4 hours.

point strati£ed shear layers break into growing bil-         the altitude of the layers and the billows since the
low structures. However, exceptions to the rule are          conductivities play an important role in the dynam-
common. The Kelvin Helmholtz instability is in-              ical processes. Cosgrove and Tsunoda (2001) and
herently nonlocal, and growth depends on the shape           Cosgrove and Tsunoda (2002b) analyzed the polar-
of the entire velocity and temperature pro£le rather         ization electric £elds produced by vertical displace-
on their values at any given point. Analysis is fur-         ments in the ionization driven by neutral oscillations
ther complicated when vector horizontal winds and            including shears in the neutral ¤ows. Their results
turning shears are taken into account.                       indicated that large £elds can be explained by such
    Larsen (2000) reviewed neutral wind pro£le               dynamics. The large electric £elds, plasma density
measurements in the MLT region for a number of               gradients, and neutral winds present in this scenario
cases in which QP echoes were observed. His con-             would be available to drive the plasma instabilities
clusion was that the atmospher was shear unstable in         ultimately responsible for QP echoes.
each case. However, this conclusion was based on                  However, Cosgrove and Tsunoda (2002a) and
modeled rather than measured temperature pro£les             Cosgrove and Tsunoda (2003) recently proposed a
since temperature and horizontal wind pro£les have           new, large-scale E-layer instability that could also
never been measured together in the context of a QP          break up the sporadic layers. The driver is a shear
echo experiment. Furthermore, the conclusion was             in the zonal wind which creates an unstable equi-
based on evaluation of the Richardson number cri-            librium in the sporadic E layer in which imposed
terion rather than analyses of complete wind pro£le          plane wave vertical perturbations can grow if they
shapes. In is therefore unknown whether dynamical            have favorable orientations. The dynamics involved
instability is a common feature of QP echo events            are reminiscent of the Perkins instability in the F
and quasiperiodic sporadic E layers or not.                  region. The mechanism is especially interesting be-
                                                             cause the instability growth rate maximizes when
Are neutral shear instabilities or large-scale               the wave fronts are aligned from northwest to south-
plasma instabilities mainly responsible for the              east, consistent with the observed preferred align-
formation of patches and quasiperiodic struc-                ment and propagation direction of the QP echo
tures in sporadic E layers?                                  structures (see, e.g. Tsunoda et al., 2000; Hysell
                                                             et al., 2004).
The mechanism responsible for breaking up spo-                    Both the nuetral shear instability and the plasma
radic E layers and initiating the process that leads         instability appear to be able to account for the patch-
to QP echoes has not been identi£ed, but neutral dy-         iness and large-scale organization of the sporadic
namics are thought to play a crucial role. Miller and        E layers. Distinguishing between the two mecha-
Smith (1978) and Smith and Miller (1980) showed              nisms necessitates in situ measurements of nuetral
that sporadic E layers are often collocated with un-         and plasma parameters.
stable neutral shears, suggesting that K-H billows
should be present. Larsen (2000) further argued that
                                                             What additional, intermediate-scale plasma in-
Kelvin-Helmholtz billows associated with shear in-
                                                             stabilities are involved in the production of QP
stability in the neutral atmosphere can produce the
                                                             radar echoes?
large vertical displacements in the sporadic E layers
needed to initiate plasma instabilities. Some direct         Kilometric plasma waves thought to signify pri-
evidence that K-H billows are present in the volume          mary plasma waves and instabilities have been de-
from which QP echoes are received emerged from               tected in sporadic E layers in radio sounding data
the SEEK-2 experiments.                                      (Barnes, 1992), airglow experiments (Djuth et al.,
     Bernhardt (2002) simulated the distortions in           1999), radio scintillations (Maruyama et al., 2000),
an initially planar E layer produced by Kelvin-              and in rocket investigations from Wallops (Kelley
Helmholtz billows. He found that the nonlinear re-           et al., 1995). The small-scale irregularitites respon-
sponse of the ionization is complicated but essen-           sible for coherent scatter are thought to be pro-
tially mimics the structure in the neutral ¤ow that is       duced mainly by mode coupling to such primary
acting as the driver. The response also depends on           waves. While a number of theories have been pre-

sented over the years, consensus regarding the par-         conducted fully three-dimensional simulations of
ticular intermediate-scale primary plasma instabil-         plasma patches in the midlatitude E region iono-
ity at work is elusive. The Cosgrove and Tsunoda            sphere and found that another instability existed
(2003) mechanism described above appears to op-             with a faster growth rate than gradient drift. The
erate at too long a wavelength to explain the kilo-         collisional drift waves that grew were longitudi-
metric waves. A gradient drift instability similar          nal (gradient drift waves are transverse), existed
to the one occurring in the equatorial electrojet is        throughout the plasma cloud (rather than on just
widely considered a likely candidate. However, £-           one side), and required a small but £nite paral-
nite parallel gradient scale length effects at middle       lel wavenumber component (gradient drift waves
latitudes were originally thought to stabilize gradi-       are preferentially £eld aligned). The primary
ent drift instabilities (Woodman et al., 1991). While       waves emerging from the simulations had kilomet-
Seyler et al. (2004) recently showed with a non-            ric wavelengths and propagated in the direction par-
local analysis that intermediate-scale gradient drift       allel to the main polarization electric £eld in the
instabilities with £nite parallel wavenumbers may           plasma cloud.
form in uniform sporadic E layers at mid latitudes,              Distinguishing between gradient drift and col-
their theory does not account for the quasiperiodic-        lisional drift instabilities in the layers can only
ity of the radar echoes or for the kilometric wave-         be accomplished using instruments on sounding
lengths observed and also suggests that laminar lay-        rockets which alone can resolve £ne structure in
ers should produce echoes, contradicting the results        the plasma density and electric £elds and compare
from Arecibo.                                               measured plasma wave parameters (densities and
     The association of QP echoes with patchy spo-          electric £elds) with the dispersion relations of the
radic E layers removes many theoretical objections          plasma instabilities in question. Chemical release
for gradient drift instabilities. This is because the       experiments in nearby regions of space make it pos-
patches have ¤ux-tube-integrated density gradients          sible to collocate the plasma waves with regions of
all around their periphery, as opposed to the local-        neutral shear and neutral instability. Appropriate
only density gradients that exist on the topside and        radar support of the experiments can guarantee that
bottomside of a horizontally homogeneous layer.             the in situ measurements corresponed to the regions
This obviates the need for either the nonlocal theory       of space from which QP echoes emerge.
of Seyler et al. (2004) or the reversed wind forcing
theory of Kagan and Kelley (1998) to understand             2.2   Experiment overview
how gradient drift instabilities can operate. (These
various mechanisms may yet function, but neither            We propose to address the three questions posed
seems strictly necessary.) Furthermore, Shalimov            above with an experiment involving instrumented
et al. (1998), Hysell and Burcham (2000), and Cos-          and chemical release rockets along with ground-
grove and Tsunoda (2002b) showed that patchy spo-           based support, data reduction, and modeling. The
radic E layers can spontaneously produce very large         experiment will be performed from the Wallops
polarization electric £elds and currents when forced        Flight Facility in the summer of 2007 during a
by transverse electric £elds or winds, readily able         moon-down nighttime interval. We are proposing
to drive gradient drift and Farley-Buneman instabil-        two identical sets of measurements to be performed
ities. We would expect primary gradient drift waves         on separate nights. Each set of launches will in-
to form on one side of the patchy layers only, since        volve three rockets — one with an instrumented
only one side should be linearly unstable. The pri-         payload and two with chemical release payloads.
mary wave wavefronts could be aligned in any di-            Two sets of launches are proposed for the purpose
rection perpendicular to B, depending on the direc-         of redundancy and to con£rm the universality of
tion of the forcing winds and background electric           our results. Conditions for launch include the ob-
£eld. The waves would propagate in the direction            servation quasiperiodic (QP) echoes by a portable
normal to the main polarization electric £eld in the        imaging coherent scatter radar, a sporadic E layer
patchy layer.                                               detected by the Wallops Island ionosonde, and clear
     Recently, however, Hysell et al. (2002b)               skies for optics.

                                                             from sites on the ground, yielding two altitude pro-
                                                             £les of the horizontal wind velocity. From this in-
                                                             formation, absolute wind shear pro£les will be de-
                                                             rived. Having wind pro£les at the upleg and down-
                                                             leg locations allows us to assess the importance of
                                                             spatial gradients in the winds during the experi-
                                                                  A second chemical release rocket carrying a sin-
                                                             gle TMA cylinder will be launched to an apogee
                                                             of 120 km. The chemical release from this rocket
                                                             will occur along a nearly horizontal path pass-
                                                             ing through the altitude strata associated with QP
                                                             echoes and patchy sporadic layers. Triangulation of
                                                             the trail will permit the determination of the hori-
                                                             zontal wind speed along its arc, providing more in-
                                                             formation about horizontal gradients in the wind ve-
                                                             locity. Moreover, the horizontal trail will serve as
                                                             a tracer of neutral instability, highlighting the pres-
                                                             ence of Kelvin Helmholtz billows developing in the
Figure 3: Map showing the location of the Wallops            region. This strategy worked successfully during
Flight Facility (W) and the coherent scatter radar to        the SEEK II experiments when such billows were
be located in Virginia.                                      clearly detected.
                                                                  An instrumented sounding rocket will be
     The geometry of the experiments is illustrated          launched just after the chemical release rockets
in Figure 3, in which the Wallops Flight Facility is         along the same trajectory. The instrument package
indicated by a “W” and the Cornell coherent scatter          will include a photometer and an ionization guage
radar, to be located in Fort Macon, Virginia, is indi-       (IG) from the Aerospace Corporation and a plasma
cated by its radiation pattern, which will point back        impedance probe (PIP) and an electric £eld instru-
toward Wallops. Concentric circles give the dis-             ment from Utah State University. The IG will mea-
tance from the radar to a target at 100 km altitude,         sure neutral temperature pro£les required for as-
and the remaining arcs in the £gure give the altitude        sessing atmospheric stability. The photometer will
where the condition for £eld aligned radar backscat-         measure neutral density ¤uctuations. The PIP mea-
ter is met. The straight dashed line is the nominal          sures plasma density and density ¤uctuations along
¤ight path of the instrumented rockets crossing the          with electron and ion temperature and the electron-
radar £eld of view.                                          neutral collision frequency. The electric £eld instru-
     The radar and the Wallops ionosonde will in-            ment will measure vector electric £elds in the plane
dicate when sporadic E layers and quasiperiodic              perpendicular to the geomagnetic £eld and will de-
echoes are present in the E region southeast of the          tect both the background £elds as well as ¤uctua-
rocket range. By extension, we will detect the pres-         tions associated with plasma waves and irregulari-
ence oof the sporadic E layer patches, which are             ties. Upleg and downleg data will be recorded for
now known to be collocated with the coherent scat-           all the instruments.
ter.                                                              The stability of the neutral MLT region during
     In each experiment, a chemical release rocket           QP echo events will be assessed on the basis of
will be launched to an apogee of 200 km. The                 the shapes of the vector wind pro£les and the at-
rocket will carry two cylinders of liquid trimethyl          mospheric temperature pro£le. Similar calculations
aluminum (TMA) from Clemson University to be                 have been performed in the past using measured
deployed on the upleg and downleg, respectively,             wind pro£les and modeled temperatures, but this
between about 70–150 km altitude. The chemilu-               will be the £rst time that measured wind and tem-
minescent trails that develop will be photographed           perature pro£les will be available for evaluating the

stability of the neutral atmosphere during a QP echo            3 Technical Approach — science
     If the complete criteria for shear instability are
met at the altitude where sporadic layers and QP                Figure 4 shows a radar image derived from the co-
echoes are detected, that will constitute strong evi-           herent scatter radar data from St. Croix presented
dence that neutral Kelvin Helmholtz instabilities are           earlier. Radar imaging has been used to deter-
responsible for breaking up and structuring the lay-            mine echo power versus bearing in three dimen-
ers. The evidence will be stronger still if Kelvin              sions. From range and bearing, we compute the lati-
Helmholtz billows are evident in the photographs of             tude, longitude, and altitude. Altitude information is
the horizontal TMA trails during the QP echo event.             not depicted in this £gure, however. The brightness,
Finally, close correlation between the neutral den-             hue, and saturation of the colored regions indicate
sity ¤uctuations observed by the photometer and the             the signal-to-noise ratio, Doppler shift, and spectral
plasma density ¤uctuations detected by the PIP will             width of the coherent echoes. Signal-to-noise ratios
signify that neutral dynamics, as opposed to plasma             are plotted on a scale between 8 and 38 dB here.
processes, are mainly responsible for the generation                 The image in Figure 4 depicts a number of dis-
of the patchy sporadic layers. In the event that bil-           crete, localized, patchy scatterers. Over time, an-
lows are not observed or that the neutral and plasma            imations show that the patches drift mainly west-
density ¤uctuations are poorly correlated, we will              ward or southwestward, although some on the west
be able to use the measured neutral wind, plasma                side of the image drift eastward toward St. Croix.
density, and electric £eld pro£les to evaluate if the           Each patch corresponds to a striation in the RTI di-
large-scale instability mechanism proposed by Cos-              agram in Figure 2, and the range rates of the stri-
grove and Tsunoda (2002a, 2003) is instead respon-              ations correspond to their line-of-sight drifts with
sible for generating the patches.                               respect to St. Croix. Differences in the motions of
     Finally, the plasma density and electric £eld              the patches give the striations in Figure 2 different
data acquired by the Utah State instruments will                slopes and even permit them to cross. We associate
provide the plasma wave diagnostics necessary to                the patches in the images with drifting, polarized E
distinguish between the two rimary plasma instabil-             region ionization patches. Evidence of polarization
ity candidates thought to be viable within the layers           is found in the circulation evident in the patches,
and to be producing the small-scale £eld aligned ir-            where the hue (Doppler shift) is typically different
regularities responsible for the QP echoes, gradient            in the center and at the periphery.
drift and collisional drift. The vector electric £eld                The “X” in the £gure in Northwest Puerto Rico
data are particularly incisive, indicating the direc-           represents the location of the Arecibo radar. Yellow
tion in which the electrostatic plasma waves prop-              and green lines radiating from this point represent,
agate. We will also be able to determine from the               respectively, the regions where the beams from the
density and electric £eld data the wavelength of the            Arecibo linefeed and Gregorian antennas intercept
primary waves and the phase relationship between                the E region between 100-110 km in altitude at the
the density and potential ¤uctuations. If kilomet-              time indicated. Images like Figure 4 make it possi-
ric and small-scale waves and irregularities are only           ble to collocate sources of coherent scatter with the
found to be present where the sign of the transverse            sporadic E layer features responsible for them. Note
plasma density gradient is favorable for gradient-              for instance that, in Figure 4, the Gregorian beam in-
drift instability, that will constitute strong evidence         tercepts a coherent scatter “bright spot” whereas the
that gradient drift instabilities are operating. If kilo-       linefeed system does not. Consulting Figure 1 and a
metric primary waves exist throughout the sporadic              similar £gure with data from the Gregorian feed re-
layer patches and propagate in the direction of the             veals that both of the Arecibo beams were detecting
main polarization electric £eld, than collisional drift         sporadic E ionization layers at 2330 LT (1930 UT).
waves are most likely at work. Further evidence                 However, the Gregorian beam passed through an ir-
would come from the phase relationship between                  regular, moderately thick patch of ionization at this
potential and density ¤uctuations, a £ngerprint of              moment while the linefeed observed only a very thin
the instability mechanism at work.                              layer. The £rst collision of the linefeed beam with

           Figure 4: Radar image of coherent echoes received at 2332 UT. Note that UT=LT+4.

a region of strong coherent scatter occurred at about       shifts alternate from large positive to negative val-
2348 LT, at which time a thick, irregular patch was         ues. These are the line of sight phase speeds of the
observed with the linefeed. By animating sequences          small-scale irregularities observed from St. Croix.
of images like Figure 4, we were able to compare                 At the time that the large-scale wave was pass-
the entire incoherent and coherent scatter datasets         ing over Puerto Rico, Arecibo observed a num-
for the June 14-15 event in this way. Overall, we           ber of very thick and intense plasma irregularities
demonstrated a very high correspondence between             or patches with both of its feed systems. There
the coherent backscatter patches and the ionization         is a high degree of correspondence between these
patches.                                                    clouds and the wavefronts in the coherent scatter
     A remarkable feature of Figure 4 is the tendency       radar images. It is when the yellow and green ra-
for the scattering patches to fall along lines. Sim-        dial markers in the images pass through he bright-
ilar linear formations of scatterers were observed          est parts of the wavefronts that the corresponding
with the MU and Clemson University radars (Hysell           Arecibo beam systems detect the patches. As the
et al., 2002a). Over Puerto Rico, however, the lines        wavefronts are very elongated, we surmise that the
were more evident and prone to coalesce into multi-         ionization clouds are either very elongated or occur
ple, unbroken, frontal structures propagating across        in long chains which move with the waves.
the ionosphere.                                                  Radar imaging allows us to understand the mor-
     Figure 5 shows an example of these quasiperi-          phology of QP echoes and their relationship to the
odic structures or waves. We regard the structures          underlying, patchy sporadic E layers. However,
as waves, both because wavefronts are evident in the        we do not know what causes the sporadic layers
image brightness and because the observed Doppler           to break up. Previous sounding rocket investiga-
shifts also vary periodically between wavefronts. In        tions have suggested that the neutral atmosphere
Figure 5, a large-scale wave with wavefronts run-           in the vicinity of the layers is shear unstable, but
ning from northwest to southeast is shown. The              since wind and temperature measurements have not
wavelength is about 30 km, and the period is about          been made together, we cannot be certain that neu-
10 min. Animated sequences of images con£rm                 tral instabilities are always present. Nor can we be
that the wave propagates to the southwest. Doppler          certain that Kelvin Helmholtz ¤ows are responsi-

                      Figure 5: Radar image of coherent echoes received at 0039 UT.

ble for breaking up the sporadic E layers as sug-            to show that sporadic E layers are often collocated
gested by Larsen (2000) and simulated by Bern-               with unstable neutral shears, thus suggesting that
hardt (2002), since Cosgrove and Tsunoda (2002a)             K-H billows are likely to also be a common fea-
have also shown that the layers may be inherently            ture of the layer dynamics. The number of obser-
unstable in the presence of wind shear absent neu-           vations in which the neutral wind pro£les across the
tral instability. Finally, we do not know what pri-          sporadic E layer were measured is limited, but ev-
mary plasma instability is responsible for producing         ery case that has been reported in the literature has
kilometric waves known to exist in the layers and,           shown the presence of large wind shears that were
ultimately, for generating small-scale irregularities        either unstable or close to instability in the vicinity
and £eld-aligned backscatter. Below, we examine              of the layer. Larsen (2000) summarized much of the
each of these science issues in detail.                      earlier data. More recent measurements from Wal-
                                                             lops Island in July 1999 and from the SEEK-2 ex-
3.1   Neutral stability                                      periment in Japan in August 2003 produced similar
3.2   Sporadic E layer patches and waves                          Bernhardt (2002) has simulated the distortions
That large-scale patchy or wavelike structures in            in an initially planar E layer produced by Kelvin-
sporadic E layers are associated with QP radar               Helmholtz billows. He found that the nonlinear re-
echoes is clear, but the mechanism producing the             sponse of the ionization is complicated but essen-
structures remains contentious. Larsen (2000) sug-           tially mimics the structure in the neutral ¤ow that is
gested that Kelvin-Helmholtz billows associated              acting as the driver. The response also depended on
with shear instability in the neutral atmosphere can         the altitude at which the billows were located since
produce the large vertical displacements in the spo-         the conductivities play a critical role in the dynam-
radic E layers required to initiate plasma instabili-        ical processes. Cosgrove and Tsunoda (2001) and
ties more effectively than gravity waves, the other          Cosgrove and Tsunoda (2002b) investigated the de-
main candidate at that time. Miller and Smith                velopment of polarization electric £elds associated
(1978) and Smith and Miller (1980) were the £rst             with vertical displacements in the ionization driven
                                                             by neutral oscillations, such as gravity waves, or by

shears in the neutral ¤ows. The results indicated                  Meanwhile, the Cosgrove and Tsunoda E-layer
that large £elds can be explained by such dynamics.            instability mechanism is appealing because it pro-
     Recently, however, Cosgrove and Tsunoda                   vides an explanation for the preferred propagation
(2002a) and Cosgrove and Tsunoda (2003) have                   direction, as well as being consistent with other
investigated a new E-layer instability. The driver             known characteristics of the echo structures. A
is a shear in the zonal wind which creates a situ-             potential problem with the theory is that Kelvin-
ation in which the zonal wind driven polarization              Helmholtz billows or other small-scale wave struc-
becomes unstable to an imposed plane wave per-                 ture would tend to disrupt the instability (Cosgrove
turbation in the vertical. In particular, wind shear           and Tsunoda, 2003). Since large and often unsta-
causes the winds to drive oppositely-directed hor-             ble shears seem to be a consistent feature of the
izontal ion currents at the crests and troughs of              QP environment, the billow structure is expected to
a seed wave in a sporadic E layer. This gives                  be present, at least part of the time. Cosgrove and
rise to ion enhancements and depletions in differ-             Tsunoda (2003) therefore suggested that the shear-
ent phases of the wave. The wavelike polariza-                 driven polarization instability may be the operative
tion electric £elds that form arrest the horizontal ion        mechanism at certain times and that one of the other
drifts but cause vertical drifts that enhance the pre-         mechanisms, such as the K-H billows modulation
existing seed wave. Electrons preserve quasineu-               mechanism, may be important at other times.
trality by streaming along B. The dynamics in-
volved are reminiscent of the Perkins instability in           3.3   Other Plasma instabilities
the F region. The mechanism is especially interest-
ing because the instability growth rate maximizes              Direct observations of small-scale £eld aligned ir-
when the wave fronts are aligned from northwest                regularities and indirect observations of kilometric
to southeast, consistent with the observed preferred           plasma waves in patchy sporadic E layers point to
alignment and propagation direction of the QP echo             the existence of intermediate-scale primary plasma
structures (see, e.g. Tsunoda et al., 2000; Hysell             instabilities that operate apart from whatever mech-
et al., 2004).                                                 anism causes the layers to break up in the £rst place.
     Both of the mechanisms described above are vi-            While the particular instability at work has not been
able, although all have potential shortcomings. In-            identi£ed, gradient drift instabilities have received
voking Kelvin-Helmholtz billows as the mechanism               the most attention in the literature. Gradient drift in-
for modulating the layers has the advantage that the           stabilities occur when an E region Hall current ¤ows
perturbations are associated with the shear itself,            in the direction normal to a background plasma den-
which is not only a common feature of the QP en-               sity gradient. Density perturbations in along the
vironment but is likely to be a part of the mecha-             ¤ow of the current become polarized in accordance
nism that generates the ionization layer in the £rst           with the constraint of quasineutrality, with the re-
place. The billows therefore would be intrinsic to             sulting convection enlarging the initial perturbations
the layer dynamics and the appropriate spatial rela-           and leading to instability. In the equatorial E region
tionship between the altitude of the layers and the            where background parameters are nearly invariant
altitude of the billows is automatically maintained.           along B, a linear, nonlocal dispersion relation for
Yokoyama et al. (2003) recently showed that polar-             the growing waves can be derived:
ized E region plasma clouds can induce drastic per-
                                                                         k · (ve − vi )
turbations in the densities and electric £elds in the          ωr =                         + k · vi
                                                                   (1 + ψ)(1 + (k◦ /k)2 )
plasma lying above them on common £eld lines,
                                                                                  k◦      1
suggesting that relatively shallow modulations cre-            γ = (ωr − k · vi ) −                  2
                                                                                              (D⊥ k⊥ + D k 2 )
ated by billows can still account for plasma struc-                               k      1+ψ
tures that are more extended in the vertical. A short-           − 2αn◦
coming of the billows mechanism is that it does not                 νe νi        Ω2 k
                                                               ψ =           1+ 2 2
explain the preferred propagation direction for the                Ωe Ωi         ν e k⊥
QP structures since the K-H instabilities should not                     νin β 2                    1
have a preferred direction.                                    k◦ ≡                     β2 ≈ 1 −
                                                                         Ωi L 1 + ψ                kd

where ve and vi are the background electron and                  faster growth rate than gradient drift. This insta-
ion drift velocities, respectively, ψ is the anisotropy          bility derives free energy from the parallel conduc-
factor, k is the wavenumber, D⊥ and D are the                    tivity gradient in the layers rather than the perpen-
perpendicular and parallel ambipolar diffusivities,              dicular conductivity gradient and favors a small but
α is the recombination rate, and where β is a nonlo-             £nite parallel wavenumber, whereas gradient drift
cal parameter re¤ecting the fact that the background             waves grow most readily for k = 0. The instabil-
density gradient length scale L is con£ned to a re-              ity is like a collisional drift instability and involves
gion of space of size d.                                         electrons streaming along B to preserve charge neu-
     At midlatitudes, this dispersion relation can be            trality as the waves propagate across B. Because
generalized by treating the magnetic £eld lines as               of parallel gradients in the parallel conductivity, the
equipotentials and replacing all the parameters with             electrons are out of phase with the ions and over-
appropriately computed ¤ux-tube-integrated terms.                shoot, and instability results.
In the case of a sheet-like sporadic E layer, the                     Figure 6 depicts several two-dimensional cuts
integrated gradient scale length L becomes essen-                through the three dimensional simulation volume
tially in£nite, and so the growth rate vanishes.                 for different simulation timesteps. In the £rst
Seyler et al. (2004) recovered growth by consider-               timestep, a sporadic E layer patch immersed in a
ing intermediate- and small-scale primary waves,                 meridional background electric £eld generated by
for which geomagnetic £eld lines are not equipo-                 the F region dynamo is shown to become polarized,
tentials, and Kagan and Kelley (1998) considered                 with a polarization electric £eld directed northeast
how variable wind forcing along £eld lines could                 (upward and to the right). Over time, the layer be-
also restore wave growth. We need not take recourse              comes structured and produces regular, longitudi-
to these theories in the case of patchy sporadic lay-            nal waves propagating in the direction of the orig-
ers, however, since the ¤ux-tube-integrated gradient             inal polarization electric £eld. The growth time
scale length is already £nite and potentially small.             for the waves in the linear regime was found to be
     Instability growth is proportional to ve − vi , i.e.        about 80 s. The wavelength of the waves, which
to the current density, where the electron and ion               is closely tied to the parallel conductivity gradient
velocities are given approximately by the E × B                  length scale, is always close to 1 km. Waveforms
drift velocity and the neutral wind velocity, respec-            exist throughout the layer rather than on just one
tively. The density gradient that matters in calcu-              edge and propagete in the direction of the main po-
lating k◦ is the component transverse to the cur-                larization electric £eld in the cloud. Another im-
rent density and to the geomagnetic £eld, i.e., γ ∝              portant, distinguishing characteristic of the waves is
(ve − vi ) · ⊥ lnn × B. Consequently, we expect                  that the potential and density ¤uctuations associated
primary wave growth to occur only on the side of                 with them differ in phase by precisely 135◦ , as op-
a sporadic E layer patch with a favorable transverse             posed to 90◦ for gradient drift waves. The instabil-
density gradient. The nonlocal term represented by               ity appears to be robust and can grow under widely
the β 2 factor above limits the wavelength of the                varying conditions in terms of the background elec-
wave to be less than the dimension over which the                tric £elds and winds. The waves ultimately satu-
gradient exists, which we may take to be the radius              rate at large amplitudes approaching 100% relative
of the patch, typically a few kilometers. The waves              amplitude. The waveforms emerging from simula-
should propagate in the direction normal to the main             tion were similar to those observed by Kelley et al.
polarization electric £eld in the patchy layer and to            (1995). If collisional drift instabilities are found to
the magnetic £eld.                                               exist in sporadic E layers, they would represent an
     Gradient drift instabilities offer a viable expla-          important new mechanism for hastening their dissi-
nation for both the kilometric and the small-scale               pation.
irregularities observed in sporadic E layers just as
they do in the equatorial electrojet. However, using
a fully three-dimensional simulation of patchy spo-
radic layers in the midlatitude E region, Hysell et al.
(2002b) identi£ed another plasma instability with a

Figure 6: Simulation of the evolution of an unstable plasma cloud. Simulation timesteps 1, 14, 16, 18,
20, and 22 are shown from left to right, top to bottom. Each time step corresponds to 32 s elapsed time.
Grayscales depict plasma number density in cuts through a patchy sporadic layer. Contours are equipoten-
tials. The horizontal and vertical dimensions of the plots are about 25 km. Horizontal lines are surfaces of
constant altitude.

4 Methodology – experiment de-                               repetition of the measurements will provide redun-
                                                             dancy and will allow us to appraise the universality
  tails                                                      of our £ndings. Conditions for launch will include
To summarize, our proposed experiments pertain to            the observation of sporadic E layers on the Wallops
patchy sporadic E layers and quasiperiodic struc-            ionosonde and of QP echoes on the coherent scatter
tures in the postsunset midlatitude ionosphere and           radar imager located at Fort Macon, Virginia. We
seek to answer to the following questions:                   will also require clear skies for the ground-based
                                                             optical instrumentation involved in the experiments.
   1. Is the MLT region dynamically unstable dur-            Below, we summarize the methodology to be em-
      ing QP echo events?                                    ployed in our study.

   2. Are neutral shear instabilities or large-scale         4.1   Rocket measurements
      plasma instabilities mainly responsible for the
      formation of patches and quasiperiodic struc-          One instrumented and two chemical release rockets
      tures in sporadic E layers?                            will be launched together on two different nights.
                                                             The instrumented rocket will measure plasma den-
   3. What additional, intermediate-scale plasma             sities and vector electric £elds along with electron
      instabilities are involved in the production of        and ion temperatures and electron collision frequen-
      QP radar echoes?                                       cies in the E region. Low frequency density and
                                                             electric £eld data will be used to identify regions
    We will address these questions with an exper-           where patchy sporadic E layers are encountered and
iment involving instrumented and chemical release            to diagnose the background electrodynamics, i.e.,
rockets, ground based measurements, and analysis.            the background forcing and main polarization elec-
We propose to ¤y two sets of three rockets. The              tric £elds in the layers. High frequency density and

electric £eld observations will be useful for iden-              • Plasma frequency probe (PFP)
tifying the plasma waves and instabilities present
in the layers and for determining their dispersion               • Plasma sweeping probe (PSP)
characteristice, e.g. the propagation direction of the           • DC probe (DCP)
waves, their dominant wavelength, and the phase re-
lationship between the density and potential ¤uctua-             • Sweeping Langmuir probe (SLP)
tions. The plasma temperature and the electron col-
lision frequency measurements are useful byprod-              These types of probes have been ¤own by USU
ucts of the PIP instrument and will be useful for             on numerous ionospheric diagnostic sounding rock-
improving our estimates of the plasma conducitiv-             ets and satellites since the early sixties, including
ity. Neutral parameters will also be measured by the          nearly twenty sounding rockets launched at low,
photometer and the ionization gauge on each instru-           mid, and high latitudes since 1987 (Jensen and
mented payload. The photometer will detect ¤uc-               Baker, 1992). Flight heritage includes COPE II,
tuations in the neutral density which can be corre-           Thunderstorm II, CRIT II, SISSI, EQUIS I, SAL, E-
lated with plasma density ¤uctuations to determine            winds, and EQUIS II. The plasma impedance probe
to what extent the neutral ¤ow in¤uences the plasma           techniques are optimized when the probing anten-
structuring. The IG measures temperature pro£les              nas are extended from the payload and out of the
necessary for assessing the stability of the MLT re-          vehicle wake. This will be achieved by placing the
gion. Neutral and plasma temperatures can be com-             sensors along the vehicle spin axis and aligning the
pared in the region where the measurements overlap            payload, via the attitude control system, with the lo-
as a means of validation and calibration.                     cal magnetic £eld (perpendicular to the ¤ight direc-
     The chemical release rockets will provide neu-           tion). Antenna lengths should ideally be a factor of
tral wind pro£les for assessing stability while also          ∼10 or more greater than the characteristic plasma
serving as tracers of the neutral dynamics. One of            Debye length to insure that the impedance of the an-
these will be launched to a high apogee and will              tenna is the dominant factor of a collective plasma
make nearly vertical upleg and downleg releases               response. This is always the case for reasonable
between 70–150 km altitude. The other will be                 probe dimensions in the lower ionosphere.
launched to a low apogee and will make a nearly                   The PFP and PSP techniques are based upon the
horizontal release through the sporadic E layer re-           analysis of the impedance pro£le of an electrically
gion. Analysis of the three trails will provide ver-          short antenna immersed in ionospheric plasma. The
tical pri£les of the horizontal winds and permit us           most common type of antenna used in theory and
to assess the importance of horizontal wind gradi-            experiment in the past decades has been that of a
ents. The evolution of the trail deployed by the low-         short dipole or monopole. To £rst order, the antenna
apogee release will indicate the absense or presence          impedance ZANT as a function of frequency, when
of K-H billows in the region where the instrumented           driven at approximately a volt or less, is only de-
rocket ¤ew and the region from which QP echoes                pendent on the average dielectric properties encom-
arise. The particular instruments to be ¤own are de-          passed by the near £eld of the antenna. It is indepen-
scribed below and summarized in Table 1.                      dent of the surface characteristics of the antenna and
                                                              instrumentation ground reference. The antenna res-
                                                              onates with the characteristic plasma frequencies,
4.1.1 Instrumented rockets
                                                              producing distinct magnitude and phase transitions
Utah State Plasma Impedance Probe The USU                     in the impedance pro£le. The PFP technique deter-
Plasma Impedance Probe (PIP) is a suite of in-                mines Ne by tracking a resonance associated with
struments capable of making simultaneous mea-                 the upper hybrid frequency, ωh . ZANT closely re-
surements of several ionospheric parameters includ-           sembles that of a parallel resonant electrical circuit
ing electron density, electron-neutral collision fre-         near ωh with the phase of ZANT abruptly passing
quency, and electron and ion temperatures. The PIP            through 0◦ from ±90◦ . The 0◦ phase crossing point
is a combination of four plasma diagnostic instru-            is easily tracked using a closed loop control system,
ments:                                                        resulting in absolute Ne measurements accurate to
                                                              within 1% with sampling rates greater than 10 kHz

  Instrument                         Parameter                            Provider                    Payload
  PIP/SLP                            Ne , Te , Ti , νen                   USU (Swenson)               1
  Electric £eld probe                DC/AC Electric £elds                 USU (Swenson)               1
  Photometer                         Neutral density                      Aerospace (Hecht)           1
  IG                                 Neutral temperature                  Aerospace (Hecht)           1
  Chemical releases                  Neutral wind pro£les                 Clemson (Larsen)            2&3

                                       Table 1: Rocket Instrumentation

possible. Relative variations as small as 0.05% have          an issue in the ion portion of the Langmuir curve.
been routinely measured at altitudes greater than             The primary disadvantage of Langmuir probe tech-
150 km. The PSP makes impedance magnitude and                 niques is that the experimenter must accurately ac-
phase measurements at multiple points over a wide             count for the varying parametric in¤uences of the
bandwidth (typically 0.1-15 MHz) and at frequency             probe surface conditions and overcome such exper-
values near ωh , the electron-neutral collision fre-          imental dif£culties as ground reference variations,
quency νen is derived from the Q of the parallel-             contamination, and probe geometry effects. One
type resonance as observed in the frequency mea-              way to mitigate these problems is to control the
surements. The PSP impedance curves are also £t               probe surface after integration and to clean the sur-
against theory to recover absolute electron density           face, through an integrated heating source, on as-
in the strongly collisional D and E regions of the            cent. The PFP and PSP techniques will not be sub-
ionosphere where the ωh resonance is damped and               ject to such concerns since the surface conditions of
can not be tracked using a PFP. Ionospheric elec-             the antenna are not crucial. Variations of the surface
tron density measurements in the range of 103 –106            work function potential do not affect the RF char-
cm−3 in collisional and collisionless plasmas are             acteristics of the antenna and hence do not affect
easily attainable by employing a PFP/PSP combi-               the accuracy of the measurement. By using both
nation.                                                       impedance and Langmuir probe techniques, abso-
    The DCP and SLP techniques are based on                   lute Ne , Te , Ti , and νen can be measured with con£-
the DC response of plasma to an applied potential.            dence. With these measurements, coupled with neu-
They are part of the family of probes that are of-            tral density models and known collisional cross sec-
ten just called Langmuir Probes. A Langmuir probe             tions, the Hall and Pedersen conductivities can be
consists of a small conducting electrode immersed             determined.
in a plasma. This electrode is connected to a source
that can be biased at various potentials with respect         Photometers Each instrumented payload will
to some appropriate ground potential reference. The           carry a set of three photometers provided by The
method then consists of interpreting the current col-         Aerospace Corporation to provide upward-looking
lected by the probe as a function of the applied volt-        zenith brightnesses for three emission features. The
age. The DCP functions in the electron saturation             photomultiplier tubes are all mounted in an alu-
region of the Langmuir curve and provides a rela-             minum block, as shown in Figure ?? that is cooled
tive measurement of Ne in complement to the PFP               prior to ¤ight by liquid nitrogen to a temperature of
measurements. Ne measurements can be made in a                approximately −30◦ C to reduce the dark count rates
plasma with densities less than 5 × 103 cm−3 . The            to acceptable levels. The LN2 lines are cut at launch
SLP will employ a different sensing surface than the          and the block stays at low temperature throughout
DCP and will be used as a diagnostic of Te , Ti , and         the ¤ight. Three £lters used to isolate channels cen-
Ne at a 40 Hz rate. These parameters will be deter-           tered on the following features. One is at 762.0 nm
mined from the electron and ion retarding regions of          with a FWHM of 8 nm and therefore isolates the
the Langmuir curve using two different gain chan-             O2A(0,0) band. The second is at 773.5 nm with
nels. Since these will be night launches, photoelec-          a FWHM of 5 nm and therefore isolates part of
tron emissions from the probe surface will not be             the OH(9,4) band. The third is at 820.0 nm with

a FWHM of 5 nm and isolates background (non-                       Clemson University will provide the four chem-
OH or O2A airglow). All data will be corrected for             ical release payload sections and the photographic
the van Rhijn effect. Following the ETON approach              and imaging equipment required to support for the
atomic O densities can be obtained from these data.            chemical release experiment.
This experiment was successfully ¤own in the Co-
qui Dos campaign and TOMEX 2000 experiment.                    4.1.3 Vehicle requirements
Atomic O could be obtained at vertical resolution
of about 1 km from 83 km to 105 km.                            The experiment objectives require that three rock-
                                                               ets be launched along the same azimuth within a
                                                               few minutes. The £rst vehicle will carry a chemical
4.1.2 Chemical Release Rockets
                                                               tracer payload with an estimated weight of 100 kg
Trimethyl aluminum (TMA) reacts with oxygen to                 to an apogee near 200 km. The second vehicle
produce chemiluminescence that makes it visible at             will carry a chemical tracer payload with an esti-
night. The liquid can be released as a trail to provide        mated payload weight of 100-kg to a low apogee
a tracer for the motion of the neutral component of            also near 120-km altitude. The third vehicle will
the atmosphere that can be tracked with cameras                carry an instrumented payload with an estimated
from the ground. Triangulation of the images pro-              payload weight of 150 kg to an apogee near 120
vides wind pro£les with typical accuracies of 3–5              km. A Terrier-Improved Orion is suf£cient for all
m s−1 over the altitude range from 70 to 180 km alti-          three ¤ights. The measurements will be repeated on
tude. A height resolution for the measurements of 1            a second night, giving a requirement for six rock-
km can easily be achieved. The technique has been              ets. Each instrumented payload requires a magnetic
used extensively by Clemson University in recent               attitude control system (ACS) to align its spin axis
years and by various other investigators prior to that.        with the ambient magnetic £eld. Telemetry is not
Each payload will incorporate a nine-inch diameter             required for the chemical tracer payloads. No other
canister with 6 kg of TMA. The canister design is              special requirements are anticipated.
shown in Figure ??. A solenoid valve system is used
to modulate the release, so that a single canister             4.2   Ground based radar imager
can be used to provide both the upleg and downleg
trails, as well as generating the puffs which improve          The 30 MHz coherent scatter radar for this inves-
the accuracy of the triangulation and make it pos-             tigation is the portable radar interferometer used to
sible to obtain vertical velocities from the motion            observe QP echoes from Clemson, South Carolina
of the releases. The puffs will be 1.0 s in duration           and St. Croix, USVI. A description of the radar sys-
with a 1.0 s delay between puffs. The puffs will be            tem hardware can be found in Hysell et al. (2002a).
tracked to determine horizontal and vertical veloc-            This radar differs from more conventional systems
ities. TMA has a larger molecular weight than the              in its use of interferometry with multiple baselines
background atmospheric constituents so that down-              to construct true images of the radar targets illu-
ward velocities are expected for the puffs. The dif-           minated by the transmitting antenna beam. Inter-
ference between the expected and observed vertical             ferometric imaging was implemented £rst at Jica-
drift can be used to determine the local atmospheric                                     uu u
                                                               marca by Kudeki and S¨ r¨ c¨ (1991) and was an-
velocity as described by Rieger (1974). Based on               alyzed in detail subsequently by Woodman (1997).
past experience with puffed TMA, accuracies of a               It is well known that interferometry using a single
few meters per second for localized vertical veloc-            antenna baseline yields two moments of the radio
ity estimates will be achieved, which should be suf-           brightness distribution, the distribution of received
£cient to identify regions of large vertical ¤ow. The          power versus bearing (Farley et al., 1981). Inter-
focus of our experiment is not the vertical veloci-            ferometry with multiple baselines yields multiple
ties per se, but accounting for the vertical velocities        moments, and the totality of these moments can
explicitly gives a more accurate estimate of the hor-          be inverted to reconstruct the brightness distribu-
izontal winds. When a continuous trail is used, no             tion versus azimuth and zenith angle. The inversion
vertical velocity information can be extracted.                essentially amounts to performing a Fourier trans-
                                                               form of the interferometry cross spectra (Thomp-

son, 1986). However, since the cross-spectra are             than has been possible in the past. The Digisonde
inevitably incompletely sampled due to the limited           has many new capabilities lacking in traditional
number of interferometry baselines available, and            ionosondes. For our experiments, however, we
because of the presence of statistical ¤uctuations           are interested merely in identifying the presence of
in the data, the inversion must generally be per-            patchy sporadic E layers near the launch zone. In
formed using statistical inverse methods to achieve          ionograms, the patchy layers we seek to investigate
satisfactory results nearly free of artifacts (Ables,        typically have very high critical frequencies and
1974; Jaynes, 1982). For our imaging work, we                much lower bottom frequencies and are not blan-
have employed the MAXent algorithm pioneered                 keting. Note that our primary diagnostic tool for as-
for applications in radio astronomy (see for exam-           sessing launch conditions will be the coherent scat-
ple Wilczek and Drapatz (1985)). This is consid-             ter radar imager.
ered a “super-resolution” method since there reso-
lution of the images it produces is not limited by
the Nyquist sampling theorem. Our problem differs
                                                             5 Scenarios Leading to Closure
from that in radio astronomy mainly in that radar
                                                             We will answer the science questions posed in this
range gating adds the third dimension to the images
                                                             proposal depending upon which of several scenar-
(Hysell, 1996). The time evolution of the scattering
                                                             ios the experiments unfold. Some of the possible
medium is moreover revealed by comparing images
                                                             scenarios are listed below.
from successive integration times.
    The 30 MHz radar will provide range-time-                   • The complete, nonlocal analysis of the mea-
intensity information like that plotted in Figure 2               sured neutral winds and temperatures will in-
along with Doppler spectrograms in real time, pro-                dicate that the MLT region is dynamically un-
viding the investigators with information about the               stable during the QP echo experiments, and
occurrence, intensity, approximate location, and dy-              instability will be veri£ed by the presence
namics of the QP echoes upon which to base the                    of K-H billow structures in the low-apogee
launch decision. During post-¤ight analysis, we                   TMA trail deployment. This result would be
will generate sequences of plots like those in Fig-               consistent with earlier but less conclusive ex-
ure 4 and Figure 5 in order to place the rocket data              periments and would con£rm the importance
in context. It will be clear from the images when the             of shear instabilities in mixing and atmo-
rocket was penetrating patchy layers and quasiperi-               spheric/ionospheric coupling in the region.
odic layer structures, what the propagation direc-
tion of the structures was, and where the strongest             • The MLT region would be found to be dy-
£eld aligned irregularities were. In combination                  namically stable during QP echo events, ei-
with the in situ plasma and neutral measurements,                 ther because of insuf£cient shear strength or
the TMA-derived wind velocity pro£les, and photo-                 because of the inherent convective stability of
graphic images of K-H billows, the radar data will                the region. This £nding would call into ques-
provide a complete description of the QP events and               tion the importance of shear instability in at-
thereby allow us to resolve the science questions we              mospheric mixing and sporadic E layer struc-
have posed.                                                       turing.

                                                                • QP echoes occur when there is strong shear
4.3   Wallops Island ionosonde                                    and evidence of billows modulating the height
The Digisonde at Wallops Island is being replaced                 of the sporadic layer with a close correlation
by a more modern Dynasonde unit which is ex-                      between billow scales and ¤uctuation scales
pected to be operating by December, 2005. The up-                 in the plasma density and electric £elds. The
grade will increase the sensitivity of the ionosonde              low apogee rocket will be critical in map-
and also lower the minimum effective operating fre-               ping out the horizontal structure in the neu-
quency by virtue of the increased antenna size, mak-              tral atmosphere. The high apogee rocket will
ing it possible to detect weaker sporadic E layers                extend the wind pro£les and stability infor-
                                                                  mation above the altitudes covered by the li-

  dar. If the structure in the plasma density and           6 Impact of Proposed Work
  electric £elds closely matches the structure in
  the billows and the neutral density ¤uctua-               7 Relevance to Past, Present, and
  tions, that will suggest that altitude modula-
  tions created by those structures are important             Future NASA Programs
  in initating the QP structures.
                                                            NASA has a long history of supporting investi-
• QP echoes occur when there are neutral                    gations of plasma irregularities in the midlatitude
  shears but no evidence of billow or signi£-               ionosphere with sounding rockets. Past campaigns
  cant wave structure. This represents the sim-             include Coqui and Coqui II as well as numerous ex-
  plest result and would indicate that the Cos-             periments ¤ows out of Wallops. Our study com-
  grove/Tsunoda mechanism is of primary im-                 plements rather than repeats earlier work by focus-
  portance. The low-apogee rocket data will                 ing for the £rst time on the role of neutral atmo-
  be critical in verifying the lack of horizon-             spheric stability in creating conditions necessary for
  tal neutral structure. The electric £eld mea-             QP echoes. Atmospheric ionospheric coupling is a
  surements on the instrumented rocket will                 key component of the Geospace Sciences discipline
  be critical for determining if strong polar-              of NASA’s Sun-Earth Connection science theme.
  ization electric £elds are present. The high-             Our focus on the structure and effects of lower ther-
  apogee rocket will be critical in extending the           mospheric winds moreover help to support the sci-
  wind pro£les to higher altitudes to determine             enti£c goals of the imminent TIMED mission.
  the stability characteristics above the altitudes
  covered by the lidar.
                                                            8 Management Approach
• Primary plasma waves are found to exhibit
                                                            The overall project planning and post-¤ight data
  characteristics of gradient drift waves. If the
                                                            analysis will be conducted under the direction of the
  primary plasma waves occur on the side of
                                                            P.I., D. L. Hysell, with participation by all the co-
  the plasma patches favorable for gradient drift
                                                            investigators and collaborators. The P.I. has exten-
  instability, propagage in the direction normal
                                                            sive experience with radar experiments in the mid-
  to the main polarization electric £eld in the
                                                            latitude ionosphere and has been the P. I. of a rocket
  layer, and have density and potential ¤uctua-
                                                            experiment previously. All of the investigators have
  tions that are 90◦ out of phase, that will sug-
                                                            extensive experience with the analysis of data from
  gest that gradient drift instabilities play a cru-
                                                            the mesosphere/lower thermosphere/ionosphere re-
  cial role in producing the £eld aligned irregu-
                                                            gion. Several of the Co-Investigators (Larsen,
  larities responsible for coherent radar scatter
                                                            Swenson) and the P.I. (Hysell) have extensive expe-
  and QP echoes.
                                                            rience with sporadic E and QP studies using radars,
• Primary plasma waves behave like collisional              rockets, or both. The launches are planned for
  drift waves. If the primary plasma waves oc-              June/July 2007. A planning meeting will be ar-
  cur throughout the patchy sporadic E layers,              ranged soon after the proposal is approved and again
  propagate in the direction parallel to the main           in the fall prior to the experiment. A data analysis
  polarization electric £eld, and and have den-             workshop will be set up for the fall of 2007 to fa-
  sity and potential ¤uctuations closer to 135 ◦            cilitate the comparison of the various data sets. The
  out of phase, that will constitute strong evi-            roles of the individual co-investigators and collabo-
  dence that collisional drift instabilities instead        rators is summarized below.
  predominate.                                              D. L. Hysell (Cornell) – P. I. Radar data acquisi-
                                                            tion and analysis, overall data analysis and synthe-

                                                            M. F. Larsen (Clemson) – Co-I. Chemical re-
                                                            lease payloads; ground-based optical support for
                                                            chemical releases.
C. Swenson (Utah State University) – Co-I. PIP           References
and electric £eld instruments.
J. Hecht (Aerospace) – Co-I. Payload photome-            Ables, J. G., Maximum entropy spectral analysis,
ters and ionization gauges.                                Astron. Astrophys. Suppl. Ser., 15, 383, 1974.

                                                         Barnes, R. I., An investigation of the horizontal
Timeline                                                   structure of spread Es, J. Atmos. Terr. Phys., 54,
                                                           391, 1992.
   • January 2005: grant approved
                                                         Bernhardt, P. A., The modulation of sporadic-E lay-
   • February 2005: mission initiation conference
                                                           ers by Kelvin Helmholtz billows in the neutral at-
   • June 2005: comprehensive design review                mosphere, J. Atmos. Sol. Terr. Phys., 64, 1487,
     conference                                            2002.

   • Summer 2006: payloads complete                      Chau, J. L., and R. F. Woodman, Low-latitude
                                                           quasi-periodic echoes observed with the Piura
   • Fall 2006: testing and environmental at Wal-          VHF radar in the E region, Geophys. Res. Lett.,
     lops                                                  26, 2167, 1999.
   • Spring 2007: radar deployed to Fort Macon           Choudhary, R. K., and K. K. Mahajan, Tropical
                                                           E region £eld-aligned irregularities: Simultane-
   • Summer 2007: experiments conducted
                                                           ous observations of continuous and quasiperiodic
   • Fall 2007: data analysis workshop                     echoes, J. Geophys. Res., 104, 2613, 1999.

                                                         Cosgrove, R. B., and R. T. Tsunoda, Polariza-
9 Budget justi£cation                                      tion electric £elds sustained by closed-current dy-
                                                           namo structures in midlatitude sporadic E, Geo-
                                                           phys. Res. Lett., 28, 1455, 2001.

                                                         Cosgrove, R. B., and R. T. Tsunoda, A direction-
                                                           dependent instability of sporadic-E layers in the
                                                           nighttime midlatitude ionosphere, Geophys. Res.
                                                           Lett., 29, 10.1029/2002GL014669, 2002a.

                                                         Cosgrove, R. B., and R. T. Tsunoda, Wind-
                                                           shear-driven, closed-current dynamos in mid-
                                                           latitude sporadic E, Geophys. Res. Lett., 29,
                                                           10.1029/2001GL013697, 2002b.

                                                         Cosgrove, R. B., and R. T. Tsunoda, Simula-
                                                           tion of the nonlinear evolution of the sporadic-
                                                           e layer instability in the nighttime midlati-
                                                           tude ionosphere, J. Geophys. Res., 108, 1283,
                                                           doi:10.1029/2002JA009728, 2003.

                                                         Djuth, F., P. A. Bernhardt, C. A. Tepley, J. A. Gard-
                                                           ner, M. C. Kelley, A. L. Broadfoot, L. M. Kagen,
                                                           M. P. Sulzer, J. H. Elder, B. Isham, C. Brown, and
                                                           H. C. Carlson, Production of large airglow en-
                                                           hancements via wave-plasma interactions in spo-
                                                           radic E, Geophys. Res. Lett., 26, 1557, 1999.

Farley, D. T., H. M. Ierkic, and B. G. Fejer, Radar                               uu u
                                                            Kudeki, E., and F. S¨r¨c¨, Radar interferometric
  interferometry: A new technique for studying                imaging of £eld-aligned plasma irregularities in
  plasma turbulence in the ionosphere, J. Geophys.            the equatorial electrojet, Geophys. Res. Lett., 18,
  Res., 86, 1467, 1981.                                       41, 1991.

Hysell, D. L., Radar imaging of equatorial F region         Larsen, M. F., A shear instability seeding mecha-
  irregularities with maximum entropy interferom-             nism for quasi-periodic radar echoes, J. Geophys.
  etry, Radio Sci., 31, 1567, 1996.                           Res., 105, 24,931, 2000.

Hysell, D. L., and J. Burcham, HF radar observa-            Maruyama, T., S. Fukao, and M. Yamamoto, A pos-
  tions of quasiperiodic E layer echoes over North           sible mechanism for echo-striation generation of
  America, J. Geophys. Res., 104, 4361, 1999.                radar backscatter from midlatitude sporadic E,
                                                             Radio Sci., 35, 1155, 2000.
Hysell, D. L., and J. Burcham, The 30 MHz radar
  interferometer studies of midlatitude E region ir-        Mathews, J. D., Sporadic E: Current views and re-
  regularities, J. Geophys. Res., 105, 12,797, 2000.         cent progress, J. Atmos. Sol. Terr. Phys., 60, 413,
Hysell, D. L., M. F. Larsen, and Q. H. Zhou,
  Common volume coherent and incoherent scat-               Miller, K. L., and L. G. Smith, Incoherent scatter
  ter radar observations of mid-latitude sporadic E-         radar observations of irregular structure in mid-
  layers and QP echoes, Ann. Geophys., 22, 2004,             latitude sporadic E layers, J. Geophys. Res., 83,
  in press.                                                  3761, 1978.

Hysell, D. L., M. Yamamoto, and S. Fukao, Imag-             Ogawa, T., O. Takahashi, Y. Otsuka, K. Nozaki, M.
  ing radar observations and theory of type I and             Yamamoto, and K. Kita, Simultaneous MU radar
  type II quasiperiodic echoes, J. Geophys. Res.,             and ionospheric sounder observations of midlati-
  107, 1360, 2002a.                                           tude E region irregularities and sporadic E layer,
                                                              J. Geophys. Res., 107, 1275, 2002.
Hysell, D. L., M. Yamamoto, and S. Fukao, Simula-
  tions of plasma clouds in the midlatitude E region        Pan, C. J., and M. F. Larsen, Observations of QP
  ionosphere with implications for type I and type            radar structure consistent with neutral wind shear
  II quasiperiodic echoes, J. Geophys. Res., 107,             control of the initiation mechanism, Geophys.
  1313, 2002b.                                                Res. Lett., 27, 867, 2000.

Jaynes, E. T., On the rationale of maximum-entropy                                     o
                                                            Pan, C. J., C. H. Liu, J. R¨ttger, S. Y. Su, and J. Y.
  methods, Proc. IEEE, 70, 939, 1982.                         Liu, E region observations over Chung-Li dur-
                                                              ing the SEEK campaign, Geophys. Res. Lett., 25,
Jensen, M. D., and K. D. Baker, Measuring iono-               1805, 1998.
  sphere electron density using the plasma fre-
  quency probe, J. Spacecr. Rockets, 29, 91, 1992.          Pan, C. J., and P. B. Rao, Low altitude quasi-
                                                              periodic radar echoes observed by the Gadanki
Kagan, L. M., and M. C. Kelley, A wind-driven gra-            VHF radar, Geophys. Res. Lett., 29, 25–1, 2002.
  dient drift mechanism for midlatitude E-region
  ionospheric irregularities, Geophys. Res. Lett.,          Rieger, E., Neutral air motions deduced from bar-
  25, 4141, 1998.                                             ium releases experiments - I. Vertical winds, J.
                                                              Atmos. Terr. Phys., 1974.
Kelley, M. C., D. Riggin, R. F. Pfaff, W. E. Swartz,
  J. F. Providakes, and C. S. Huang, Large ampli-           Seyler, C. E., J. M. Rosado-Roman, and D. T. Far-
  tude quasi-periodic ¤uctuations associated with             ley, A nonlocal theory of the gradient-drift insta-
  a midlatitude sporadic E layer, J. Atmos. Terr.             bility in ionospheric plasmas at mid-latitudes, J.
  Phys., 57, 1165, 1995.                                      Atmos. Sol. Terr. Phys., 2004, in press.

Shalimov, S., C. Haldoupis, and K. Schlegel, Large            Woodman, R. F., M. Yamamoto, and S. Fukao,
  polarization electric £elds associated with midlat-          Gravity wave modulation of gradient drift insta-
  itude sporadic E, J. Geophys. Res., 103, 11,617,             bilities in mid-latitude sporadic E irregularities,
  1998.                                                        Geophys. Res. Lett., 18, 1197, 1991.

Smith, L. G., and K. L. Miller, Sporadic- layers and          Yamamoto, M., S. Fukao, T. Ogawa, T. Tsuda,
  unstable wind shears, J. Atmos. Terr. Phys., 42,              and S. Kato, A morphological study of mid-
  45, 1980.                                                     latitude E-region £eld-aligned irregularities ob-
                                                                served with the MU radar, J. Atmos. Terr. Phys.,
Swartz, W. E., S. C. Collins, M. C. Kelley, J. J.               54, 769, 1992.
  Makela, E. Kudeki, S. Franke, J. Urbina, N.
  Aponte, S. Gonzalez, M. P. Sulzer, and J. S.                Yamamoto, M., S. Fukao, R. F. Woodman, T.
  Friedman, First observations of an F-region tur-              Ogawa, T. Tsuda, and K. Kato, Mid-latitude E re-
  bulent upwelling coincident with severe E-region              gion £eld-aligned irregularities observed with the
  plasma and neutral atmospheric perturbations, J.              MU radar, J. Geophys. Res., 96, 15,943, 1991.
  Atmos. Sol. Terr. Phys., 64, 1545, 2002.
                                                              Yokoyama, T., M. Yamamoto, and S. Fukao, Com-
Thompson, A. R., Interferometry and Synthesis in                puter simulation of polarization electric £elds as a
  Radio Astronomy, John Wiley, New York, 1986.                  source of midlatitude £eld-aligned irregularities,
                                                                J. Geophys. Res., 108, 1054, 2003.
Tsunoda, R. T., J. J. Buonocore, A. Saito, T. Kishi-
  moto, S. Fukao, and M. Yamamoto, First obser-
  vations of quasi-periodic radar echoes from Stan-
  ford, California, Geophys. Res. Lett., 26, 995,

Tsunoda, R. T., W. L. Ecklund, and P. E. John-
  ston, Radar measurements of electric £elds in the
  topside of the equatorial electrojet: First results,
  Geophys. Res. Lett., 27, 2861, 2000.

Tsunoda, R. T., S. Fukao, and M. Yamamoto, On
  the origin of quasi-periodic backscatter from spo-
  radic E, Radio Sci., 29, 349, 1994.

Whitehead, J. D., The structure of sporadic E from
 a radio experiment, Radio Sci., 7, 355, 1972.

Whitehead, J. D., Recent work on mid-latitude and
 equatorial sporadic E, J. Atmos. Terr. Phys., 51,
 401, 1989.

Wilczek, R., and S. Drapatz, A high accuracy al-
 gorithm for maximum entropy image restoration
 in the case of small data sets, Astron. Astrophys.,
 142, 9, 1985.

Woodman, R. F., Coherent radar imaging: Signal
 processing and statistical properties, Radio Sci.,
 32, 2373, 1997.