1 Introduction. considerably in the 1990 s due to the discovery

Transcription

1 Summary This is a proposal to investigate midlatitude sporadic E layers, layer irregularities, and so-called quasiperiodic (QP) echoes using the Arecibo incoherent scatter radar, its instrument cluster, and a new coherent scatter imaging radar to be deployed near the University of the Virgin Islands (UVI) on St. Croix. We are proposing to expand upon earlier observations from Arecibo and St. Croix made in the summer of 2002 which showed that the QP echoes arose from elongated, wavelike structures in sporadic E layers of the kind observed previously at Arecibo and tentatively associated with convection driven by neutral shear instabilities. We propose to deploy an autonomous radar in St. Croix for long-term, remote observations in collaboration with the Arecibo facility aimed at assessing the stability of the neutral mesosphere-lowerthermosphere (MLT) region, identifying the true cause of the layer structuring, and studying the plasma physics involved in producing the QP echoes. The intellectual merit of the proposal lies in the investigation of neutral/plasma coupling and other processes in this complex environmental system. Sporadic E layers and irregularities drastically affect HF and VHF radio wave propagation, and our project also addresses this often neglected aspect of space weather. Finally, our emphasis on MLT wind measurements supports that aspect of the TIMED mission. The broader impact of our proposal is related to the educational experiences we propose to bring to the diverse undergraduate student body at UVI, as well as to graduate students at Cornell and Clemson University. Furthermore, the imaging radar we propose to place on St. Croix will add long-overdue coherent scatter capabilities to the Arecibo Observatory instrument cluster, accessible by the community for studies of sporadic E, midlatitude spread F, and meteor echoes, among other topics of general interest. The imaging radar is an innovative observing system. It will become part of the facility and add to the capabilities of the Arecibo ISR and optical instrumentation. Our proposal has the endorsement of Arecibo. This is a multi-institution proposal with co-investigators and students from Cornell, Clemson, and UVI taking part. 1

2 1 Introduction The existence of midlatitude sporadic E layers has been known almost as long as HF and VHF radio communications have been used (see, e.g., the comprehensive review by Whitehead (1972)). In the 1970 s and 1980 s, a considerable effort was made to improve our understanding of the mechanisms responsible for generating the enhanced ionization layers and the instabilities responsible for small-scale irregularity structures within those layers. Considerable progress was made although many questions remain, as discussed in the reviews by Whitehead (1989) and Mathews (1998), for example. Interest in the dynamics of the layers increased considerably in the 1990 s due to the discovery by Yamamoto et al. (1991, 1992) that quasiperiodic (QP) radar echo structure is often associated with sporadic E layers observed by the MU radar in Japan, but it was still not known until recently if the phenomenon was peculiar to the Japanese sector or if it was a more universal phenomenon. In the intervening years, more extensive observations were made, and the QP structures have now been detected in the American sector (Hysell and Burcham, 1999; Tsunoda et al., 1999; Chau and Woodman, 1999; Swartz et al., 2002) and in other locations in the Asian sector (Pan et al., 1998; Choudhary and Mahajan, 1999; Pan and Larsen, 2000; Pan and Rao, 2002), thus demonstrating that the structure is tied to the sporadic E layers rather than to a specific geographic location. The continued level of interest in quasi-periodic echo structures is indicated in part by the SEEK-2 sounding rocket campaign carried out in August, 2003, in Japan and by the special session of the international URSI meeting to be held in October in New Delhi devoted to the subject. The SEEK-2 results have been collected in submissions for a special section in Ann. Geophys., including contributions by Yamamoto et al. (2005); Larsen et al. (2005); Pfaff et al. (2005); Maruyama et al. (2005); Saito et al. (2005). Much more is now known about the occurrence and morphology of QP echoes than was known even just a few years ago, and it is currently believed that QP radar echoes arise from patchy irregularities in sporadic E layers like those first observed at Arecibo by Miller and Smith (1978) and Smith and Miller (1980). Larsen (2000) has argued that the patches could be produced by Kelvin Helmholtz instabilities in the neutral atmosphere driven by the same steep wind shears thought to create the layers, and Bernhardt (2002) has presented numerical simulations of the process which seem to demonstrate its viability. At the same time, however, Cosgrove and Tsunoda (2003) have argued that large-scale plasma instabilities associated with neutral shear alone can cause sporadic E layers to break up. In either case, the plasma density gradients, electric fields, and wind shears within the patches would provide the free energy for intermediate-scale plasma instabilities to occur which could in turn drive the smallscale field-aligned irregularities detected by radar. Intermediate-scale plasma waves are observed in the layers, but the particular instabilities responsible have not been identified. It is clear in any case that sporadic E and QP echoes are the result of complex interactions between the neutral atmosphere and ionization in the mesosphere and lower thermosphere region, but in spite of the growth in the observational database and of numerous modeling efforts, fundamental questions remain unanswered, including: 1. Is the MLT region dynamically unstable during QP echo events? 2. Are neutral shear instabilities or large-scale plasma instabilities mainly responsible for the formation of patches and quasiperiodic structures in sporadic E layers? 3. What additional, intermediate-scale plasma instabilities are involved in the production of QP radar echoes? We propose to address these questions using common volume data from the Arecibo incoherent scatter radar and a coherent scatter imaging radar to be constructed and deployed on the island of St. Croix. Our research is intended as a multi-year extension of a particularly revealing CEDAR experiment performed on St. Croix in the summer of We propose not only to advance our understanding of QP echoes and sporadic E layers but 2

3 also to give Arecibo a long-overdue coherent scatter capability. The St. Croix radar will be remotely accessible and open to the community for studies of midlatitude E and F region plasma irregularities in tandem with Arecibo s existing instrument cluster, including the incoherent scatter radar, ionosonde, and resonance lidar. An important element of the proposal is to extend the observational period significantly so that we not only obtain more interesting case studies but also obtain a more representative statistical and morphological picture of the behavior of the QP structures. The earlier CEDAR experiment that we carried out in 2002 focused on a three to four week period in June and July. The events that were observed clearly showed a number of common features and some common behavior, but there was also considerable diversity in the strength and characteristics of the echoes, including the motion that was observed. The limited extent of the observations in the earlier experiment make it difficult to determine how representative those observations were. By building on the work that was done in connection with the 2002 experiment, especially the experience that we now have in developing the radar instrumentation and the experience with the site in St. Croix, we expect to start the observations quickly after funding is approved and to obtain observations routinely by making the instrument autonomous and remotely accessible. Sporadic E layers, layer patches, and E region plasma irregularities can both enhance and degrade HF and VHF radio communications and studying the geophysical factors that cause and influence them contributes to the National Space Weather Program. Since atmospheric-ionospheric coupling appears to be an important controlling factor, we will also be addressing that key aspect of the CEDAR Phase III science theme. Finally, our emphasis on the structure and effects of winds in the lower thermosphere supports the mission of the TIMED satellite. 2 Results from Prior Support ATM /15/01 02/28/05 $202,000 This award primarily supported an observational study of sporadic E layers and quasi-periodic (QP) echo structure in the Caribbean above Puerto Rico. A coherent scatter imaging radar system was developed and installed on the island of St. Croix and was operated together with the Arecibo Observatory incoherent scatter radar on selected nights for a period of approximately one month during June and July This proposal is a direct follow-up to that experiment, and the results are described in more detail in the introductory material and main text. In addition, several studies of sporadic and turbulent layers in the E region or mesopause and lower thermosphere region were carried out. Some results were based on lidar observations of the neutral winds and stability in that part of the atmosphere. Other studies were based on numerical model results and dealt with tidal effects on sporadic layers and the interaction between the neutral atmosphere and the ionization in the E region to form intermediate layers. The topics covered by this award, namely radar imaging, parameter estimation from coherent scatter and incoherent scatter techniques, lidar measurements, and analysis of sporadic layers and turbulent neutral layers are all germane to the current proposal. The publications that have resulted from the previous award include: Larsen, M. F., A. Z. Liu, R. L. Bishop, and J. H. Hecht, TOMEX: A comparison of lidar and sounding rocket chemical tracer wind measurements, Geophys. Res. Lett., 30, /2002GL015678, 1375, Bishop, R. L., and G. D. Earle, Metallic ion transport associated with midlatitude intermediate layer development, J. Geophys. Res., 108, /2002JA009411, 1019, Bishop, R. L., M. F. Larsen, J. H. Hecht, A. Z. Liu, and C. S. Gardner, TOMEX: Mesospheric and lower thermospheric diffusivities and instability layers, J. Geophys. Res., 109, /2002JD003079, D02S03, Larsen, M. F., A. Z. Liu, C. S. Gardner, M. C. Kelley, S. Collins, J. Friedman, and J. H. Hecht, Observations of overturning in the upper mesosphere and lower thermosphere, J. Geophys. Res., 109, /2002JD003067, D02S04,

4 Hecht, J. H., A. Z. Liu, R. L. Bishop, J. H. Clemmons, C. S. Gardner, M. F. Larsen, R. G. Roble, G. R. Swenson, and R. L. Walterscheid, An overview of observations of unstable layers during the Turbulent Oxygen Mixing Experiment (TOMEX), J. Geophys. Res., 109, /2002JD , D02S01, Hecht, J. H., A. Z. Liu, R. L. Walterscheid, R. G. Roble, M. F. Larsen, and J. H. Clemmons, Airglow emissions and oxygen mixing ratios from the photometer experiment on the Turbulent Oxygen Mixing Experiment (TOMEX), J. Geophys. Res., 109, /2002JD003035, D02S05, Liu, A. Z., R. G. Roble, J. H. Hecht, M. F. Larsen, and C. S. Gardner, Unstable layers in the mesosphere region observed with Na lidar during the Turbulent Oxygen Mixing Experiment (TOMEX) campaign, J. Geophys. Res., 109, /2002JD003056, D02S02, Hysell, D. L., M. F. Larsen, and Q. H. Zhou, Common volume coherent and incoherent scatter radar observations of mid-latitude sporadic E-layers and QP echoes, Ann. Geophys., 22, /ag/ , , Bishop, R. L., G. D. Earle, M. F. Larsen, C. M. Swenson, C. G. Carlson, P. A. Roddy, C. Fish, and T. Bullett, Sequential observations of the local neutral wind field structure associated with E-region plasma layers, J. Geophys. Res., in press, Figure 1: Electron densities measured with the Arecibo linefeed antenna. The azimuth of the antenna feed, which was scanning, is plotted beneath. ations in its altitude would accompany and be synchronized with the azimuth scan. Moreover, there is no significant E region ionization above the layer. Several preliminary theories of QP echo generation relied on multiple dense layers, tilted layers, deeply modulated layers, and density perturbations extending high in the E region (e.g. Woodman et al. (1991), Tsunoda et al. (1994), Maruyama et al. (2000), Cosgrove and Tsunoda (2001), Ogawa et al. (2002), and Yokoyama et al. (2003)). However, none of those features are present in Figure 1 despite the fact that intense QP echoes were observed from the same ionospheric volume by a coherent scatter radar beginning at 1930 LT. Figure 2 shows coherent scatter data obtained from a radar located on St. Croix, USVI, and situated so as to observe field-aligned irregularities in the ionosphere above Puerto Rico, including the region immediately over Arecibo. The quasiperiodic echoes shown in Figure 2 started just as the sporadic E layer over Arecibo began to break into patchy structures. The five clusters of QP echoes which appear starting at 2330, 0015, 0100, 0200, and 0230 UT in the coherent scatter data are coincident with the passage of five groups of patches passing over Arecibo. Figure 3 shows a radar image derived from the same coherent scatter radar data. Radar imaging has 3 Science overview Figure 1 shows a sporadic E layer observed by the Arecibo incoherent scatter radar in the summer of The plasma densities are plotted in units of on a logarithmic scale. The zenith angle of the line feed was 15, and the azimuth angle, which scanned in time, is plotted below the density data. Local sunset was at 1904 LT, and most of the E region can be seen to have recombined by about 1920 LT. Similar (but different) data were recorded simultaneously using the Gregorian feed. Before about 20 LT, the scene was dominated by a strong, narrow layer at about 105 km with only a very weak layer beneath it at the base of the E region. The stronger layer appears to be neither tilted nor deeply modulated. If it were, regular vari- 4

5 Figure 2: Range-time-intensity (RTI) plot of coherent echoes received by a 30 MHz radar coherent scatter radar sharing Arecibo s scattering volume. Grayscales depict signal-to-noise ratio in db. Note that UT=LT+4 hours. been used to determine echo power versus bearing in three dimensions (Woodman, 1997; Hysell and Woodman, 1997). From range and bearing, we compute the latitude, longitude, and altitude. Altitude information is not depicted in this figure, however. The brightness, hue, and saturation of the colored regions indicate the signal-to-noise ratio, Doppler shift, and spectral width of the coherent echoes. Signal-to-noise ratios are plotted on a scale between 8 and 38 db here. The image in Figure 3 depicts a number of discrete, localized, patchy scatterers. Over time, animations show that the patches drift mainly westward or southwestward, although some on the west side of the image drift eastward toward St. Croix (which is off the right side of the figure). Each patch corresponds to a striation in the RTI diagram in Figure 2, and the range rates of the striations correspond to their line-of-sight drifts with respect to St. Croix. Differences in the motions of the patches give the striations in Figure 2 different slopes and even permit them to cross. We associate the patches in the images with drifting, polarized E region ionization patches. Evidence of polarization is found in the circulation evident in the patches, where the hue (Doppler shift) is typically different in the center and at the periphery. The X in the figure in northwest Puerto Rico represents the location of the Arecibo radar. Yellow and green lines radiating from this point represent, respectively, the regions where the beams from the Arecibo linefeed and Gregorian antennas intercept the E region between km in altitude at the time indicated. Images like Figure 3 make it possible to collocate sources of coherent scatter with the sporadic E layer features responsible for them. By animating sequences of images like Figure 3, we were able to compare the entire incoherent and coherent scatter datasets for the June event. Overall, we demonstrated a very high correspondence between the coherent backscatter patches and the ionization patches. A remarkable feature of Figure 3 is the tendency for the scattering patches to fall along lines or fronts. Similar linear formations of scatterers were observed with the MU and Clemson University radars (Hysell et al., 2002a). Over Puerto Rico, however, the lines were more evident and prone to coalesce into multiple, unbroken, frontal structures propagating across the ionosphere. Figure 4 shows an example of these quasiperiodic structures or waves. We regard the structures as waves, both because wavefronts are evident in the image brightness and because the observed Doppler shifts also vary periodically between wavefronts. In Figure 4, a large-scale wave with wavefronts running from northwest to southeast is shown. The wavelength is about 30 km, and the period is about 10 min. Animated sequences of images confirm that the wave propagates to the southwest. Doppler shifts alternate from large positive to negative values. These are the line-of-sight phase speeds of the small-scale irregularities observed from St. Croix. At the time that the large-scale wave was passing over Puerto Rico, Arecibo observed a number of very thick and intense plasma irregularities or patches with both of its feed systems. There is a high degree of correspondence between these clouds and the wavefronts in the coherent scatter radar images. It is when the yellow and green radial markers in the images pass through the brightest parts of the wavefronts that the corresponding Arecibo beam systems detect the patches. As the wavefronts are very elongated, we surmise that the ionization clouds are either very elongated or occur in long fronts which move with the waves. 5

6 Figure 3: Radar image of coherent echoes received at 2332 UT. Note that UT=LT+4. Figure 4: Radar image of coherent echoes received at 0039 UT. 6

7 Images of nm emissions recorded by the Boston University all-sky imager during the November 14 event show wavelike structuring very similar to what we find in the radar images (personal communication, Steve Smith). The similar features were observed late in the event when the sporadic E layer descended to the nm emission height. The wavelength was about 30 km in both the radar and the optical data, and the wavefronts were similarly oriented. The propagation direction of the waves in the optical images was northeast rather than southwest, but this is likely due to aliasing, as the optical images were recorded at 6 min. intervals, whereas the Nyquist period was 5 min. To our knowledge, this dataset represents the only common-volume observation of sporadic E layer structuring by incoherent scatter, coherent scatter, and passive optics. The morphology of the QP sporadic E structures and the associated coherent echoes is therefore understood: QP echoes arise from sporadic E layer patches, and the striations in coherent scatter RTI plots are mainly indicative of the proper motion of the patches toward or away from the radar. The patches are frequently elongated and organized along chains or frontal bands. However, this picture is incomplete; we do not not know the role played by shear and neutral instability or whether neutral or plasma processes are mainly responsible for the large-scale structuring. Nor do we know what plasma processes at intermediate scales are at work and responsible for creating the small-scale irregularities visible to coherent scatter radars. The picture will remain incomplete until three fundamental questions can be answered: Is the MLT region dynamically unstable during QP echo events? The altitude region where the QP echoes predominantly occur, i.e., the range between 95 and 110 km, is known to be characterized by large shears in the neutral wind. Large shears are frequently present statistically (Larsen, 2002) and, as shown by Larsen (2000) and numerous references therein, every rocket flight into sporadic E or QP conditions which included wind profile measurements has shown large shears in connection with the sporadic layer that were either shear unstable or close to shear instability. The neutral shears are undoubtedly important for the electrodynamics of the region, but if the neutral shears are also unstable, the vertical displacements produced in the neutrals could be responsible for generating new scale-length structure in the plasma that might not otherwise be present. It is important therefore to characterize any neutral shear instabilities as completely as possible. Two types of neutral instabilities are known to occur in the mesosphere, namely convective instabilities in which the lapse rate is close to the adiabatic lapse rate, and shear instabilities in which the Richardson number is less than The recent papers by Hecht et al. (2004), Larsen et al. (2004a), and Bishop et al. (2004) discuss recent observations obtained during the Turbulent Oxygen Mixing Experiment (TOMEX) and earlier observations and modeling studies related to such instabilities. Expressions for the Richardson number and the related Brunt Väisälä frequency are given by The Brunt Väisälä frequency is a measure of the neutral stability and becomes smaller, i.e., closer to zero as the atmospheric temperature lapse rate approaches the adiabatic lapse rate. The Richardson number also decreases as the magnitude of the shear increases. In the mesopause and lower thermosphere region where the QP layers are expected, the temperature generally increases with height so that increases with height. The main source of destabilization in the region of interest is therefore the shear. Nonetheless, there is sufficient variability in the temperature profiles in that altitude range, as shown by lidar measurements, for example, that an accurate assessment of the stability characteristics will require a measurement of the temperature profile in addition to the neutral wind velocity profile. The basic theory related to shear instabilities has been summarized in the texts by Gossard and Hooke (1975) and Scorer (1978), among others. Although the evolution of the shear instability is a highly nonlocal and nonlinear process, the predictions of the linear, local theory have been surprisingly useful in terms of predicting the onset of the instability and 7

8 the scale sizes and general characteristics of the initial perturbations. Scorer (1978), in particular, has discussed the agreement between theory and observations, although primarily in the context of the troposphere and the lower stratosphere. Larsen (2002) also summarized some of the more recent work. One shortcoming of many of the analytic and modeling treatments is that they assume the wind shear is a pure magnitude shear. In fact, every midlatitude wind profile measurement from the height range of primary interest to us ( km) shows significant rotation of the wind vector with height, with the wind direction often changing by as much as 360 in km. The presence of rotational shear has to be accounted for in order to properly characterize the effect of any neutral shear instabilities that may be present. Brown (1980) dealt with this issue in his review paper. More recent literature on the subject was summarized by Larsen (2002). In effect, a shear instability in the Richardson number sense that occurs in a rotational shear flow is maximally unstable to perturbations with wave vectors aligned with the vector direction represented by the difference in the wind vectors at the top and bottom of the unstable layer, i.e., wave vectors aligned with the shear vector. In a shear flow without rotation, the shear vector is parallel to the wind direction, and the standard result is obtained. In a flow with significant rotational shear, however, the propagation direction of the instabilities can be significantly different from the wind direction. Such effects need to be accounted for in fully characterizing the neutral wind drivers in an environment that is generating QP structure. Are neutral shear instabilities or large-scale plasma instabilities mainly responsible for the formation of patches and quasiperiodic structures in sporadic E layers? The mechanism responsible for breaking up sporadic E layers and initiating the process that leads to QP echoes has not been identified, but neutral dynamics are thought to play a crucial role. Miller and Smith (1978) and Smith and Miller (1980) showed that sporadic E layers are often collocated with unstable neutral shears, suggesting that K-H billows should be present. Larsen (2000) further argued that Kelvin-Helmholtz billows associated with shear instability in the neutral atmosphere can produce the large vertical displacements in the sporadic E layers needed to initiate plasma instabilities. Some direct evidence that K-H billows are present in the volume from which QP echoes are received emerged from the SEEK-2 experiments (Larsen et al., 2005). Bernhardt (2002) simulated the distortions in an initially planar E layer produced by Kelvin- Helmholtz billows. He found that the nonlinear response of the ionization is complicated but essentially mimics the structure in the neutral flow that is acting as the driver. The response also depends on the altitude of the layers and the billows since the conductivities play an important role in the dynamical processes. Cosgrove and Tsunoda (2001) and Cosgrove and Tsunoda (2002b) analyzed the polarization electric fields produced by vertical displacements in the ionization driven by neutral oscillations including shears in the neutral flows. Their results indicated that large fields can be explained by such dynamics. The large electric fields, plasma density gradients, and neutral winds present in this scenario would be available to drive the plasma instabilities ultimately responsible for QP echoes. Recently, however, Cosgrove and Tsunoda (2002a) and Cosgrove and Tsunoda (2003) investigated a new E layer instability. The instability shares some of the characteristics of the Perkins instability, which is often thought to be partly responsible for midlatitude spread F (Perkins, 1973). In the Perkins instability, a wavelike perturbation in the F layer height gives rise to an accompanying perturbation in the current density, since the Pedersen conductivity is height dependent. The quasineutrality condition forces the current density perturbations to be parallel to the wavefronts. Associated perturbations in the vertical component of the force modify the (quasi)equilibrium layer height in such a way as to reinforce the initial perturbation, and instability results. The unstable wavefronts may not be aligned magnetic north-south, since the perturbed force has no vertical component in that case. Nor may they be aligned magnetic eastwest, since that configuration entails no inhomogeneity in the height-integrated Pedersen conductivity, the quantity that matters for waves with wavelengths larger than about a kilometer. The Cos- 8

9 grove and Tsunoda E region instability is conceptually similar, except that perturbations in the current density arise from vertical neutral wind shear rather than from Pedersen conductivity height variations. Since neutral wind shear is a prerequisite for sporadic E layer formation, the instability seems to follow naturally. As with the Perkins instability, the growth rate maximizes when the wave fronts are aligned from northwest to southeast, consistent with the observed preferred alignment and propagation direction of the QP echo structures (see, e.g. Tsunoda et al., 2000; Hysell et al., 2004). Both of the mechanisms described above are viable, but both have potential shortcomings. Invoking Kelvin-Helmholtz billows as the mechanism for modulating the layers provides a simple and intuitive explanation for sporadic E layer structuring. The billows are intrinsic to the layer dynamics believed to be present in the MLT region, and since the shear driving the billows is the same shear generating the layers, the appropriate spatial relationship between the altitude of the layers and the altitude of the billows is automatically maintained. Yokoyama et al. (2003) recently showed that polarized E region plasma clouds can induce drastic perturbations in the densities and electric fields in the plasma lying above them on common field lines, suggesting that relatively shallow modulations created by billows can still account for plasma structures that are more extended in the vertical. A shortcoming of the billows mechanism is that it does not explain the preferred propagation direction for the QP structures since the K-H instabilities should not have a preferred direction. Meanwhile, the Cosgrove and Tsunoda E layer instability mechanism is appealing because it provides an explanation for the preferred propagation direction, as well as being consistent with some other known characteristics of the echo structures. A potential problem with the theory is that Kelvin- Helmholtz billows or other neutral atmospheric wave structure would tend to disrupt the instability (Cosgrove and Tsunoda, 2003). Since large and often unstable shears seem to be a consistent feature of the QP environment, the billow structure is expected to be present, at least part of the time. Moreover, the QP echo structures are known sometimes to have wavefront alignments in the directions excluded by the Cosgrove and Tsunoda mechanism. Cosgrove and Tsunoda (2003) therefore suggested that the shear-driven polarization instability may be the operative mechanism at certain times and that one of the other mechanisms, such as the K-H billows modulation mechanism, may be important at other times. Weighing the merits of the two mechanisms necessitates more incisive experimental investigation. What additional, intermediate-scale plasma instabilities are involved in the production of QP radar echoes? Kilometric plasma waves thought to signify primary plasma waves and instabilities have been detected in sporadic E layers in radio sounding data (Barnes, 1992), airglow experiments (Djuth et al., 1999), radio scintillations (Maruyama et al., 2000), and in rocket investigations from Wallops (Kelley et al., 1995). The small-scale irregularities responsible for coherent scatter are thought to be produced mainly by mode coupling to such primary waves. While a number of theories have been presented over the years, consensus regarding the particular intermediate-scale primary plasma instability at work is elusive. Farley-Buneman instabilities have been shown to occur in midlatitude sporadic E by Haldoupis and Schlegel (1994) and others, but these are rare and leave the kilometric primary waves and the majority of the radar echoes to be explained. The Cosgrove and Tsunoda (2003) mechanism described above appears to operate at too long a wavelength to explain the kilometric waves. A gradient drift instability similar to the one occurring in the equatorial electrojet is widely considered a likely candidate. However, finite parallel gradient scale length effects at middle latitudes were originally thought to stabilize gradient drift instabilities (Woodman et al., 1991). While Seyler et al. (2004) recently showed with a nonlocal analysis that intermediate-scale gradient drift instabilities with finite parallel wavenumbers may form in uniform sporadic E layers at mid latitudes, their theory does not account for the quasiperiodicity of the radar echoes or for the kilometric wavelengths observed and also suggests that laminar layers should produce echoes, contradicting the results 9

10 from Arecibo. The association of QP echoes with patchy sporadic E layers removes many theoretical objections for gradient drift instabilities. This is because the patches have flux-tube-integrated density gradients all around their periphery, as opposed to the localonly density gradients that exist on the topside and bottomside of a horizontally homogeneous layer. This obviates the need for either the nonlocal theory of Seyler et al. (2004) or the reversed wind forcing theory of Kagan and Kelley (1998) to understand how gradient drift instabilities can operate. (These various mechanisms may yet function, but neither seems strictly necessary.) Furthermore, Shalimov et al. (1998), Hysell and Burcham (2000a), and Cosgrove and Tsunoda (2002b) showed that patchy sporadic E layers can spontaneously produce very large polarization electric fields and currents when forced by transverse electric fields or winds, readily able to drive gradient drift and Farley-Buneman instabilities. We would expect primary gradient drift waves to form on one side of the patchy layers only, since only one side should be unstable. The primary wave wavefronts could be aligned in any direction perpendicular to, depending on the direction of the forcing winds and background electric field. The waves would propagate in the direction normal to the main polarization electric field in the patchy layer. Recently, however, Hysell et al. (2002b) conducted fully three-dimensional simulations of plasma patches in the midlatitude E region ionosphere and found that another instability existed with a faster growth rate than gradient drift. The collisional drift waves that grew were longitudinal (gradient drift waves are transverse), existed throughout the plasma cloud (rather than on just one side), and required a small but finite parallel wavenumber component (gradient drift waves are preferentially field aligned). The primary waves emerging from the simulations had kilometric wavelengths and propagated in the direction parallel to the main polarization electric field in the plasma cloud. Distinguishing between gradient drift and collisional drift instabilities in the layers can be accomplished using the Arecibo incoherent scatter radar, which has a beam sufficiently narrow to resolve and observe kilometric waves directly in the E region. Simultaneous coherent scatter imaging radar observations can guarantee that the Arecibo wave measurements correspond to the regions of space from which QP echoes emerge. 4 Methodology The centerpiece of this proposal is a new 30-MHz coherent scatter radar imager to be deployed near the University of the Virgin Islands (UVI) on St. Croix with its field of view in the E region over Arecibo. The capabilities of the radar will be similar to the one deployed in the earlier experiment except that this one will be autonomous and remotely accessible to experimenters at Arecibo or elsewhere across the internet. The radar will be used for our investigation and also made available to the community for support of other E and F region experiments at Arecibo. It will utilize some equipment left behind from our previous experiments, namely most of the antennas, but will incorporate new solid-state transmitter, digital receivers, digital waveform synthesis, and a high-speed data acquisition and processing system. Local support of the radar from our collaborators at UVI will assure long-term system reliability. We have identified a site for the radar adjacent to the campus of the University of the Virgin Islands on St. Croix and are making plans for power and internet connections, shelter, climate control, access, and security. QP echo events are coherent scatter events by definition, and the only way to know that the conditions observed with the Arecibo ISR are associated with the QP phenomenon is with common volume coherent scatter measurements. Radar imaging moreover is necessary to relate the coherent scatter returns to the specific volume probed by the ISR unambiguously. The radar imager uses interferometry with multiple baselines to construct true images of the radar targets illuminated by the transmitting antenna beam. Interferometric imaging was implemented first at Jicamarca by Kudeki and Sürücü (1991) and was analyzed in detail subsequently by Woodman (1997). It is well known that interferometry using a single antenna baseline yields two moments of the 10

11 radio brightness distribution, the distribution of received power versus bearing (Farley et al., 1981). Interferometry with multiple baselines yields multiple moments, and the totality of these moments can be inverted to reconstruct the brightness distribution versus azimuth and zenith angle. The inversion essentially amounts to performing a Fourier transform of the interferometry cross spectra (Thompson, 1986). However, since the cross-spectra are inevitably incompletely sampled due to the limited number of interferometry baselines available, and because of the presence of statistical fluctuations in the data, the inversion must generally be performed using statistical inverse methods to achieve satisfactory results nearly free of artifacts (Ables, 1974; Jaynes, 1982). For our imaging work, we have employed the MAXent algorithm pioneered for applications in radio astronomy (see for example Wilczek and Drapatz (1985)). Our problem differs from that in radio astronomy mainly in that radar range gating adds the third dimension to the images. The time evolution of the scattering medium is moreover revealed by comparing images from successive integration times. System details can be found in the article by Hysell et al. (2002a). Neutral stability As discussed in the science overview, an analysis of the stability of the neutral atmosphere during QP echo events requires measurements of neutral temperature and vector wind profiles in the altitude region where sporadic E layers form. At night and in the altitude regime in question, we are well justified in equating the neutral temperatures and drifts with the plasma temperatures and drifts in the sporadic E layers. We can therefore measure the desired profiles essentially directly with incoherent scatter. The sensitivity of the Arecibo radar affords the possibility of rapid, accurate, high-resolution profile measurements within sporadic E layers. In order to carry out the measurements, it will be necessary to allow for the ion composition, which will be mainly metallic (Behnke and Vickrey, 1975; Tepley and Mathews, 1985). Incoherent scatter measurements of the MLT region at Arecibo are conventionally carried out using one of two approaches - the coded long pulse (Sulzer, 1986b) and the pulse-to-pulse code (Zhou, 2000). The former is similar to the alternating code employed at many of the other UAF sites but uses pseudo-random sequences, which may be advantageous in practice. This technique acquires rangegated autocorrelation functions from each transmitted pulse and is most useful at altitudes above about 100 km where the scatter may be overspread, depending on composition. The nominal range resolution for the experiment is 1.2 km, although this can be modified as conditions demand. Plasma density, temperature, composition (see below), and line-ofsight drifts can be obtained from the measured autocorrelation functions using standard nonlinear least squares analysis. Time resolution is determined by signal strength but should be of the order of minutes within strong sporadic E layers. The pulse-to-pulse code, meanwhile, involves transmitting phase coded pulses and interpreting the received signals using pulse-to-pulse analysis, i.e., standard Fourier analysis, followed by parameter fitting. This approach is feasible below about 100 km (or higher, depending on composition) where the scatter is underspread. In practice, there will be some overlap with the coded long pulse. Zhou (2000) used a Barker coded pulse with a 600 m range resolution, but longer codes with finer resolution can also be used. A third option, employed during some of our experiments in 2002, is to transmit a double pulse, with each pulse being a long coded pulse (e.g. 88x1 s bits), and with the lag spacing chosen to optimize the accuracy of line-of-sight velocity measurements throughout the km altitude regime where we expect to find the sporadic layers. This mode has proven to be effective for making rapid estimates of the line-of-sight drift velocity in trial experiments. It represents a simple implementation of a multipulse mode, the forerunner of the coded long pulse. The single lag of the autocorrelation function it measures can also be used in the parameter fitting described below. In collisionally dominated plasmas like the sporadic E layers in question, particularly low altitude ones, difficulty arises in incoherent scatter analysis from ambiguity in the incoherent scatter autocorrelation function (see Tepley and Mathews (1985) and Zhou et al. (1997) for discussions). In this case, the 11

12 shape of the ACF depends on the normalized collision frequency parameter: which in turn depends on the ion-neutral collision frequency, the ion mass, and the ion temperature. The ion-neutral collision frequency, meanwhile, also depends on ion mass as well as neutral density (Banks and Kockarts, 1973a,b). The difficulty lies in distinguishing the effects of ion mass (composition) and temperature in the ACF. The conventional approach is to adopt a boot-strapping method. A rough estimate of the temperature can be obtained directly from the systematic altitude vari parameters, obtained through curve ation of the fitting, which is indicative of the neutral scale height and hence the neutral temperature. This temperature can then be substituted back in the to estimate the mean atomic weight of the ion species throughout the layer, between 31 for purely molecular and 56 for purely iron ions, so long as the layer is in thermodynamic equilibrium. With the assumption that the composition of the layer is homogeneous, tem perature profiles can then be extracted from the parameters in the final analysis. Composition can be found first with rather poor time resolution but good statistical accuracy in the first pass through the data. In the second pass, temperature profiles can be deduced with improved time resolution. This is the strategy we propose to refine and follow. We will interleave the coded long pulse, pulseto-pulse code, double pulse, and a fourth mode optimized for F region observations (MRACF, Sulzer (1986a)) for our experiments. The duty cycle of the combined mode will be determined by trial and error. Useful velocity and temperature profiles will be produced with a range resolution of 1.2 km or better. Data will only be available when the beam intersects strong sporadic layers, but Figure 1 suggests that this condition can be met for long periods of time even as the azimuth swing spans a large fraction of a complete circle. In order to infer horizontal vector velocity profiles from line-of-sight drift measurements, we will assume the vertical winds are negligible and employ the regularization strategy pioneered by Aponte et al. (2003). This strategy selects the vector pro- Figure 5: Sodium resonance lidar measurements from Arecibo. Figure 6: Combined lidar and radar data from Arecibo. file from a family of allowed profiles that is maximally consistent with available line-of-sight data while maximizing desirable profile properties, such as smoothness. This philosophy arises from statistical inverse theory and is also the foundation of radar imaging. The incoherent scatter radar measurements are critical to the observational studies described here. The lidar and optical imager instrumentation that is part of the instrument cluster at the Arecibo Observatory will provide important additional diagnostics about the dynamics and stability of the neutral atmosphere. Examples of the type of information that can be obtained from the combination of the lidar and incoherent scatter radar are shown in Figure 5 and Figure 6, which are reproduced from the article by Larsen et al. (2004b). Figure 5 shows sodium 12

13 resonance lidar measurements from Arecibo during the night of Feb. 6, Of particular interest is the strong overturning that starts around 21 UT and continues until midnight. Several upwellings are evident during that period with vertical extents that reach above 102-km altitude, i.e., into the region where sporadic E layers are expected to form. Larsen et al. (2005) attributed features of this type to inflection point instabilities, which are neutral shear instabilities. A second example is shown in Figure 6. In that figure, the color scale shows the lidar measurements while the gray scale shows the incoherent scatter radar electron density measurements during the same period on Feb. 22, In this particular case the lidar measurements show the lower portion of the upwelling features, but the signal-to-noise ratios for the lidar data are low above 100-km altitude. The incoherent scatter radar measurements, however, show the continuation of the upwelling and overturning at higher altitudes manifested in the electron density contours. The observations in Figure 6 therefore show that the dynamical feature associated with the perturbed densities in the lower E region originated at lower altitudes where neutral dynamics would dominate. The combination of the the radar and lidar instrumentation is therefore especially useful for the project proposed here. The measurements shown here used a sodium resonance lidar. The lidar instrumentation at Arecibo now uses a different resonance line, but we expect that the measurements will be similar in vertical extent and resolution to what is shown here. Large-scale plasma structuring Both the neutral shear instability as discussed by Larsen (2000) and the plasma instability described by Cosgrove and Tsunoda (2002a) and Cosgrove and Tsunoda (2003) appear to be able to account for the breakup and large-scale organization of sporadic E layers. Determining which, if either, of these mechanisms is actually at work over Puerto Rico is a central objective of this proposal, and we are seeking evidence in favor of one or the other. Some evidence suggests that both mechanisms may be operating at different times, in which case our objective will be to learn to discriminate between Figure 7: Radar image of QP structures from June 30, the two kinds of events. The outcome of the investigation of neutral stability outlined above will have an important bearing on our investigation; if QP echo structures only appear when the atmosphere is dynamically unstable, that will constitute strong evidence for the Larsen/Bernhardt picture. If neutral stability does not accompany QP echoes, the Cosgrove and Tsunoda picture will be favored. In addition, there are other tests that can help rank the importance of the two candidate mechanisms. The orientation of the frontal sporadic E layer structures is indicative of the mechanism producing them. If they are created by neutral instabilities, they should be aligned with the neutral wind shear vector at the most unstable altitude. If they are produced by intermediate-scale plasma instabilities, their orientation will depend on the background wind and electric field in a predictable way. Structures oriented parallel to the magnetic equator and magnetic meridian are forbidden by the plasma instability model. The presence and orientation of waveforms in the nm all-sky images available at Arecibo will also be helpful in this part of the investigation. Existing data from St. Croix suggest that the plasma instability theory cannot by itself explain the QP echo phenomenon entirely. In Figure 7, we see frontal structures forming to the west of Puerto Rico that are aligned with the magnetic meridian. Animated sequences of images show that the structures propagate to the northeast, rotating counterclockwise as they propagate. The data are compelling since they seem to defy the plasma insta- 13

14 bility model. However, they are drawn from a small database, and we have no appreciation of how representative they are. The data are also difficult to interpret in the absence of supporting wind and electric field observations from Arecibo or imagery from the all-sky imagers on site. More numerous and comprehensive experiments are called for. Another indication of the mechanism responsible for the sporadic E layer structuring is the state of the F region ionosphere during the event. Tsunoda et al. (2004); Cosgrove et al. (2004) recently showed that their instability mechanism becomes more vigorous when electrically coupled to Perkinslike instabilities in the F region. If a coupled instability is responsible for QP echoes and midlatitude spread F, then the two phenomena should be highly correlated. Incoherent scatter data from Arecibo as well as coherent scatter data from St. Croix can be used to assess the stability and structuring of the F region during intervals when QP echoes and E layer structuring are detected. Other perspectives on the role of the F region ionosphere during QP echo events will also be tested. Haldoupis et al. (2003) see E layer instabilities as the driver for midlatitude spread F, and Hysell et al. (2002a) regard the F layer strictly as an electrical load. Variations in F layer conductivity associated with the kind of medium scale traveling ionospheric disturbances documented by Saito et al. (1998) should modulate QP echoes if the latter perspective is correct, whereas both the Cosgrove and Tsunoda and Haldoupis perspectives imply a close resemblance in the morphology of E and F layer disturbances (wavelengths, propagation speeds and directions, locations, etc.) These are all testable hypothesis which can be used to guide our research. Intermediate-scale plasma waves Direct observations of small-scale field aligned irregularities and indirect observations of kilometric plasma waves in patchy sporadic E layers point to the existence of intermediate-scale primary plasma instabilities that operate apart from whatever mechanism causes the layers to break up in the first place. While the particular instability at work has not been identified, gradient drift instabilities have received the most attention in the literature. Gradient drift instabilities occur when an E region Hall current flows in the direction normal to a background plasma density gradient. Density irregularities along the flow of the current become polarized in accordance with the demands of quasineutrality, with the resulting convection enhancing the initial perturbations and leading to instability. A number of theoretical treatments of intermediate-scale gradient drift waves have appeared in the literature, some including crucial nonlocal and nonlinear effects (Kudeki et al., 1982; Riggin and Kadish, 1989; Ronchi, 1990; Hu and Bhattacharjee, 1999; Hysell and Burcham, 2000b; Hysell and Chau, 2002). On the basis of these studies and of interferometric observations made at Jicamarca, we can expect the following behavior from the waves: Waves should emerge in and be mainly confined to regions where the cross product of the plasma density gradient and the Hall current is parallel to the geomagnetic field. Waves should propagate in the direction opposite the Hall current. The wavelength of the waves must be small compared to the plasma density gradient scale length. The phase velocity of the waves will be smaller than but of the same order of magnitude as the electron convection speed. Gradient drift instabilities offer a viable explanation for both the kilometric and the small-scale irregularities observed in sporadic E layers just as they do in the equatorial electrojet. However, using a fully three-dimensional simulation of patchy sporadic layers in the midlatitude E region, Hysell et al. (2002b) identified another plasma instability with a faster growth rate than gradient drift. This instability derives free energy from the parallel conductivity gradient in the layers rather than the perpendicular conductivity gradient and favors a small but finite parallel wavenumber, whereas gradient drift waves grow most readily for = 0. The instability is like a collisional drift instability and involves electrons streaming along to preserve charge neutrality as the waves propagate across. Because 14