Proceedings with Extended Abstracts (single PDF file) - Radio ...

map in Figure 2b (and Figure 3a) is labeled as “altitude”,the measurement of backscattered power is infact performed as a function of radar range, i.e., theradial distance of scattering targets from the radarantenna. The altitudes marked on the vertical axiswere derived from radar range assuming that radartargets at 50 MHz are perfectly field aligned 3 m(Bragg scale) density waves propagating perpendicularto the geomagnetic field ⃗ B. Since Bragg scaledensity waves responsible for coherent backscatter at50 MHz are known to be field aligned [e.g., Huanget al.[1995]], the labeling shown in Figure 2b is welljustified and facilitates straightforward comparisonsof ISR and RI data shown in Figures 2a and 2b. Forinstance, the mean altitude of the LQP layer in Figure2b, z ≈ 93 km, coincides very closely with thealtitude of density filaments seen in Figure 2a.In further comparisons of Figures 2a and b we alsoneed to take into account the height resolutions ofISR and RI data. The ISR data were collected witha range resolution of 150 m which is also the effectiveheight resolution in Figure 2a since the ISRbeam was pointed in the vertical direction. RI measurementswere conducted using an antenna beamwith θ = 49 o zenith angle and 1.05 km transmitterpulse lengths yielding an effective height resolution of1.05 cos 49 o ≈ 0.69 km for Figure 2a. assuming thatfield aligned target hypothesis is perfectly valid. Noticethat the height resolution in Figure 2b, ∼ 700 m(or more if waves are not perfectly field aligned), iscoarser than few hundred meter vertical extents of thetopside density structures seen in Figure 2a. Thereforewe cannot expect to see the structures of Figure2a repeated in Figure 2b. The question is then, whatis the cause of the LQP structures seen in Figures 2band 3a? This will be addressed in the next section.3 Data interpretationRejecting the possibility that quasi-periodic fluctuationsportrayed in Figures 2a and b are uncorrelated— i.e., independent phenomena taking place at twodifferent locations — we will describe in this section aradar backscattering scenario that explains the LQPsignatures shown in Figure 2b. One reason why wethink the structures are correlated is the fact thatISR filaments disappear at a later time than RI structures— that ordering is consistent (causal!) with∼100 m/s westward drift inferred from interferometrydata. Furthermore, at the given drift speed the ∼65km distance between the observation volumes is only∼11 minutes apart in time, which is shorter than theoverall duration of the LQP event. Most importantly,however, we have seen such a large number of unstructuredlayers during the measurement campaigndescribed in Urbina et al.[2000] that when exceptionsoccur simultaneously in both ISR and RI data withvery closely matching periodicities we can’t help butthink that the exceptions are not coincidental.We envision a horizontally extended tidal ion layer[e.g., Urbina et al.[2000]] centered about 92.5 km altitudewith a rippled topside boundary. The topsideripple is periodic, has a ∼7 km periodicity in southwestdirection, and travels in the same direction with∼70 m/s speed. As the ripples pass through the verticalpointed Arecibo beam quasi-periodic topside filamentsof Figure 2a are measured with ∼100 s period.We also envision that the rippled surface “carries”localized regions of enhanced Bragg scale waves in away that reconciles the common periodicities of Figures2a and b. In other words a specific phase of theperiodic ripple structure is unstable for the Braggscale density waves observed by the RI system.The periodic ripple structure described above willhave ∼100 m/s trace velocity and ∼10 km apparentwavelength measured in southward as well aswestward directions. These numbers agree with uand λ x estimates obtained from interferometry dataconcerning the east-west dynamics of scattering regions.They are also compatible with north-southdynamics of the scattering regions as follows: Asthe topside ripples travel in the southwest directionthrough ∼northward pointed beam of the RI system(see Figure 1) the radar range to the rippleswithin the beam will decrease at a rate proportionalto the southward trace speed. Using a north-southtrace velocity v = −100m/s we find that range willchange at a rate drdt= v sin θ ≈ −100 sin 49o ≈ −75m/s. This range rate applies to the ripples as wellas any other quantity phase locked to the rippledsurface including the localized regions of enhancedBragg scale waves. Since the vertical axis of thepower map in Figure 2b is in effect z ≡ r cos θ, itfollows that dzdt= drdtcos θ ≈ −75 cos 49o ≈ −50m/s, which is approximately the observed slope ofdescending LQP structures shown in Figure 2b (aswell as Figure 3a). Likewise, an apparent horizontalwavelength of λ y ≈ 10 km in north-south directiontranslates into an equivalent “vertical wavelength”of λ z = λ y cos θ sin θ ≈ 5 km which matches quiteclosely the vertical separations of neighboring structuresin Figure 2b. In above calculations θ = 49 o correspondsto the observation zenith angle of E-regionreturns as illustrated in Figure 4.92