David A. Holdsworth

A simple model of atmospheric radar backscatter: Description and application to the full correlation analysis of spaced antenna data

A model has been developed for simulating the effects of backscatter from scatterers advected with a mean background wind. This model has been designed to be as simple yet as realistic as possible, allowing the simulation of both spaced antenna and Doppler radars, and includes features such as aspect sensitivity, gravity wave perturbations, and turbulent motions. The model simulates the characteristics of radar backscatter very well at both MF and VHF. Results of the application of the full correlation analysis to model data generated for the spaced antenna configuration are presented, revealing good agreement with the model input velocity. The effects of the sampling time upon the full correlation analysis are investigated, suggesting an upper limit for the successful application of the technique. The triangle size effect of the full correlation analysis is also investigated, confirming that the major cause is the failure to properly compensate for the effects of noise. The application of techniques for the estimation of turbulent velocities and aspect sensitivity from model-generated data have also proven successful.

Meteor observations with an MF radar

We conducted meteor echo observations using the Buckland Park MF radar (35°S, 138°E) at 00:40–05:45 LT on October 22, 1997. In addition to the usual full correlation analysis (FCA) technique to measure horizontal wind velocities from 60 to 100 km MF radars have a potential to detect meteor echoes and infer winds through their Doppler frequency shifts. Because of the relatively low radio frequency employed MF radars have a great advantage of providing meteor wind well above 100 km altitude, where very few techniques can measure wind velocities. There is a limitation which should be noted as well. The observations are possible only during night time when the electron density of E-region is low enough for the radio wave to penetrate into the upper region. We detected 233 underdense meteor echoes from 80 km to 120 km with a mean height of 104.4 km. Although the transmitting antenna beams were steered toward off-zenith angles of 25°, almost all the echoes were received outside of the main lobe, indicating that conventional MF radar systems with a broad transmitting beam can work well for meteor observations. Bi-hourly wind profiles were obtained from 94 to 114 km altitudes. The profiles revealed a clear wave structure with a downward phase progression with time. FCA winds from 80 to 100 km were also estimated, and a continuous wind structure was obtained from FCA to meteor heights. Note that the present observations happened to be conducted during a major meteor shower activity. However, a majority of the underdense echoes were from non-shower meteors, and observations during non-shower periods will also yield enough information.

Influence of instrumental effects upon the full correlation analysis

The “full correlation analysis” (FCA) of spaced antenna (SA) data is sometimes found to be susceptible to the “triangle size effect” (TSE), whereby the magnitude of the FCA “true” velocity increasingly underestimates the actual wind velocity with decreasing antenna spacing. The TSE has been attributed to a number of instrumental effects, including coarsely digitized data, receiver characteristic differences, noise, and antenna coupling. In this paper the sources of the TSE are discussed, and the influence of various previously undocumented instrumental effects upon the FCA are investigated using simulated data. These additional effects are also shown to produce the TSE. The possibility that these instrumental effects may contribute to the lack of agreement exhibited between winds estimated using medium frequency SA radars employing the FCA and alternative in situ, satellite, and other radar techniques is discussed.

Spaced antenna analysis of atmospheric radar backscatter model data

A simple computer model of atmospheric radar backscatter is used to investigate and compare various spaced antenna (SA) techniques, including the full correlation analysis (FCA) and a number of imaging interferometry techniques. The results illustrate that the FCA true velocity proves to be an excellent estimate of the model input velocity for all magnitudes of model turbulent motions. On the other hand, the interferometric velocities increase as the magnitude of the turbulent motions are increased, providing a poor estimate of the model input velocity for large turbulent motion magnitudes. Maps illustrating the interferometry-estimated scattering positions superimposed upon the actual model scatterer positions reveal the estimated positions to be inaccurate. The estimated positions are shown to approach the zenith as the magnitude of the turbulent motions increases, confirming the volume scatter derived suggestions of Briggs (1995). Furthermore, the estimated positions have a preferred azimuth angle agreeing with volume scatter derived suggestions. The effects of receiver noise upon the interferometric techniques are investigated, revealing an effect analogous to the FCA triangle size effect whereby the estimated velocity decreases with increasing noise level. The effects of the thresholds used to preclude individual Doppler frequencies from the analyses are illustrated. The results are in excellent agreement with the experimental results obtained in a number of different studies.

Collision frequencies in the D-region

Abstract Partial reflection differential absorption and differential phase observations have been used to systematically study collision frequencies (ν) in D-region of the lower ionosphere. The observations made with the large MF radar located at Buckland Park (35° S , 138° E ) near Adelaide in the period September 1996–December 2000 show ν values larger than predicted by previously used models. The new estimates are compared with values calculated using new collision frequency momentum cross-sections for N2(σ) measured in the laboratory. The two types of measurements are found to be in good agreement, with a moderate seasonal variation of ν at a constant height. No change due to solar cycle variations is found.

Differential absorption measurements of mesospheric and lower thermospheric electron densities using the Buckland Park MF radar

Abstract The differential absorption experiment (DAE) was first proposed in the 1950s for the estimation of mesospheric and lower thermospheric electron density using MF/HF radars. The technique was used extensively until the late 1970s, when interest in the technique declined, due to experimental limitations and questions regarding the assumptions of the technique. This paper describes the application of the DAE within the online observations of the Buckland Park MF (BPMF) radar. The experimental limitations of the technique for the BPMF radar are discussed, with particular attention paid to effects of complex gain differences between receiving channels used to decompose the linearly received signals into circular components. Hourly and monthly averaged midday DAE electron densities are presented, revealing good agreement with IRI model estimates. Monthly averaged midnight DAE electron densities are also presented, revealing good qualitative agreement with the IRI model estimates.

An investigation of biases in the full correlation analysis technique

Abstract A study of the biases inherent in the application of the full correlation analysis (FCA) is presented using a simple computer model of atmospheric radar backscatter. The “radar backscatter model” allows the simulation of the effects of anisotropic scatteres, transmitter and receiver polar diagrams, turbulent motions, monochromatic gravity waves, and scatterer lifetimes. The model is used to investigate the velocity biases introduced by a number of instrumental effects. The FCA is applied to data recorded using the recently upgraded Buckland Park MF radar, revealing that the radar is now capable of obtaining usable returns down to 54 km.

Spatial correlation analysis revisited: Theory, and application to “radar backscatter model” data

Three novel correlation techniques for the estimation of atmospheric wind velocities using spaced antenna (SA) radars are introduced. These techniques apply the concepts of spatial correlation analysis, providing an alternative to the conventional full correlation analysis (FCA). The techniques are implemented using simulated data. The first technique is applicable to three or more noncolinear antennas. This technique is shown to produce velocities in excellent agreement with the FCA “true” velocity and is also subject to the “triangle size effect” (TSE), whereby the velocity decreases below the actual wind velocity with decreasing antenna spacing. The remaining two techniques are applicable to four antennas arranged in either a square or a “Y” configuration. These two techniques are shown to be independent of the TSE and may therefore provide a valuable alternative to the FCA in situations where the FCA may underestimate the actual wind velocity.

A comparison of tropospheric VHF Doppler beam steering and full correlation analysis measurements of aspect sensitivity

Tropospheric VHF radar data collected using both spaced antenna (SA) and Doppler beam steering (DBS) methods are presented. The SA data are analyzed using full correlation analysis (FCA) on vertical beam data, while DBS analysis is applied to vertical beam data and off-zenith beam data at 3.6°, 7.2°, and 10.8°. A comparison of meridional velocities using SA FCA and DBS analysis is presented. These results indicate good agreement between the DBS and FCA velocities on all off-zenith beams, particularly at 10.8°. The main aim of the study is to compare the aspect sensitivity parameter θs,using the spatial correlation method for the FCA data [e.g., Lesicar and Hocking, 1992] and the power method of Hocking et al. [1986] for the DBS data. This comparison reveals good general agreement between isotropic values of θs with both methods, and aspect sensitive values of θs values with both methods, although there are times when one method produces isotropic values and the other produces aspect sensitive values and visa versa. During aspect sensitive periods the FCA values agree best with the DBS method using the 0.0°–3.6° beam combination.

Evidence of tilted layers in angle of arrival and Doppler beam steering power measurements

The results of an experiment combining Doppler beam steering (DBS) and angle of arrival (AOA) measurements are presented. Nonzero vertical beam AOAs and off zenith DBS power measurements have been attributed by various authors to the tilting of horizontal layers in the atmosphere. This study aims to directly compare vertical beam AOAs and off zenith beam power data to determine whether the relationship between the two sets of measurements is consistent with tilted layers. The results indicate that the power differences in beams at 3.6° off zenith are related to the vertical beam AOAs and that the variation of the two quantities is consistent with tilted layers being observed in all of these beams. In contrast, the power differences at 10.8° off zenith do not appear to have the same relationship to the vertical beam AOAs.