CUSTAR

Introduction

The Canterbury University Stratosphere Troposphere Radar (CUSTAR) is a powerful atmospheric radar based at the Birdlings Flat Field Site (near Christchurch , New Zealand ). The University of Canterbury has just completed the first phase of the development of this ST radar. This radar operates at 42.5MHz, has an array area of 3000m2 and the peak transmitted power is 100kW. This radar measures returns from clear air echoes and examines the returned signal power and Doppler shift from ranges between 3 and approximately 14km. At this stage in the radar’s development only the returns from a single vertically pointed beam are made. Therefore, only vertical signal power and vertical velocity measurements can be examined at this time. The addition of six spaced antennas in the near future will allow the system to measure the horizontal velocity between 3 and 14km.

CUSTAR is a clear air radar (radars of this class are also often called wind profilers since they are most often used to measure the vertical and horizontal wind speed) which means that it primarily observes clear air echoes which are produced by fluctuations in the atmospheric radio refractive index. It is important to note that these radar are also sensitive to Rayliegh scatter from hydrometeors, but that the signal from this source is significantly smaller than that from radio refractive index fluctuations at the operating frequency of the radar (42.5MHz). The radio refractive index is a function of absolute temperature, atmospheric pressure and the partial pressure of water vapour and while the variations of atmospheric temperature, pressure and humidity only give rise to variations of refractive index with magnitudes of the order of <10-6, these are sufficient to cause detectable radar signals to be observed up to 14 km.

In general these small perturbations in the radio refractive index are produced by turbulence caused by dynamic and convective instabilities. Turbulent mixing across a region gives rise to refractive index gradients across a wide range of scale sizes which produce returned signals. Fresnel reflection also occurs from sharp vertical changes in the refractive index which are horizontally coherent over a large spatial scale (the first Fresnel zone). Theoretical analysis of both these return mechanisms indicates that the returned signal power observed is directly related to the vertical gradient of the radio refractive index. Thus examination of the returned signal allows us to determine –

         information about humidity

         information about temperature

         information about clear air turbulence and other scattering mechanisms

 

Current CUSTAR system

The first phase of the development of the CUSTAR system has now been completed and a single near-vertically directed antenna has been in continuous operation since August 2002. This radar operates at 42.5MHz and has a large antenna formed from an array of dipoles (Figure 1), the area of the array being approximately 3000m2 (7.5l by 7.5 l). The transmitter utilized produces pulses with a peak transmitted power of 100kW and can produce pulses with a maximum duty cycle of 1.4% with a pulse repetition frequency of nearly 2000Hz.. The returned signals are at present processed using a simple Doppler spectral processing scheme. The first three moments of the derived Doppler spectra allow the returned signal power, Doppler shift and spectral width to be derived. The spectra can also be examined to determine the noise level and the signal to noise ratio of the returns. It should be noted that at this stage in the radar’s development because only the returns from a single vertically pointed beam are measured only vertical signal power and vertical velocity measurements can be examined.

 

Figure 1: Antenna array observed from the corner.

The antenna array has been designed to maximize gain and directivity while minimizing sidelobes and the overall cost. To achieve these aims, in the diagonal direction, the dipoles are spaced by 0.707l, and in the N-S and E-W directions the rows are spaced by half a wavelength. Thus, when looking either North-South or East-West the power polar diagram is that of a half-wavelength spaced array (with no ground lobe) and when looking along either diagonal the pattern is that of a 0.707l array. The theoretical antenna power polar diagrams for the array along the North-South axis, the NW-SE axis and the NE-SW axis shows that the ground lobe is more than 20dB below the value at the mainlobe. To ensure that the theoretical polar diagram and the antennas actual polar diagram are consistent reverse radio astronomy has been used. Reverse radio astronomy usually uses a reference sky temperature map to determine radar reflectivity calibrations and allow system performance monitoring. However, the presence of strong radio stars and the galactic centre can also be used to verify the beam diagram, beam width and beam pointing direction of the antenna array.

 

Phase 2 system development

Several developments are to be made in the near future which will allow horizontal wind speed to be determined. A set of five (possibly six) extra spaced array antennas are to be built, the area of each of these antennas being 3l by 3l. The returns from each of these antennas will be received by a new receiver design. This design will reduce the saturation effect observed at low altitudes at present so that measurements can be made to 1km. The inphase and quadrature components from each of the six receivers is then passed to a new state-of-the-art data acquisition system which can sample the twelve channels simultaneously at a rate greater than one million samples per second.  It should be noted that the throughput of this system is expected to be in the region of several GigaBytes per hour and thus data reduction and signal processing schemes will be of great importance. By examining the  returns from each of these antenna using a scheme known as Full Correlation Analysis the horizontal wind speed can be determined. Full Correlation Analysis estimates the atmospheric wind velocity from the ground diffraction pattern resulting from the backscatter of a transmitted signal by atmospheric refractive index irregularities. The analysis assumes that contours of equal spatiotemporal correlation of the ground diffraction pattern can be approximated by a family of ellipsoids. The pattern is then generally sampled at three antennas, and the magnitudes of the temporal autocorrelation  function and cross correlation functions calculated from the complex signals recorded at the antennas are used to parameterize the spatiotemporal correlation function and thereby the wind speed.

Initial Results

The first results from the radar are in the process of being examined. Initial results produced by Trevor Carey Smith can be found here. 

A PDF version of  a poster describing some initials results from the CUSTAR system which was presented at the 2002 Conference of The Meteorological Society of New Zeland can be found here



Back to Home Page