• Angevine, W. M., , White A. B. , , and Avery S. K. , 1994: Boundary layer depth and entrainment zone characterization with a boundary-layer profiler. Bound.-Layer Meteor., 68, 375385.

    • Search Google Scholar
    • Export Citation
  • Balanis, C. A., 1982: Antenna Theory: Analysis and Design. John Wiley & Sons, 790 pp.

  • Balsley, B. B., , and Gage K. S. , 1982: On the use of radars for operational wind profiling. Bull. Amer. Meteor. Soc., 63, 10091018.

  • Barth, M. F., , Chadwick R. B. , , and van de Kamp D. W. , 1994: Data processing algorithms used by NOAA’s wind profiler demonstration network. Ann. Geophys., 12, 518528.

    • Search Google Scholar
    • Export Citation
  • Bianco, L., , and Wilczak J. M. , 2002: Convective boundary layer mixing depth: Improved measurement by Doppler radar wind profiler using fuzzy logic. J. Atmos. Oceanic Technol., 19, 17451758.

    • Search Google Scholar
    • Export Citation
  • Bianco, L., , Wilczak J. M. , , and White A. B. , 2008: Convective boundary layer depth estimation from wind profilers: Statistical comparison between an automated algorithm and expert estimations. J. Atmos. Oceanic Technol., 25, 13971413.

    • Search Google Scholar
    • Export Citation
  • Carter, D. A., , Gage K. S. , , Ecklund W. L. , , Angevine W. M. , , Johnston P. E. , , Riddle A. C. , , Wilson J. , , and Williams C. R. , 1995: Developments in UHF lower tropospheric wind profiling at NOAA’s Aeronomy Laboratory. Radio Sci., 30, 9771001.

    • Search Google Scholar
    • Export Citation
  • Detrick, D. L., , and Rosenberg T. J. , 1990: A phased array radiowave imager for studies of cosmic noise absorption. Radio Sci., 25, 325338.

    • Search Google Scholar
    • Export Citation
  • Ecklund, W. L., , Carter D. A. , , and Balsley B. B. , 1988: A UHF wind profiler for the boundary layer: Brief description and initial results. J. Atmos. Oceanic Technol., 5, 432441.

    • Search Google Scholar
    • Export Citation
  • Ecklund, W. L., , Carter D. A. , , Balsley B. B. , , Currier P. E. , , Green J. L. , , Weber B. L. , , and Gage K. S. , 1990: Field tests of a lower tropospheric wind profiler. Radio Sci., 25, 899906.

    • Search Google Scholar
    • Export Citation
  • Fabry, F., , and Zawadzki I. , 1995: Long-term radar observations of the melting layer of precipitation and their interpretation. J. Atmos. Sci., 52, 838851.

    • Search Google Scholar
    • Export Citation
  • Farley, D. T., 1985: On-line data processing techniques for MST radars. Radio Sci., 20, 11771184.

  • Gage, K. S., , Williams C. R. , , and Ecklund W. L. , 1994: UHF wind profilers: A new tool for diagnosing tropical convective cloud systems. Bull. Amer. Meteor. Soc., 75, 22892294.

    • Search Google Scholar
    • Export Citation
  • Hashiguchi, H., , Fukao S. , , Tsuda T. , , Yamanaka M. D. , , Tobing D. L. , , Sribimawati T. , , Harijono S. W. B. , , and Wiryosumarto H. , 1995: Observations of the planetary boundary layer over equatorial Indonesia with an L-band clear-air Doppler radar: Initial results. Radio Sci., 30, 10431054.

    • Search Google Scholar
    • Export Citation
  • Hashiguchi, H., , Fukao S. , , Moritani Y. , , Wakayama T. , , and Watanabe S. , 2004: A lower troposphere radar: 1.3-GHz active phased-array type wind profiler with RASS. J. Meteor. Soc. Japan, 82, 915931.

    • Search Google Scholar
    • Export Citation
  • Hildebrand, P. H., , and Sekhon R. S. , 1974: Objective determination of the noise level in Doppler spectra. J. Appl. Meteor., 13, 808811.

    • Search Google Scholar
    • Export Citation
  • Imai, K., , Nakagawa T. , , and Hashiguchi H. , 2007: Development of tropospheric wind profiler radar with Luneberg Lens Antenna (WPR LQ-7). Sumitomo Electric Industries Technical Review, No. 64, Sumitomo Electric Industries, Osaka, Japan, 38–42. [Available online at http://global-sei.com/tr/pdf/energy/64-06.pdf.]

  • Krishnan, P., , Kunhikrishnan P. K. , , and Nair S. M. , 2005: Time-height evolution of intraseasonal oscillations in the tropical lower atmosphere: Multilevel wind observations using UHF radar. Geophys. Res. Lett., 32, L07805, doi:10.1029/2004GL022019.

    • Search Google Scholar
    • Export Citation
  • Kumar, K. K., , and Jain A. R. , 2006: L band wind profiler observations of convective boundary layer over Gadanki, India (13.58°N, 79.28°E). Radio Sci., 41, RS2004, doi:10.1029/2005RS003259.

    • Search Google Scholar
    • Export Citation
  • Law, D. C., , McLaughlin S. A. , , Post M. J. , , Weber B. L. , , Welsh D. C. , , Wolfe D. E. , , and Merritt D. A. , 2002: An electronically stabilized phased array system for shipborne atmospheric wind profiling. J. Atmos. Oceanic Technol., 19, 924933.

    • Search Google Scholar
    • Export Citation
  • May, P. T., , and Strauch R. G. , 1998: Reducing the ground clutter on wind profiler velocity measurements. J. Atmos. Oceanic Technol., 15, 579586.

    • Search Google Scholar
    • Export Citation
  • May, P. T., , Cummings F. , , Koutsovasilis J. , , Jones R. , , and Shaw D. , 2002: The Australian Bureau of Meteorology 1280-MHz wind profiler. J. Atmos. Oceanic Technol., 19, 911923.

    • Search Google Scholar
    • Export Citation
  • Mead, J. B., , Hopcraft G. , , Frasier S. J. , , Pollard B. D. , , Cherry C. D. , , Schaubert D. H. , , and McIntosh R. E. , 1998: A volume-imaging radar wind profiler for atmospheric boundary layer turbulence studies. J. Atmos. Oceanic Technol., 15, 849859.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., 1995: Using radar-measured radial vertical velocities in rain and snow to distinguish precipitation scattering from clear-air scattering. J. Atmos. Oceanic Technol., 12, 257267.

    • Search Google Scholar
    • Export Citation
  • Rao, T. N., , Rao D. N. , , Mohan K. , , and Raghavan S. , 2001: Classification of tropical precipitating systems and associated Z-R relationships. J. Geophys. Res., 106 (D16), 17 69917 711.

    • Search Google Scholar
    • Export Citation
  • Rao, T. N., , Kirankumar N. V. P. , , Radhakrishna B. , , Rao D. N. , , and Nakamura K. , 2008: Classification of tropical precipitating systems using wind profiler spectral moments. Part I: Algorithm description and validation. J. Atmos. Oceanic Technol., 25, 884897.

    • Search Google Scholar
    • Export Citation
  • Reddy, K. K., and Coauthors, 2001: Lower atmospheric wind profiler at Gadanki, tropical India: Initial results. Meteor. Z., 10, 457468.

    • Search Google Scholar
    • Export Citation
  • Reddy, K. K., and Coauthors, 2002: Planetary boundary layer and precipitation studies using lower atmospheric wind profiler over tropical India. Radio Sci., 37, 1061.

    • Search Google Scholar
    • Export Citation
  • Rinehart, R. E., 2004: Radar for Meteorologists. Rinehart Publications, 482 pp.

  • Rogers, R. R., , Ecklund W. L. , , Carter D. A. , , Gage K. S. , , and Ethier A. , 1993: Research applications of a boundary layer wind profiler. Bull. Amer. Meteor. Soc., 74, 567580.

    • Search Google Scholar
    • Export Citation
  • Sato, T., 1989: Radar principles. Handbook for MAP, Vol. 30, S. Fukao, Ed., SCOSTEP Secretariat, 19–53.

  • Srinivasulu, P., , Padhy M. R. , , Yasodha P. , , and Rao T. N. , 2006: Development of UHF wind profiling radar for lower atmospheric research applications. Department of Space, National Atmospheric Research Laboratory Preliminary Design Rep., 74 pp. [Available online at http://www.narl.gov.in/PDFs/TechnicalReport.pdf.]

  • Srinivasulu, P., , Yasodha P. , , and Narayana Reddy S. , 2007: Modified Butler beam forming network for UHF wind profiling radar application. Proc. of Int. Radar Symposium India (IRSI), Bangalore, India, IETE Bangalore Centre and Cosponsors, 406–410.

  • Srinivasulu, P., , Yasodha P. , , Reddy S. N. , , Satyanarayna S. , , and Jayaraman A. , 2011: Simplified active array L-band radar for atmospheric wind profiling: Initial results. J. Atmos. Oceanic Technol., 28, 14361447.

    • Search Google Scholar
    • Export Citation
  • White, A. B., , Jordan J. R. , , Martner B. E. , , Ralph F. M. , , and Bartram B. W. , 2000: Extending the dynamic range of an S-band radar for cloud and precipitation studies. J. Atmos. Oceanic Technol., 17, 12261234.

    • Search Google Scholar
    • Export Citation
  • Williams, C. R., , Gage K. S. , , and Ecklund W. L. , 1995: Classification of precipitating clouds in the tropics using 915-MHz wind profilers. J. Atmos. Oceanic Technol., 12, 9961012.

    • Search Google Scholar
    • Export Citation
  • Woodman, R. F., 1985: Spectral moments estimation in MST radars. Radio Sci., 20, 11851195.

  • View in gallery

    Block diagram of the LAWP.

  • View in gallery

    Photographs of (a) microstrip patch (left) ANT element and 16-element linear array panel (top right) without and (bottom right) with FRP radome, (b) extruded building roof for mounting the planar array, (c) partially filled planar antenna array, and (d) planar array (left) covered with secondary FRP radome and (right) surrounded by aluminum clutter suppression fence.

  • View in gallery

    (a) Block diagram of the transceiver module, and (b) photographs of the (left) transmit and (right) receive side.

  • View in gallery

    (a) Schematic of the 16-port linear-modified Butler network, (b) photograph of the 16-port linear-modified Butler network with BS switch, (c) measured (top) phase and (bottom) amplitude distributions at the output ports.

  • View in gallery

    Measured zenith beam radiation pattern of a 16-element linear array fed with linear BFN.

  • View in gallery

    (a) Schematic of the 2D BFN and (b) half-power cross section of all 81 beams in 2D angular space. The five beams used for DBS operation are superscribed (dashed lines).

  • View in gallery

    Photograph showing the X-BFNs and the TMs suspended below the ground plane of the roof-mounted antenna array ground plane.

  • View in gallery

    (a) Original (gray) and detrended (black) time series data for a single range bin at 1.95-km range, (b) corresponding spectrum (top) before and (bottom) after detrending and (c) stacked range Doppler spectra, in (Z, R3) beam, for (left) original and (right) processed data. The mean Doppler computed using adjacent peak picking algorithm is shown (gray line).

  • View in gallery

    Typical range Doppler spectra in five beam directions used for DBS mode.

  • View in gallery

    (a) Comparison of winds observed by the LAWP and GPS sonde at 1700 LT 3 Aug 2010 and (b) scatterplot comparing winds observed by the LAWP and GPS sonde during the month of April 2011.

  • View in gallery

    Diurnal variation of horizontal wind on 11 Jun 2011.

  • View in gallery

    Diurnal variation of radar echo intensity represented as SNR (dB), showing the evolution of ABL.

  • View in gallery

    Typical range Doppler spectra observed in the zenith direction during convection on (left) 15 Apr 2011, and stratiform precipitation on 14 Sep 2010 with a range resolution of (middle) 150 and (right) 37.5 m.

  • View in gallery

    LAWP observations during convection on (a) 15 Apr 2011 and during stratiform precipitation on 14 Sep 2010 with (b) 150- and (c) with 37.5-m range resolution.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 11 11 7
PDF Downloads 11 11 9

1280-MHz Active Array Radar Wind Profiler for Lower Atmosphere: System Description and Data Validation

View More View Less
  • 1 National Atmospheric Research Laboratory, Gadanki, India
  • 2 S. V. U. College of Engineering, S. V. University, Tirupati, India
  • 3 Vikram Sarabhai Space Centre, Trivandrum, India
© Get Permissions
Full access

Abstract

An L-band radar wind profiler was established at National Atmospheric Research Laboratory, Gadanki, India (13.5°N, 79.2°E), to provide continuous high-resolution wind measurements in the lower atmosphere. This system utilizes a fully active array and passive beam-forming network. It operates at 1280 MHz with peak output power of 1.2 kW. The active array comprises a 16 × 16 array of microstrip patch antenna elements fed by dedicated solid-state transceiver modules. A 2D modified Butler beam-forming network is employed to feed the active array. The combination of active array and passive beam-forming network results in enhanced signal-to-noise ratio and simple beam steering. This system also comprises a direct intermediate frequency (IF) digital receiver and pulse compression scheme, which result in more flexibility and enhanced height coverage. The scientific objectives of this profiler are to study the atmospheric boundary layer dynamics and precipitation. Observations made by this profiler have been validated using a collocated GPS sonde. This paper presents the detailed system description, including sample observations for clear-air and precipitation cases.

Corresponding author address: Parvatala Srinivasulu, National Atmospheric Research Laboratory, Gadanki-517 112, Andhra Pradesh, India. E-mail: pslu@narl.gov.in

Abstract

An L-band radar wind profiler was established at National Atmospheric Research Laboratory, Gadanki, India (13.5°N, 79.2°E), to provide continuous high-resolution wind measurements in the lower atmosphere. This system utilizes a fully active array and passive beam-forming network. It operates at 1280 MHz with peak output power of 1.2 kW. The active array comprises a 16 × 16 array of microstrip patch antenna elements fed by dedicated solid-state transceiver modules. A 2D modified Butler beam-forming network is employed to feed the active array. The combination of active array and passive beam-forming network results in enhanced signal-to-noise ratio and simple beam steering. This system also comprises a direct intermediate frequency (IF) digital receiver and pulse compression scheme, which result in more flexibility and enhanced height coverage. The scientific objectives of this profiler are to study the atmospheric boundary layer dynamics and precipitation. Observations made by this profiler have been validated using a collocated GPS sonde. This paper presents the detailed system description, including sample observations for clear-air and precipitation cases.

Corresponding author address: Parvatala Srinivasulu, National Atmospheric Research Laboratory, Gadanki-517 112, Andhra Pradesh, India. E-mail: pslu@narl.gov.in
Keywords: Wind profilers

1. Introduction

In recent years, radar wind profilers (RWPs) operating close to the 1-GHz band have been extensively used for atmospheric research and operational meteorology. It is also referred as the lower-atmospheric wind profiler (LAWP) because it is used to probe the lower part of the atmosphere (up to about 5 km). Many LAWPs have been developed in the recent past by research and commercial groups for meteorological applications, in particular to understand the dynamics of the atmospheric boundary layer (ABL), air quality studies, and precipitation (Balsley and Gage 1982; Rogers et al. 1993; Gage et al. 1994; Ralph 1995; Fabry and Zawadzki 1995; Rao et al. 2008). The pioneering work in making the RWPs practical is accomplished in 1980s at National Oceanic and Atmospheric Administration (Carter et al. 1995). Initial LAWPs (Ecklund et al. 1988, 1990; Hashiguchi et al. 1995) are configured with either dish antenna or passive array for simplicity and commercial viability. Advances in data acquisition and processing systems and radar hardware led to the development of more complex and advanced systems in later years for a better understanding of boundary layer turbulence, wind variability, and synoptic systems (Mead et al. 1998; May et al. 2002; Law et al. 2002; Hashiguchi et al. 2004; Imai et al. 2007).

The National Atmospheric Research Laboratory (NARL) has operated a 1357.5-MHz LAWP (Reddy et al. 2001) under the Indo-Japanese collaboration program during the period 1998–2000 before it became nonoperational. It has yielded valuable data for the research community with respect to the winds, ABL, and precipitation (Rao et al. 2001, 2008; Reddy et al. 2002; Krishnan et al. 2005; Kumar and Jain 2006). The scientific demand for this profiler was so strong that NARL has initiated the development of a new LAWP system (Srinivasulu et al. 2006) to sustain the research activities. While designing the radar, several scientific requirements need to be considered. For example, it is well known that the receiving system, which is configured to receive weak (Bragg) backscatter from the clear-air refractive index irregularities, often gets saturated during precipitation (White et al. 2000) resulting from the strong Rayleigh scattering. To avoid receiver saturation during the precipitation observation we need low transmit (Tx) power and/or low receive (Rx) gain (G), compared to the case of clear-air (wind) observation. These contrasting scientific requirements were taken into consideration while designing the new system. Similarly, high-resolution measurements are required for a better understanding of the melting layer, a signature of stratiform rain (Rao et al. 2008). The radar is designed to be operated in multiple modes to cater to the above scientific requirements. The new system is configured with a simplified active array in which the beam-agility feature of a conventional active phased array is replaced with a constrained passive beam-forming network (BFN) that is adequate enough to generate multiple beams for wind profiling application. The combination of the active array and passive BFN results in an enhanced signal-to-noise ratio (SNR) and simple beam steering. This system also comprises a direct intermediate frequency (IF) digital receiver and pulse compression scheme, which result in more flexibility and enhanced height coverage. This is the first wind profiler in its class employing a fully active array as well as a 2D Butler BFN. The design concepts are validated by successfully operating a smaller (8 × 8 array) prototype transportable wind profiler (Srinivasulu et al. 2011) as a precursor to the development of a new LAWP. The present paper discusses technical details and capabilities of the new LAWP, which is bigger in size and more powerful in terms of transmitted power and antenna aperture compared to the prototype transportable system. The paper is organized as follows: A detailed description of the LAWP system is presented in section 2. Sample observations and wind validation are presented in section 3, and conclusions are given in section 4.

2. System description

LAWP operates at 1280 MHz and employs the Doppler beam-swinging (DBS) technique for measuring the wind vector. The functional block diagram of the system is shown in Fig. 1. It consists of an active array, a 2D Butler BFN, an exciter and receiver unit, and a direct IF digital receiver (DRx), which also acts as radar controller. Active array (antenna elements with integral transceiver modules) and BFN, which are distributed in nature, have been mounted at the roof level of the instrumentation room. The centralized subsystems, like the exciter and receiver DRx, are housed in an instrumentation rack located at the ground level. Exciter generates the local oscillator (LO), IF, and radio frequency transmitter (Tx-RF) signals with reference to a high stable oven-controlled crystal oscillator (OCXO). The Tx-RF pulse is fed to the 2D BFN through the Tx/Rx switch. BFN distributes the signal among the output ports with appropriate phase gradient and feeds the antenna array elements via transceiver modules (TMs). In Rx mode, the signals received by the antenna array are amplified in the front-end section of TMs, passed through the BFN, and fed to the downconverter via the Tx/Rx switch. The 70-MHz Rx IF output from the downconverter is amplified, band limited, and then fed to the direct IF digital receiver, which performs the digital quadrature detection, baseband-matched filtering, coherent averaging, and raw data transfer to the host PC. The host PC performs the data cleaning, fast Fourier transform (FFT), incoherent integration, and computation of spectral moments and wind vector. Specifications of the profiler are shown in Table 1. The following subsections present a detailed description of the design, realization, and functioning of the various subsystems of LAWP.

Fig. 1.
Fig. 1.

Block diagram of the LAWP.

Citation: Journal of Atmospheric and Oceanic Technology 29, 10; 10.1175/JTECH-D-12-00030.1

Table 1.

Specifications of the wind profiling radar.

Table 1.

a. Active array

Active array consists of a planar array in which the antenna (ANT) elements are fed directly by dedicated solid-state TMs. The planar array consists of 256 microstrip patch ANT elements arranged in a 16 × 16 matrix over an area of 2.8 m × 2.8 m. Figure 2 shows different phases of array construction and installation. The patch elements are fabricated using 125-mil RT/Duroid 5870 substrate. The patch element, designed for linear polarization, is rectangular in shape with dimensions of 73.3 mm × 73.1 mm with a ground plane of 92 mm × 92 mm, and is incorporated with mounting holes at the four corners in order to be fitted onto an aluminum ground panel. Designed with a coaxial probe feed, the antenna elements are found to have return loss better than 15 dB over the bandwidth of 15 MHz. The cross-polarization level of the patch elements is measured to be better than 23 dB. Gain of the elements is measured to be 6.5 dB with a half-power beamwidth (BW) of more than 70° in cardinal planes. Sixteen elements, fitted onto an aluminum ground panel along the H plane, form the basic linear array panel. A preliminary radome, fabricated with a fiber reinforced plastic (FRP), is fitted on to the linear array panel for environmental protection. Figure 2a shows the picture of the patch element (left) and the 16-element linear array panel with (right top) and without (right bottom) preliminary radome. The effect of the FRP radome is found to be negligible on the radiation characteristics. An interelement spacing of 0.73λ, where λ is the operating wavelength, is chosen to have an optimal compromise between the required minimum beamwidth and maximum grating-free steer angle. Sixteen such linear panels are laid along the E plane to construct the full planar array. Dummy narrow aluminum panels are used as spacers between the linear antenna panels in order to realize the same interelement spacing along the E plane. The roof of the instrumentation room is extruded and incorporated with inverted reinforced cement concrete (RCC) beams and a square iron frame over which the planar array is installed. Figure 2b shows the picture of the extruded roof, and Fig. 2c shows the picture of the planar array partially filled with linear array panels. The bottom side of the linear array panels and dummy panels, which are contiguously laid, acts as a firm composite ground plane for the entire planar array. Figure 2d shows the picture of the entire planar array covered by a secondary FRP radome (left) and surrounded by the grounded aluminum clutter suppression fence (right). The mutual coupling between the closest antenna elements is measured to be better than 27 dB. After installation, the array axes are found to have azimuthally rotated 11° clockwise with respect to the principal geographical directions.

Fig. 2.
Fig. 2.

Photographs of (a) microstrip patch (left) ANT element and 16-element linear array panel (top right) without and (bottom right) with FRP radome, (b) extruded building roof for mounting the planar array, (c) partially filled planar antenna array, and (d) planar array (left) covered with secondary FRP radome and (right) surrounded by aluminum clutter suppression fence.

Citation: Journal of Atmospheric and Oceanic Technology 29, 10; 10.1175/JTECH-D-12-00030.1

Two-hundred-fifty-six solid-state TMs, each with a peak output power of 10 W, are used to feed the antenna elements directly. These TMs, placed below the antenna ground plane, are designed to have uniform amplitude and phase characteristics. The functional block diagram of a single TM is shown in Fig. 3a. Each TM consists of a Tx/Rx (T/R) switch at the input, Tx section, Rx front-end (FE) section, and a circulator (CIR) and bidirectional coupler (BDC) at the output. The Tx section consists of a phase trimmer, a driver amplifier (DA), and a final gain-controlled power amplifier (PA) module. The Rx FE section comprises a passive diode limiter (LIM), a blanking (BLK) switch, a low-noise amplifier (LNA) with a gain adjustment pad, and phase trimmer. Commercially available off-the-shelf components are used to build the TMs to make them low cost. Mitsubishi RA18H1213G metal oxide semiconductor field-effect transistor (MOSFET) continuous wave (CW) 15-W amplifier is modified for pulsed application and used as PA. The gate voltage is switched in synchronization with the Tx pulse to avoid the continuous drain current flow there by avoiding the heat dissipation. The Advanced Semiconductor Business Solutions ALN 1280, which has a noise figure (NF) of 0.7 dB, is employed as LNA. This device, which has a gain of 28 dB and output compression point of 16 dBm is used for DA also. Varactor diodes are used in the phase trimmers of both the Tx and Rx sections. RCPL E3NG and Anaren 11305-20 are used as CIR and BDC, respectively. CIR connects the output signal from the Tx section to the antenna and the antenna to the Rx section. The Tx–Rx isolation achieved is at about 90 dB (25 and 65 dB resulting from circulator and blanking switch, respectively). BDC is used to monitor the Tx output during Tx time and to inject a test signal into the Rx section. The loss resulting from the circulator and BDC is about 1 dB. Though TMs, produced in the mass production, are expected to have uniform phase and amplitude characteristics, differential errors do exist due to the assembly process. A provision is made to adjust the phase and gain to the tolerable limits in both Tx and Rx paths to compensate for these errors. Phase trimmers are kept to achieve the uniformity in phase among all the TMs in both Tx and Rx paths. Gain can be adjusted in the Tx section by varying the gate bias to the power amplifier, and in the Rx section by fixing attenuator pads with the appropriate value at the output of LNA. Both Tx and Rx sections are realized on FR4 substrates and mounted on the top and bottom sides, respectively, of a milled aluminum structure. Figure 3b shows the picture of Tx (left) and Rx (right) sides of the TM. A control and safety circuit, located on the Rx side, ensures safe operation of TM. The entire TM, realized in a compact aluminum enclosure weighing about 400 g, with dimensions of 98 mm × 102 mm × 33 mm, operates with a single 12.5-V direct current (dc) power supply. It has three RF interface ports, namely, the Tx-in/Rx-out port, ANT port, and test port. The control port receives the timing signals to control the on/off PA gate, BLK, and T/R switches.

Fig. 3.
Fig. 3.

(a) Block diagram of the transceiver module, and (b) photographs of the (left) transmit and (right) receive side.

Citation: Journal of Atmospheric and Oceanic Technology 29, 10; 10.1175/JTECH-D-12-00030.1

A Tx pulse at −10 dBm level is fed to the TM to generate 10 W at the ANT port. The pulse width (PW) and the interpulse period (IPP) are in the range of 0.25–4.0 and 20–200 μs, respectively, with duty ratio (DR) not exceeding 10%. The rise and fall times of the Tx-RF pulse are measured to be in the range of 50–70 ns. In the Rx mode, the signal received from the ANT element is amplified by the FE section with a net gain of 25 dB. Coupled ports of the BDC are monitored for forward and reflected signal power levels. A fraction of the forward path voltage is brought to outside for testing Tx and Rx paths. The control and safety interlock circuit, in addition to controlling different switches, continuously monitors the temperature, output power, and voltage standing wave ratio (VSWR) and generates interlock if any parameter goes beyond the preset limits, and cuts off the gate bias to the PA, thus ensuring safety. The 24-cm-long coaxial cable, used between the TM and antenna element, which introduces a loss of about 0.2 dB, is considered as an integral part of the TM in Tx and Rx path characterization. TMs are set for the same Tx power level and insertion phase with maximum deviations of ±0.5 dB and ±5°, respectively. Similarly in the Rx path, the gain and insertion phase of the TMs are adjusted to be uniform with maximum deviations of ±0.5 dB and ±2°, respectively. NF values of the Rx path with reference to the ANT port are measured to be in the range 3.5–4 dB though the design goal was 3 dB. TMs are suspended below the antenna panels by fixing them to the array ground plane adjacent to the corresponding ANT feed points. This configuration, where the TMs are located very close to the ANT elements, will result in the enhanced Tx power availability (at the antenna plane) and keep the overall system noise figure at minimum value. The net result is the significant enhancement in the SNR at the receiver output when compared to a passive array system.

b. Beam-forming network

The BFN is built with passive networks instead of using active phase shifters to reduce the cost and complexity of the system. In addition to performing the beam formation, this network also acts as a distribution network for the active array. It is realized with a 2D 256-port modified Butler matrix and beam selection switch to feed the elements of the 16 × 16 active array. In this scheme, the traditional Butler matrix is modified to generate a zenith (broadside) beam, which is essential for a wind profiler for measuring the vertical wind component and studying precipitation. The 2D Butler network is built by connecting two sets of linear (1D) networks arranged in orthogonal directions.

A standard 16-port Butler network generates 16 beams, of which 8 each are located symmetrically on both the left and right sides of normal. Because this network does not generate a zenith (broadside) beam, which is important for RWPs, it needs to be modified to realize the zenith beam. Figure 4a shows the schematic of 1D 16-port modified Butler network, which is configured as logical extension to the 8-port network presented by Detrick and Rosenberg (1990) for a riometer application. This network generates the zenith beam at the cost of one of the outermost beams. It consists of four-port 90° hybrid junctions (squares superscribing the cross) and fixed transmission line sections (small circles) as phase shifters. A signal fed at any input port i (i = from −7 to +8) will be distributed equally among the 16 output ports, but with a phase gradient equal to /8 rad. For the chosen interelement spacing of 0.73λ, this network generates nine grating-free beams; one zenith (Z) beam, four beams (L1, L2, L3, and L4) toward the left, and four beams toward the right (R1, R2, R3, and R4) of the zenith direction. The remaining seven input ports (R5, R6, R7, L5, L6, L7, and L8), which generate grating lobes, are terminated with characteristic impedance. The oblique beams are positioned at ±i5° from zenith. A nonreflective-type beam selection (BS) switch is kept before the input ports to select any one of the nine beams. The 90° hybrid junctions are fabricated over FR4 substrate and sections of transmission lines are used as crossover interfaces between the hybrid junction layers. Fixed phase shifters, indicated by small circles in the schematic, are realized by adjusting the lengths of the respective crossover transmission lines. The low-power BS switch is custom built using PIN diodes to achieve good isolation between the ports and also to safely handle power levels up to 30 dBm. Fixed attenuator pads are kept at the output ports for realizing Taylor amplitude distribution (Balanis 1982) to suppress the sidelobe level (SLL) down to the desired level of 17 dB with respect to the main lobe. Figure 4b shows picture of the linear BFN, comprising 16-port 1D-modified Butler matrix and BS switch. Figure 4c shows the measured phase (top) and amplitude (bottom) distributions at 16 output ports of the linear network for all the nine beams. The rms differential errors in phase and amplitude distributions are found to be about 3° and 0.3 dB, respectively. Figure 5 shows the measured radiation pattern for the Z beam of a 16-element linear array (covered with an FRP radome) fed with a linear BFN. The array, mounted on an azimuth-over-elevation positioner, receives RF signal transmitted by a horn antenna located in the bore-sight direction at a distance of 120 m. The array is rotated in horizontal plane for pattern measurement. A Flam & Russel digital pattern recorder is used to record the radiation pattern. The G and BW are found to be close to the theoretical values, which are 17 dB and 5°, respectively. The SLL, however, is found to be 16 dB against the theoretical value of 17 dB, probably resulting from the residual differential errors in phase and amplitude distributions.

Fig. 4.
Fig. 4.

(a) Schematic of the 16-port linear-modified Butler network, (b) photograph of the 16-port linear-modified Butler network with BS switch, (c) measured (top) phase and (bottom) amplitude distributions at the output ports.

Citation: Journal of Atmospheric and Oceanic Technology 29, 10; 10.1175/JTECH-D-12-00030.1

Fig. 5.
Fig. 5.

Measured zenith beam radiation pattern of a 16-element linear array fed with linear BFN.

Citation: Journal of Atmospheric and Oceanic Technology 29, 10; 10.1175/JTECH-D-12-00030.1

The schematic, shown in Fig. 6a, is adopted to construct the 2D BFN, which employs the 1D BFNs as basic building blocks (Srinivasulu et al. 2007). The 16 linear networks (X-BFNs) are kept along the y axis with their output ports arrayed parallel to the x axis. These X-BFNs are controlled by a common BS control signal to select beam in x–z plane. The input ports of these 16 X-BFNs, now, arrayed along the y axis, are fed by another identical linear network (Y-BFN), which is kept in orthogonal (y) direction. The Y-BFN is controlled independently to select the beam in y–z plane. Phase distribution, at the output ports of the resultant 2D network, is given by the following equation:
e1
where m (1, 2, …, 16) and n (1, 2, …, 16) are the output ports along the orthogonal directions, and i (−4, −3, −2, −1, 0, 1, 2, 3, and 4) and j (−4, −3, −2, −1, 0, 1, 2, 3, and 4) are the input ports of X-BFNs and Y-BFN, respectively, selected by the BS switches. The above phase distributions generate beam positions in 81 fixed directions ij, ϕij) given by
e2
where d is the interelement spacing along x and y directions, θij is the zenith angle, and ϕij is the azimuth angle. Selection of input port/beam (i, j) is accomplished by two independently controlled BS switches located in X-BFNs and Y-BFN, respectively. Figure 6b shows the half-power cross sections of all 81 beams in the 2D angular space. It is evident from the figure that these beams fill almost the entire probing volume within 20° × 20°. It may also be noted that only one beam at a time can be generated whose direction depends on the positions of the BS switches. The combination of Y-BFN and 16 X-BFNs functions as a 2D BFN, with 256 output ports and a common input port, and facilitates beam formation in the 2D angular space. The network is passive except for the fact that the beam switching is performed by controlling low-power BS switches. The Y-BFN is connected to the radar exciter and receiver through a Tx/Rx switch, whereas the 16 X-BFNs feed the active array elements. X-BFNs are suspended close to the antenna array by fixing them to the array ground plane at the center of the linear array panels to avoid the long cables for feeding TMs, whereas the Y-BFN is kept in the instrumentation rack with 16 interfacing cables running in between.
Fig. 6.
Fig. 6.

(a) Schematic of the 2D BFN and (b) half-power cross section of all 81 beams in 2D angular space. The five beams used for DBS operation are superscribed (dashed lines).

Citation: Journal of Atmospheric and Oceanic Technology 29, 10; 10.1175/JTECH-D-12-00030.1

Figure 7 shows the picture of the X-BFNs and TMs suspended below the antenna ground plane. The one-way insertion loss and bandwidth of the entire 2D BFN are measured to be 7 dB and 15 MHz, respectively. This loss, though it appears to be on the higher side, does not affect the NF of the overall system because the BFN is preceding the PA in the Tx path and follows the LNA in the Rx path. The realized bandwidth is adequate for transmitting/receiving the shortest pulse. The rms errors of the measured phase and amplitude distributions of the entire network for all beams are measured to be within 6° and 0.5 dB, respectively. While integrating the TMs and 2D BFN with the antenna array, the placement of TMs is made such that the differential (phase and amplitude) error distributions of the resultant (BFN + TMs) system are brought down to a minimum value. For example, a TM with differential errors +5° and −0.3 dB is connected to a BFN port with averaged differential error close to −5° and then +0.3 dB, so that the net errors are brought down close to zero or a minimum. This exercise is carried out in both Tx and Rx paths with more priority given to the central portion of the array, where the amplitude weighting is significant. Though the errors are minimized, some errors do exist toward the end, but within the tolerable limits. A simulation study confirmed that these errors do not degrade the radiation pattern except for a slight increase in the SLL. After integration, the total Tx peak power is reduced to about 1.2 kW resulting from amplitude tapering. The designed 2D amplitude distribution provides a conical beam, with a G of 27.5 dB, BW of 5°, and SLL of 17 dB, which can be switched into any one of the 81 (fixed) directions. There are several advantages of having multiple beams, not only in principal directions but also in different azimuths (Fig. 6b). It allows us to choose the appropriate beam (normal to acoustic wave front) for Radio Acoustic Sounding System (RASS) operation during windy conditions, when the acoustic wave front moves away from the radar. Further, merging of atmospheric signal and ground clutter during weak wind conditions can be avoided by choosing beams with large tilt angles.

Fig. 7.
Fig. 7.

Photograph showing the X-BFNs and the TMs suspended below the ground plane of the roof-mounted antenna array ground plane.

Citation: Journal of Atmospheric and Oceanic Technology 29, 10; 10.1175/JTECH-D-12-00030.1

c. Exciter and receiver unit

The exciter and receiver unit consists of a high stable frequency reference OCXO source, phase-locked oscillator (PLO), upconverter, power amplifier, T/R switch, downconverter, and IF chain. The 10-MHz output from the OCXO is given to PLOs as a reference to generate the 1210-MHz LO and 70-MHz IF signals. The IF signal is pulse code modulated (Farley 1985) and mixed with LO to generate the 1280-MHz Tx-RF pulse, which can be amplified up to 25 dBm in steps of 2 dB using a digital attenuator (Tx-ATT) and power amplifier and fed to the BFN and TMs via the T/R switch. In the Rx mode, the echo received from the ANT array and BFN is mixed with the same LO in the downconverter, and the resultant IF signal is fed to the IF chain consisting of a 30-dB amplifier, BLK switch, digital attenuator (Rx-ATT), high-gain amplifier, and filter bank. The gain of the receiver unit can be varied in the range of 20–90 dB with 5-dB resolution by controlling the Rx-ATT. The BLK switch provides extra isolation for the Tx leakage to avoid receiver saturation. The filter bank consists of three parallel bandpass filters (BPFs) corresponding to 0.25-, 0.5-, and 1-μs pulses, which are preceded and followed by SP3Tswitches for bandwidth selection. The maximum signal level of the IF chain output, which is fed to the DRx, is +10 dBm. The exciter and receiver unit, realized in a single enclosure, receives the timing and control signals like Tx pulse, code, Tx-ATT, BLK pulse, bandwidth select, Rx-ATT control, etc., from the radar controller. The 10-MHz reference is also fed to the DRx/radar controller as a clock to generate timing and control pulses for different subsystems.

d. Digital receiver/radar controller

The direct IF digital receiver (DRx), implemented with a commercially available digital board, is fed with the analog Rx IF output signal. DRx performs analog-to-digital conversion (ADC), digital down conversion (DDC) to baseband, decoding/pulse compression (Farley 1985), coherent integration, and data transfer to the PC. It is built around the Analog Devices AD6654 “IF-to-baseband receiver,” ADSP-TS201S Tiger SHARC DSP processor, and Xilinx VIRTEX II (1.5 V) XC2V500 FPGA. The AD6654 is a mixed-signal IF-to-baseband receiver consisting of a 14-bit, 92.16 MSPS ADC, and a four-/six-channel, multimode DDC. The DDC stage has six channels, each consisting of a frequency translator, a fifth-order cascaded integrated comb filter, two sets of cascaded fixed coefficient finite impulse response (FIR) and half-band filters, three cascaded programmable sums of product FIR filters, and an interpolating half-band filter (IHB). The functions of downconversion, filtering, and sample-rate reduction are performed by DDC to reduce the load of software processing considerably. The ADSP-TS201S Tiger SHARC processor is an ultrahigh performance digital signal processor (DSP) optimized for large signal processing tasks in real-time computing. The DSP combines very wide memory widths with floating-point (32 bit and extended precision 40 bit) and fixed-point computation processing capability. The SHARC processor performs pulse compression, coherent averaging, and FFT (optional) on the baseband data. The timing and control signal generator (TCSG), realized with FPGA, provides timing pulses and control signals required by different subsystems. This board is interfaced to a PC through a USB port for control, data archive, and display purposes. Operational parameters such as pulse width (PW), complementary code words, IPP, BLK, gate, T/R pulses, BS control, Tx-ATT, Rx-ATT, bandwidth select control, number of coherent integrations (NCI), number of FFT points (NFFT), incoherent integrations (NICI), observation range window, etc., are set through VC++ graphic user interface (GUI) in the radar control (RC) PC. The timing and control signals are given directly for the centralized subsystems like exciter and receiver and Y-BFN, whereas they are split and buffered to feed the distributed subsystems like TMs and X-BFNs. Multiple experiments (up to four) can be run in a cyclic manner to facilitate observations in different modes like clear air, precipitation, etc.

The PW of the system can be varied from 0.25 to 4.0 μs in binary steps. Both uncoded and coded transmission is possible. For the case of coded transmission the baud length can be selected from 0.25 to 2.0 μs in binary steps and the code length can be selected from 2 to 16 in binary steps. A complementary pair of binary (A and B) codes are transmitted in A, , B and sequence. The 70-MHz Rx IF signal is sampled at 80 MSPS rate by the ADC and the baseband signal is extracted from DDC by setting its NCO to 10 MHz. DDC, incorporated with series of filters, performs pulse width–based (0.25/0.5/1 μs) Bessel filtering and decimation to reduce the data rate to 16 MSPS at the output. The digitized baseband data are stored in first-in/first-out (FIFO) format implemented in the FPGA, where the data rate is further decimated down to 4/2/1 MSPS (for PWs of 0.25/0.5/1 μs, respectively) and read by the DSP at the end of every sampling window. DSP acquires the data, for each range bin, for the consecutive pulses and decodes the data (pulse compression) by the following steps: (i) subtracting the second pulse data from first pulse (code A data − code data), (ii) computing the autocorrelation function (ACF-A) of this data with code A, (iii) subtracting the fourth pulse data from the third pulse (code B data − code data), (iv) computing ACF-B of this data with code B, and (v) adding the two ACF data (ACF-A + ACF-B). The resultant ACF data give the compressed (decoded) pulse for each range gate with signal power enhanced by a factor equal to the code length. This will be repeated till NCI pulses are received and the decoded data for each range bin are added coherently. This process will continue till specified points (NFFT) are acquired. The digital receiver is validated with a range-gated and Doppler-shifted simulated pulse and found to have a dynamic range of 75 dB. The time domain data are transferred to the host PC, where physical parameters are extracted.

e. Data processing

The host PC performs data cleaning and processes the time domain data to estimate the physical parameters like the mean Doppler, Doppler width, and 3D wind vector. Time domain data are first subjected to DC removal to eliminate any fixed DC offset in the complex time domain signals. The in-phase and quadrature phase signals are averaged separately and the respective mean values are subtracted to eliminate the DC offset. However, the slowly varying (low Doppler) clutter resulting from undesired echoes like reflections from trees, power lines swaying in the wind, etc., will be still present. A technique called detrending (May and Strauch 1998) is used in the time domain to remove the slowly varying (low Doppler) clutter signals. In this technique, a long series (1024 points) of time domain data is divided into small segments of length (either 256 or 128 points). A linear trend is approximated to each segment as a straight line and is removed from the original data. This removes the clutter effectively. Figures 8a and 8b illustrate the detrending process for a single range bin and its efficacy in cleaning the time domain data. It may be noted from Fig. 8a that the original in-phase (top) and quadrature (bottom) components, shown in a gray shade, are associated with a strong DC shift as well as strong but slowly varying clutter signals. It can be clearly seen that the DC and clutter, which are present in the original signal, are completely removed (black lines) after the detrending process. The corresponding Doppler spectrum is shown in Fig. 8b before (top) and after (bottom) detrending. The strong peak with narrow width, which corresponds to DC and clutter, is removed in the detrending process. After detrending, the time series data are subjected to either rectangular, Hamming, or Hanning windowing. In the routine operation, a Hanning window is selected for its optimal performance. The complex time domain data, after windowing, is converted into the Doppler spectrum by applying complex FFT and computing the power spectrum. Incoherent integration is then performed, if necessary, where several successive spectra are averaged to improve detectability. Clutter, alternately, can be removed in the frequency domain also by taking out a significant number of points on both sides of the zero Doppler (Barth et al. 1994), and these points are replaced by the value obtained by averaging the two points bracketing the area being removed and by interpolating. The number of points to be replaced is dynamic for each range bin. Any of the two clutter removal techniques (either in the time domain by detrending or in the frequency domain by interpolating) can be selected. The Gadanki site is surrounded by mountains, which contaminate the data by strong DC and clutter. Figure 8c shows the stacked range Doppler spectra for the uncoded mode before (left) and after (right) the DC and clutter removal. It can be clearly seen that the clutter signal is masking the atmospheric signal. In the processed spectra the atmospheric signal is clearly visible and a continuous trace can be seen throughout the height range.

Fig. 8.
Fig. 8.

(a) Original (gray) and detrended (black) time series data for a single range bin at 1.95-km range, (b) corresponding spectrum (top) before and (bottom) after detrending and (c) stacked range Doppler spectra, in (Z, R3) beam, for (left) original and (right) processed data. The mean Doppler computed using adjacent peak picking algorithm is shown (gray line).

Citation: Journal of Atmospheric and Oceanic Technology 29, 10; 10.1175/JTECH-D-12-00030.1

The technique given by Hildebrand and Sekhon (1974) is used to compute the mean spectral noise for each range bin and is subtracted from the respective Doppler power spectra. The zeroth, first, and second moments, which represent the received signal power, mean Doppler frequency, and Doppler variance, respectively, and also SNR, are computed using the formulation given by Woodman (1985) and Barth et al. (1994). The adjacent peak pick algorithm is implemented, which improves the moment computation even when multiple signal peaks are present in the spectral data. This is illustrated in processed spectra of Fig. 8c (right). The estimated mean Doppler frequency is shown with a gray line. It can be noted that the algorithm is working well also in the height region of 4–5 km, where the SNR is relatively weak. Radial velocities obtained from either three or five beams are used to compute U (zonal), V (meridional) and W (vertical) components of the wind vector (Sato 1989). Because the antenna array of the radar is not aligned along the true north direction, the azimuthal bias angle (11°) is used to correct U, V, and W values to obtain the actual component values. The consensus averaging algorithm (Barth et al. 1994) is used to average the wind (hourly), which examines the data to identify the largest subset of velocities that are within a given window. VC++-based GUI facilitates the data displays in different formats as per the selection.

3. Observations

The profiler was ready for trial runs by August 2010 and it was operated since then for both clear-air wind profiling and precipitation observations. For clear-air observations, the radar is operated with five beams: one zenith and four off-zenith beams pointed 15° down toward azimuth directions of 11°, 101°, 191°, and 281°. The zenith beam is named as (Z, Z) and the oblique beams are named as (L3, Z), (Z, L3), (R3, Z), and (Z, R3). LAWP is operated in low (300 m – 4 km) and high (750 m – 6 km) modes alternately for continuous wind profiles from 300 m to about 6 km. In the low mode the radar is operated with an uncoded 1-μs pulse, whereas in the high mode it is operated with coded 4-μs pulse with a baud length of 1 μs. The typical IPP used for wind profiling is 60–80 μs. For precipitation observations, only the (Z, Z) beam is used and the Rx gain is set to a lower value, to avoid the saturation of the receiver. The IPP is set to 100 μs, to cover a height range of 15 km. Two modes, the low- and high-resolution mode, are used for precipitation. The low-resolution mode uses a PW of 1 μs (150-m resolution) whereas the high-resolution mode uses a PW of 0.25 μs (37.5-m resolution). The low-resolution mode is to study the convection/precipitation up to about 13 km. The high-resolution mode is used to study the radar bright band in detail. Table 2 shows the experimental parameters used for different modes of observations shown in this paper.

Table 2.

Experimental parameters for different observations.

Table 2.

Figure 9 shows the typical clear-air-stacked range Doppler spectra measured by LAWP for the coded mode in five beam directions observed on 3 August 2010. Here, positive Doppler frequency corresponds to downward motion. It may be noticed that the echoes can be traced up to a height of about 9 km in all directions. To validate the horizontal winds derived from LAWP, they are compared with those obtained by a standard instrument, in our case a GPS sonde (Meisei RS01GII). A typical comparison of LAWP winds with those obtained from a GPS sonde, launched from NARL at 1700 LT, is shown in Fig. 10a. Because the GPS sonde takes ~30 min to reach a height of 9 km, 30 min of the profiler data starting from the time of balloon launch were averaged for comparison. A very close agreement may be noticed between LAWP and GPS sonde measurements, in spite of the fact that the sonde drifted horizontally up to about 14 km from the LAWP site. Winds derived by the LAWP and GPS sonde during the month of April 2011 are compared for validation. Figure 10b shows scatterplots for the U and V wind components measured by LAWP and GPS sonde. The large cluster of points along the diagonal indicates close agreement between LAWP and GPS sonde-derived winds. The correlation coefficient is found to be 0.93 and 0.94, respectively, for U and V components, validating the LAWP wind measurements. The capability of new LAWP in monitoring the lower atmospheric winds continuously is depicted in Fig. 11. Here, winds are averaged over 38 min. It shows westerly-to-southwesterly winds, which are typical during the southwest monsoon season (June–September), and large diurnal variation of the horizontal wind vector. Figure 11 also shows that the profiler is capable of measuring wind vector consistently up to ~4–6 km.

Fig. 9.
Fig. 9.

Typical range Doppler spectra in five beam directions used for DBS mode.

Citation: Journal of Atmospheric and Oceanic Technology 29, 10; 10.1175/JTECH-D-12-00030.1

Fig. 10.
Fig. 10.

(a) Comparison of winds observed by the LAWP and GPS sonde at 1700 LT 3 Aug 2010 and (b) scatterplot comparing winds observed by the LAWP and GPS sonde during the month of April 2011.

Citation: Journal of Atmospheric and Oceanic Technology 29, 10; 10.1175/JTECH-D-12-00030.1

Fig. 11.
Fig. 11.

Diurnal variation of horizontal wind on 11 Jun 2011.

Citation: Journal of Atmospheric and Oceanic Technology 29, 10; 10.1175/JTECH-D-12-00030.1

LAWP is a potential tool for measuring the height of ABL (Angevine et al. 1994; Bianco and Wilczak 2002; Bianco et al. 2008). The intensity of the backscattered echo depends on turbulence intensity and gradients in the refractive index. Near the boundary layer, the relative humidity and therefore refractive index shows large gradients. This results in large SNR near the boundary layer top, and this property can be utilized to identify the ABL height, as shown in Fig. 12. This observation corresponds to 16 April 2011 and shows the typical echo intensity (SNR) indicating the evolution of ABL, which starts rising from 1100 local time (LT) and reaches maximum height at ~1400 LT and then collapses after 1600 LT.

Fig. 12.
Fig. 12.

Diurnal variation of radar echo intensity represented as SNR (dB), showing the evolution of ABL.

Citation: Journal of Atmospheric and Oceanic Technology 29, 10; 10.1175/JTECH-D-12-00030.1

LAWP is very sensitive to hydrometeors and therefore receives strong backscatter from precipitation (rain and snowflakes) resulting from Rayleigh scattering. Figure 13 shows the typical stacked range Doppler spectra in zenith direction for two different precipitation events. It is clear from the figure that the height coverage of the profiler is significantly high (~9–11 km) during the precipitation. Figure 13a corresponds to a convective event, observed on 15 April 2011, where the Doppler velocity of hydrometeors (resulting from the fall velocity of hydrometeors and vertical air motion) shows large positive and negative Doppler frequencies, indicating the presence of strong updrafts and downdrafts. In particular, the negative Doppler shift in the height region of 3.6–5.6 km indicates the presence of strong updrafts. The updrafts are so strong that they overcome the fall velocity of hydrometeors and carry the hydrometeors to higher altitudes. Figures 13b,c show the Doppler spectra during stratiform precipitation observed on 14 September 2010 with low and high resolutions (i.e., 150 and 37.5 m), respectively. In both high- and low-resolution modes, only the zenith beam was used for this experiment to improve the temporal resolution. Both modes were run sequentially, with each mode taking nearly 20-s beam dwell time. A sudden increase in Doppler may be noticed in the height region of 4.8–3.9 km, which is a signature of stratiform rain (Williams et al. 1995; Rao et al. 2008). This increase is mainly due to the aggregation of snowflakes and melting of the snow to ice (Rinehart 2004). The melting layer (the radar bright band) is seen much more clearly in time–height sections of profiler moments discussed below.

Fig. 13.
Fig. 13.

Typical range Doppler spectra observed in the zenith direction during convection on (left) 15 Apr 2011, and stratiform precipitation on 14 Sep 2010 with a range resolution of (middle) 150 and (right) 37.5 m.

Citation: Journal of Atmospheric and Oceanic Technology 29, 10; 10.1175/JTECH-D-12-00030.1

Figure 14 shows the time–height profiles for echo intensity (SNR) and the mean Doppler velocity of hydrometeors for the precipitation cases shown in Fig. 13. Figure 14a, resembles a classical convective cloud with high reflectivity, large up-/downdrafts (during 1430–1520 LT), and an anvil structure (during 1520–1930 LT above the height of 4.5 km). The radar brightband and strong reflectivity region at about 4.3 km (below the 0°C isotherm level) may be noticed in Figs. 14b,c, where ice is melting into raindrops. Although the bright band appears clearly in the low-resolution mode, estimation of its thickness will have some uncertainty because of low resolution. On the other hand, high-resolution brightband profiling not only reduces this error, but also is useful to understand the microphysics of the processes occurring near the melting layer (Fabry and Zawadzki 1995).

Fig. 14.
Fig. 14.

LAWP observations during convection on (a) 15 Apr 2011 and during stratiform precipitation on 14 Sep 2010 with (b) 150- and (c) with 37.5-m range resolution.

Citation: Journal of Atmospheric and Oceanic Technology 29, 10; 10.1175/JTECH-D-12-00030.1

4. Discussion and conclusions

A state-of-the-art RWP was developed for ABL monitoring, lower atmospheric wind profiling, and understanding precipitation processes. This is the first system in its class to be configured with fully solid-state active array and 2D-modified Butler BFN. Active array configuration, in which antenna elements of the array are fed directly by dedicated TMs, maximizes the SNR significantly (at L band) by eliminating the feeder loss in both Tx and Rx paths. Improvement in SNR results in better height coverage and offers better range resolution. The size of the array will be significantly smaller when compared to a passive array profiler with equal SNR performance. Another major advantage of the true active array is the graceful degradation, where the array performance is not affected significantly even when a few TMs fail to function. Solid-state TMs are realized with commercially available off-the-shelf communication components to make them low cost. The 2D-modified full Butler BFN provides multiple beams, which fill almost the entire radar probing volume. It offers simple beam-steering mechanism and avoids any need for calibration because it is passive in nature. Pulse-to-pulse beam switching is possible since the beam is switched by means of low-power solid-state BS switches. BFN, realized with 90° hybrid junctions fabricated over FR4 substrate and coaxial cable sections, is a low-cost one and operates at low power. Tx-ATT in the radar exciter can be used to reduce the Tx power level up to 20 dB in steps of 2 dB. In a similar manner, Rx-ATT in the Rx chain can be controlled to vary the receiver gain in the range 20–90 dB in steps of 5 dB. These two features (Tx power and Rx gain control) can be exploited in observing the disturbed weather conditions, like convection/precipitation events of varying intensities.

Initially, a full-fledged active phased array, with beam agility, was planned to be built for studying the 3D atmospheric turbulence in the entire radar probing volume. This idea was dropped due to the high cost and complexities involved. As a compromise, the low-cost 2D Butler BFN is built to generate 81 fixed beams to fill the 2D angular space (20° × 20°). In principle, the radar beam can be switched on a pulse-to-pulse basis to image the radar probing volume in a near-simultaneous manner. However, we have restricted the present profiler to operate in five beams (DBS mode) because the present DRx is not capable of processing the echoes on a pulse-to-pulse basis. A custom-designed digital receiver, which will be able to process echoes on a pulse-to-pulse basis, is under development, to perform the 3D radar imaging at a later date.

A special campaign was conducted to validate the LAWP data and it was found that the performance of LAWP was exceptionally good. Typical examples, presented in section 4, illustrate the potential of LAWP for studying wind variability, ABL dynamics, and diagnosing the tropical precipitating systems.

Acknowledgments

We sincerely thank the Governing Council, NARL, for granting necessary financial resources for the development of LAWP. We gratefully acknowledge the support rendered by Mrs. A. Triveni, Mr. J. Raghavendra, and Mr. K. Jayaraj in integrating the radar and Dr. M. Venkatratnam for providing GPS sonde-derived wind profiles. We also thank the local industries for developing the subsystems of LAWP.

REFERENCES

  • Angevine, W. M., , White A. B. , , and Avery S. K. , 1994: Boundary layer depth and entrainment zone characterization with a boundary-layer profiler. Bound.-Layer Meteor., 68, 375385.

    • Search Google Scholar
    • Export Citation
  • Balanis, C. A., 1982: Antenna Theory: Analysis and Design. John Wiley & Sons, 790 pp.

  • Balsley, B. B., , and Gage K. S. , 1982: On the use of radars for operational wind profiling. Bull. Amer. Meteor. Soc., 63, 10091018.

  • Barth, M. F., , Chadwick R. B. , , and van de Kamp D. W. , 1994: Data processing algorithms used by NOAA’s wind profiler demonstration network. Ann. Geophys., 12, 518528.

    • Search Google Scholar
    • Export Citation
  • Bianco, L., , and Wilczak J. M. , 2002: Convective boundary layer mixing depth: Improved measurement by Doppler radar wind profiler using fuzzy logic. J. Atmos. Oceanic Technol., 19, 17451758.

    • Search Google Scholar
    • Export Citation
  • Bianco, L., , Wilczak J. M. , , and White A. B. , 2008: Convective boundary layer depth estimation from wind profilers: Statistical comparison between an automated algorithm and expert estimations. J. Atmos. Oceanic Technol., 25, 13971413.

    • Search Google Scholar
    • Export Citation
  • Carter, D. A., , Gage K. S. , , Ecklund W. L. , , Angevine W. M. , , Johnston P. E. , , Riddle A. C. , , Wilson J. , , and Williams C. R. , 1995: Developments in UHF lower tropospheric wind profiling at NOAA’s Aeronomy Laboratory. Radio Sci., 30, 9771001.

    • Search Google Scholar
    • Export Citation
  • Detrick, D. L., , and Rosenberg T. J. , 1990: A phased array radiowave imager for studies of cosmic noise absorption. Radio Sci., 25, 325338.

    • Search Google Scholar
    • Export Citation
  • Ecklund, W. L., , Carter D. A. , , and Balsley B. B. , 1988: A UHF wind profiler for the boundary layer: Brief description and initial results. J. Atmos. Oceanic Technol., 5, 432441.

    • Search Google Scholar
    • Export Citation
  • Ecklund, W. L., , Carter D. A. , , Balsley B. B. , , Currier P. E. , , Green J. L. , , Weber B. L. , , and Gage K. S. , 1990: Field tests of a lower tropospheric wind profiler. Radio Sci., 25, 899906.

    • Search Google Scholar
    • Export Citation
  • Fabry, F., , and Zawadzki I. , 1995: Long-term radar observations of the melting layer of precipitation and their interpretation. J. Atmos. Sci., 52, 838851.

    • Search Google Scholar
    • Export Citation
  • Farley, D. T., 1985: On-line data processing techniques for MST radars. Radio Sci., 20, 11771184.

  • Gage, K. S., , Williams C. R. , , and Ecklund W. L. , 1994: UHF wind profilers: A new tool for diagnosing tropical convective cloud systems. Bull. Amer. Meteor. Soc., 75, 22892294.

    • Search Google Scholar
    • Export Citation
  • Hashiguchi, H., , Fukao S. , , Tsuda T. , , Yamanaka M. D. , , Tobing D. L. , , Sribimawati T. , , Harijono S. W. B. , , and Wiryosumarto H. , 1995: Observations of the planetary boundary layer over equatorial Indonesia with an L-band clear-air Doppler radar: Initial results. Radio Sci., 30, 10431054.

    • Search Google Scholar
    • Export Citation
  • Hashiguchi, H., , Fukao S. , , Moritani Y. , , Wakayama T. , , and Watanabe S. , 2004: A lower troposphere radar: 1.3-GHz active phased-array type wind profiler with RASS. J. Meteor. Soc. Japan, 82, 915931.

    • Search Google Scholar
    • Export Citation
  • Hildebrand, P. H., , and Sekhon R. S. , 1974: Objective determination of the noise level in Doppler spectra. J. Appl. Meteor., 13, 808811.

    • Search Google Scholar
    • Export Citation
  • Imai, K., , Nakagawa T. , , and Hashiguchi H. , 2007: Development of tropospheric wind profiler radar with Luneberg Lens Antenna (WPR LQ-7). Sumitomo Electric Industries Technical Review, No. 64, Sumitomo Electric Industries, Osaka, Japan, 38–42. [Available online at http://global-sei.com/tr/pdf/energy/64-06.pdf.]

  • Krishnan, P., , Kunhikrishnan P. K. , , and Nair S. M. , 2005: Time-height evolution of intraseasonal oscillations in the tropical lower atmosphere: Multilevel wind observations using UHF radar. Geophys. Res. Lett., 32, L07805, doi:10.1029/2004GL022019.

    • Search Google Scholar
    • Export Citation
  • Kumar, K. K., , and Jain A. R. , 2006: L band wind profiler observations of convective boundary layer over Gadanki, India (13.58°N, 79.28°E). Radio Sci., 41, RS2004, doi:10.1029/2005RS003259.

    • Search Google Scholar
    • Export Citation
  • Law, D. C., , McLaughlin S. A. , , Post M. J. , , Weber B. L. , , Welsh D. C. , , Wolfe D. E. , , and Merritt D. A. , 2002: An electronically stabilized phased array system for shipborne atmospheric wind profiling. J. Atmos. Oceanic Technol., 19, 924933.

    • Search Google Scholar
    • Export Citation
  • May, P. T., , and Strauch R. G. , 1998: Reducing the ground clutter on wind profiler velocity measurements. J. Atmos. Oceanic Technol., 15, 579586.

    • Search Google Scholar
    • Export Citation
  • May, P. T., , Cummings F. , , Koutsovasilis J. , , Jones R. , , and Shaw D. , 2002: The Australian Bureau of Meteorology 1280-MHz wind profiler. J. Atmos. Oceanic Technol., 19, 911923.

    • Search Google Scholar
    • Export Citation
  • Mead, J. B., , Hopcraft G. , , Frasier S. J. , , Pollard B. D. , , Cherry C. D. , , Schaubert D. H. , , and McIntosh R. E. , 1998: A volume-imaging radar wind profiler for atmospheric boundary layer turbulence studies. J. Atmos. Oceanic Technol., 15, 849859.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., 1995: Using radar-measured radial vertical velocities in rain and snow to distinguish precipitation scattering from clear-air scattering. J. Atmos. Oceanic Technol., 12, 257267.

    • Search Google Scholar
    • Export Citation
  • Rao, T. N., , Rao D. N. , , Mohan K. , , and Raghavan S. , 2001: Classification of tropical precipitating systems and associated Z-R relationships. J. Geophys. Res., 106 (D16), 17 69917 711.

    • Search Google Scholar
    • Export Citation
  • Rao, T. N., , Kirankumar N. V. P. , , Radhakrishna B. , , Rao D. N. , , and Nakamura K. , 2008: Classification of tropical precipitating systems using wind profiler spectral moments. Part I: Algorithm description and validation. J. Atmos. Oceanic Technol., 25, 884897.

    • Search Google Scholar
    • Export Citation
  • Reddy, K. K., and Coauthors, 2001: Lower atmospheric wind profiler at Gadanki, tropical India: Initial results. Meteor. Z., 10, 457468.

    • Search Google Scholar
    • Export Citation
  • Reddy, K. K., and Coauthors, 2002: Planetary boundary layer and precipitation studies using lower atmospheric wind profiler over tropical India. Radio Sci., 37, 1061.

    • Search Google Scholar
    • Export Citation
  • Rinehart, R. E., 2004: Radar for Meteorologists. Rinehart Publications, 482 pp.

  • Rogers, R. R., , Ecklund W. L. , , Carter D. A. , , Gage K. S. , , and Ethier A. , 1993: Research applications of a boundary layer wind profiler. Bull. Amer. Meteor. Soc., 74, 567580.

    • Search Google Scholar
    • Export Citation
  • Sato, T., 1989: Radar principles. Handbook for MAP, Vol. 30, S. Fukao, Ed., SCOSTEP Secretariat, 19–53.

  • Srinivasulu, P., , Padhy M. R. , , Yasodha P. , , and Rao T. N. , 2006: Development of UHF wind profiling radar for lower atmospheric research applications. Department of Space, National Atmospheric Research Laboratory Preliminary Design Rep., 74 pp. [Available online at http://www.narl.gov.in/PDFs/TechnicalReport.pdf.]

  • Srinivasulu, P., , Yasodha P. , , and Narayana Reddy S. , 2007: Modified Butler beam forming network for UHF wind profiling radar application. Proc. of Int. Radar Symposium India (IRSI), Bangalore, India, IETE Bangalore Centre and Cosponsors, 406–410.

  • Srinivasulu, P., , Yasodha P. , , Reddy S. N. , , Satyanarayna S. , , and Jayaraman A. , 2011: Simplified active array L-band radar for atmospheric wind profiling: Initial results. J. Atmos. Oceanic Technol., 28, 14361447.

    • Search Google Scholar
    • Export Citation
  • White, A. B., , Jordan J. R. , , Martner B. E. , , Ralph F. M. , , and Bartram B. W. , 2000: Extending the dynamic range of an S-band radar for cloud and precipitation studies. J. Atmos. Oceanic Technol., 17, 12261234.

    • Search Google Scholar
    • Export Citation
  • Williams, C. R., , Gage K. S. , , and Ecklund W. L. , 1995: Classification of precipitating clouds in the tropics using 915-MHz wind profilers. J. Atmos. Oceanic Technol., 12, 9961012.

    • Search Google Scholar
    • Export Citation
  • Woodman, R. F., 1985: Spectral moments estimation in MST radars. Radio Sci., 20, 11851195.

Save