THEORY

Description of Sodar Installation

The SODAR consists of three main sub-systems;

  • Antenna sub-system, consisting of a concrete foundation, with three antennae, made from fibreglass acoustic shields, and housing three high-powered acoustic drivers (one per shield). This is connected to the control electronics via a set of cables, which may be run overhead or underground. These cables do not carry mains power. An optional automatic weather station can also be used, to provide pressure, temperature, relative humidity, wind speed and direction at 2 metres above ground level. The automatic weather station is not part of the baseline instrument and is not required as an input to the SODAR data processing.
  • Control electronics sub-system, consisting of a rack of low frequency analogue electronics, connected to a control and processing computer with analogue acquisition card, digital controller card, and communications interface to the display computer. This sub-system can be housed in a user-supplied building or caravan within about 100 metres of the antenna site. The control electronics requires mains power, and is compatible with a standard 10A GPO. The floor area required for the unit is approximately that required for a standard 19″ rack. The computer may be rack-mounted, or desk mounted, and the details can be determined with the user. The computer can also provide the interface for the optional automatic weather station, and to other meteorological instruments, formatting and recording the data along with the SODAR data, and transmitting it to the display computer.
  • A display computer sub-system, which serves to provide a display of the horizontal wind velocities, archive data, and communicate with the user’s central air quality database computer. It can also provide a display of the current weather conditions measured by the optional automatic weather station. The computer communicates with the control computer via a serial communications link. The nature of the link varies with the installation, and may be a hard-wired serial link (eg RS232, or RS488 for long distance), a local area network (if one is available at the site), a leased line, or a pair of radio modems (which do not require a license for a range of up to 2 km). The baseline approach is the use of a hard-wired RS232 serial link between the control and display computers, with options as noted in Section 5 of this response. The display computer is a Pentium computer with colour display. If desired, the display computer functions can be incorporated into the control computer.

Key Features of the Atmospheric Research Sodar

The ARPL Sodar is a state-of-the-art design, with the following features;

  • Low Maintenance. The ARPL design, with separate transducers and antennae for the three directions, has a lower parts count, and resulting higher reliability, than a phased-array system. In addition, the acoustic shields are well protected from accumulated dirt and leaves, so that frequent cleaning is not necessary, as is the case in a directly-illuminated phased array system.
  • High Accuracy. The ARPL Sodar provides a real-time display of data quality, so that the user is instantly aware of the altitude to which data is valid. The algorithm used to determine the upper limit of valid data is a conservative one, and the model used can be altered (by recompilation of the code) to user-defined criteria.
  • Access to Raw Data. The data stream from the instrument consists of both the raw spectral data, and the processed wind fields. The user can employ the raw data for post-processing, trialing different algorithms, or extracting additional information from the data stream.
  • Storage of Data. The raw and processed data can both be stored on the SODAR computer hard disks, and written to CD-ROM for archiving and backup.
  • Access to Source Code. The ARPL Sodar can be supplied with a partial set of source code, to allow the user to change data processing routines, quality criteria etc, and recompile the resultant code. (Note that access to this source code requires that the user sign a non-disclosure agreement with Atmospheric Research Pty Ltd. Some low-level signal processing routines are not provided as source code, due to the nature of the intellectual property, and the linkage between details of the hardware design and the signal processing codes. The user is required to provide their own copy of the compiler).
  • Built-in Test. The ARPL Sodar has built-in test (BIT) functionality, to identify when any of the antenna chains fails (typically due to a faulty transducer diaphragm). The BIT electronics monitors the peak voltage and current to each transducer, and can thus identify open or shorted diaphragms, as well as failed power amplifiers, relays or power supplies.
  • Support from Current Research. The ARPL Sodar is based on ongoing research undertaken by the Lower Atmosphere Research Group within the School of Physics at the Australian Defence Force Academy. This current and ongoing link between ARPL and ADFA ensures that the Sodar is a state-of-the-art unit.
  • Direct Support by the Developers. The ARPL group who install and support the instrument are from the team that undertakes the design and development of the instrument, and are involved in ongoing research into atmospheric remote sensing with ADFA. They can also provide post-installation support to data interpretation.
  • User-Selectability of the Central Frequency. If there is a specific environmental noise source, this allows the Sodar to be tuned away from the noise source.
  • Smart Frequency Dithering. The Sodar emits sound pulses in a small range around its nominal value, and offsets the signal processing for each frequency appropriately, so that any fixed-frequency noise source is averaged out in signal processing. Because the frequency range is small, the transducer is operated within its band of highest efficiency, and the bandwidth of the electronic chain is kept low, to further minimise noise.
  • Operating System. The SODAR can run under Linux as the operating system for the instrument control computer and instrument display and communication computer. Linux is a widely supported, robust version of the UNIX operating system, with extensive support for data communication, including web serving. (More web servers run under Linux than any other operating system).
  • Remote Operation. The display computer may be remotely located from the processing computer, via a radio modem link, RS422, or via a web-based interface. For mobile operations (eg parachute operations), the display computer can be a portable or palm-sized computer.

Modifications to the configuration of the hardware, or to the software (eg to the user display), may be undertaken by ARPL.

Tradeoffs Between Sodar Designs

The design and operation of a SODAR is, fundamentally, set by physical constraints. For a SODAR, the key operating variables are;

  • The operating frequency (which in turn impacts the range, through the frequency dependent attenuation of sound in air);
  • The height resolution;
  • The pulse length (which in turn impacts the range, since short pulses mean less energy available for the measurement);
  • The Doppler shift resolution.

Once the frequency is set, there is only one free variable available between the height resolution, the pulse length or the Doppler shift resolution (ie once one is fixed, the other two follow). For example, as the height resolution is increased, the pulse length must be shorter, which decreases the time available for Doppler shift determination, which leads to decreased Doppler shift resolution, and in turn to decreased wind speed resolution.

There are two key technical trade-offs in the development of a Sodar system;

  • Use of three single-transducer antennae versus a phased array;
  • Use of a single frequency for operation, versus use of several simultaneous frequencies.

1: Phased Arrays versus Three Antennae

Key trade-offs in the selection of a phased array system versus a 3-antenna system are shown in Table 2.4-1. The range of the systems is driven by background noise levels, which are in turn driven (for a common site) by the acoustic isolation of the transducers from the environmental noise, and by the acoustic power from the transducers. Phased array systems require additional shielding to achieve the same range as 3-Antennae systems. The acoustic power delivered by the phased array system depends on the nature and number of transducers. Phased array systems typically use many, low-powered transducers, while a 3-Antenna system uses a single, high-powered transducer per axis.

The power consumption of the systems is generally dominated by their associated processing, display and storage computers, which have high power consumption in their backplanes (for ISA-bus PCs), hard disk drives, and monitors.

In general, unless the system is required to be mobile, the 3-antenna design is preferred.

2: Single-Frequency Versus Multi-Frequency Operation

It is, in principle, possible to operate the Sodar at more than one acoustic frequency, in order to obtain data at several frequencies to improve the signal processing on the returned data. This can be done in two ways;

  • Simultaneous pulsing at several frequencies;
  • Sequential pulsing at several frequencies.

In practise, the performance of the Sodar is driven by the acoustic power that can be launched into the atmosphere. This is set by the power handling capability of the transducers used in the Sodar. In a simultaneous multi-frequency system, the power handling capability is split between several frequencies, so that the net energy of the acoustic pulse is reduced. In addition, the receive chain has to have a wider bandwidth, and hence greater susceptibility to environmental noise and internal electronic noise (although this would typically be low).

The overall effect of these design trade-offs is that the theoretical improvements in performance that ought to be achievable with a multi-frequency system are not seen in practise. ARPL use a single-frequency system, optimised for range (which decreases with acoustic frequency) with two refinements, smart frequency dithering, and user-selectable central operating frequency.

ARPL offer a simultaneous pulsing option, in which the three transucers are fired simultaneously, rather than in succession, and processed in parallel. This increased the number of data points per channel by three, improving the signal to noise of the data, and hence the range of the instrument. Simultaneous pulsing is not possible with a phased-array configuration, because only one beam can be synthesised optimally at a time.

ParameterPhased Array3 Antenna
Electronic ComplexityHigh, due to the need for 16 - 256 transducers and associated electronicsLow, using 1 - 3 channels of transducers and processing electronics
ReliabilityElectronic reliability reduced by more complex circuitry and more interconnectionsHigh reliability due to low parts count
SizeSmall (for system without enclosure to avoid rain and dust)Large
Susceptibility to NoiseHigher than 3 Antenna approach due to sidelobesLow, due to good beam pattern and shielding
External Noise PollutionCan be problematic if sidelobes are present near the horizonLow, and fixed, due to shielding
Frequency of CleaningHigh, since transducers are open to the elementsLow, since shields protect the transducers
TransportabilityCan be transported on a small trailer.Requires skid mounting and a truck
Ability to Simultaneously measure multiple componentsCannot be easily done, since only one beam is generated at once.Can be done, for higher data rates
RangeFundamentally limited by frequency. Can be higher due to potentially higher acoustic power.Fundamentally limited by frequency.
Power ConsumptionHigh due to large number of transducers and associated amplifiers and processing circuitsLow, due to fewer electronics.