On last post I explained briefly some features related to this Remote Sensing (RS) apparatus, but I consider that it should be explained further. Therefore, I am going to detail other considerations related to Sodar’s technology through two posts based on the academic paper of S. Bradley, I. Antoniou et al. (2005). By now, physical principles and uncertainties will be described. Next post will focus on its calibration methods.
Sodar is a Remote Sensing apparatus that measures 3D wind speeds at high altitudes. By emitting vertical sound beams of sound, it is possible to interpreter the backscattering frequencies due to the Doppler Effect and thus, wind components can be decoded (See Fig. 1). Usually, three or five beams are necessary to obtain reliable raw data measurements. Each of them is usually tilted 15-20º (ϕ) to the vertical (See Fig. 2). Though the emitted signal produces a continuous backscattering echo after crossing the infinite turbulent layers in the atmosphere, the echo generated at the studied altitude (Z) can be recognized according to the following formula. This means that among the continuous echo signal received, the specific signal generated at the Scattering Volume at Z height is generated at t (time) moment. The same principle is used by pulsed lidars.
Where Z is the altitude of the scattering volume, c is the speed of the sound (340 m/s), and finally, t is the time that requires the sound pulse to go and back. Since c is known and t can be set, the altitude can be calculated.
Furthermore, this Scattered Volume (V) can be calculated according to the following formula, where τ is the acoustic time duration.
After receiving and sampling the backscattered signal in a particular beam direction, it is emitted another acoustic pulsed in the next angle. Usually, acoustic beams are automatically oriented. If the Sodar disposes a tweeters array of NA x NA size, sound beam can be oriented by incrementing the phase angle of each row in the array (∆φ), and transmitting different frequencies (fT) to obtain the transmitted wavefront in the beam (ϕ). This is represented in the following formula where d is the distance between tweeters.
The frequency of the backscattering signal contains different sound frequencies that come from electronic and background noises, echoes from nearby objects (mast, walls, buildings, etc.), and the echo signal from the Scattering Volume at Z height. The mechanism to separate noises from the Doppler Shift frequency is based on applying the Doppler FFT (Fast Fourier Transform) but the algorithm used by each manufacturer is different.
However, all of them are developed in order to identify the wind parameters of each sound beam and then, calculating the Doppler Swift frequency (∆fj) according to following formula. Vrj is the wind component for each sound beam.
After describing the main principles than Sodars are based on, let see which are the main sources of errors before studying how they should be calibrated:
- Wrong calibration: Usually, Sodars are calibrated by using measurements from cup or sonic anemometers as standards. Therefore, if anemometers measure wrongly, Sodar will be affected by these mistakes.
- Different calibration methods in relation with Cup Anemometers: Cup anemometers measure different wind speeds and then, they divide the results by the averaging time to obtain the average wind speed. However, Sodar’s approach is different. Sodars calculate the wind speed by summing the average of each wind components (u and v). This different approach between anemometers and sodars mean that though measurements between both apparatuses are compared, they are the result from different procedures.
- Variance errors of average wind speeds (σ2v) and direction (σ2θ): Generally, Sodars provide wind speed values (and directions) after averaging a pack of “acceptable” values. Thus, it is possible to determine the averages of wind components (u and v), mean wind speed (V) and direction (θ), and their variances (σu, σv). However, further analysis should be done if we want to know the average variance for wind speeds and direction. Therefore, the number of recorded values that determines the mean wind speed and direction is not provided usually by sodars.
- Errors to estimate the altitude (Z): Since velocity of sound is not a constant and it relies on the atmospheric temperature, some manufacturers take in consideration the outside temperature to determinate more accurately the altitude of the backscattered signal.
- Wind itself: Since velocity of acoustic pulses rely on wind (it is the basis of ultrasonic anemometers), transmitted and backscattered signals are affected by the wind velocity. Therefore, some corrections should be applied for reaching higher reliabilities. Sodars with three beam sodars are more prone to this uncertainty.
- Incomplete data: Outside noises can affect the reliability of measurements and thus, pieces of data could be dismissed. In order to avoid this issue, some sodars apply filters to cope with external frequencies that they are called noises. Since 5-beam sodars takes more time to provide the all three wind components than 3-beam sodars, these ones are more convenient when data availability is higher.
- Estimating the Doppler Swift Frequency: Though each manufacturer has a different approach to determine the Doppler Swift, there are some uncertainties should be considered to increase the sodar’s reliability.
- Distance and meteorological limitations: Sodars’ reliability drops when the scattering volume is too far (several hundred meters). Under those conditions, sodars are not the best Remote Sensing apparatus for measuring wind profiles. On the other hand, the data availability is not the best one when atmospheric profiles are under neutral conditions (temperature at ground levels is equal (or similar) to higher distances).
- Separation of spatial and temporal samples: Time gaps between different sample measurements must be correlated with the separation of scattering volumes. If this issue is not considered, reliability of Sodar will be affected.
- Tilt angles misalignments (∆ϕ): It has been estimated that 1° misalignment error in the tilt angle represents 5% error in wind speed measurements for a typical 18° beam tilt angle. This misalignments can be caused by several sources:
- Temperature: Typically, variation of 33ºC represents 1° beam tilt misalignment.
- Sodar Out-of-level: For obvious reasons, sodar apparatus must be leveled on the ground.
- Beam spread: Errors are directly proportional to the beam width.
– Cover picture: Ducksters.com
– Fig. 2 and text: SODAR calibration procedure (final reporting on WP3, EU WISE project NNE5-2001-297). S. Bradley, I. Antoniou, et al. (2005).