Calibration and zero reference 



The acoustic instruments are not calibrated in the common sense of the 

 word, i.e., by finding precise estimates of one or two coefficients. Raw data 

 are not immediately meaningful, but must be processed substantially before 

 physical interpretations are possible. 



Range normalization of backscatter data. As stated in the introduction, 

 acoustic techniques have obvious advantages over single-point measurement 

 methods because they are both profiling and non-intrusive. The profiling ca- 

 pability, however, makes the measurement more complex because the strength 

 of the outward propagating pulse is reduced as it propagates through the water 

 column. This variation in incident energy must be accounted for before the 

 data can be analyzed. This is referred to as range normalization of the data. 



Raw acoustic backscatter data, hereafter referred to as echo level (EL), 

 were collected with both acoustic systems throughout the first 2 weeks of 

 wave action at SUPERTANK. The uncorrected echo level varies as a func- 

 tion of depth even if the concentration is constant along the path of the acous- 

 tic pulse. Range normalization includes the following elements: 



a. Water absorption. 



b. Geometrical spreading. 



c. Attenuation due to particles in the water column. 



The first part, water absorption, can be accounted for by adding a linear 

 term (in logarithmic units) to the echo level. The magnitude of the coefficient 

 is determined mainly by the acoustic frequency but varies with water tempera- 

 ture and salinity. 



The second part of the range normalization, geometrical spreading, is 

 simple when the acoustic scattering volume is located in the acoustic far field, 

 defined as the volume beyond the Rayleigh distance r=X 2 /rf, where d is the 

 diameter of the transducer and X is the acoustic wave length. For the 

 600-kHz and the 2.4-MHz systems, this distance is about 4.0 and 1.5 m, re- 

 spectively. In SUPERTANK, the distance from the transducers to the bottom 

 was 1 to 1.2 m, and all the data were collected in the near field. A numerical 

 model must be used to correct for spreading effects. The model is described 

 in Ma, Vardan, and Vardan (1987), and the results for the 2.4-MHz system 

 are shown in Figure 8-6. 



It is difficult to correctly account for attenuation provided by particles in 

 the water column. This is especially true when the suspended concentration is 

 high and a combination of fine and coarse particles are present. To correctly 

 estimate the attenuation, we ideally need to know the suspended load as a 

 function of particle size. Since this is not possible, the attenuation can be 

 estimated using the following information: 



Chapter 8 ADCP Measurements at SUPERTANK 



145 



