Pre-processing and filtering 



A 10-Hz, fifth-order anti-aliasing Bessel filter was applied to both the 

 resistance wave gauge and electromagnetic current meter analog data to elimi- 

 nate noise and avoid aliasing. 



Data Analysis 



Analysis and interpretation methods 



Frequency domain and time domain analyses were performed on the 

 SUPERTANK hydrodynamic data. The analysis routines for the resistance 

 wave gauge and electromagnetic current meter data were nearly the same, 

 with only minor differences in the initial processing of the data and output 

 parameters. The data were filtered to separate the incident-band wave motions 

 from the low-frequency motions due to the channel seiche, and then spectral 

 and zero up-crossing analyses were performed on the raw, low-pass, and high- 

 pass signals. 



Preprocessing. The electromagnetic current meter data required two types 

 of processing prior to analysis. First, a low-pass filter with a 2-Hz cutoff was 

 applied to remove high-frequency noise. High-frequency noise was present in 

 some data records for deployments with the current meters spaced less than 

 30 cm apart. Second, the elevation of each current meter was compared with 

 the water surface elevation (estimated from a resistance wave gauge at the 

 same cross-shore location) to determine if the current meter was submerged. 

 The criterion for submergence was that the water level was 5 cm or more 

 above the surveyed current meter elevation. Current meter readings were set 

 to zero (for that portion of the record) when the gauge was out of the water. 

 The current meters deployed on the carriage and current meters shoreward of 

 resistance wave Gauge 1 could not be checked for submergence. These 

 gauges were analyzed assuming 100 percent submergence; thus, these results 

 should be used with care. 



Filtering. To separate incident-band wave motions from low-frequency 

 motions due to channel seiche, a non-recursive, low-pass filter was applied. 

 The period cutoff for the filter was set to twice the peak period of the incident 

 waves (the peak period of the long-period motion was generally 20 sec or 

 longer). The period response of the filter is shown in Figure 3-2 for a cutoff 

 period of 6 sec. The filter was applied beginning with the 1,009th data point. 

 The first 1,008 points were skipped to ensure that the waves had reached the 

 most shoreward gauge. Two hundred points were sacrificed at the beginning 

 (in addition to the 1,008 points that were skipped) and end of the data record 

 in constructing the filter. The low-pass time series was output directly from 

 the filter. The high-pass time series was obtained by subtracting the low-pass 

 time series from the preprocessed data. 



Chapter 3 SUPERTANK Hydrodynamics 



61 



