(1) Input data for this program are in the form of digital 

 magnetic-tape records of 4,100 values. The first 4 values of the 

 records are the gage number, month, day, and time of the observa- 

 tions; the last 4,096 values are the time-series pressure values of 

 the wave gage. In the present program the wave gage pressures are 

 stored in thousandths of a foot (head) water at 0.25-second inter- 

 vals. Subroutine BUF reads time-series data into array CNTL, 

 where it is averaged to provide 1,024 time-series values of At = 1 

 second spacing. Units are also divided by 1,000 to convert values to 

 feet (head) of water. 



(2) The date groups of record 1 and record 2 are compared to 

 ensure that times of records are simultaneous; if the times are not, 

 the program searches the record file until this condition is met. 

 The two records are than checked for proper sequence to ensure that 

 gage 1 is analyzed first. Subroutine SWITCH switches arrays if they 

 are not in proper order. 



(3) Each of the two 1,024 value time series is then analyzed for 

 average values which are printed out along with the average depth of 

 water at each gage site. The average value of each of the time- 

 series records is again averaged and is added to the height of the 

 gages above the bottom to obtain the water depth: 



DEPTH = AVERAGE 1 + AVERAGE 2 + fi 



in which AVERAGE 1 is the average of time series 1 = a (0), AVERAGE 2 

 the average of time series 2 = a (0) , and B the height of sensors 

 above the bottom. 



An option to apply a weighting function w( j ) (= W(I) in program) 

 has been incorporated before the FFT subroutine is called. In this 

 particular program a cosine bell weighting function has been incorpo- 

 rated. If the data window option is selected, the two time-series 

 data records, which are read into FIR and F2R arrays, are multi- 

 plied by the following weighting function (cosine bell) 



«o>-i[i-~(¥)] 



where j is the time step number and N the number of data points 

 in series. If no weighting function is desired in analysis set 

 w(j) = 1.0, which is the "box car" weighting function. 



As the cosine bell function reduces the total energy content of 

 the waves, the final energy obtained from the FFT must be rescaled up 

 to the proper value. This is accomplished by scaling up the time- 

 series pressure values by the ratio 



Unwindowed energy /<p > 

 R = 



Windowed energy y <p • 2 > 

 as discussed in equation (4). 



19 



