rare. Although changing water temperature is the predominant cause of 

 gain change, it appears that, in outdoor testing in the SPTB, there were 

 other causes, including perhaps the chemicals in the water. Sufficient 

 information is not available to determine the nature and extent of these 

 other causes. When records are to be analyzed for wave heights, the meas- 

 urements should be adjusted to account for the change in gain. Since there 

 is no method for determining at what time between the initial calibration 

 and the calibration check the change in gain occurred, the maximum error, 

 as shown by the calibration check, should be assumed for all data during 

 this time interval. 



d. Recording of Wave Envelopes . All instrumentation was placed on a 

 carriage which moved along the tank and recorded a profile of the water 

 surface. The envelope was recorded during each run (except for the first 

 10-minute run) by turning on the recorder at a slow chart-paper speed 

 while the carriage was propelled slowly along the tanks with a handcrank 

 and cable system. One envelope was recorded from station +15 to the gen- 

 erator, the second envelope on the return. The desired carriage speed 

 was 10 feet per minute. However, due to mechanical problems early in each 

 testing season and since the carriage was propelled by hand, this velocity 

 was only approximate. To assist in determining the speed of the carriage, 

 a magnetic relay mechanism was installed in each test area so that the 

 event marker on the strip-chart recorder would mark the position of the 

 carriage at 5-foot (1.52 meters) intervals. During the recording of each 

 envelope, the appropriate station identification was written by the instru- 

 ment operator on the chart paper adjacent to these pen marks. 



2. Data Reduction . 



A manual method and an automated method were used in the reduction of 

 wave reflection data. In 1970, the manual method was used to analyze all 

 the wave records and the automated method for 20 percent of those wave 

 records. In 1971, each method was used to analyze all data. In 1972, 61 

 percent of the data were analyzed by the automated method, 28 percent by 

 the manual method, and 11 percent by both methods. 



a. Manual Method . Using the 5- foot event marks as an approximate 

 scale and knowing that nodes and antinodes should be one-quarter wave- 

 length apart, the approximate locations of nodes and antinodes were deter- 

 mined over the constant depth section of each tank. Wave heights were 

 measured from the preceding trough to the crest. The high wave and the 

 low wave along these intervals were then precisely determined by choosing 

 the highest or lowest wave having at least three increasing or three 

 decreasing waves immediately adjacent to the wave. Pairs of nodes and 

 antinodes were then measured by increments (0.065 inch or 1.66 millimeters) 

 on the chart paper, to the nearest 0.2 increment (0.013 inch or 0.33 

 millimeter), and a K^ determined for each pair using: 



% = 



m - Htf ^ m 



H^ + H^ ~ Hj 

 31 



