Investigations showed good agreement between electrical and direct 

 measurements of peak water. Figure A-4 displays measurements on the 

 3-inch circular pile by three techniques in nominally the same test 

 conditions. The three patterns show good basic agreement with slight 

 deviations. Differences could occur because the powder erosion measure- 

 ment is for one wave crest, and because the paper sleeve is rougher than 

 the Plexiglas pile and might cause a different flow boundary layer. 

 Also, the paper might dry slightly before peak water could be marked and 

 the measurement completed. 



The importance of another possiMe effect of materials on the meas- 

 urements can be assessed from the W(a) data in Figure A-5. Here a 

 test was repeated with 40 parts per million of wetting agent, Edwal 

 Kwik-Wet, added to the water to reduce effects of surface tension; ob- 

 served average peak water was marked on the Plexiglas pile with a crayon. 

 The two patterns were very similar, except that the minimums were higher 

 and more symmetrically located without the wetting agent. From these 

 data, it was concluded that the smoothness of the pile material caused 

 no anomalous effects on the data obtained in the model tests. 



A final possible source of variability in W(B) measurements is the 

 variable probe placement near the pile. The waveforms in Figure A-6 and 

 the associated measurements in Table A-1 provide an estimate of the 

 extreme range in this effect. Four waveforms were obtained deep within 

 and just within the 2x2 H-pile channel at 3 = 0° and g = 90°; the two other 

 waveforms were obtained relatively near and far from the flat plate 

 oriented at g = 90°. Although the waveforms show definite qualitative 

 effects of gage placement, the peak water level measurements show less 

 than a 4-percent change, which is considered insignificant compared to 

 the variation in W(g) over the range of orientation angle. 



The major conclusions are: 



(a) The waveforms recorded by electrical water level gages 



are somewhat suspect, because the gage response to changing 

 water level is imperfect in an undefined way; 



(b) measurements of crest dimensions (peak water level) from 

 gage records have adequate accuracy; the agreement between 

 direct and electrical measurements of peak water levels 



is within about 5 percent. 



77 



