the water level was at the top of the void. As the 

 neutron source nesired the water level, the 

 moisture content increased linearly until the 

 center of measurement reached the water level; 

 then the true soil moisture content approached 

 the saturated value of 45 percent (maximum 

 pore volume) asymptotically. Thus, the neu- 

 tron meter indicated that the water level (total 

 saturation) was at 26 inches depth when the 

 water level was actually at 18-3/4 inches depth. 

 By comparing the control curve and the sample 

 curve, it can be seen that there was no error 

 when the probe was at 14 inches, but the posi- 

 tive error increased with depth as the fully satu- 

 rated void strongly influenced the neutron 

 measurement in the sample tank. The maxi- 

 mum positive error was +18.5 percent, or 

 slightly higher than when the water level was at 

 the surface of the medium. The reason for this 

 was that the peak reading in the sample tank 

 had not changed while the curve for the control 

 tank flexed sharply under the effect of the 

 lowered water table. 



In figure 7C, the water level was down to 

 21-1/4 inches, and the void was three-fourths 

 full. At 14 inches depth, there was no meas- 

 urement error; at 23-3/4 inches depth, the 

 measurement was +19.2 percent. The increase 

 in positive error, when the void was not fully 

 saturated, was caused by the lowering of the 

 rounded portion of the control curve to a 

 point opposite the peak in the sample curve. 

 The total measurement error was +1.38 

 inches. 



Figure 7D shows the situation when the 

 water level was 23-3/4 inches below the surface 

 and the void was half full. With the center of 

 measurement at 14 inches depth, the source 

 was on a level with the upper half of the void 

 which acted as a large air gap and gave a nega- 

 tive measurement error. The water level just 

 below the source acted to reduce the magni- 

 tude of the negative error. The total measure- 

 ment error from 14 inches to 33 inches was 

 +0.50 inch of water and was the algebraic sum 

 of the negative and positive measurement 

 errors. 



Figures 7E and 7F show the reduction in the 

 amount of positive measurement error as the 

 water level dropped. When the void was en- 

 tirely empty, the largest measurement error 



was -4.1 percent. The total measurement error 

 was -0.35 inch. 



Large Void-Coarse Sand 



The measurement error caused by the large 

 void was very similar in shape to that of the 

 small void except that the magnitude of the 

 errors was quite different. Figure 8A shows the 

 water level above the large void, which means 

 the medium surrounding the large void was 

 saturated completely. The maximum 

 measurement error was +55.0 percent mois- 

 ture by volume, and the total measurement 

 error through the void was +5.64 inches. In 

 figure 8B the water level was 42 inches below 

 the surface and at the top of the large void; the 

 soil moisture curve for the control tank showed 

 the same sharply reflexed shape as seen in fig- 

 ure 7B for the small void. The maximum meas- 

 urement error was +55.0 percent and the total 

 error was +5.52 inches. Figures 8C through 8F 

 show the progressive reduction of the positive 

 error and the increase of the negative error as 

 the water-filled void became an air gap. 



Small Void-Fine Sand 



Figure 9A shows the results of measure- 

 ments from the two tanks when the small void 

 was fully saturated. The error in measurement 

 with the probe at 24 inches below the surface 

 was +16.5 percent by volume. The total meas- 

 urement error from 13 inches to 33 inches 

 below the surface was +1.54 inches of water. 



Figure 9B shows the two curves with the 

 water level dropped to 18-3/4 inches below the 

 surface; the water level was at the top of the 

 void. The maximum measurement error was 

 +16.6 percent by volume with a total measure- 

 ment error of +1.42 inches of water from 16 

 inches below the surface to 33 inches below the 

 surface. 



Figure 9D shows the two curves with the 

 small void half full of water, a condition which 

 produced only a small amount of negative 

 measurement error. The maximum negative 

 error was -3.1 percent moisture by volume, and 

 the maximum positive error was +8.4 percent 

 by volume. The total measurement error for 

 the void was +0.32 inch of water. 



9 



