Effects of Detector Distortion
The typical symptom of detector degradation (due to icing, vacuum leak, lowered bias, ete.) was
reduced resolution, i.e., wider photopeaks. The simple analysis program used in the field could not
accommodate such an effect. Photopeak counts would be spread into the background windows
resulting in an erroneously low value for net counts and, therefore, soil concentration. Window
limits in the program could have been changed in the field if one cared to analyze peak shapes for a
detector that was degraded but stable. The philosophy at Enewetak, however, was to correct the
problem rather than attempt to correct the data.
Brush Correction Factor
Most of the islands surveyed were covered with a dense growth of Messerschmidia and Scaevola
scrub vegetation, ranging in height from 1 to approximately 4 meters. A series of measurements
were performed in October - November 1977 on Pearl to determine the effect of this vegetation on
the 60 keV gamma ray from 24lam. Ten representative areas with brush covering 70-80 percent of
the IMP field-of-view were measured. (The access road cut through the brush accounted for most of
the open area.) Brush in each area was then earefully cleared by hand to prevent any soil
disturbance and the measurements repeated. The results of this experiment gave a brush correction
factor of 15 pereent for 100 percent brush cover (see Tech Notes 1.0 and 1.1 for details). No
correlation was observed between the brush height and the brush attenuation factor. This was
attributed to the fact that the vegetation normally encountered on the northern islands typically
grew in the form of a canopy rather than solid cover.
At each measurement location, an estimate of the percent brush cover
within the detector
field-of-view was made by the operator. This value was then used to provide a correction factor for
brush attenuation.
The estimate of brush cover was somewhat subjective and could have been in
error by as much as 20 percent for some locations. Even a 20 percent error in the brush cover
estimate, however, would only introduce a 3 percent error in the 41am concentration value. Thus,
although some uncertainty was inherent in the method used to determine a brush attenuation
correction factor, the uncertainty was less than would result from neglecting brush attenuation
effects completely.
Measurement Reproducibility
A repeatability experiment was conducted on Pearl at location 3-N-0.5 in May 1979 to determineif
any systematic variation could be observed in the IMP measurements over the course of a typical
day. A total of 17 measurements were made, each for the standard 900-sec measurement time, with
the detector fixed in position. The sample standard deviation for the series of measurements was 5
percent of the mean value.
For the same set of measurements, the average one sigma error due to
counting statistics was 6.7 percent.
No systematic variation was observed between the early
morning measurements, made when the ground was damp due to an early morning rain, and the
afternoon measurements made during the hottest part of the day.
One location on Janet was remeasured five times over a two-month period in the fall of 1977. The
standard deviation for this set of measurements was 7.8 percent of the mean value. During the same
period of time, two locations on Pearl were remeasured three times over a period of one month. The
standard deviation was 4.4 percent of the mean for one location and 6.6 percent of the mean for the
other location.
These data indicate that the primary source of error in measurement reproducibility was associated
with counting statistics, which generally ran from 5-7 percent. Additional details on measurement
reproducibility can be found in Tech Note 21.
3.2.8 Detector Field-of-View
The detector field-of-view is of some practical concern for an in situ measurement. However, as
shown in Figure 3-3, even with a collimator the detector response does not drop abruptly to zero.
Thus the "field-of-view" has an edge which is somewhat fuzzy.
97
The field-of-view ean only be