Colonel Charles J. Treat ae hye “A an . : -2- Nov 10 1977 De For clarity, any detector measuring 2han in-situ (which includes the aerial survey), will not detect material below 3cm. 6. The discussion by Madaline Barnes should be amplified to include L - the point that low density measurements or less frequent measure- ,- error term, ak “ments result in a lower confidence in the estimate and a greater oo " In Tom Crites’ discussion the point was left out that the calculated grid size was a direct variable in relation to the chosen contami- nation level. This hypothetical case only applies to homogeneous distributions which we don't have on RUNIT. I also wish to reem- phasize that the half-distance technique also implies a somewhat 4 homogeneous distribution, This means the greater the distance between sample points, the less confidence and the more you have to be prepared for surprises. I very strongly support the master grid concept, and that every sample must lie on the grid, or sub-component, 7-9. 10. 11-13. 14, nae - No comment, I believe this method has some utility for the quick and dirty answer, but caution that it also contains high variability in individual results. No comment. In my opinion the choice of grid size or intensity of investigation ‘is where you really trade cost and effort of scoping against the desire to gain high confidence of estimates, One must carefully watch the data here, because the situation exists where . a little effort may increase the confidence a great deal, or it may take a great deal of effort to increase the confidence a little. . 15-17. No comment, I believe we will all be surprised at the amount of resources it ‘+. will take for the job outlined, and that there will be no such mg _ thing as minimal resource expenditure. I agree with the concer that without giving priority to the RUNIT task that it will take a very long time. - me Looking at the time spent on RUNIT during the 1972 survey, the “oe A ws 20. . Pre vq 18,19. . < ne ve