Laserfiche WebLink
2.0 SAMPLING METHODOLOGY <br />The sample layout protocol for reclamation success evaluation used a systematic procedure <br />designed to better account for the heterogeneous expression of the multiple seedings and <br />physiognomic features within the sampling area while precluding bias in the sample site selection <br />process. By design, the procedure was initiated randomly, and thereafter, samples were located in a <br />systematic manner, e.g., at grid coordinates spaced at consistent intervals for each sampled area. In <br />this manner, "representation" from across the entire area was "forced" rather than risking that <br />significant pockets are entirely missed, or overemphasized as occurs with other sample distribution <br />techniques. <br />The procedure for sample location within each unit was as follows. First, a fixed point of <br />reference that could be relocated from year to year (e.g. road - fence gate) was selected. Second, a <br />computer generated systematic grid of appropriate dimensions (e. g., 900 ft. x 900 ft.) was selected <br />to provide sufficient coordinate intersections to be used for sample sites (as described above). Third, <br />a scaled representation of the grid was overlain on maps of each area and the coordinates <br />transferred to handheld GPS receivers. Transect orientation (direction) at each selected grid node <br />was randomly predetermined during the office planning phase. <br />Sampling within the Phase II Bond Release Block was conducted to a minimum of 40 cover <br />samples. From these preliminary efforts, sample means and standard deviations for total non - <br />overlapping vegetation ground cover was calculated. The Cochran formula (below) for determining <br />sample adequacy was used to calculate nmin, whereby the population was estimated to within 10% <br />of the true mean (u) with 90% confidence. <br />When the inequality (nmin 5 n) was true, sampling was deemed adequate; and nmin was <br />determined as follows: <br />nmin = (t 2S 2) (0.1 X )2 <br />where: n = the number of actual samples collected (initial size = 40) <br />t = the value from the one -tailed t distribution for 90% confidence with n-1 <br />degrees of freedom; <br />S Z = the variance of the estimate as calculated from the initial samples; <br />7 <br />