2009-04-01_REPORT - C1982057 (6)

Home Browse Search

• Statistical Methods STATISTICAL DESIGN OF SAMPLING Cover and herbaceous production data were collected in all areas. Woody plant density samples were collected in the 2004 and 2006 Reclaimed Areas and in the Shrub Concentration and Random Sampling Areas as described above, but woody plant density samples were not collected in any of the reference areas or the 2007reclaimed area. Sampling proceeded using a single stage statistical design. Sample locations in reclaimed areas were randomly plotted by use of an internal script routine within ArcGIS 9.2 software. Minimum spacing between sample points was 170 feet (one 50 m transect length). In the extended reference areas, each reference unit was gridded and numbered. Grid cell size was 200 feet X 200 feet. These grid cells were placed contiguously by hand on maps of 1:4,800 scale. Sample locations were selected as the center points of grid cells selected using random numbers generated by computer. Sample points were located in the field using Garmin III handheld GPS units and topographic • maps. Cover and woody plant density transects were located and oriented using, at a glance, the seconds reading on a wristwatch x 6 as an azimuth or a blind throw of a meter stick . The final location of each herbaceous production plot was chosen with a blind throw of the sampling hoop made from the origin of the associated cover/woody plant density transect. At each cover/woody plant density sampling location, the slope (percent) and aspect of the site were determined and recorded, as was the orientation of the transect (in degrees not corrected for magnetic declination). STATISTICAL METHODS Sample adequacy calculations for parameter estimation were carried out using the formula as prescribed by the Colorado Mined Land Reclamation Board for parameter estimation without hypothesis testing (DMG 1995): t 2 s2 Nmin= _ (d X )2 where: • t« = one-tailed t-value with (n-1) degrees of freedom, a=0.10 4