Laserfiche WebLink
plant species present in only one of the samples accounted for, at most, 7.4% of the <br />shrubs /sapling density. <br />There were 729 step -point samples that fell in Phase 1 with the Program -level <br />design as compared to the 1581 samples that were established with the Phase 1 sampling <br />design (Table 6). The Program -level step -point data includes step - points from several <br />land cover types (forest and emergent wetland) in Phase 1. Total percent plant cover in <br />Phase 1 was similar for both sampling efforts; total plant cover was estimated at 79.04% <br />with the Program -level sampling and 76.27% with the project specific sampling. <br />However, comparison of the percent plant cover of each species varied between the two <br />surveys due to seasonal changes in the development of vegetation between the sampling <br />periods (the Program - levels monitoring on Cottonwood Ranch began June 21 and was <br />completed July 16). Kentucky bluegrass was the most abundant species from both <br />sampling periods (28% versus 21 %). However Downy brome (Bromus tectorum) <br />dropped from 11.85% plant cover for the Program -level sample to 0.34% for the project - <br />specific sample, and common ragweed increased from 3% to 17% from the Program - <br />level sample to the project - specific sample. <br />We also computed the Ochiai Index (a similarity coefficient) for the two species <br />lists obtained in each of the two samples (Ludwig and Reynolds 1998). The index is <br />equal to 1 if every species is present in both samples and 0 if there is no overlap of <br />species in the two samples. The index is calculated as the number of species in both <br />samples divided by the square root of the number of species in each sample multiplied <br />together. There were 53 species detected using the Program -level sample and 65 species <br />detected using the Management -level sample. Thirty-one species were common between <br />the two samples for a similarity index of 0.528. The index suggests that nearly half the <br />species from the two samples were the same and nearly half were different. This <br />difference may again be attributable, in part, to seasonal variation. <br />Management -level Sampling Design Evaluation <br />Phase 1 PCQ Level of Effort <br />We evaluated the bias, confidence interval coverage, and precision of density <br />estimates under alternate sampling intensities. The tree and shrub density estimates were <br />recalculated using a dataset where the point- centered quarter data (species and distance) <br />had been randomly re- sampled with replacement (Manley 199 1) from the complete <br />dataset (17 points). The re- sampled datasets were created to simulate from 2 (1 point per <br />44 acres) to 35 (1 point per 2.5 acres) points. The sampling procedure and calculation of <br />the estimates was repeated 1000 times. <br />Relative bias was calculated as the mean (over 1000 iterations) of the difference <br />between the re- sampled estimate and the best estimate made using all the data (tree <br />density for Cottonwood Ranch and Jeffrey Island combined was 29.13 stems per acre and <br />shrub /sapling density was 79.18 stems per acre), divided by the best estimate. The best <br />estimate is also referred to the `true' parameter value in re- sampling theory because the <br />original dataset is viewed as the population and the re- sampled datasets are viewed as <br />samples from the population (Manley 1991). Confidence interval coverage was defined <br />as the proportion of iterations (out of 1000) where confidence interval estimates <br />contained the true parameter value. Relative precision of density estimates was estimated <br />