2016-08-08_PERMIT FILE - C1981019A (15)

Home Browse Search

RULE 4 PERFORMANCE STANDARDS Sample Adequacy Determination Sampling within each monitored unit shall be conducted to a minimum of 5, 15 or 20 samples as necessary. Sampling within each unit under consideration for bond release shall start with a minimum of 15 (reference area) or 20 samples (revegetated area) and continue until a statistically adequate sample (if necessary) has been obtained in accordance with Section 4.15.11 (2) (a). From initial sampling efforts, sample means and standard deviations for total non -overlapping vegetation ground cover, production, and woody plant density will be calculated. For bond release applications, the typical procedure is that sampling continues until an adequate sample, nmin, has been collected in accordance with the Cochran formula (below) for determining sample adequacy, whereby the population is estimated to within 10% of the true mean (p) with 90% confidence. For woody plant density, the estimate is to within 15% of the true mean. When the inequality (nmin < n) is true, sampling is deemed adequate; and nmin is determined as follows: nmin = (t2 s2) / (d X )2 where: n = the number of actual samples collected (initial size = 15 or 20) t = the value from the one -tailed t distribution for 90% confidence with n-1 degrees of freedom (a value of approximately 1.3); S2 = the variance of the estimate as calculated from the initial samples; d = precision (0.10 for cover and production or 0.15 for woody plant density; x = the mean of the estimate as calculated from the initial samples. If the initial samples do not provide a suitable estimate of the mean (i.e., the inequality is false), additional samples should be collected until the inequality (nmin :S n) becomes true. However, where sampling is for managerial (monitoring) information, adequacy is not necessary and is calculated for informational purposes only. If reverse -null testing will be utilized to document success, then in accordance with Rule 4.15.11 (2) (c) a minimum of 30 samples must be collected and a demonstration of sample adequacy is not necessary. In this circumstance a two -sample reverse null t-test is mandated along with Satterthwaite approximated degrees of freedom and standard error. However, if an adequate sample can be obtained from the reference area, then a less complex one -sample t-test may be utilized. With the reverse null test, the smaller the variance (given by extra sampling) the better the chances of passing closely matched parameters. For certain statistical demonstrations of woody plant density, documentation of sampling adequacy is often problematic, hence Rule 4.15.11 (3) may be used in lieu of Rule 4.15.11 (2). Rule 4.15.11 (3) (a) is a reverse -null approach based on the median and requires a minimum of 30 samples. Rule 4.15.11 (3) (b) allows direct comparison with standards if a statistically adequate sample cannot be demonstrated in accordance with Rule 4.15.11 (2) (a), however, a minimum of 75 samples with a minimum quadrat size of 100 m2 is required (equivalent to total enumeration of 1.85 acres). Rule 4.15.11 (3) (c) is a standard -null approach based on determination of a "running mean" and a minimum of 40 samples is required. Success Evaluation To summarize, success evaluations involve either a direct or a statistical t-test comparison of appropriate parameters for each variable of interest (cover, production, diversity, or woody plant density). For monitoring efforts, comparisons shall be made directly with either the reference area parameters or the permitted standards to facilitate a determination of the progress of revegetation. In the case of ground cover and production, comparisons shall be made against reference area data of the same year. Diversity and woody plant density variables shall be compared against the standards defined above. Collom — Rule 4, Page 31 Revision Date: 4/18/16 Revision No.: PR -04