Laserfiche WebLink
BIO-Logic, Inc. www.bio-geo.com <br />Comparison of Measured Parameters. The following summarizes CDRMS regulations and <br />guidance on comparing vegetation measurements at reclaimed lands to standards for purposes <br />of determining adequacy for Phase 3 bond release. (rec) = reclaimed a7_;.'(_r.e0f) = reference <br />area. 1. If sample adequacy is achieved at?rec) and (ref),and if the m4ean (rec) equals or <br />exceeds Q, then revegetation is successful. If the mean (rec) is less than Q, then <br />statistical hypothesis testing is required, as f Ilows. <br />2. If sample adequacy is achievecrat (r6c) and (ref), and the mean (rec) is less than Q, use <br />a one-sample "standard null" Hest (Equation 2). A one-sample test assumes that the <br />mean (ref) is a "standard" rather than a sample mean, and does not require an <br />evaluation of the variance of the mean (ref) or other statistical complications inherent in <br />two-sample tests. The null hypothesis being tested is that the mean (rec) is equal to or <br />greater than Q. If the null hypothesis is NOT rejected, revegetation is successful. <br />Equation 2 40 <br />Where: <br />X = Bond Release Block Sample Mean <br />Q = 90% of Standard <br />S- = Standard error of mean [s / n J <br />S = Sample standard deviation <br />n = Sample size <br />t, = Calculated t value <br />t, = Table t value <br />If sample adequacy is NOT achieved at (rec) but is achieved at (ref), use a one-sample <br />"reverse null" Mest, using the t-test shown in Equation 2. CDRMS requires a minimum <br />of n = 30 at (rec), and that mean (rec) is greater than Q. The reverse-null t-test is one- <br />sided with alpha = 0.20. The null hypothesis being tested is that the mean (rec) is less <br />than or equal to Q. If the null hypothesis is rejected, vegetation is successful. <br />4. If sample adequacy is NOT achieved at (ref), use a two-sample "reverse null" t-test with <br />Satterthwaite approximation of standard error and degrees of freedom. The test <br />requires a minimum of 30 samples in the (rec) and (ref), and that mean (rec) is greater <br />than Q. To use this test, first the Satterthwaite standard error is calculated as <br />Equation 3