GENERAL32698 - Laserfiche WebLink

Home Browse Search

census method is more stringent than sampling methods specified in the regulations; any error in counting is more likely to result in "under-count" than "over-count". Year 1 census results in the grassland community indicated 258 woody stems per acre, compared to the success standard of 90 stems per acre. Although the regulations do not specifically address the total census method, it is actually more precise than sampling, and I would recommend that we accept the census as a valid demonstration of success for the small grassland area woody stem density standard. For the riparian community, proper sampling procedures were used, and results cleazly document that woody stem density in the reclaimed area exceeds the approved standard (reclaimed area sample mean of 8,175 stems per acre vs. standard of 900 stems per acre.) Cedar Creek used a reverse null t-test to demonstrate success, and for practical purposes the test demonstrates that the reclaimed area mean exceeds the standard by an extremely wide margin, at the 90% level of statistical confidence. Unfortunately, Rule 4.15.11(2)(c) requires a minimum sample size of 30, if the reverse null demonstration is employed, and a sample size of only 20 was obtained. An alternative approach that Cedar Creek could have employed would be to use a direct comparison if sample adequacy is obtained in the reclaimed area. Cedar Creek's results indicate that sample adequacy was not obtained (Nmin of 24 indicated on Table 16). However, it appears that Cedar Creek used a precision level of 0.10 in the sample adequacy formula, and in addition, they used a table "t" value for atwo-tailed test rather than the value for aone-tailed test, Both the precision level and the table t value used by Cedar Creek were more stringent that required by regulation, and had the effect of inflating the required sample size. Rule 4.15.11(2)(a) allows for use of the t-tail value for aone-tailed test, and also allows for use of a precision level of 0.15 in the sample adequacy formula, specifically for woody plant density. Using the formula with a precision level of 0.15, and table "t" value for aone-tailed test results in a minimum required sample size of 6, rather than 24 (see spreadsheet calculations attached, "2006 Blue Ribbon Riparian Woody Plant Density"). Table 16 of the report will need to be revised, using the less stringent table "t" and 0.15 precision -evel allowed by 4.15.11(2)(a), and success should be demonstrated by direct comparison of the sample mean to (90% of) the approved standard. Summary Table i will need to be revised accordingly. Section 3.5 narrative on page 16 of the report should be amended to reference the direct comparison success demonstration, based on comparison of the sample mean of a statistically adequate sample to the standard. For 2007 sampling, the operator will need to ensure that a minimum sample size of 30 is obtained (if reverse null t-test is employed), or that sample adequacy is demonstrated (with minimum sample size of 15), if direct comparison of sample mean to standard is to be employed. Let me know if you have any questions.