Laserfiche WebLink
the 2005 data that production in the reclaimed area is much higher than in the reference <br />area. This situation is ideally suited for use of the "reverse null" t-test approach allowed <br />by new Rule 4.15.11(2)(c). A minimum sample size of 30 is required, in order to use the <br />reverse null success demonstration approach. The general form of the reverse null t-test <br />given in the cited rule is based on the situation where sample adequacy is achieved in the <br />reference area. When sample adequacy is not achieved in the reference area, a two <br />sample reverse null t-test is necessary, incorporating Satterthwaite standard error <br />term and adjusted degrees of freedom. <br />The narrative sections of the submittal include discussion of "adjusted" production data, <br />obtained by removing the contribution of "undesirable species" (defined as annual grasses <br />and fortis, biennial fortis, and undesirable perennial fortis). The undesirable perennial <br />fortis should be defined. Also, for the purpose of statistical success testing, the <br />undesirable species component will need to be removed from each sample <br />observation prior to calculation of sample statistics (mean, standard deviation, <br />variance). This would be the case for cover as well as production, in cases where <br />statistical testing is required to demonstrate success. <br />If the standard can be met with all annuals and biennials removed from the reclaimed and <br />reference azea data, this is the simplest procedure to follow. However the 1995 Bond <br />Release Guidelines offer some alternatives to elimination of all annuals and biennials, <br />which the operator may wish to consider if success demonstration using the "perennial <br />only" comparison is problematic (see Item IV.A.3, on page 11, third paragraph of the <br />Guidelines). <br />2005 Production Samoline <br />AVF Reclaimed and Reference Area Data <br />Despite use of small quadrats, the 2005 production data for the AVF reclaimed and reference areas <br />did achieve sample adequacy based on the appropriate sample size adequacy formula, and quadrats <br />were appropriately treated as subsamples. The sample mean did (bazely) exceed 90% of the <br />reference area sample mean. Although the quadrats were smaller than allowed under the current rule, <br />there was no minimum quadrat size specified in the rules in effect at the time of sampling. A <br />statistically valid demonstration of success was made, based on the 2005 production data. <br />Rangeland Reclaimed and Reference Area Data <br />In contrast to the AVF area, statistical sample adequacy was not demonstrated for the (much more <br />variable) rangeland production data, in either the reference area or the reclaimed area. Probably due <br />to the sample adequacy problem, Greystone chose to define individual quadrats as sample <br />observations, even though they were arranged as groups of 6 quadrats along 15 transects. Based on <br />this terminology, 90 sample observations were tallied, rather than 15. Still, sample adequacy was not <br />achieved in either the reference area or the reclaimed area. The validity of this approach could be <br />questioned on the basis that 6 quadrats placed at systematic intervals along a 25 meter transect might <br />2 <br />