2013-02-04_REVISION - C1981008 (3)

Home Browse Search

JoEllen Turner 970 - 864 -7682 p.8 Borth Environmental Pollution Consulting, LW October 2, 2012 Comments to "disturbed" soils, handling of the A and B lifts, statistical approach, and selection of data: X concede that it is appropriate to change the reference in my report from "dirt piles" to "disturbed soils ". However, my point is that soils stored in stockpiles are not per definition a soil according to e.g., the Soil Taxonomy, 2" Edition (see my initial report). For instance, a soil is characterized by one or both of the following: horizons, or layers, that are distinguishable from the initial material as a result of additions, losses, transfers, and transformations of energy and matter or the ability to support rooted plants in a natural environment. In my opinion disturbed soils are in no way comparable to well established agricultural soil. Just to clarify, I did not mean to say that the A and B lifts should be placed in the same geographical location. I meant to say that the B lift should be placed below the A lift and not mixed together. I agree with Mr. Dunker that the term "significant" is a term to denote the results of statistical data analysis. The major problem with the obtained data for the substitute soil is that the number of replicates (n =3) is so low that no statistical test can be used with confidence to compare the different soils. However, Aaron Dejoia claimed that, for instance, the salinity is STATISTICALLY HIGHER (at the 95% confidence interval) in the substitute soil as compared to Lift A, Lift B, Barx Profile2, Barx Lift A3, and Barx Lift B4. Please refer to figure 4 in Mr. Aaron Dejoia's report (however the mean values calculated for each soil can not be compared by simply calculating the confidence interval in fact a t- test/ANOVA test will not even work for this dataset for reasons given below). In addition, to prove that the Darvey -Barx Subsoil has statistically fewer coarse (rock) fragments than the Bench I substitute subsoil please refer to the previous analysis of these soils which showed that the average percent coarse fragments in the Darvey -Barx subsoil was 0% but 6.37% in the Bench 1 substitute subsoil based on the dataset presented in the Walsh report on Subsoil suitability, February 2008 (revised July 2008). With respect to pH no statistical test can be used for the given data sets to compare the mean values (see comments below). However, all but the Lift B soil had numerical pH values lower than the substitute soil (since no statistical test can be reliably applied to these data sets we can only rely on measured (empirical) data points. To prove that the organic matter content of the Substitute Soil is STATISTICALLY lower than any of the other soils i.e., Lift A, Lift B, Barx Profile2, Barx Lift A3, and Barx Lift B4 please refer to figure 5 in Mr. Aaron Dejoia's report (though I do not concede the validity of Mr. Dejoia's statistical approach). The Key Ag Services data set is not easy to interpret for the simple reason that some data are presented in cm (e.g., sample TPO4) others in inches (e.g., sample TP05) and some even in feet (e.g., sample TP06). During my site visit I was told by Aaron Dejoia that they "always" use SI units, though this is contradicted by the data sets. This represent a good example of an analytical lab that may not pay attention to important details and that does not have a good QA /QC in place. Another example of lacking QA/QC was when Key Ag had to retract a dataset due to problems with their pH meter. With good QA/QC in place the dataset would never have been send to the client. 7 1 Page PLTF 002481