Abstract
The development of high-speed computers has enlarged the scope of data processing and reduction. Many statistical methods, particularly those involving multivariate analysis, can be routinely used for the organization of data into workable quantities. A study of the chemical quality of ground water from aquifers of the Woodbine Formation was done to illustrate the use of selected statistical methods. The Woodbine Formation (basal Upper Cretaceous) crops out in northeastern Texas and dips toward the center of the East Texas Basin. In the area of study the Woodbine contains the shallowest aquifers that can provide reasonably large amounts of ground water. The formation consists of sand and sandstone (some of which is tuffaceous), clay and shale (some of which is carbonaceous or gypsiferous), lignite, and locally some gravel. The gross chemical quality of the ground water is described by hydrochemical facies which are defined as areally segregated portions of a genetically related body of ground water-within each facies thedominant dissolved constituents are in similar proportions. From the definition a linear statistical model was proposed, and data from 221 samples were fitted to the model. Samples were compared to each other by computing a similarity coefficient (cosine θ) based on the cosine of the angle between sample vectors whose components were the values for the major dissolved ions: calcium, magnesium, sodium plus potassium, bicarbonate plus carbonate, sulfate, and chloride. Samples were grouped into hydrochemical facies by a simple clustering process and tested for homogeneity by correlation analysis. A simple regression analysis for each facies provided workable relationships between each of the dominant ions and dissolved-solids content, which was then used as the single parameter to map the areal distribution of the facies..
Keady, Donald Myron (1971). Application of selected statistical methods to a study of the chemical quality of water in the woodbine aquifers of Texas. Doctoral dissertation, Texas A&M University. Texas A&M University. Libraries. Available electronically from
https : / /hdl .handle .net /1969 .1 /DISSERTATIONS -178478.