USGS Home
Contact USGS
Search USGS

U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2008–5167

Back to Table of Contents

Introduction

The eastern Snake River Plain (ESRP) aquifer, Idaho’s largest fresh-water source, lies within the ESRP mafic volcanic province. Pahoehoe basalt and intercalated sediments make up most of the aquifer, with andesite and rhyolite as minor components. Transmissivities and hydraulic conductivities in the aquifer range over six orders of magnitude, and linear ground-water flow velocities are as much as 10 ft/d (Ackerman, 1991; Anderson and others, 1999). Recent ground-water flow modeling has shown that these hydraulic properties are sensitive to the proportion of sediment intercalated within the basalt.

Analysis of sedimentary occurrences by Welhan and others (2006) indicated that the abundance of sediment relative to basalt was statistically stationary: the local medians and variances of sediment abundance do not vary in response to spatial trends, either geographically or time-stratigraphically. Evidence of this was used to justify the application of the kriging interpolation method for modeling sediment abundance in the aquifer. Nonstationary data can be kriged by accounting for a trend, but more importantly, the absence of a trend reveals one of the underlying properties of the geologic system (in this case, a relatively constant sedimentation rate over the entire study area for at least the past 800 thousand years [Ka]).

Testing for stationarity is relatively straightforward. In this study, as in the work of Welhan and others (2006), nonparametric tests of similarity were applied because the distributions were non-normal and variably skewed. The tests evaluate the degree of statistical similarity between paired and multiple sample populations, in terms of their similarity in shape, medians, and variances.

Numerous well borings, geophysical and lithologic logging records, and drill cores have been collected in the INL region to help define the geohydrologic regime in this part of the ESRP aquifer (fig. 1). Improved understanding of the subsurface stratigraphy has led to increasingly sophisticated conceptual and digital models of the aquifer and of ground-water flow and contaminant transport at various spatial scales. A subregional-scale flow model (approximately 25 by 75 mi; fig. 1) is currently being tested and refined at the U.S. Geological Survey (USGS) INL Project Office. Relative sediment abundances within the aquifer have been shown through model calibration to significantly affect hydraulic conductivity values. A more realistic model of the spatial and stratigraphic variability of sediment may be required to further refine the ground-water flow model.

This study was undertaken by the USGS, in cooperation with the U.S. Department of Energy, to test fundamental assumptions, namely spatial and temporal stationarity, of a geostatistical model of sediment abundance at INL that was developed by Welhan and others (2006). Any future ground-water flow model calibrations will hinge on the accuracy and reliability of such models, and this study was designed to test statistical assumptions in the sediment abundance model. The study also was designed to identify problems or biases in data sets that could affect subsequent studies of a larger scope, such as the subregional scale ground-water flow model being developed by the USGS (Ackerman and others, 2006).

Purpose and Scope

This report describes the results of statistical tests that were used to evaluate the conclusion of Welhan and others (2006) that sedimentary interbeds at INL are distributed in a statistically stationary manner, in a spatial and a temporal sense, using new and more detailed data available subsequent to their study. Welhan and others used stratigraphic data interpreted from geophysical logs of 333 wells to demonstrate that sediment abundances of the stratigraphic units below the water table appear to be statistically invariant in different areas of the INL and through geologic time. The data used in this study are of higher quality than those used by Welhan and others (2006) because detailed lithologic logs of drill cores were available to compliment interpretations based on natural-gamma logs by confirming the existence, position, and thickness of sedimentary interbeds.

The lithologic information used in the previous study of Welhan and others (2006) was based only on natural-gamma log interpretations, so the reliability and quality of the stratigraphic data were unknown; therefore, this study also evaluated the accuracy of the lithologic information that can be obtained solely from natural-gamma logs. High quality stratigraphic data compiled from lithologic logs in this study were compared with stratigraphic inferences based only on geophysical logs to determine if the natural-gamma logs consistently failed to identify interbeds less than about 2–3 feet thick. The relative effort required to interpret geophysical data in conjunction with lithologic logs versus natural-gamma logs alone was also evaluated.

Stratigraphic data, including 122 sedimentary interbeds, from 11 coreholes within the INL boundary were entered into a stratigraphic database. The 11 holes included eight “USGS series” holes and three “Middle series” holes. All new holes are in the southwestern part of the INL. Stratigraphic information was gathered using lithologic, photo, and natural-gamma logs.

Previous Work

Geophysical logging has been the primary method for characterizing lithologies and distinguishing sediment from basalt in the subsurface at the INL (Barraclough and others, 1976; Anderson, 1991; Anderson and Liszewski, 1997). Drill cores have been selectively analyzed for properties such as geochronology, paleomagnetism, and bulk geochemistry (Kuntz and others, 1980; Anderson and others, 1997). Systematic interpretation of these and other data sets has led to the development of a stratigraphic framework and a conceptual model of ground-water flow in the ESRP aquifer (Ackerman and others, 2006). Recent work has focused on characterizing the geostatistical distribution of sediment within the aquifer to aid in future model calibrations (Welhan and others, 2006).

Statistical Analysis

Welhan and others (2006) modeled the distribution of sediment in the aquifer to better understand the variability in hydraulic conductivity within the aquifer and to increase the accuracy of simulation made with the subregional-scale ground-water flow model being developed by the USGS INL Project Office. A main conclusion drawn from their work was that sediment abundance (relative to basalt) in the aquifer is spatially and temporally stationary (statistically invariant). This fundamental assumption allowed Welhan and others (2006) to krige borehole data as two-dimensional spatial data, characterize the geographic distribution of sediment abundance, and parameterize hydraulic conductivity with respect to individual flow model layers (Welhan and others, 2006, p. 13–15).

Interpreted Stratigraphy

Because fewer than 10 percent of all boreholes on the INL have been cored, almost all stratigraphic correlation analyses have relied on geophysical and geochemical indicators. Early subsurface stratigraphic work in the area of the Radioactive Waste Management Complex (RWMC) (Barraclough and others, 1976) used predominantly geophysical logs to characterize sedimentary interbeds and individual basalt-flow groups (defined as packages of flows derived from a single shield volcano). Studies conducted by Anderson and coworkers (Anderson and Lewis, 1989; Anderson, 1991; Anderson and Bowers, 1995; Anderson and Bartholomay, 1990) resulted in the first and most carefully documented stratigraphic framework for the INL subsurface. This work characterized subsurface lithologic variability and offered stratigraphic correlations across the INL on the basis of natural-gamma log records that had been calibrated to available core properties. Interpreted subsurface stratigraphy from 333 boreholes across the INL and vicinity (fig. 2) were compiled into a database (Anderson and others, 1996; Anderson and Liszewski, 1997). These data were used by Welhan and others (2006) to analyze and geostatistically model sediment abundances in the aquifer beneath the INL.

Composite-Unit Stratigraphy

Anderson and Liszewski (1997) proposed a stratigraphic classification based on groups of units of similar age to show principal subsurface stratigraphic features. They grouped individual basalt and sedimentary units into 14 composite stratigraphic units, each made up of 5 to 90 individual stratigraphic units of similar age (Anderson and Liszewski, 1997, p. 14, table 4).

Composite units recognized in the subsurface of the INL are summarized in table 1. Composite unit 1, the youngest, consists of 78 basalt-flow groups and as many as 12 sedimentary interbeds. Composite unit 14, the oldest, consists of 4 identified basalt-flow groups and 1 sedimentary interbed. The decreasing number of individual units within successively older composite units may be either the result of larger and less-frequent volcanic eruptions or possibly a function of the decreasing availability of borehole data with depth on which to base stratigraphic correlations. Composite units 2–7 are the primary focus of this study because the ESRP aquifer is hosted primarily in these units within the study area. Composite units 8–14 were grouped together and evaluated as a single sample because of the limited number of deep coreholes, resulting in small sample size and the difficulty of making individual composite-unit assignments.

Back to Table of Contents